RIJKSUNIVERSITEIT"TE GENT

WERKEN UITGEGEVEN DOOR DE FACULTEIT VAN DE

WIJSBEGEERTE EN LETTEREN

I IS" AFLEVERING

UITTREKSEL UIT HET REGLEMENT

De werken van de professoren en docenten, oud-professoren en oud­
~/ z. 3;/7­
docenten worden uitgegeven onder de persoonliJlte verantwoordelijkheid van
~/7r
de schrijvers. De andere krachtens een beslissing ....an de Faculteit.

DR R. DEROLEZ
I

»::t
GEA8S0CIBERDB VAN BET N. F.
----
w. o.

RUNICA
----- '

MANUSCRIPTA .,....-.-­

THE ENGLISH TRADITION

1954­
"DE TEMPEL", TEMPELHOF 37, BRUGGE (BELGIE)

 1//
I {
f.:lJ
,
20/7
'])7

CONTENTS

PREFACE IX

INTRODUCTION. . . . . . . . . . Xlll

Runes and Runica Manuscripta . XIV

Research on Runica Manuscripta xxxiii

Plan and Scope of the Present Study Iv

CHAPTER I. The Fuporcs (I) I

The Manuscripts and Their Runes. . I

Conclusion . . . . . . . . . . . . 86

VAN DlT WERK Wl!RDBN GlIDaUKT OP DB PBRSBN

CHAPTER II. The Fuporcs (II) : The Isruna Tract . 89

VAN DB SINTB-CATHARINA DRUKKERIJ TB BRUGGB,

OP HOUTVRII VELIJN, 600 I!XEMPLARBN WAAR­
The Manuscripts. 90

The Isruna Tract . . . . . 120

VAN 50 VOORBBHOUDBN /\AN Oil SCHRIJVBR,

G£NUMMERD I TOT L
The.Runes . . . . . . . . 122

BN 550 BXBMPLARBN VOOR DB HANDEL,

The Values and Rune-Names 124~
GENUMMBRD [ TOT 550. The Examples and the Texts 13 1 ~~,.
Runic Cryptography . . 137

Epigraphical Evidence . . 13 8

Drr IS NUMMHR -404

The Ogham . . . . . . . 146

Excursion: Nemruvus's Alphabet. 157

Conclusion . . . . . . . . . . 160

Appendix I : Related Latin Cryptography 161

. Appendix II : Norse Runic Cryptography 16 5

CHAPTER III. Runic Alphabets (I). . 17 1

The Manuscripts and Their Runes. . . 174

Conclusion . . . . . . ~ . . . . . . 273

Appendix: Spurious Alphabets and the Last Phase of

the Runica Manuscripta. ..... 274

-CHAPTER IV. Runic Alphabets (II) : De Inventione

Litterarum . 279

The Manuscripts. 290

The Tradition .
345
The Texts
349
vii
 The Text on the Runes . . . 354
The Runes and Their Names CORRIGENDA
I~
359
Conclusion . . . . . . . . 371
Appendix:. Illustrative Materials p. xxiii. 1. 21 : should read: ...the actual usage of runes and
378 Latin letters shows ...
CHAPTER V. Non-Alphabetic Runica Manuscripta 385 p. 14, 1. 18: ... by having the lateral strokes ...
(a) Runes as Additional Letters . . . . 387 . 1. 23: ...not unparalleled form...
(b) Runes Written for Their Names . . . 390 p. 26, 1. 20: ... the Epilogue of his Manual?
(c) Runes Used as Reference Marks, etc. . 402 p. 40, 1. 5: ... covers some appalling lacunae.
(d) Runes in Scribal Signatures, Notes, etc. 40 3 p. 105. 1. 28: unparalleled and may be due...
Conclusion . . . . . . . . . 422 p. 130, 1. 2: The latter may go back...
GENERAL ESTIMATE AND OUTLOOK . . . . . 42 5 1. 3: ... a concession to...
BIBLIOGRAPHY . . . . . . . . . . . . . . 433 p. 145, 1. 16: ... is identical with...
PERIODICALS QUOTED BY ABBREVIATED TITLES 449 p. 195, 1. 6: This is hardly surprising ...
INDICES: A. Manuscripts . . . . . . p. i16, 1. 4: ... can still be recognized.
45 1
B. Inscriptions . . . . . . . p. 224, 1. 3: ... Rabani monachi de computo liber ...
454
FOLDING PLATE: Key to Transliteration p. 258, 1. I: ...the form of the rune... the vertical stroke ...
457
p. 301, 1. 33: ...y and z of the Latin alphabet...
p. 327. 1. 13: ...ad Secundinum episcopum...
p. 335, 1. 27: B = Passus CCC, etc.
In fig. 12 (p. 78) a rune m is missing to the upper left of the third· '.,;"
rune in the last line. Fig. 15 (p. 120) should be as follows:

feh ur dorn oos . ra:h cen gebo huun

~f rlu ~d ~oo Rr ~c
Yis ~uu ,
hagal nod iis ger .h perd elux S19' '

~h In
ti berg
Ii
eh
+9
man
Jk fp Xx '1s
la9 u Inc tag odtl
tt
eLC
&b Me Mm
eLse .ur aer
~l iin Ht ~o
~~ ~~ ~q vrz Variants: {n(S) rrz (U,S)
viii
 PREFACE

Although many runic inscriptions in manuscripts have been

examined before, no study covering all material has been
untertaken thus far. Some scholars have propounded theories
taking account of all material known to them, but nobody has
tried to find out whether these reconstructions of the background
of the Tunica manuscripta agree with the available material.
My task was consequently a double one. First I had to examine
the individual items and their relations to their environment;
then I had to see whether they could really be integrated into an
overall picture, and whether this picture agreed with that present­
ed thus far. Some readers will no doubt feel that I have given
too much space to details; others may object that my scepticism
went too far. To the former I would say: how can we know
which details are relevant? Those which look irrelev~1Ut now _'"
may become significant ten or twenty years hence. When I'" '
examined the studies on the subject published till now I found
that progress 'had been slow not because too many details had
been taken into account, but because important details had
been overlooked. The readers who should have wished a more
synthetic treatment will see that the more material is examined,
the greater the number of question marks. They will also see
that premature syntheses are more harmful than helpful,
because they prevent the student from reali?:ing where the
problems lie.,
The subject treated here raises problems in various fields.
To treat it in a way that satisfies from all points of view, one
should be a runologist, a philologist, a historian, a' palaeo­
grapher, an art historian and some sort of a detective all in one.
I should never have been. able to complete it without help
from many parts. It has been a most pleasant experience to
find so much willingness to help and advise.
In the first place my gratitude goes to the Members of the
IX
Faculteit der Wijsbegeerte en Letteren of the Rijksuniversiteit whose suggestions made my work at the Stadt- und Hoch­
in Ghent, whom I found always ready to put their knowledge schulbibliothek in Berne the more fruitful; to Dr. J. Duft,
and experience at my disposal. I am most grateful to Professor whose helpfulness I could fully appreciate while examining
F. De Backer, under whose supervision it was a pleasure to manuscripts at the Stiftsbibliothek in St. Gall; and to Professor
work, and whose help and encouragement made my task much K. Jackson (Edinburgh) for his advice and kind encouragement
easier. I also wish to thank him for having corrected those when my study brought me to examine matlers in the field. of
inaccuracies and stilistic blemishes which easily pass unobserved Celtic.
when one does not write one's mother tongue. A special word of thanks is due to my friend Dr. G. De Smet
Among the other Members whose assistance I gratefully (Oudenaarde), who also read a chapter of the first draft and
remember I have t.o mention especially Professor E. I. Strubbe, discussed a number of problems with me.
whose vast experience in palaeographical matters made him a As I indicate in the Introduction, lowe thanks to many
safe guide in that field; Professors F. L. Ganshof and J. Dhondt, library authorities for having provided me with photostats and
whose counsel in historical questions proved most enlightening; microfilms of manuscripts I could not examine de vim, for
Professor P. Van de Woestijne who helped me to solve some allowing me to reproduce them here and for giving me all sorts
problems of Latin; Professor G. De Poercq, who kindly drew of information in a most gracious way. I beg the staffs of the
my attention to Mandeville's alphabets and assisted me in other following libraries to find here the expression of my gratitude :
matters as well; finally Professor P. De Keyser, who helped me Avranches, Bibliotheque Municipale.
to find a suitable illustration for the discussion of the hahalruna. Bamberg, Staatliche Bibliothek.
My thanks also go to a far wider circle. Amol.1g the scholars Basle, Universitatsbibliothek..
abroad to whom I am much indebted I must mention first of Berne, Stadt- und Hochschulbibliothek. ,:-..
all Professor B. Bischoff (Munich), without whose numerous Cambridge, Corpus Christi College.
suggestions, and the valuable information which he gave in a Chartres, Bibliotheque Municipale.
most unselfish way, I would hardly have been able to complete Dublin, Trinity College.
this work. It was most gratifying, too, to find one of the Exeter, Cathedral Library.
foremost students of palaeography and Mediaeval culture pre­ Florence, Biblioteca Medicea Laurenziana.
pared to read a chapter of the first draft in typescript and to St..Gall, Stiftsbibliothek.
offer his advice; and still more to find him expressing the' hope Heidelberg, U niversitats- Bibliothek.
that this work might soon appear in print. Karlsruhe, Badische Landesbibliothek.
I also owe thanks to Professor Bruce Dickins (Cambridge), Kassel, Landesbibliothek.
who not only enabled me to examine the manuscripts of Corpus Leyden, Bibliotheek der Rijksuniversiteit.
Christi College in the best conditions, but with whom I had London, British Museum.
also the privilege of discussing the plan of my work. Munich, Bayerische Staatsbibliothek.
I am no less indebted to Dr. C. E. Wright (London), whose Niirnberg, Germanisches Nationalmuseum.
work in this field of study I had full opportunity to appreciate, Oxford, St. John Baptist College.
and who also very kindly read a chapter of the first draft in Bodleian Library"

typescript. Moreover lowe him special. thanks for having Paris, Bibliotheque N ationale.

assisted me whilst I was working at the British Museum. Bibliotheque de I' Arsenal.

I wish to express my gratitude to Professor O. Homburger, Salzburg, Stiftsbibliothek.

x XI
 the more chances there are that two alphabets are related. The
transition from one alphabet to another must be as • natural'
as possible. This procedure lies open to many attacks : how
can we decide whether one transition is more natural than
RUNES AND RUNICA MANUSCRIPTA
another, whether one alphabet is more easily transformed into
runes than another?
The earliest runic inscriptions, viz. those of 0vre Stabu, Of course all explanations of the runic alphabet must be
Fr0ihov, Kowel, Vi, Dahmsdorf, Mos, Torsbjrerg and perhaps based into aI. on a comparison with possible models, but most
one or two others, are mostly dated in the third or the late scholars have looked for support in other fields, especially in
second century A. D. Whether runic writing had been practised archaeology, prehistory and history. They connect the origin
for any length of time before those earliest inscriptions, as some and early diffusion of the runes with archaeological phenomena
authors maintain, we do not know. The origin of the runes is and historical events. B. Salin and O. von Friesen, e. g.
still debated. Many have forms which closely resemble letters associate the genesis of the runes with the settlement of the
in contemporaneous alphabets of Southern Europe (Greek, Goths in the Black Sea area, and their diffusion with the spread
Latin, North Etruscan); so much so that no doubt is permitted of a special form of culture which can be traced in the Germanic
about the runes being derived from one of those alphabets. No zoomorphic ornamental style (1). F. Altheim and others have
solution proposed thus far can be considered entirely satis­ been struck by the similarity between rock carvings in S. W.
factory, but the general trend is in favour of derivation from the Sweden and in some valleys of the Alps, and between the runes
Latin alphabet. As I have discussed the various hypotheses and the ~o-called North Etruscan alphabet; they connect the .
elsewhere (I), I give only· a few general observations on this origin of the runes with the raid of the Cimbrians and the' ,;-,
question, which may serve to justify the present study. Teutons in Northern Italy at the end of the second century'
The material with which scholars have tried to solve the B. C. (2).
genetic problem is heteroclitic. The solution is usually sought Archaeology, however, is not the perfect assistant of runology
in a combination of various arguments, but the preference of which some scholars believe it to be (3). We may question
scholars for one or other field seems to play a part in the actual their right e. g. to place runes and ornamental styles on one
demonstration, if not in the choice of the solution itself. level. Runic writing can hardly be considered as an article of
Some authors have tried to solve the problem mainly from trade and export comparable to other elements of material
the angle of the alphabet and its history. Thus L. F. A. Wimmer Civilization. At the origin it was probably a secret craft, if the
and H. Pedersen derive the runes from Roman capitals (2). interpretation of OCmc. "'rii.n- as 'secret, mystery', etc.
They depend in the first place on the resemblance between the is to have any value. It was never used by many: it is no
two alphabets, assuming tacitly that the greater the resemblance, paradox to say that Germanic culture remains illiterate down

(1) Problemen, 193 fr.; Richtingen, 5 fr. Cf. also O. VON FRIIlSEN, De senast
(1) B. SALIN, ThierDr1Ulmentik, '145 fr.; O. VON FRIESEN, Runskrijtens hiir­
jramstiillda meningarna; A. NORDLING, Runskrijtens ursprung; F. AsKEBERG,
kamst, and other works by the same author.

Norden och kontinenten. 38 fr.; A. BJEKSTBD, Mdlnmer, 132 fr. Judging only
(2) F. ALTHEIM-E. TRAUTMANN; Vom Ursprung tIer Runen; H. ARNT"L,

by a short summary, I suspect that the theory adVllIlced by V. WANSCHI!.R,

Hmulbuch.1, 52 fr.• Hmulhuch s, 30 if.; G. BABSI!CKE, Vor- tmd FriiJrgesc:lrichte I,

La langu.e eltUsIjIUI rellatt (Copenhagen, 1952): Egypt> Etruria> Rhetia> 96 fr.; F. ALTHEIM-E. TRAUTMANN-NEHRING, Kimlwen und Runen.

N. Eur~, overstrains the evidence and will find little approval. (3) See e.g. F. AsKEBERG'1I defence of archaeology in Norden och kontinentl'1t,

(2) L. F. A. WIMMBR. Die Runenschrijt, 139; H. PlIDRRSIlN. L'origine.

38 f.

XlV xv
'with no other alphabet" is still obscure : not one of the many tively small., Some 150 are recorded in ,the Scandinavian
attempts to solve this riddle can be considered convincing. The countries, and about 30 on the Continent (1).
name fupark, which is often given to the runic alphabet, con­ At an early date two developments in opposite directions set
sists of the first six letters in the original order. in. In Scandinavia the'number of run~&.was drastically reduced
There is another peculiarity the origin of which is likewise to sixteen by the end of the eighth century (2). At the same
obscure. The fupark was divided into three groups of eight time some new forms appeared, as can, be seen from the two
runes each' (1). These subdivisions are indicated on the basic types in fig. 2.
Vadstena and Grumpan bracteates (sixth century), but we have
no earlier evidence. In Chapter II I shall have to return to
this peculiarity (p. 89 if.).
Doni~h
Swed.. Nol"'W.
rf)~ ~ kY *
: i I .,. '1 tB'f't!
Pl D~ 1\ r f· ~ r ~ I 1 ~ t~
,I I
For a third feature of· the fupark we have no direct epigraph­
ical evidence at all. Yet there can be no doubt that it goes fupQ.rk hnias ~bmlR
back to the first centuries of runic writing. Each rune had a Frc.2
name, usually a word (either a common noun or a proper name) The old system with three ~r~ups 'of eight run~s' was 'ad~pted
beginning with the sound indicated by the rune. The rune to the new total of sixteen runes : one' group of six and two of
for m e. g. was called *manna- ' man '. Instead of writing such five runes each. '.In later Icelandic literature on this subject
a name in full when it happened to occur in his text, the carver these subdivisiorlsare called altt (pi., alttir; , tribe, family', or
might be satisfied to engrave only the rune. Thus the f-rune , group of eight' ?), and this has b,ecome the, technicJll term in
on the Gummarp stone probably stands for *fehu 'riches, runological literature, just like· fu~ark. They are sometimes
property '. The runes were not only abstract sound-symbols ,distinguished by the later Icelandic names of the ,first rune in , ~~'.
such as we are used to, but they had also something of ideo­ each group: Freys altt, Hagals altt,' Tys altt group of Frey,.,
graIns. They were learned and called by their names. When Hagall, Tyr}. Later on other forms were developed on the
the initial sound of such a name was affected by a phonetic same basis, some still increasing the difficulty of reading (the
change, the value of the rune itself was affected. Thus in OE., Halsinge runes) (3). But gradually the small number of char­
where *ansu-, the name of the a-rune, became Os, a variant. acters must have proved a handicap. New runes were added; by
form of the original rune took the place of a in the fupark and 1300 there were runes corresponding to all letters in the Latin
received the value c. Only two rune-names make an exception alphabet, and some extra symbols as well. Inscriptions with
to ,the acrostic, principle, i. e. the principle that the rune-name these younger runes are far more numerous than those in the
begins with the sound indicated by the rune : nos. IS and 22.
Since the sounds they stood for were never found initially,
their names begin with another sound (IJ = 'Ing ').
(1) O. VON FRlBSEN, Runonra, map no. 1; H. ARNTZ-H. ZEISS, Runendenk­
The number of inscriptions using the 24-rune fupark is rela­ mIiJer, 474 j most of the Frisian inscriptioDS are in the later alphabet, cf. ibid.,
t07 f. and P. C. I. A. BOELllS, Friesland, 338 ff. On the whole runic writing
seems never to have gained a sure footing On the Continent. If we leave 'out
of account the Frisian inscriptions and those ascribed to East Gennanic tribes,
very little remains indeed, and that little belSlDgs to a relatively short period:
A. D. 400-700.
(I) J. LP.JEGREN, Run-Lara, 52, assumes that the fupark could also be
(2) H. ANDElISEN, Ikt yngre Ru7U!alfahets Oprinrklse. Arkiv 6z (1948)"
divided into two, four, etc. groups of runes, but these systems must be late 203-227.
elaborations. ' " ""
(3) O. VON FR1I!SI!N, Rrmoma,' 84, rOl, 140 C.,' Ilj.6'.'149. zZ9.z3I, 244.
xviii
xix
 original fupark : some 350 in Denmark (I), 500 in Norway and in form or value. The runes f u j;) r g w h nip s t b e mId
•1.': its colonies (2), and well over 2000 in Sweden (3). require no special comment (I). The change in no. 4 (*ansu- >
In Frisia and England the number of runes gradually in­ {JS,- hence value a> 0) has been mentioned on p. xviii. The
creased. As new phonemes arose in the language, new runes new form of the rune has been explained as the result of a
were created. Thus we find inscriptions pointing to an al­
phabet with twenty-eight and even with thirtyAtwo or thirty­
ligature a + n. Run~s which did not have full height in the
.old fupark received a full-length shaft : compare Gmc. k and OE.
three runes. As the value of the third rune had become 0, c (no. 6), Gmc. and OE. j (no. 12). No. 22 reaches the normal
and c was the usual OE. spelling for k, I propose to caU this height in a different way: the four sides of the square are
expanded alphabet juPOTC (as opposed to jupark for the dGmc. lengthened. No. 13 is not well defined: it stands for h (Ruth­
alphabet, and fupqrk for the shorter Northern varieties)
well allme3tig, U rswick toro3tredre), for i (Dover ;3s1heard)
(fig. 3). and for g, or rather j (Thornhill III eate3nne). No. IS is also
12345(' 78 ,101111131+151& 171817t011122314 puzzling. In the oldest inscriptions it indicates final voiced
rrl~~khXt M1-1 +.t~T'" t~ Mf1t~H~ s (z), usually transcribed -R, which develops into ON. -r
fupol"c~w hnij3pxs lbeml!)doo (=Go. s). It occur" very rarely in English inscriptions. At
an early date (St. Cuthbert's coffin, 698) it is used as the equi­
2S U '1.1 t8 if 30 31 32 33. 124 U. 1'1, 21. :S:SQ.
P'P tf1'~*.*~ +Ii r, ~~ valent of Lat. x. The name of no. 24 was affected by i-muta­
tion, and the value of the rune was changed accordingly :
a. Ie yea.k 9 k J S~ *opil- > repel (still in Anglian texts) > WS. and Ke. eper.
FIG. 3 No. 25 is derived from the a of the original fupark. Its name,
OE. ac, goes back to a Gmc. *aik-, and therefore it is supposed "<~"
This type of alphabet occurs on about half a dozen Frisian
inscriptions (4), and some 30 or 40 English inscriptions. Un­
that the new a-rune is a contraction of an old a +
i. The
original a has become no. 26. Its value is now re, as in OE.
fortunately we still lack a full and up-to-date edition of the
*a was palatalized to re in many positions. No. 27 is au-rune
English inscriptions; the material lies scattered in more or less
with a subscript i, used to represent OE. y, i. e. the i-mutation
accessible publications, and some items· at least need a careful
of u. The origin of no. 28 is obscure. It renders the OK
re-examination (5).
sound which corresponds to Gmc. *au. Nos. 29-31 were
Most of the runes of the old fupark show no important changes
created to distinguish various types of gutturals. Nos. 6 and 7
have names beginning with palatal gutturals: cen, giefu. No. 29
(I) L. JACOBSBN-E. MOLTKE, Ru7leiruJskrijter (Tex.t), uno if.
(2) O. VON FRIESEN, Runorna, 83 (M. OlSl!N): about 350 inscriptions in is the velar corresponding to no. 6, no. 30 that corresponding
NoI'WllY. a few in Bohuslin and Jimtland (which formerly belonged to to no. 7. No.' 31", finally, seems.to be used for rendering a
Norway), and the rest in Ireland, the Isle of Man, Cumberland, the Orimeys, back stop before a front vowel, as in kyniy. For no. 12 I give
Iceland and Greenlan4.· .
(3) O. VON FRIESEN, RU1IOT1Ul, 174. the form directly_ developed from the OGmc. type; it is. the
(4) H. ARNn-H. ZEISS, Ru:nendenkI'lU'ileT, nos. 6, 9, 14; 20 f., 37-39. form which occurs most frequently in manuscripts. In in­
P. C. J. A. BoELl!S, Friesland, 338 fl. scriptions its place is usually taken by no. 32, the origin of which
(5) The edition announced by B. DICKINS and G. B. BROWN (Atheneum
1920, 874; Notes and Queries (s. 12), 7, 500; Scottish Historical Review 18,
will be discussed later. No. 33 is best known from Frisia
156) has not yet materia~; nejther has H. ARNTZ'S (Gesamtausgobe i:hr (Westeremden), but it is also found in England. It stands for
dlteren RunendenkmiiJer, Band III: Gesamtausgobe tier einheimischen englisclum
~ : d. his Bibliogrtl/>hie no. 116). . (I) I mainly follow B. DICKINS's system of transliteration; see p.llilii .for
further details.
xx
XXi
-,
the combination st. At the end I give some variant forms : no convincing proof has come down of such a usage. Their
nos. 12a and 24a from the Thames sword, I6a from the Ruth­ main usefulness lay in other fields than in those of commu­
well and Dover inscriptions, 16b from St. Cuthbert's coffin nication and recording. It is usually assumed that they played
and the Thames sword, 33a from Westeremden B (Frisia). a part in magic: they conveyed a special power to the inscribed
The chronology of the English runes is still uncertain, and object. They protected against the evil eye, they acted. as
will probably remain so. The number of inscriptions is very love-charms and brought victory or defeat, abundant crops or
small, and very few of them can be dated within narrow limits. disease and misfortune. It did not matter whether the inscrip­
Since a number of changes in the runes and their values and tion was read: as soon as it was carved, it became efficient.
names reflect the phonetic evolution of the language, the latter A. Breksted has tried quite recently to disprove this assumption.
affords some basis for dating the appearance of new runes. He believes that runic writing may be regarded as " an actually
Thus the new 0 can only have been created after the change abortive attempt at creating a cultural factor not yet required
an > 0 had reached a point where it could be felt as a phonematic for practical purposes", without' relation to "profane or
change; and similarly the y-rune may be dated by the i-muta­ religious (magic) requirements " ... " The fact that this attempt
tion. But the phonetic changes themselves permit of hardly took the shape of a 'home-made' alphabet instead of the
more than a relative chronology': absolute dates must needs be simple adaption of the Latin letters, may be explained by a
hypothetical. W. Keller, who made a synthetic study of OE. conscious, or unconscious, aversion in the author of the runic
runic chronology, proposed the following dates: the new a system towards the culture of the 'Herrenvolk', an urge to
and 0 (besides re, which only changed its value) : saec. V; j demonstrate Germanic character and independence" (1). Of
before 600; ea and y in the period 650-670, .g before 670, course, when we consider the runes as " a luxury and a play­
k after that date (1). But the foundations of this chronology thing", there is no need for magical or other backgrounds. ,~,
are quite weak. The inscriptions referred to have been assigned At any rate the actual usage of runes and Latin letters show
to different dates. When older spellings are used as points of that they belonged to different worlds. To be sure, at a very' .
reference, one should not forget that we have very little material late period the runes are used much in the same way as Latin
by which to decide at what rate spellings (and runes) followed letters for inscriptions (even in Latin), but even Breksted will
the changes in pronunciation. Therefore it is not surprising point out that runes were not used for book writing (except in
that some scholars propose entirely different chronologies. rare, obviously archaizing instances). There was at any rate
H. Arntz e. g. believes that the OE. fuporc was extended in nothing that destined runes to be written in manuscripts; there
two movements: first to twenty-eight runes, and later to thirty­ appears to have been " a radical difference of function between
three. One movement should be dated before 800-850, the runic and bookhand" (2). When we find runes in manu­
other after (2). But we must in all probability also reckon scripts, we must' conclude that the people who normally wrote
with regional differences, even if the small number of inscrip­ Latin had become int-erested in them. This could only happen
tions hardly gives a clear picture. after the runes had left their narrow barbaric, or worse, pagan
At the origin, and for quite a long time, the runes. were surroundings. It also meant the transfer of an epigraphical
essentially an epigraphical and non-utilitarian form of writing. style to bookhand, but this will occur only in very rare, late
They seem not to have been used systematically for writing instances, with obvious archaizing tendencies (e. g. the Codex
down law books, poetical texts, accounts, etc., or if they were, Runicus). Usually the runes we meet in manuscripts retain
. (I) W. KEu.J!.R, Zu.r Chronologie. (1) A. BA!KBTBD, MtIlru1U!T, Ja8•
(2) H. ARm:z, Handbuch1 , 146.
(2) J. BLOMFIELD, Runes, 182.

xxii '
xxiii
their special character in some measure : they may be used as mann (monn) and (Bpel (epel). Similarly the Norse In-rune

an ornamental script, or for cryptograms; or else they become is occasionally written for the word maar., The poet

an object of study. Cynewulf invented an intricate variant of this usage for

What I have called TUnica manuscripta covers in fact a wide signing several of his works.

range of runic writing :

In some manuscripts runes are used as reference marks,
(a) First of all there is the use of runes as additional letters. e. g. for indicating the passage to which a marginal note
The latin alphabet proved a poor instrument for rendering refers, or for numbering quires. The former usage does not
Gmc. sounds, as a glance at the earliest OE. or OHG. , require any knowledge of the runes beyond that of the
texts will show. Besides various other devices (digraphs, mere forms: the scribe could as well have used Latin
modified letters), scribes also introduced runes to indicate letters, or yarious sorts of conventional marks. In the
sounds for which they found no satisfactory equivalents in latter case the runes appear to have been arranged in the
the Latin alphabet. Many OE. manuscripts e. g. use the order of the Latin alphabet. As such runic alphabets
p-rune to render the dental spirant and the tV-rune for the enjoyed some popularity, in the ninth, tenth and eleventh
bilabial spirant. The English example had some influence centuries, written models are probably at the origiJ;l of this
on ON. and, though rarely, on OlIGo spelling (e. g. the usage.
Hildebrandslied). As far as OE. is concerned, this usage is
Runes were occasionally used for writing short notes, such
rather more limited than is usually, assumed (1). Our
as the scribe's or a reader's name, an expression of pride
information is still fragmentary; a full palaeographical and
or gratitude at the completion of a manuscript, a remark
philological investigation .of OE. (and ON.) usage would
on the text, a' riddle; sometimes also for an ornamental ,'",
probably be most rewarding. Of course this sort of runic ~.

title page. As far as the material goes, these notes are

writing does not tell us much about the character of the
invariably in Latin. They are hardly proof of the scribe's
runes. These are completely assimilated to Latin letters,
knowledge of the runes: runic alphabets were circulated
and we have some reason to doubt that the scribes always
so widely that many scribes must have had access to them.
knew what sort of characters they were using. But the
With such an alphabet before them (which might at times
introduction of these runes may be an important indication
bear misleading inscriptions: 'Arabic', 'Syriac' etc.)
of the general attitude towards runic writing.
they could write any text. For the same purpose they also
(b) A more intimate knowledge of the runes is implied by the used Greek letters, and even the so-called Scythian alphabet
following variety, which is also much rarer. We saw that of Aethicus Ister.
each rune had a name, and that, instead of writing that Two considerations seem to have led scribes to practise
name in full, if it happened to occur in a text, it would be this sort of runic writing: either they were attracted by the
sufficie~t to write the rune by itseJf. . This usage is also ornamental character of runic script, which, as far as style
found in OE. manuscripts. In some texts, words which is concerned, could compete with Roman capitals; or else
were also rune-names' would occur so frequently that a the cryptic nature of the runes, which would be intelligible
scribe who knew the names would feel justified to sub­ only to those who had a key alphabet, was the decisive factor.
stitute the corresponding runes for them. Thus we find the
runes d, m and (2 taking the place of the words d:eg (deg), Fuporcs and runic alphabets. In a whole series of manu­
scripts, going from the end of the eighth to the early fifteenth
(1) J. BLOMP'IELD, Runes, 185. Cf. Chapter V. century. we find lists of OE. runes, often with their Latin
xxiv xxv
 equivalents and their names. In some the runes have order and the division into three sections. All scholars who

II,».., retained the original ful>orc order, but mostly they have discussed the runes in a general way, and especially those

been rearranged so as to follow the order of the Latin have dealt with the problem of the origins, have'·referred

alphabet. I have reserved the term ' runic alphabets' for the material included in manuscripts. But their attitude

the latter. A number of manuscripts also contain short this material is far from uniform; in fact it varies from

remarks on runic writing. In one group we find a treatise scepticism to uncritical acceptance.

on cryptographical systems derived from the divisions of L, F. A. Wimmer, who laid the basis for a scientifically sound

the ful>orc; in another various alphabets, ranging from , did not discuss the Tunica manuscripta by themselves.

Hebrew to runic, are briefly discussed. a treatment would hardly have fitted into the plan of his

Comparable Scandinavian material belongs to saec. lX­ the first part of which discussed the origin of the runes,

XVIII, cf. p. lvii. the second~ their further evolution in Scandinavia. More­

Wimmer devoted little space to the peculiarities of the

(f) Finally, there are the so-called runic poems. There are
; he believed that their origin could not be explained.

four poems of this sort: one OK, one Norwegian, one

et manuscript runes play some part in his work: he relied on

Icelandic, one mixed ON. and Continental Germanic. The

for the interpretation of the inscriptions on the Vadstena

first three are made up of short stanzas, each of which deals

the Charnay brooch· and the Thames sword (1);

with one rune. The OE. poem therefore has twenty-nine

I>mUl<U1Y he believed the material in Vienna MS. 795 to be proof

stanzas, the Norwegian and Icelandic only sixteen. The

Goths' having known runic writing (2); etc.. '

OE. poem is usually dated saec. XI or X ex., but it goes

Still Wimmer's work also forebodes a more extensive useDf

back to a much older original (eighth or early ninth cen­ :-~~,

these data, He saw the origin of the runes as a creative act

tury). The Scandinavian poems are of a later date: the

(based, it is true, on a' thorough knowledge of the Latin al­

Norwegian may be of the twelfth or thirteenth century,

phabet). The features which distinguish the runes from the

the Icelandic of the fifteenth. But all three have, besides

Latin alphabet can only be understood if we see in them the

the. general structure, a number. of details in common,

work of one man, who proceeded in a logical and systematic

which can only point to a common source, although the

'way. If we accept this solution, we must conclude that the

character of that source may be disputed. The fourth,

. study of those special features may throw some light upon the
the Abecedarium Nordmannicum, hardly deserves to be
creative act itself, and upon the cultural and ideological sphere
called a poem. It has occasionally been praised as the
in which it occurred. But Wimmer considered his solution of
most remarkable of all four, which may explain why it has
the genetic problem sufficiently well established, and preferred
received so much attention; but it rather looks like .mnemo­
not to weaken his case by doubtful arguments.
technical doggerel. Owing to its mixture of ON. and
The data derived from manuscripts played a much greater
Continental Gmc.-the runes and their names are mainly
ON., the text itself is OLG. with some admixture of OHG. (1) L. F. A. WIMMER, Runenschri/t, 88 : Die bedeutung der einzelnen zeichen

- it is of greater interest for the study of ninth century der runenreihe auf dem brakteaten von Vadstena (der spange von Chamay

culture than for runology. und dem Themsemesser) ist zum grl>ssten teile kIar. teUs durch vergleichung

mit den handschriftlichen altenglischen alphabeten, wo die bedeutung and

Groups (e) and (f) provide us with an amount of essential die ruunen der runen hinzugefiigt sind, teih and vomehmlich durcb die

bedeutung, welche die zeichen in den lIUlI der ilteren und mittleren eisenzeit

evidence on the runes : they contain all we know about the iiberlieferten inschriften selbst aufweisen.

names of the runes, and some welcome add'itional evidence on (:&) L. F. A. WIMMER, Runenschrift, 71.

XXVI xxvii
part in S. Bugge's introduction to the ,edition of the Norwegian captives, or Goths serving in the Roman army, learned the
inscriptions with the older fupark (I). His explanation of the alphabet. From the latter they derived an alphabet of
rune-names actually determined his' solution of the genetic own, with the addition of a few Roman letters which
~endered Gothic sounds more closely than Greek letters could
problem. He believed that the runes had originated with the
Goths, who had passed them on, by different ways, to the' For the rest von Friesen relies mainly on archaeological
Scandinavians and to the WGmc. tribes. Formally the runes l!V1i1l!nl'p.j the manuscript material has no influence on his
were derived partly from the Greek, partly from the Roman of the genetic problem, although he interprets the rune­
alphabet. But for their names Bugge found models in Georgian as being in favour of a Greek rather than of a Latin
and Armenian : cf. Georgian man, Armenian men the name In fact his treatment implies a further step: the special
~hllTacteristics of the runes may result from a later development.
of the letter m) and Gmc. *manna- the name of the m-rune).
This rather fanciful theory should not be judged too severely : Friesen laid-the foundations for this theory when he 'pos­
the author did not live to revise this part of his text, which tulated that the runes were used for magical purposes only at
actually consists only of fragmentary notes. a later stage of their development.
Bugge's extensive use of manuscript material led him to It is obvious that von Friesen's attitude' hardly encourages
subject it to a more thoroJlgh criticism than it had usually thorough study of the TUnica manuscripta. The same applies
received thus far. The special study of this material had made H. Pedersen, who advocates a strict adherence to Wimmer's
some progress in recent years, especially at the hands of von method (I). He occasionally discusses the names of the runes,
Grienberger (p. xlii). This we find reflected in Bugge's but only from the point of view of the possible relationship
treatment; his text on the runic alphabets written on the Con­ between the fupark and the Old Irish ogham.
tinent contains valuable suggestions for a closer investigation (2). With C. J. S. Marstrandet we find again a wider use of the
But it looks rather as if Bugge's uncritical use of the manuscript manuscript evidence (2). He believes that the ruries go back
material was felt as a warning by the next generation of runo­ to a Celtic prototype, and that' they originated in the Rhine­
logists. Danube area. This implicates that the ogham and the fupark
At any rate the TUnica manuscripta play no such part in have a common origin. The two alphabets are characterized
O. ,von Friesen's theory (3). He supposes that Gothic slaves by long and significant letter-names. The Old Irish letter­
names form homogeneous groups (names of plants, of rivers;
(i) S. BUGGE, Norgu lndskri/ter med de mldre Runer, Indledning. of saints, etc.), whilst the rune-names seem at first sight to be
(a) S. BUGGE,' lndledning, 39: "I flere paa Fastlandet hjemmeh8rende unconnected. Marstrander, however, explains the latter as
Haandskrifter,, som meddeler ags. RunenaVIie, er disse forandrede ved
Paavirkning fra eller Omsetning til 'tyak Sprogform (hvilket tidligst synes at names of gods and other mythical beings, and so has to discuss
have fundet Sted henimod Slutningen sf 8de Aarh.). I mange af disse paa them in detail.
Fastlandet skrevne HaandSkrifter har Runenavnene ogsaa undergsaet uvil· M. Hammarstrom derives the runes from one of the Alpine
kaarlige eller vilkaarHge Forvanskninger. Men ogaaa de i disse Haandskrifter
forekonunende Navneformer fol1;jener Opma:rksomhed, dels fordi nogle af
alphabets (3). He is not convinced by Marstrander's Celto,,:
dem f<;lr Runenavnene henviser til a:ldre ags. Spmgformer end de, der Germanic reconstruction: the relations between ogham and
foreligger i de rent ags. Haandskrifter, dels '80m interessante Bidrag til fupark belong rather to a late date. Therefore the manuscript
Kundskab om Runemes og Runenavnenes Udbredelse og Forandringer...
material is of little importance.
Ved Sanunenstilllngen af Runenavnene i det f01gende har jeg i Regelen kun
ruevnt rent angelsaksiske N,vneformer, men ikke de Navneformer, i hvilke
(I) H. PIIDBRSEN, L'origine.

de ags. Fonner er va:sentlig omdannede enten ved Indflydelae fra h8itysk

, (a) c. J. s. ~1lR, RuntJntJ.

Spmg eller ved Forandringer sf anden Art."

(3) M. HAMMARSTROM, Om runskri/tens hdTkomst.
(3) O. VON FRIESEN, Runskrift;ens h4rkomst, and, other works.

xxviii XXIX
, S.' P. Agrell, on the other hand, depends mainly on the rune­ runes which we shall examine later (p. Iii):
names to connect the fupark (or rather the upark: he shifts way of handling the manuscript material can hardly
the first rune to the end of the alphabet) with. the mystery termed satisfactory: every single item is mad,e to fit into
religions of late Antiquity, especially with Mithraism(I). He general theory.
tries to show that the. magic use of the runes is a direct con­ In his work on the relations between Scandinavia and the
tinuation of that esoteric system in which numbers and religious F. Askeberg, with whom we may end this survey,
concepts are closely connected : • I ' • bull " the first being a stand against Arntz's (and other German scholars')
created by Ahura Mazda = the u-rune, the name of which is ur wproach (I). He stresses the importance of archaeology as a
, bison' (in the up ark u is the first rune). Here the manuscript for exploring prehistoric periods. With von Friesen
evidence (our only source for the rune-names) is decisive~ Bugge) he holds that runic writing was first practised
With H. Arntz we find again a very extensive use of the rune­ the Goths. ~ At the same time, however, his comparative
names, especially in the second edition of his manual (2). of alphabets convinces him that Wimmer's solution is
Following Hammarstrom, Altheim and Trautmann, he lays most probable. Askeberg right ly stresses the remarkable
the origin of the runes in Northern Italy. In the valleys of 'phonematic analysis behind the fupark : the 'striking adaptation
the Alps through which they marched on their way south, the the runes to OGmc. phonology proves that they cannot be
Cimbrians and the Teutons would have found tribes of Ger­ outcome of a gradual evolution or of haphazard borrowing;

manic origin, who practised rock-carving much in the same must go back to an individual creation.

way as it had been done e. g. in S. W. Sweden (Bohuslan) in A. Breksted's recent solution was briefly mentioned on p. xxiii.
the Bronze Age. These tribes had learned the North Etruscan This short survey enables us to characterize the part of the

alphabet from their neighbours, and thus it came into the manuscripta in general runic studies.

T.!"',
hands of the invaders. These had already a sort of rudimentary First of all, there is no question of a study of the runica

ideographic script which they used for cultic purposes. After manuscripta as an autonomous subdivision by the side of epi­

their defeats some of the survivors would have returned to graphical runes. The reason for this is obvious: manuscript

their homeland, and thus an alphabet used on the southern runes are secondary; they imply a 'break' in the tradition,

slopes of the Alps served as a model for the fupark. Arntz even an adoption by a different world; The importance of this

tries to reconstruct the religious and cultural circumstances of break, however, should not be exaggerated. The latest views

this creation, and here he has to lean heavily on the rune­ on the conversion of the Germanic peoples to Christianity are

names (3). As a by-product of this theory there is a study on a long way from the • conversion or death' conception (z).

The conversion meant a break, to be sure; but just as the old

(x) s. P. AGl\Eu., Herkurift tier R:u.nenschrift; R:u.nomar ta/mystik, etc. poetical style and the traditional ornamental types would be

(:z) H. ARNTz, Hand1Juch (first and second edition). called upon to serve the new religion, so would the runes.

(J) "Die Beziehungen zu unserer hronzezeitlich-eisenzeitlichen get:­

manischen Religion, die wit bereits (oben S. x3+ff.) von den Sinnzeichen sus
featsteUten, haben der Gegenprobe in verbliiffender Weise standgebalten.
Wir wollen versuchen, sie 'mit der Behandlung der einzelnen Runerinamen
zu ergllnzen. Diese Behandlung fuszt auf den literarischen Quellen, den (x) F. AsKl!BERG, Norden och kontimmten, 38-94­
Runenliedem und Handllchriften; nur in seltenen F"allen kann aus lnschriften (:z) See e.g. H. LJUNGBERG, Den lIardiska 1'tIligionen och kristendomen.

die Gegenprobe auf unsere Schliisse gemacbt werden. Die Uberlieferung St.u.dier IJver det nordiska religWnskiftet under 'lJikingatiden. (Nordiska texter nch

gestattet uns die sprachliche Rekonstruktion der Namen, und .die Frage undel'8Okningar II), Stockholm, etc. 1938; and W. BAI!TKE, Die Aufnahme

lautet, was die Germanen um die Zeitwende sich dsrunter vot:ge8~lIt haben ". du Christentums dUTCh die Germanen. Die Welt als Geschichte 9 (1:943),

(HandJmeh", x86 f.). 143-166.

xxx xxxi
 The Ruthwell cross. and the great J rellinge stone are eloquent
, for a good deal at its late nineteenth-century level. The
witnesses of this transition (I).
following account of research on manuscript runes may help
Chronological considerations may also have contributed to
to show how the present situation arose.
this neglect of the manuscript material as such. Whereas the
oldest runic inscriptions were carved in the third or even in
the second century A. D., the earliest date for the TUnica manu- .
scripta is the late eighth century. But we must not forget that the REsEARCH ON RUNICA MANUSCRIPTA
vast majority of all runic inscriptions are also later than the
eighth century. Our information about the earliest scholarly work on Tunica

Second, the authors we reviewed use only a fragment of the manuscripta is probably not complete (I). The work of some

manuscript material. The situation is hardly different if we early Scandinavi~ runologists, like Johannes Thomas Bureus,

consider the whole of runic studies in the same period, and has been made the object of careful study (2), but scholars in

not in the first place the genetic problem. The names of the England and on the Continent have not been so fortunate.

runes are the only aspect on which the inscriptions give no Yet It is in their works, rather than in those of Danish, Swedish

information. They alone can induce most runologists to refer and Norwegian .runologists, that we have to look for the first

to manuscript material (and vice versa: scholars who discuss appearance of the material which interests us here. I do not

TUnica manuscripta will almost automatically dlrect their attention intend to give a full survey of the use to which manuscript

to the wne-names, and neglect the forms of the runes and runes have been put, still less to sketch a history of runology

their order). This, however, applies only to the period covered in the last four centuries. The coming pages are only meant

by our brief survey. 'Before ca. 1875, when the study of the to show how the study of the TUnica manuscripta reached its

inscriptions had not reached its full development, there was no present state, how the material was gradually collected, and
...."",
such restriction in the use of manuscript material. So we may why a new examination of the whole field is necessary.

expect to find that the study of the TUnica manuscripta is 'still In Johannes Trithemius's Polygraphia (first edition 1508 ) we

find the earliest printed runes which go back to a manuscript

(1) Moreover it is not so simple to uphold a distinction between primary

and secondary usage. Certain inscriptions maY also be tennOO secondary,
e.g. the Falstone inscription (M. D. FORBES-B. DICKINS, TIuJ RII.'thweU and the
BelJICI1St1e Crosses, 34). The following remark, on the distinction between (1) References may be found in R. VON RAUMIlR, Geschichte tier genmmischen
• genuine' and • spurious' inscriptions should be kept in mind: "Del Philologie; H. PAUL, Guchichte tier germanischen PhilDiogie; T. BIBDlIR, Guchichte
lInZgtelig noget uklsre skel melle.m ' a:gte • og • ua:gte , runeindskrifter drages dI!r Germantmforschung. G. JAFFE'S Geschichte tier R~orschung, which
for det meste 80m en temmelig skarp tidsgnense,mel!em middelalder og nyere discusses runic studies from the 16th to the 18th century, is far from satisfactory
tid. Runeskrift era fer 1.500 betragtes regeima:ssigt som a:gte runemindes­ for our purpose. It examines those studies mainly from the point of view
ma:rker, der staal' i direkte forhold til a:ldr~ tiders traditi~nelle' runeak.ik,mooens of general principles, and pays little attention to the actusl approach, to the
indskrifter fra tiden efter c. 1.500 ligesas regelma:ssigt betragtes som en art material which was available and to the way in which it was used. Moreover
kuries renaissancebeva:gelse, der i ferste ra:kke skylder la:rd interesse for JAPri's text is sometimes so confused, that one can hardly disentangle his
criticism from the views of the authors he studies.
fortidens skrift ain tilblivelse. Ssmtidig har man dog holdt en der asben
. til begge sider: de enkelte bevarede' middelalderlige ' runehaandskrifter (2) His RunakamJlanu LtBrupdn (1.599) was reprinted by G. KLBMMING,
karskterisefes - uden tvivl moo rette·--'. som en bevidst modebeva:gelse, who also devoted several papers to his runological work (H. ARNTz, Bibli.o­
altaaa 80m en art ua:gte rune.minder frs selve middelalderen,mOOeng; adski11ig graphU, nos. 1685 If.); Cf. also E. SVJ\RDSTROM, JDhannes BlI.reus' arbete:n om
folkelig runeskrift ers nyere tid, i Cyprianus-bIilgex 0.1. -,mOO~ere tvivlsom stlmska nminskrifter. Stockholm, 1936 (Kgl. Akad. Handlingsr 4 2 ,3).
ret - stiltiende er blevet betrsgtet som folkelig tradition med..rod'i niiddel­ the Continental scholars BONAVENTURA VULCANIUS would no doubt
alderen" (A. BIEKSTED, Vorli' yngste ,unejndskrifter, ,I II). deserve a similar treatment; cf. his De litem et lingua Getarum sive Gothorum
and P. H. MouttlY8l!N, Eenru:nendicht. T. N. T. L. 27 (1908), 32-37.
xxxii
xxxiii
 source (I). In his sixth book Trithemius prints two alphabets alphabet' OCcurs in a collection of calli graphical
~I'''''''' which he connects with the Normans and Bede. One is given
by the Swiss schoolmaster Urbanus Wyss (1549) (I).
under the heading De alphabeto et modo scribendi Nortmannorum
seems to have been the first to identify these char­
mirabili eo tempore quo GaUiam sua incurnone uastabant. The
as runes. On one Of the plates in his Thesaurus he
name of Bede, who is mentioned as a source in this connexion,
such an alphabet with a remarkable peculiarity: the
would be sufficient to make Trithemius's reconstruction sus­ of the letters are in Greek script. Unfortunately he
picious. Actually this alphabet has nothing to do with runes (2). at give any details on hiS source (2). Wyss and Ambrosius
The origin of the second' Norman' alphabet, however, must had their alphabet from Mandeville (p. 27 8).
be runic, even if we may be sure that Bede is again mentioned first author of real importance for the study of the runica

erroneously : Praeter illud quod in principio huius libri sexti

was Wolfgang Lazius, the Austrian physician,

posuimus alphabetum Nortmannorum, etiam quod sequitur apud and folklore student. The ninth book of his work

Bedam axaratum (sic) inuenimus, quod litem commendamus.

,cenhum aliquot migrationibus (1555) is devoted to the Marco-

The alphabet itself is closely related to that of the De inventione By a curious confusion, Lazius connects them with

group. Mercians; these appear as Martii, gens Martiorum; their

Not all of those early editors were aware of the real character Penda is rex Martiorum; their name is related to that of

of their material. What they printed as runes may Occasionally

hoc est Teutonici Meni'i (3). To prove that the

have had a totally different origin, whilst, on the other hand,

or Danes are descendants of these Marcomanrii, he

some rumc material had been circulated under such fancy a brief text in which the runes are described as the letters

names as • Arabic' or 'Syriac'. Thus, in the lntroductio in by the Marcomanni, quos nos Nordmannos vocamus. Then

Chaldaicam /inguam (1539) by Theseus Ambrosius, a canon a runic alphabet with the names of the runes. It .7~'.
and grammarian of Padua, we find an alphabetum Saracenorum off after p, either because the rest was missing in the
(fol. 204 v). The names of the letters are borrowed from the used by Lazius, or because it had become· illegible
, Istrian • or • Thracian • alphabet of Aethicus Ister (cf. p. 276) 297 f.). .
but the letters themselves are runes (3). A closely related complete version of Lazius's alphabet was discovered
(I)H. ARNTz, Bibliographie, no. 2800 lists a work by GUILLAUME POSTEL,
Li:n.guanmI dtl.Odecim. clrmacterihus differentium alphabetum introdm:tio, ac gotisches Alphabet von IS39. Z. f. d. Ph. 32 (1900), 140 f.;

legendi modw longe facilimus••• Guillielmi POltelli Barentonii diligentia. ParUiis GEBHARDT, Ein angeblic:h gotisches Alphabet. Ibid" 564-566; .F. BURG,

apud Dionysius Lescuier 1538. This work, however, contains no runic alphabet. del Theseus Ambrosius. Z. f. d. Ph. 36 (1904), 124 f. .

Lilnllus valde doctus, elegans, & utilis, muita & uario ·scribtmdarum
Cf. also C. P. BURGER JR, VreenuJe AJphGbetten. Het Boek 18 (1928), 321-334­
(2) TRITlU!MIUS h.imself points out the connexion between this alphabet
genera complectem, fol. NiiijV; facsimile edition by H. KIENZLE,
1927· Another edition of the same work appeared in Antwerp in 1576 :
and the Greek numerals. It is in fact a system of secret writing in which the
KRUITWAGIIN, Laat-Middeleeur.osche Paleograjica. 's Gravenhage 1942,
letters are indicated by Greek numerals: a = ", b = fl, c - y ••• i = tJ, note 3.
k " I = .... m = ,p ... t = ,8, v 1(, :It = r«<, y = Kfl, z = KY, W = 1(8 (I).

Reprinted by G. HICKES, ThesatnuS, Grammatic:a Franco-Theotisca, 3, Tabella; Granmuzticae Isltmdicae Rudimenta, Tabella HI. Collections of alphabets.

may have led to this confusion will be mentioned in connexion with

CC. also the cryptogram in BrullSeis MS. 9565-9566, mentioned on p. 97. MS. 14436 and Arsenal MS. n69.

(J) On fol. 206' AMBROSIUS also gives a • Gothic' alphabet. It goes back
King Offa of Mercia is said to have founded the town of Offenburg in
to the shorter Norse fupQl'k and had been communicated to h.im by the
(first mentioned in the twelfth century,. This is probably not the only
secretary of JOHANNES MAGNUS, the last Catholic bishop of Uppsala. It is
where Offa I, legendary king of the Continental Angles (latter half of the
closely related to that in JOHANNES MAGNUS'S Gotlwrom Sveonu.mque historio
(1554) and in his brother OLAUB MAGNUS'S Historia de gtmtihru septtmtrionalibu.s centul'y?) is mixed up with Offa II, king of the Mercians (757-796), cf.
Kr..mallll, Beowulf, p. 197; LAzIUS'S toponymiesJ fantasy is of course entirely
(1555), cC. O. VON F'RlI!IiEN, Runonra, 240 f. (with facaimile); E. NESTLE, Ein the style of his time.
xxxiv
xxxv
'some fifty years later by Melchior Goldast (cf. p. 303 if.). He
.found it part of a treatise on the history of the alphabet, to Danicarum e variis manuscriptis desumptae IS ab amicis

tmunicatae; it contains into al. an alphabet e codici M.S.

which he gave the not very appropiate tide De inventione

linguarum ab Hebraea usque ad Theodiscam, IS now antiquis and
~liotkecae Florentinae ad D. Laurentium, qui alphabeta diver­
which he edited in his Alamannicarum rerum scriptores aliquot linguarum sexaginta sex continet. In fact Wormius does

make a distinction between manuscript and. epigraphical

vetusti (1606). He attributed the treatise to Hrabanus Maurus,

He even prints the text of the Bjarkamal and of Egill's

on what grounds is unknown. The paragraph on the runes

.£."lIl",m. with runes. On the whole, and for obvious reasons,

is almost identical with that in Lazius; there can be no doubt

that the manuscripts they used were closely related. Goldast's work is of greater importance for the study of the

text was reprinted in Colvenerius's edition of Hrabanus Mau­

TUnica manuscripta. The same holds for the material

rus's Opera (1626) and again by Migne in his Patrologia Latina. e. g. by Ion Olafsson, whose Runologia (firs~ written

173 2, but revised in 1752) contains some excerpts from

Whilst historians and philologists were digging up such bits

sources, into al. from a Flatey manuscript (1). Some of

of runic material on the Continent (and also in England (I»,

material, however, will have to be used for solving problems

led by chance rather than by systematical research, and without present study (cf. p. 165 11'.).

knowing very well what to do with this material, Scandinavian

the activity of George Hickes and Humphrey Wanley we

scholars were laying the foundations for an autonomous study

one of the most important collections of runica manuscripta

of the runes. In their immediate surroundings they found

made (2). The first Anglo-Saxon scholars had been

numerous runic inscription8'; newly discovered inscriptions

It"'rested mainly in religious, legal and historical texts; now

were being communicated to them all the time; they could rely
field was widened to include all antiquities. Runic inscrip­

on a tradition, which, by however tortuous ways, went back to

,in England were few, and it would take a long time before ..~.
the runic age. In some' remote regions runes were still being"
could be interpreted in a satisfactory way. Therefore it

used; antiquarian interests led even to a sporadic revival (2).

hardly surprising that the grammarian Hickes turned to'

The runic inscriptions, however, were not easily deciphered;

nuscript material which, as far as it was preserved in fuporc

therefore runologists were not averse to include manuscript

alphabet order (as happened in most cases), could be easily

runes in their material.

In his task he recejved invaluable assistance from
We find a good survey of the material collected down to 1636
, whose catalogue of Old and Middle English texts in
in Olaus Wormius's Danica literatura antiquissima. In his
British Isles has not been superseded in two and a half
of Auctores quorum testimoniis ope usi sumus we meet Trithemius,
Wanley recorded fourteen manuscripts containing
Lazius and Goldast, but also Arild Huitfeld, who, discovered;
Not all items listed by him are' equally valuable, but
runes in the Royal Danish archives; Arngrim Jonsson (Am.
all the important items known nowadays, are there.
grimus Jonas), who collected some items ex ... membranis
two very remarkable bits, Wanleyand Hickes are our only
vetustissimis, etc. Wormius has a whole chapter on the
lthorities : for the runic collection in Cotton MS. Galba A 2,
(I) Little material on this activity in England seems to be available, to
for the Rune Poem in Cotton MS. Otho B 10; both manu­
from E. N. ADAMS, Old E1I/llUh Scholarship i1l E1I/llmul from I566-rBoo. '
are now lost, and we know them .only through Hickes's
Haven, etc. 1917 (Yale Studies in English SS). In the manuscript notes
,
sixteenth and seventeenth century historians, students of ancient law,

more data could no doubt be gathered.

(I) I. LINDQUIST, En skittnhandskri/t fran Flatey. Cf. also P. G. ThoRSEN,
(z) Cf. e.g. C. 8. Pl!':nmsBN, Btn damke Litt.eratu:r fra FolketumtJ:ringstiJtl B'I'IIIJ, and especially A. BJEKSnID, Islands Rfmei:ndskri/ter, 220.
(z) G. HICXI!S, Thesaunu.
irultil Holberg. 561 f.; O. VON FRlBSBN, Runoma. 240ft'. H. WANLIlT, Catalogus.
xxxvi
xxxvii
facsimiles. Hickes also facsimiled other material and repro­ fuporcs and runic alphabets from Vienna MSS. 64 (now 1761),
duced some from Trithemius, Lazius, Goldast and others. 277 (75 1) and 828 (1609); St. Gall MSS. 270 and 878, Munich
Apart from this, the eighteenth century would bring few MS. 19410 (his" Tegernsee "), etc. He was not satisfied with
additions to the stock of TUnica manuscripta. Only Eccardus's copying the information given by his predecessors. In a number
edition of a page with alphabets (two or three of which are of cases he offered valuable suggestions, e. g. on the manu­
runic) from Munich MS. 14436, and the runes from Paris MS. scripts which Lazius and Goldast may have used (I). Whenever
Grec 375 printed by Montfaucon must be mentioned here (I). the possibility existed, he checked his sources or had them
The systematic search for historical and literary texts begun checked by others. Thus his brother Jacob made a copy of
in. the early nineteenth century led to the discovery of a truly the three Vienna manuscripts. He had some notion of the
astounding amount of rumea manuscripta. At the same time filiation of runic knowledge on the Continent: he stated that
the impetus given to Germanic studies by Rasmus Rask, Jacob the runes of St. Gall MS. 270 are of English origin, and so are
Grimm and others directed the attention of Continental scholars those of St. Gall MS. 878; the latter, however, were copied by
to runology. But hardly any inscriptions had as yet been an A1emannic scribe who understood little, if anything· at all,
found on the Continent, and so scholars there turned mostly of what he was writing (2). To his text Grimm added a set
to.the literary sources (Tacitus, Venantius Fortunatus, Egill) of eleven plates, the first three of which contain TUnica manu­
and to manuscript material. . Moreover the study of the epigraph­ scripta in facsimile.
ical runes was still unsettled: the shorter Norse fUPllrk was Grimm's work came at the right moment to direct the attention
usually believed to be the original alphabet. This made it of philologists and historians, who were searching libraries and
difficult for Continental scholars to find their way among the archives, to manuscript runes which they might otherwise have
often contradictory opinions. overlooked. New discoveries were communicated to Grimm; :t~'"

As a first token of this renewed interest there is Wilhelm jf possible he secured facsimiles through his correspondents.
Grimm's remarkable work Ueber deutsehe Runen (1821). On Pertz and Kopitar found a fourth manuscript in Vienna, " cod.
a variety of questions Grinun held to the opinions of his time : membr. Salisb. no. 140 oHm Salisb. LXXI", the so-called
the shorter fUPlilrk he considered older than the original fupark Alcuin manuscript (Nationalbibliothek MS. 795). Pertz also sent
with twenty-four runes; the Kylver stone had' Anglo-Saxon a copy of Vatican MS. Regin. 338, Graff one of Paris MS.
runes', etc. But the sections on manuscript runes still deserve 5239, Docen one of Munich MS. 19410. In St. Gall von Arx
(jur attention : they contain suggestions by which later scholars made new attempts to read part of MS. 878, etc. Only seven
might have profited (2). years after his first work appeared Grimm published Zur
Besides the manuscript runes which he knew through Lazius, Litteratur tier Runen, with a number ,of new items, corrections
Goldast, Montfaucon, Wormius and Rask, Grimm discussed to the older material, and six facsimiles.
One is struck by the relative degree of finality Grimm reached.
(1) J. G. EccARDus, De origine Germanorum, tab. XIV, 188. His exposition is of course influenced by the opinions of his
B. DE MOm'FAUCON, Palaeographia grasca, 2<}2 f. time, but the descriptive portions of his works have hardly
(2) I believe W. GRIMM'S significance as a runologist lies here rather than
been improved upon during the next century or more. One
in· his vindication of the existence of Gernum runes, which H. ARNTZ, Wilhelm
Grimm und die deuuchen Runen, praises so highly. When GRIMM wrote his has only to read his discussion of Vienna MS. 795 to. realize
works, none of the Gennan inscriptions edited by H. ARNTz and H. Zmss how carefully he probed the evidence before him. His material
(Runendenkmaler) had been discovered; and not one of the runic alphabets
found in Gennan manuscripts seems to have anything to do with runic writing (I) W. GRIMM. Ueber deuuche Runen. 80; 8x footnote.
as known from Gennan inscriptions. (2) W. GRIMM, Ueber deutsche Runen, 120 fr., 147.

:xxxviii XXXIX
was reprinted some forty years later by George Stephens, and Runen-Fudark (I8S7), a work which most students in

since the latter's collection has become the main source of most field seem to have overlooked. Yet it contains interesting

modern authors, we may safely say that the study of an important on a number of Munich codices (MSS. Lat. 6250.,

part of the material had reached its present level more than a 6291, 13067). Some authors discussed manuscript

century ago. in a more general way, e. g. A. Kirchhoff and J. Za­

J. M. Kemble's On Anglo-Saxon Runes (1840) is closely (I). The most important contribution of this sort seems

related to Grimm's works. Kemble wanted to make Grimm's R. von Liliencron's criticism of Kirchhoff's work (2).

results accessible to his countrymen; at the same tiJ;ne he hoped Liliencron is the first to stress the secondary character of

to furnish Grimm with new matter for a second edition of his manuscript runes. He also makes a sharp distinction

book. The importance of Kemble's contribution lies rather in fuparks and runic alphabets. He gives a list of the

the field of the English inscriptions than in that of the manu­ and alphabets edited down to his day, to which he

script runes (I). To the alphabets collected by Grimm he only some sound remarks on the value of this material.

added tha. of Harley MS. 3017. But he also extended the points out the danger of ignoring the relationship between

notion of runica manuscripta to include non-alphabetic material. versions of the same source. This can only give a

Grimm had concentrated on fuparks l!lld runic alphabets. idea of the amount of evidence actually available~ That

Only incidentally did he refer to the use of a runic symbol in right in stressing this point appears from his own survey.

the Wessobrunn Prayer, or to. the signature of the scribe Ercren­ comparative purposes Hickes had arranged some fuporcs

frit. Kemble examined this material systematically, and this order Qf the Latin alphabet (3); von Liliencron believed

enabled him to decipher Cynewulf's runic signatures in E/fne, runic alphabets were also found in the manuscripts (but
Y"~.
Andreas and Christ. rejected them becau,se their evidence had no value). Errors
Kemble's appeal to his, countrymen was not very successful, this kind happened 'more often than one would expect.
at least not immediately. Only a quarter of a century later one in press-marks gave rise to fictitious doublets, as e. g.

of his countrymen would contribute materially to the study of MSS. Titus D 18 and Tiberius D 18 (4). Very few

the TUnica manuscripta, but that contribution was to be decisive. were examined more than once, and any mistake made

Meanwhile a long series of editions of manuscript runes set first editor had all chances to find its way into later

in, each devoted to, one or more, newly discovered items, F. J. Thus a misprint in an edition of 1866 can still
Mone printed runes from Brussels MS. 93 11 -93 1 9 (p. 63 ff.); in 19# (5).

C. Greith ,from Vatican MS. Regin. 338 .(p. 237 ff.), which only edition of all fuparks and runic alphabets is that

Pertz had already communicated to Grimm; G. Phillips from Stephens. He took up the, task outlined by Kemble

Salzburg MS. St. Peter a. IX. 32 (then X. 28; p. 113 fl.); H. tireless zeaL He' planned a complete edition of all runic

Hattemer from several St. Gall manuscripts (pp. 74, 90. etc.); ions. and by way of introduction he reprinted all fuparks

Sir Thomas Phillipps from a manuscript in his possession

A. KIRCHHOFF, Das gothische Runeno1phabet.
(p. 227 fl.); H. F. Massmann from Vienna and Rome manu­ J. ZACHI!R, Das gothische Alphabet.
scripts (pp. 106 ff. etc.); K. Bartsch from Heidelberg MS. R. VON LILlIlNCRON, Zur Runenlehre (with K. MOu.ENHow).
Salem 9.39 (p. 30.5 ff.); F. Losch from Berne, MS. 207 (p. 174 ff.), E.g. that of Cotton MS. Domitial] A 9: G. HICKES, Thesaurus, Gram­
islandicOil Rudimmta, Tab. II, 3.

etc. A somewhat more extensive collection, is found in F. J.

G. STBPHENB, Monuments I, 107, 112 f.
(I) B. DICKINS, J. M. Kemble and Old English Sclwlo:nhip. Proceedings G. STBPHENB, Monuments I. l~ f. ' 1443 b ' (= Munich MS. lat. 14436),

Brit. Acad. 25 (1939), 5 1- 84. AR:NTz, Runen und Runennamen, 177 f. (footnote). 193.

xl xli
and alphabets edited thus far, giving additions and corrections only source, were (and are still mostly) believed to be
in the later volumes of his work (I). But his enthusiasm could connected with the rune-.names (hence " germanische
not make up for his lack of philological training, and his ideo­ iennamen" in the title) (1). The analysis of the manu­
logical approach did not improve the situation (2). Therefore evidence is very minute, and most promising for the
his conunentaries have now o~ly historical value. His lists further investigations. Unfortunately von Grien­
of runes and rune-names, however, have become the main does not seem to have noticed Sickel's fundamental
collection of material for students of runica manuscripta. , and this neglect necessarily affects the value of his work
We may end the survey of this older period as characterized p. 54 ff.).

on p. xxxviii with P. G. Thorsen's Om RU11.ernes Brug til Shrift second article contains a discussion of the 0 N. rune­

udenfor det monume.ntale. But for our subject this work with (2). Here von Grienberger depends entirely on repro­

the promising title proves rather disappointing. It was intended made by others : for the Abecedarium Nordmannicum
as an introduction to the facsimile edition of the Codex Runicus Hattemer, Grimm and Piper; for Leyden MS. Voss. Lat.

(or Wonnianus, Copenhagen MS. AM. 28, 80 ). The author on a drawing by H. Kern, etc. He hardly discusses the

hardly ever leaves his northern sphere; moreover he depends of the runes in the manuscripts, and still less their cultural

more on literary sources than on actual manuscript runes. He round. As the title indicates, all attention is focussed

contends that runes were used extensively for all sorts of com­ the names of the runes. Only R receives a more thorough
munications and notations, going so far as to declare the epi­ (3). For the rest the genetic point of view remains
graphical usage secondary. Coming three years after the first : the discussion of the material is immediately followed
cast of Wimmer's fundamental work (3), Thorsen's attempt to U Zur ableitung der runen" (4). ...
~, "
direct the attention to the non-epigraphic runes was bound to third paper is by far the most interesting for our subject.
be unsuccessful. From now on general studies will concentl having examined the Gothic and the Old Norse material,
more and more on the rune-names, as our next author clearly Grienberger now turns to the Old English fuporcs and the
demonstrates. . Hrabanic alphabets (5). The first group is repre­
To T. von Grienberger we owe several by seven items, of which two are known only through
with manuscript runes as a starting-point. facsimiles. von Grienberger does not seem to have
who examined that material as a whole. But his attention that two of his fuporcs must at least partly go back to
went almost exclusively to the names of the runes. Only :ommon source (St. Gall MS. 270 and Vatican MS. Urbin.
occasionally did he discuss one or other type of rune. The ; a third manuscript belonging to the same group had been
value of his work is also limited by the fact that, with by E. Steinmeyer some years before (Trier MS.
exceptions, he did not examine the sources himself, but si m.I3) and a fourth had been edited (with a facsimile of the
built on the foundations laid by others. . material) more than thirty years before (Salzburg MS. a
His first study is devoted to the Gothic material in Vienna 32). Moreover his nos. I and 2 are so similar, that one
MS. 795; the names of the Gothic letters, for which this a priori suppose them to be related. The rune-names
T. VON GIURNBIlRCIlR, Die gennanischen runennamen. I. Die gothischen

G. STEPHENS, Monuments I, 100 if.; II, 82.9 if., III, I if.

(I) ibennamen.

Cf. L. F. A. WIMMER's criticism: De 8lldste _duke Runeindshrijter.'

(2.) T. VON GRI~CIlR, Beitriige zur runenlehre, tOI if.
Med s81rligt Hemyn til Prof. Stephens' V_h, etc. Aarooger 2. (1867), T. VON GRIllNBIlRCIllt, Beitriige Z1lr runenlehre, Il2. f.
(3) L. F. A. WIMMER, RtmeskriJtens Oprindehe og Udvikling i
T. VON GRIENBIlRCIllt, Beitriige zur runmlehre, 114 if.
Aarooger 9 (1874), 1-2.70.
T. VON GRIllNBIlltGllR, Die angelsiichsUcMn runenreihen, Iff.

xlii xliii
are again the most important, if not the only criterion, than the runes, which they had to imitate stroke
classifying the manuscripts; the main purpose seems again to stroke. Trifling details in the form of one or other rune
lie in the reconstruction of the original forms of those names; of greater value than a similarity of names which may
runic forms are discussed only occasionally. With these result from independent adaptations. Insufficient criti­
limitations in mind, however, we can still use von Grienberger's also created fictitious manuscripts and perpetuated rea­
study for a first general orientation. which would not have stood a renewed examination of
The way in which the runic alphabets are treated is less The blame does not attach only or mainly
satisfactory (I). Doublets are included without the slightest Grienberger. The same objections may be raised
suspicion of their being identical (Cotton MSS. Titus D 18 and most authors who made comparative studies of manu­
Tiberius D 18). von Grienberger lists 15 alphabets. Mter runes.
comparing the names of the runes he concludes that tlley go a matter of bct von Grienberger's papers ushered in a
back to two" urredactionen ", one represented by six alphabets, of stagnation in the study of the Tunica manusmpta.
the other by three. The remaining five'would be intermediate first half of the twentieth century will bring in no more
forms. Actually there are seven in the first group, whilst the half a dozen new discoveries. The general .discussion
three which von Grienberger also believes to belong together be restricted to 'an amplification of the results' obtained
are in all probability unrelated. The transmission of runic far. It will hardly develop by' a more minute study of
lore to the Continent is reduced to a very simple formula: sources.
Bede > Alcum > Hrabanus Maurus. From now on we shall progress could still be made was demonstrated by
meet this formula again and again; some authors will even try .Hempl, although he worked on a very limited material.
to connect all Continental runica manuscripta with this channel. critical analysis of the fuporc to which the OE. Rune Poem
Since von Grienberg~r's studies were to become the starting attached is a model of careful weighing of the evidence.
point for most work in Tunica manuscripta during the next fifty DE. poem on the fuporc has come down to us only as a
years, it may be worth while to examine the factors which in Hickes's Thesaurus, the manuscript having perished
influenced the value of his work in a negative way. The manu­ fire of the Cottonian library. Hempl suspected that
script material was studied only, or mainly, for the evidence on facsimile gave more than the original. He was actually
the rune-names which it was alone to provide. This caused . to prove that part of Hickes's facsimile had been borrowed
all other aspects to be neglected. The criticism of the Cotton MS. Domitian A 9 {I}. For the first time
was mostly insufficient. The forms of the runes, their peculi­ I'Mnt runes were subjected to ,a detailed investigation.
arities, the manuscript environment were ignored. Yet they
would have afforded a broader and on the whole safer' .
for the classificatio ll of the material than lists of rune-names. G. H:!iMn., Hickes', Additions, 141 ; " The Runic Poem can be appealed
evidence on runic matters only in regard of the right-hand forms (and
By referring only to the latter, some important runic alphabe most of the names) of those that have corresponding verses in the
(and fuparks) without rune-names we,re completely disregarde The remaining runes and the phonological spellings added by Hickes
Moreover: the rune-names are no doubt the most vulnerable _ only so much value as they have in the Cot. Domit. manuscript. For the

reasons, the alphabet CODStructed by Hickes from the runes given by

of the manuscript tradition: scribes were easily tempted

as appearing in the Runic' Poem, and publi8h~ in his Thesaurus (...• )

adapt them to their native dialect, or at least to copy them no greater value. In other words, I have removed from the Runic Poem
of the rubbiah that became attached to it in the process of its trans­
to us, and which has been preserved, with more or less care, by those
(I) T. VON GRIBNlIERGER, D!s angelsachsiscJum nmenreihen, 23 ff. have edited the poem".
xliv
xlv
The conclusions of Hempl's study-should have been a warning 9 (I), and C. Selmer tried to explain a runic inscription

to scholars using this material. It might also have been an in Munich MS. 13067 (2).

encouraging hint to continue work in this direction, the only Harder made a bold attempt to solve a number of difficult
which promised lasting results. But we can only regret that of rune-names by a universal rule (3). In a whole series
Hempl's example was hardly followed by other scholars. The Mnuscripts the rune-names appear in forms which differ
Old-English Futhorcs and Alphabets, which Hempl announced or less from the original OE. forms. These ~orms are
in hif; paper, seem never to have been completed. explained as due to misreading, or as reflecting the
In the study of the runica manuscnpta the runic poems have native dialect. Harder, however, believed that these
mostly had Ii privileged position. Their texts were repeatedly were conscious and methodical distortions invented by
examined with care, especially the OE. Rune Poem and the monks who wanted to amuse themselves at the expense
Abecedarium Nordmannicum. W. Grimm was the first to edit readers. He removed the letters he considered super­
two of these poems together (I). An four were printed, with and out of them he reconstructed short messages to the
introductions and full notes by B. Dickins (2). : a warning, the name of the scribe, etc., e. g. ihh iuuih
Some efforts were made to extend Hempl's criticism to other " ich mage euch necken ". Harder has failed to convince
materials, but only on a small scale. C. L. Wrenn endeavoured scholars, and for obvious reasons. He started from
to ascertain the real value of the OE. material (3). He was the runic character of which was more than doubtful :
obviously right when he pointed out that most runologists Norma (?) alphabet in Vatican MS. Regin. 338, and Nem­
ignored this aspect of the matter. To the general acceptance alphabet in Cotton MS. Titus D 18. By including these
of Stephens'S collection as a basis for further work in this in his collection, Stephens hardly meant to pass them
field, Wrenn opposed a sound ifrather exaggerated scepticism. as runes (4). Of course such alphabets lent themselves l'~,
He came to the conclusion that Cotton MS. Domitian A 9 was well to Harder's procedure and yielded appropriate warn­

the best authority, but .. a very poor and uncertain best"; to the reader. But Harder felt also obliged to • correct'

which he dated as late as the twelfth century. C. E. Wright, pedectly justifiable forms as aac, iis, raad, (in another

however, showed that Wrenn's critical evaluation was not . of the same Vatican manuscript), birca, caon, feu, naut,
quite adequate : his dating and analysis of Domitian A 9 had \J.VJ.unich MS. Lat. 14436), which are extremely interesting
to be corrected on a number of important points (4). Wright'S for OE. and ON. phonology and for runology. Moreover
own brief account of the manuscript is a model of a thorough; pattern which Harder reconstructed was not uniform. In
careful and competent analysis of manuscript runic material. case the form hagale was corrected to hagel, in another to
In the meantime some new material had. been edited, but it Of course this opens the way to the most arbitrary and
received little attention. T. Langin edited three runic extracts interpretations.
from Reichenau manuscripts (now in Karlsruhe), G.C. van study could have served to demonstrate one point,
Langenhove gave a new edition of the fuporc in Brussels MS. this : what is usually grouped together as ' runic alphabets '
T. LXNGlN, Altalemannische SprachqueUen, 699 if.

. G. C. VAN LANGENHOVE, Brussehche Runen.

C. SELMER, The Runic lmcriptitm oj Code:c.Latinus Monacenn. :13067.

(I) W. GRIMM, Ueber cieu.uche Runen, ZI7 fr., Z46 fr. P.M.L.A. 53 (t938), 645-655.
(z) B. DICKlNS, Runic and Heroic Poems, 1-34. H. HARDER, Zu:r Frage tier hrabaniJchen Alphabete.
(3) C. L. WRENN, Late Old English Rune-Name•• G. STEPaENS. Mtmuments I. 10']. states that Vatican MS. Regin. 338

(4) C. E. WRIGHT, A Postscript. to the cipher or fanciful class".

xlvi xlvii
or ' Hrabanic alphabets' is in fact a medley, the elements . in OHG. glosses and literature, in which the author
which belong to various places and different periods. no small part. But some doubt is permitted as to
G. Baesecke has repeatedly undertaken to describe the picture which he draws without hesitation really
manuscript runes in their relation to their' cultural background the facts. One cannot avoid feeling that his recon­
In one paper he sets out from the Abecedarium Nordmannicum is at times quite remote from the manuscript evi­
to explore the whole field of Tunica manuscripta (I). ). The progress lies much more in the general picture
the poem of the halo with which previous generations the discussion of the material itself, or in the integration
surrounded it. They had usually regarded it as the into the picture. The evaluation of the Abece­
.remarkable of all four runic poems, as the one which NOTdmannicum is a laudable exception. Baesecke
closest to the old Germanic spirit. Baesecke does not special credit for having realized that the study of the
to call it doggerel. He believes it to be adapted from a runes not only interests runologists, but also
original for use in ninth century German schools. In of cultural history. On the whole, however, he
St. Gall codex where it is preserved, it comes after grammaucw too boldly. He does not doubt Hrabanus Maurus's
extracts from Isidore's Etymologiae; the runes in Brussels of De inventione litterarum, although our only author­
no. 155 would be in a similar situation (2). Baesecke assumd is a not altogether trustworthy Goldast. He feels quite
that runic studies played some part in eighth and ninth this same Hrabanus Maurus promoted the study of
monastic schools. He elaborates the sketch of the transmission his schools, and this much in the same frame of mind
of runic lore given by von Grienberger. He tries to ixteenth century polyhistoJ.: or even a nineteenth century
the foundations laid by his predecessor: he classifies the . And yet' Hrabanus had no " offene Teilnahme an
~~,
uscripts with runic alphabets according to the Welt", and his eulogist Hruodolf mentions no
text. He distinguishes a longer and a shorter version of activity of his master except theology (2). My main
De inventione; he even eliminates one of von Grienberg~. against the general trend here embodied by Baesecke
doublets (Cotton MS. Tiberius D 18). But he connects : it looses sight of the relative importance of such elements
alphabets with one channel of transmission. Thus Munich alphabets and short scraps of German with on the
194 10 and Berne MS. 207 are still said to belong to the moderate literary pretentions (3). One has only to
inventione tradition. All runic alphabets that have come the leaves of Colvenerius's edition of Hrabanus's
his notice would be due to one of two scholars, the first works to understand how small a part the compiling
whom retained many OE. forms, whilst the other freely ~u-German glosses and of a short tract on the alphabet
the rune-names to OHG. phonology. The latter played in the activity of this· Carolingian scholar
whose purpose it was to create a runic alphabet fit to be
following quotation is typical : " Dasz der kiirzere Text der Runen­

in German schools, would have been no other than HrabanUi aus aga. Geleh.r$amkeit lltamme, vermuteten wir BUS dem nmstafas

Maurus. Gotenzusatz. Aus den Handschriften liszt es sich nicht erweisen :

Baesecke's reconstruction has obviously benefited by den am reinsten angelsichsischen FassuDgen (I3-1S) fehlt diese

Sicher aber ist, dasz me mit ihrem nmstafas nicht zu der fort­

Deutschheit von s, 6 und 8 gepaszt hltte" (G. BAI!SECKE, Das

(1) G. BAESECIC8, Das Abecedarium Nordmamlictnn. Nortlttunmictnn. 86: the figures lite the- numbers of the alphabets

(z) This is one of severa1 inaccurs£ies in BABSI!CKll's paper: Brussels GRIENBBRGER'S survey).

93 II -93 I 9 (BAIlSl!CKE quotes the old number in MARCHAL'S Intmltaire) BABSECKE, Die K.:orlisc1le RenaUrante, 170.

no g~tical extracts from Isidore; moreover the runes are written on most striking instances are probably to be found in his Vor- wrd

leaf which did not originally belong to the manuscript (p. 66).

xlviii xlix
(provided the tract in which the runes are mentioned Baesecke also attaches great importance (I), may hold

really be attributed to him). theological, exegetical, grammatical and mathematical

, The discussion of two recent additions to the stock of Even then they are only general statements, which

manuscripta may help us to characterize the level attained and cannot, apply to every single item of a scholar's

far. In 1941 Miss E. Raucq edited the runic material equipment. The transmission of such extra-curric­
Brussels MS. 9565-9566 for the first time, and compared as the runes must evidently have been, may have

with that in the better known St. Gall MS. 270 (I). Her far more complicated or obscure routes. Krallse tends

shows a real effort to discover the origin of this manuscri] all Continental runic alphabets with Hrabanus

tradition, but the attempt was bound to fall short of its even St. Gall MS. 270 (2). He seems hardly to have

For one thing, Miss Raucq depended almost entirely on the possibility of other scholars, besides Alcuin,

names of the runes, although she had at her disposal ~ runes to the Continent, and of other scholars, besides

text besides the fuporc. In the manuscript she changing the fuporc order to that of the Latin al­

some important details, which might have helped her to Such a trend can only lead to a fallacious picture of

its origin and elucidate its cultural background. Finally. happened.

was not aware of the existence of three more, closely we Come to two important general studies on runica

versions, two of which had been edited before. published during the last war. They differ widely

The latest discovery which led to a general discussion aims and in their ways of dealing with the material,

manuscript runes is that of a runic alphabet in a Ll<aefore it is good to examine their respective preferences

belonging to the Munich antiquary A. Weinmiiller (2).

Plassmann examined this fragment, and also a Paris maIIU:>\;U Blomfield is in the first place concerned with the .~
containing a related version (MS. 5239); W. Krause relationship between the Gothic alphabet and the

runological discussion covering other material as well. (3)· The names of the Gothic letters found in Vienna

of the two parts is really satisfactory. Plassmann's anat~slS are usually considered as the direct descendants of the

fairly accurate, but the Paris codex is at least a century names of the runes. Miss Blomfield stresses the fun­

than the date (saec. XI ex./XII in.) given by him. He difference between bookhand and runic writing; only

it to be the work of an Anglo-Saxon scribe, on very .. may runes have fouJid their way into manu­

grounds, as we shall see (p. 330). Krause's discussion is Therefore one maya priori Suppose that runes played

related to Baesecke's. Like Baesecke, Krause distinguishes part, if any, in the formation of the Gothic alphabet.

versions of the De in'fJentione, one of which he ascribes to the so-called Gothic rune-names, these "have no claim

banus Maurus. Alcuin himself would have provided the ,r~sent runic nomenclature current among the Goths

of this treatise, for the older version, represented by' They do represent alphabet lore current among some

MS. 5239 and Cotton MS. Titus D 18, would have people in VIn; and their connection with a dialect

in Alcuin's environment. Alcuin may ultimately have

his information to Bede, whose pupil he was (Alcuin was G. BAES£CKE, Die KarlitCM Renaissance, 216.
W. !C:RAUSB, Die Hrabanische Runenreih.e, 187: .. Als Hraban seine
about 735; Bede died in that year). Such spiritual genealu~\ begann, scheint jene Umbildung des angelsachsischen
In em Runen-ABC noch nicht abgeschlossen gewesen zu sein.
hat iich Hraban selbst offenbar urn diese Umsetzung ebenfalls.
(1) KRAuCQ, Brii.ueler Rtmen, 1 if. moglicherweise in mehreren Anslitzen ".
(z) 1. O. PLASSMANN-W. KRAUSE, Die HrabaniscM Runenreihtl. J. BLOMFIBLD, Runes.

Ii
which was then, as far as we know, obsolete makes it probable;
that they are riddled with antiquarian confusions. .As a source material used in that chapter (I). He intends mainly to
of runic nomencla~re they are worthless. Yet there are an account of the history and. the cultural background of
obvious points of contact. If it were possible to analyse~ the manuscript runes. His main concern is expressed in the
overlay of sophistication, some relation might be established a1; the head of his paper, a quotation from Baesecke's
between these names and rune-names" (I). By way of illus-; of the Abecedarium NOTdmannicum (2). Following
tration she examines "the transformation of the runic system he considers the efforts made to adapt OE. runic lore
in the period VIII-XII" (2). She again stresses the distance Germanic speech as the crucial problem. In
between the alphabetic lore found in manuscripts and the orig­ be wants to complete the task where von Grienberger
inal system of runic writing, which she believes to have been Starting from a discussion of Miss Raucq's article on
dominated by religious and magical considerations. MS. 9565-9566 and St. Gall MS. 270, he hopes to lay
Unfortunately this real progress in the general approach is i'foundations for a new study of the OE. rune-lists, whether
not accompanied' by a closer study of the sources themselves. or alphabets. In this introduc~on he examines the
Stephens and von Grienberger remain the principal authorities, and their values in the two related manuscripts. But
and the limitations of their works are also felt here. Of course be COmes to the rune-names he extends his field to include
we could hardly expect the author to discuss all versions in lists of names. He too is aware of the gap between
detail in these few pages--war-time conditions would anyway manuscript material and the genuine runic tradition, and
have prevented her from doing so-but in some case,s her general to classify the alterations which the material may
appreciations tend t{) become too severe, as in the case (I). His survey shows some traces of the prime im­
Cotton MS. Titus D 18. This manuscript is said to be, " H.ARNTz, Runen und Runennamen.
little value since alphabet i is .that of Paris MS. 5239, alpha.u«. " Man spilrt in diesen Schriftwerken noch das Ringen der Mission, die
ii is the Nemnivus series, while the runes of alphabet iii, Gebiete der verwandten Sprachen mit AnkIlingen durchsclilligt und die,
elsewhere found, are of the Norse type" (3). Actually es zum Schreiben kommt, keine systematische Grammatik hat, vielmehr
liber alphabetariw which forms the first part of the Cotton Fonnen und Laute oft vergeblich mit den Zeichen der eigenen
BUche" (G. BA1!lIBCIa!, Das Abecedarium Nordrruinnicum, 89).
is one of the most interesting products of the alphabetic H. ARNTz, RutIIIn una Runennamen, 188: "Von dieser gennanischen
to which Miss Blomfie1d attaches so great importance. sind, me wit sagten, die handscbriftlichen Runenreihen weit entfernt.
lowe to her work a number of valuable suggestions, .by aber fesselnd zu beohachten, wie in ihnen und den Runenliedern bis­
noch em Funke des alten Verstllndnisses aufglimmt. Freilich nur
this stu,dy has undoubtedly benefited. Denn daa Wechselverbiltnis beschrilnkt sich niOOt darauf, rlass die
H. Arntz's work is a bolder undertaking. In the first part dec drei Runenlieder aus guneinsamen Quellen sch6pften. Sondern

Runenhandschriften umfaasen: '

this introduction I have briefly sketched his. general attitll

Alte nordische (bz:w. englische) Runennamen in Nonnalschreibung.

towards the TUnica manwcripta (p. xxx f.). The tenth . Solche Namen in Schreibungen, deren Scbriftbild die Ausspracbe

of the second edition of his Handbuch was devoted to verdeutlichen solI.

names and their meaning (4). In a paper published separatelY Solche Namen in AngJeiOOWlg an Worter, die ihnlich klangen.

Ang!eiOOung emzelner Laute an dill! Lautsystun des Schreibers.

Arntz presents the apparatus criticus and a detailed discussion

Volle Umsetzung in ein lautlioo entsprechendes Wort (ohne Riicksicht

auf die Bedeutung).

(I) J. BLOMFIBLD, RUMs, all f. VoUe Obersetzung in ein Wort von gleicher Bedeutung.
(a) J. BLOM1'II!LD, Ru1fU, a09 If. Ersatz englischer dutch nordi8che Runennarnen und umgekehrt.
(3) J. BLOMFmLD, Runes, au. Missverstlindn.isse und Abschreibfchler.
(4) H. AmrrL, Handbuch l , 16,-a33. . Frage iat also zu meist nicht. ob der Schreiber sich dariiber klar war,
er schrieb. Das dilrfte (im Gegensatz z.B. zu der von' E. Raucq ver­
Iii
!iii
portance he attaches to the "Eindeutschung". Moreover carefully prepared list of possible alterations. His

his approach is not free from apriorism : before the rune-names tretation of some forms is really puzzling. Thus nod, the

have been examined, the types' of alterations which they must the n-rune in Brussels MS. 9565-9566 and St. Gall

show are enumerated. Yet Arntz's list may be a guide for the is said to be neither English nor German (I). If this

interpretation of the forms found in the manuscripts. The the Heliand and the Ludwigslied are not German

relative importance of the different types does not appear, but a conclusion which Arntz will hardly accept. He reduces

they can probably not be placed on one level. Thus the one background of the fupark to reflections of a hypo­

listed last: ' misunderstandings and copyists' errors' plays no Germanic sun cult (2).
doubt a larger part than the second' ON. and OE. rune-names Arntz's paper we must keep in mind the discussion of

in spellings which should make clear their pronunciation' of the runes and of the way in which Continental

the seventh' substitution of OE. for ON. rune-names and handled runk material. The rest should be used with

versa'. Arntz's belief that runes were a regular part of teachmg Of course a number of lacunae are due to the very

programs no doubt influenced his list (I). circumstances in which the paper was written, but I

But for all his good intentions Arntz does not pause to exanUllll whole approach hardly warranted better results. Still

his sources more closely than his predecessors. He also reduces difficult to pass a final judgment on Arntz's work, as the

the manuscripts to more or less appropriate formulae: e. g. supplement to Runen und Runennamen has not yet

runes of Brussels MS. 93II-9319 'have found their way into published (3).

the Isidore tradition' (2). The background of the manuscripts, Schneider's recent study on the rune-names has not

their historical and cultural context is treated in the same in print; I know it only from the references in F.
schematic way. Hypotheses put forward by other scholars, Handbuch der germanischen Philo[ogie(4). From what
there it seems that Schneider, too, has mainly tried to
,
,'­
sometimes in the form of a mere guess, tend to become facts.
In turn they become the foundations for further hypotheses and a world of ideas on the basis of the rune-names.
far-reaching conclusions. of textual criticism was involved the
In his discussion of the rune-names Arntz does away
morphological difficulties in a rather rudimentary way.
feels that his procedure is entirely function of his solution of
the genetic problem : almost immediately he launches upon the SCOPE AND PLAN OF THE PRESENT STUDY

reconstruction of the world of ideas which in his opinion is

reflected by the rune-names. He seems hardly to have used first two sections of this introduction have shown that:

evidence derived from runic material in manuscripts plays

tretenen MeinWlg) nur in ganz seltenen fillen zutreffen. Sondeen wir fragen,
wie die von unsem Schreibem oder ihren Vorlagen vorgenommenen Ver­
a part in iunology at large, even if that part may be a

iinderungen zu erklliren sind" Wld ob wir fiber sie ZUf ursprilnglichen Fonn matter of dispute;

RWlennamen durchstossen k6nnen ".

(1) H. ARNTz, Runen una RunenntmU!n, 2.50; " BrUssel sowohl wie SI.

H. ABNn:, Runen una Runennamen, 2z8.

gehoren zu den libn grmmnatice (?) oder IWri puerorum, d.h. zu den

H. ABNn:, Runen una Runennamen, 187.

H. ABNn:, Runen und Runennamen, 250 [footnote (z)].
rue in den rnittelalterlichen Klo8tem den sog. Extenischulen angehOrten

sich daher an ein weites Publikum wandten. Diese neue Rezeption der

K. SCHNmDllR, DU germanischen Runennameit im Lichte der vergleichenden

ihre Grenzen, Wld ihre Auswirkungen, stellen uns noch vor eine

Sprach· una &ligionswissenschaft. Versuch einer Gesamt­

Beitrag zur indogennanisch·gennanischen Kulturkunde und
ungelllster Fragen ".

(z) H. ARNTz, Runen tmd Run_men, 11)0. Jglonsgeschichte). Diss. Marburg, 19SI. ce. F.STROH, HandIJuch, 501 f.

liv Iv
 (b) research on the manuscript runes has not yet reu::hed the its full development in the ninth centuiy at the latest. We may

·level attained in the study of the runic inscriptions, al­ say that it was becoming a fossil by the end of that century, a

though a considerable amount of material has been dis­ curiosum incapable of further evolution. It found its way into

covered during the last four centuries; manuscripts during a relatively short period. All notations

I
(c) the study of the Tunica manu.scripta is not only a runological that have. come down to us belong to the eighth or the ninth

problem : it n;t.ay also contribute to the history of Mediaeval century, or are derived from originals of that period. The

culture. ON. fuPlilrk of sixteen runes lived on for centuries, and continued

to send out new shoots all the time. New runes were created

We have also found out why the study of the Tunica manuscripta
as late as the fourteenth century. Therefore it will often be

could not fully develop. From the survey in the second section hard to distinguish between original runic tradition and later

. we may conclude that progress is only possible: .

learned or playful elaborations (1). One has only to compare

(a) if the manuscripts themselves are examined more carefully. the terse Abecedarium NOTdmannicum (saec. IX) or the ON.

Special attention must be paid to the origin and date of the runes of Leyden MS. Voss. Lat. 83,40 (saec. X) with J6n

codices, to the relation of the runes to ~e rest of the con­ Olafson's Runologia (1732-1752, but using older material) and

tents, and similar aspects which may. infl.uence the value J. Liljegren's list of secret runic alphabets (2) to realize the

of the material; proportions of that growth. The cultural background of the

(b) if the forms and the values of the runes are studied,· and OE. material may therefore be expected to be more uniform,

not only the names. The material must be viewed without and consequently more easily circumscribed, than that of the

any connexion with the genetic problem or with other ON. material. Moreover the study of the former is more

problems of general runology; urgent, and in a way also more promising, than that of the

latter. As we saw in the first seCtion, the OE. inscriptions are ...~
(c) if the !elationship (or independence) of the manuscripts,
few in· number._ Additional information will consequently be
.and their cultural background,.is established with the help
of internal evidence, rather than on the basis of doubtful
welcome; it ma: y even contribute to the interpretation of some
difficult inscriptions. Compared with the ON. epigraphical
attributions to some or other famous Mediaeval scholar.
material, the ON. manuscript runes area mere drop in the
The manuscripts can be made to tell much more about
ocean. Their importance for the study of cultural history is
the meaning of these stray bits of runic writing than is
rather small and mostly restricted to Scandinavia. Therefore
commonly assumed. At the same time generalizations
I have felt justified to reserve the study of the ON. Tunica
and reconstructions for which we have no really conclusive
manuscripta for a later occasion, the more so as they have
evidence should be avoided, as they tend to hamper further
been examined fairly exhaustively· by P. G. Thorsen, K.
research, rather than promote it.
Kdlund, T; von Grienberger; I. Lindquist and quite recently
Initially I had planned to cpver the whole field of the TUnica by A. Bzksted (3).
manuscripta, and accordingly I began to collect material flOm
Norse as well as from English sources.. But gradually I foun!! * **
out that the material which goes back to the OE. fuporc differs (I) Cf. p. xxxii, note (I).
in no small way from that which is based on the ON. fup!l:rk (2) J. LILJEGREN, Rv:n-LiJra, 53 f.
or fupllCks. (J) P. G. THORSEN's Runemu Brug Wall meant as an introduction to the
facsimile edition of the Codex Wonnianus or Runiau. K. KAull'ID edited the
The OE. fuporc, with its twenty-eight or more runes, reached ON. runic poems (with notes by S. BUOOB) : Etf gamt1Ull-'fUWsk Rtme-Rm1 og
lvi
lvii
 The present work deals with all manuscript runes based on the margins, on fly-leaves and bindings. At times their

the OE. fuporc which came to my notice. For coll~cting this runic character can hardly be recognized or, worse, they

material I had three sources : assume such fancy names as ( Syriac alphabet " • Arabic "

(a) The older literature on the subject. Most items discussed etc. Finally one should not forget that for a number of

here have been edited in one form or other before. Not libraries (and some very important ones) there are no

all these editions have come to runologists' notice in due up-to-date catalogues; for a few none at all.

time; they are scattered in older works and periodicals, (c) Last, but not least, there is the information which several
some of which are quite hard to reach. A few are not scholars were so r kind to provide. Especially Professor
recorded in H. Arntz's Bibliographie der Runenkunde. B.Bischoff (Munich) put his vast knowledge of early
(b) Library catalogues enabled me to unearth several new items, Mediaeval manuscripts and archives at my disposal in a
but on the whole the results of this rather tedious search most unselfish way. As Professor Bischoff himself has
were out of proportion with the labour involved in it. When been studying Mediaeval cryptography and strange alpha­
, runes' or a northern alphabet' are indexed, the task
C bets, his hints were very helpful. The other scholars to
. is quite simple, but at times I had to proceed in an indirect whom lowe information on runica manuscripta are men­
way. Some Mediaeval collections of alphabets contain tioned in the preface.
into al. a runic alphabet, and therefore occasional references
to 'alphabeta varia' or to 'Hebrew, Greek and other But even the combination of these three sources does not
alphabets' had to be checked. In many cases librarians allow me to guarantee that the present collection is complete.
and authors of catalogues are not to blame for having On the other hand, I earnestJy hope that it may draw the atten­
overlooked runes in their manuscripts.· . Runic items may tion to unpublished or forgotten items, arid that this new ,~

be scribbled on pages which were originally left blank, in material may help to fill the blank spots which no doubt will
appear in the pictur~ given here. Although I am practically
nogle .islandske Rum-Remser. Sinastykker udg. sf Samfund til Udgivelse af
sure that new items will turn up for a long time to come, I do
gammel nordisk Litteratw: 1884-1891, I-;U, lOO-II3, and wrote also a paper not think that the chances for making staggering discoveries .
Qn the runes in the Erfurt codex of the Lund annala (Codex Erfordensis af de are very great, at least if I may judge from the new items which
Lundske Annaler og tk deriforekomrnentk Runetegn. Arkivz5 (1909), 303-309,
I was able to add to the known stock of TUnica manuscripta.
cf. E. MOLTKE, Runeindskriften i Erfurt Luntkannaler. Aarbeger 1936, 248 i.).
In N. LINQVIST'S edition of aIdeelandictreatise on magic (En isllindsk roart­ A priori we may say that the chances for runes to survive were
komtbok frdn 1500-taiet. Uppsala. 1921. 6 ff. 34 f.) there lire alao references . very slight. . It was not sufficient for them to be thought
to such runica. In the chapter on the isrrma tract I have referred til worthy of being introduced into the world of scriptoria and
1. LINDQUIST'S edition of part of J6N OLAPSSON's Runologia. A. BJ.EKSTED,
Islanih Rumintlskrifter, 213 ff. gives a survey of runic material in Icelandi~ manuscripts; .all understanding for this unpractical sort of
manuscripts. MaIlY of the maIluscripts that have to be examined are listed writing· must soon have been lost. Special circumstances,
in the catalogue of the Arnamagnaean collection in the Royal Library in however, sometimes saved them from being entirely forgotten.
Copenhagen. Here are some others to which I found references when
preparing the present study: Bamberg MS. Msc. hist. nat. 7; Cambridge· Thus we have onJy one manuscript of ca. 1400 where a runic
MS. Trinity CoUege R. 14. 34; Dublin, Royal Irish Academy, Book of alphabet is still recognized as such,b\:!t in the late fourteenth
BaUymote; Karlsruhe MS. Aug. CLXIII; Leyden MS. Voss. lat. Q. 83; and early fifteenth centuries there are a, score or more manu~
London, British Muaeum, Cotton MSS. Caligula A 15, Galba A z, DomitiaIl
scripts where the runes are called' Saracen 'letters. As I have
49, Vespaaian A 18, allrl~ MS. 2399, StoWe MS. ,57, Addit.MS. 4783;
Munich MSS. lat. 276,14436; Oxford MSS. St. John's .college 17, Bodley 572, shown in the Appendix to Chapter III, a runic alphabet and
an4the J1.!l1iusMS. Qf the Orrmulum; Paris MSS. grec 375, lat. 9666. the names of Aethicus Ister's letters were combined, and the
}viii liJ!:
whole was passed off as a Saracen alphabet. By some way or:­ p. xxiv H. (nos. (a) to (d». The plan of my work is consequently
other· this alphabet found its way into Sir John Mandeville's as follows :
TTtwels, and in this manner it received a wide circulation. Of
I. Fuporcs (1) : unrelated items and small groups.
course such Saracen alphabets can hardly be considered as
II. Fuporcs (II) : the isruna group.
runica mD.nuscripta.
The material collected in this way fell quite naturally into III. Alphabets {I} : unrelated items and small groups.
three groups : IV. Alphabets (II) : the De itroentione group.
(a) notations which retain the original fuporc order; V. Non-alphabetic runica manuscripta.
(b) runic material arranged in the order of the Latin alphabet To chapters II, III and IV I have added .appendices with such
(runic alphabets); materials as I believed might illustrate the contents. They are
(c) runes in non-alphabetic arrangements. mostly .devoted to non-runic material. Although this study
only treats of tihe . DE. runica mamucripta, I have of course
We might of course try to arrange the material according to included such re~erences to ON. material as might help t-o solve
other criteria. We might e. g. distinguish between English. some problems (e. g. runic cryptography).
and Continental manuscripts. Although this distinction will Of the isruna tract, which is of great runoiogicalimportance,
occasionally have to be made, it is of little value for a general. I have given a critical edition, and examined parallel ON. mate­
plan: some English manuscripts are derived from Continental rial. But I have not thought myself qualified at this moment
prototypes~ To distinguish between the fuporcs (and alpha­
to give what could be termed a critical edition 'of the De inventione
bets) with rune-names and those without is still less advisable : Iitterarum text. This treatise ~onsists of paragraphs on the
some alphabets· without rune-names, e. g:, are closely related Hebrew, Greek, Latin) 'Istrian' or 'Scythian' and runic
to alphabets with names, but not at all to the other alphabets alphabets. The study of the sources used for the first. three or
without. When no other criteria for classification were avail­ four paragraphs lies entirely beyond the scope of this work, as
able, the manuscripts have been arranged chronologically. it would imply a thorough examination of Mediaeval hebraica
The material of the first two groups is again subdivided~ and graeca, grammatical literature, and such material as lists of
When we examine the fuporcs, we find that besides items which nomina in'IJentorum, etc. Therefore, and notwithstanding Pro­
,'-t
are unrelated, or ean at the most be considered by pairs, there '.
fessor K. Preisendanz's encouraging invitall'Oil, I' decided to
is a group of five closely connected fuporcs. In the manu­ postpone such an edition. I have, however,ineluded the whole
scripts where they occur these fuporcs are followed by a short text in' the two or three versions which one may distinguish,
treatise on runic cryptography, which, from the name of the and I have even added some parallel texts to illustrate the
first cryptic device described in it, I have called theisruna position of dIe runic alphabet.
tract. Similarly a number of runic alphabets are incorporated Of each manuscript I have given as full a description as I
. into a brief history of the alphabet, De itroentione litterarum. eould justify within the frame of this work. These descriptions
Therefore the study of the fuporcs has been divided over two include, as far as data were available.:
chapters, and so .has that of the alphabets. The special pro­
(I) the ~istory of the manuscript in question;
blems raised by each main group (fuporcs, alphabets) have been
examined at the beginning of the first and the third chapter (z) its exterior and composition;
respectively. The matter of the fifth chapter has been sub­ (J) its date and origin;
divided according to the· t1ypological categories outlined on (4) its. contents.

Ix lxi
The descriptions are based on autopsy whenever possible; but that Alcuin is only one among many Englishmen (one hundred?
with the wide dispersion of the manuscripts to be examined. five hundred?) who may have brought over their native runes
it was of course impossible to examine them to the last. There­ to the'Continent, and that the work of converting a fuporc into
fore in the other cases I used photostats and supplemented the an alphabet may be explained more easily as the work of a
data which I could find in library catalogues with the informa­ leisurely scribe than of a Hrabanus Maurus involved in theolo­
tion which librarians kindly supplied. For that purpose I drew gical disputes.
up a questionnaire, the answers to which allowed me to give
a fairly uniform account of the manuscripts. Some readers When discussing tne runic material proper, I have always
may think this apparatus criticus somewhat out of proportion tried to do so on as , broad a basis as possible. Occasional
with the relative importance of the runological harvest. But. repetitions CQuid hardly be avoided. I tried at first to. arrange
as I have indicated before, runology was not my only object in my material as was customary till now: to list all descriptions
undertaking this study. To obtain a reliable picture of the tQgether, and have them followed by the discussion of each
cultural background of the ru.nica manuscripta no clues should rune in the order of the fupark or according to some other
be neglected. What appears insignificant now may become arrangement. But this procedure tends to obscure the ties
important if new material is added. When Phillips found runes. between the runes and their surroundings; it makes it very
in the neighbourhood of a rule for drawing letters of recom­ difficult to gain a clear view' of such problems as the relationship
mendation using Greek numerals (the so-called regula forma­ between different versions, whilst, on the other side, it hardly
taTUm) he found it hard to decide whether there was a connexion helps to avoid repetitions.
between the two (p. IIS). In my material I came across one
In transcribing the runes I have mainly followed the system
more manuscript where the runes follow immediately after the
proposed by B. Dickins; it is given with the fuporc on p. xx, .
regu.la formatarum, and a third where they come after an example ':
and on the folding plate at the end of this volume. The one
of litterae formatae. In this way new problems arise, and new
point on which I differ from Dickins is rune no. I2. The type
solutions can be attempted for old ones. I have hesitated the less
I transcribe bf' j derives directly from the OGmc. type by the
to give full descriptions because I believe they may greatly
addition of a vertical shaft (cf. p.' xxi). It is rare in inscriptions,
simplify the classification and study of future discoveries.
but quite current in manuscripts. Dickins's j is the more current
.. In the descriptions I have paid special attention to the relation
epigraphical form which occurs also in manuscripts; this I tran­
of the runes to' the rest of the manuscript. On the one hand
scribe by j (cf. g and Ii). To avoid the cumbersome repetition of
we may learn· a lot about the background of each individual
-rune in f-rune, h-rune, etc., I have adopted the system currently
item and of larger groups of TUnica manuscripta;· on the other
used in Scandinavian publications: the 'runes are rendered by
hand it may prevent us from drawing rash 'conclusions and
their transcriptions printed in heavy type (h . h-rune). This
from launching on bold- reconstructions which do not stand a
also allowed me to reduce the number of figures in the text.
critical examination. In general I believe 'to have been rather
reticent in reconstructing the·' prehistory of the manuscript The figures in the text do not claim to be facsimiles. No

runes. The study of the literature on the subject has convinced amount of careful drawing can ever take the place of a good

me that many of the brilliant accounts of that prehistory are photostat 0.1' reproduction. In the figures I have tried to render

no more than daring hypotheses, if not mere guesses. It is the essential features and as many details as the technique

tempting, indeed, to connect a fuporc with Alcuin and a runic adQpted here allowed me. In many cases it was not possible

alphabet with Hrabanus Maurus. But we must never forget fully to render the ductus, but the series of reproductions at

ldi Ixiii
...
the end of the volume gives a fair sampling of how runica
manuscripta look.
From the point of view of runology I expected no startling
findings. I have even refrained from exploiting the rune­
names as a source of information on the origin of the fupark, CHAPTER I
because I believed that such an extension would require a far
more exhaustive study of twenty-four stems in the different THE FUI>ORCS (I)
OGmc. languages than I could offer here. I feel sure, however,
that the present collection of Tunica manuscripta and the con­ Only one epigraphical English fuporc has come down to us,
clusions arrived at will be immediately helpful for the study of and that prekents the runes .in a somewhat disturbed order. It
the English. runic inscriptions. An amount of· preliminary work occurs on a scramasax found in the Thames; the runes have
has no doubt been published on individual inscriptions, il few the following forms (the figures below indicate the usual rank
may be said to have been satisfactorily studied; but a full and order of the runes in the m3.nuscripts) (I) :
up-to-date edition of all inscriptions is still the object of many
a scholar's pious wish. At times I have badly felt the lack of r" ~ ~(U" X~ f;::fi-I + 1- [:'r V 1 ~ M*?I f' M~t"' (:'(x'.'i
1 t345678'»""u«~*n~uttu~w"UUUU
such a collection : it would no doubt have enabled me to remove
soine question-marks from my text. On the other side I hope FIG. 4
that these pages may be felt as an incentive to resume the task The manuscript tradition is more abundant, but this abun­
which Stephens and Vietor did not complete. dance is not without leaving a good number of questions
unanswered. It is in fact rather disappointing: the different .
: ~,,-,

items mostly appear in a sort of vacuum, without direct con~

nexion with' their manuscript environment. Therefore we
have mainly to rely on internal evidence, and this will rarely
prove as conclusive as one might wish ..

THE MANUSCRIPTS AND THEIR RUNES.

In the manuscript tradition we may distinguish. two strains :

(I) The English tradition. In it we may a prio~i expect some
sort of continuity; the existence of Christian runic inscriptions
is there to prove that some amount of runic lore was accepted
into the new' cultural pattern; there was no break, but only an
adaptation. Of course we shall have to take into account a
certain amount of purely antiquarian" interest, kept alive both
by native poetry and by inscriptions; Qnthe other hand, the
original lore was probably
.
preserved" only partially, divested
"

(I)Cf. O. VON F'RmsEN, Rumwna, 5Z, 57 (figs. 45, 49). W. KRAUSE, Runen­
inschriften, 438 if. (with bibliography).

lxiv 1
from its more obviously pagan implications, and gradually
3· Brussels, MS. 9565-9566 ; id.
decaying t{) a fossil state.
At one time four different manuscripts with fuporcs written 4· St. Gall MS. 270: id.
in England were known to exist. Of these four, two are now 5· Vatican MS. Urbin. 290 : id.
lost; one is oniy a membrum disjectum, and the last is both late
6. Trier MS. R. III. 13 : id.
(late eleventh or early twelfth century) and influenced by
Continental lore. In all there are six fuporcs in these four 7· Sal2burg MS. a. IX. 32 : fuporc with values.
manuscripts, two manuscripts containing two fuporcs each.
8. St. Gall MS. 878 : fuj>orc without names nor values.
Of this total of six fuporcs, four offer the names of the runes,
two only their values : 9· Ohent MS. 30 6: fuporc with values (incomplete).
I. Cotton MS. Domitian A 9 : one fuporc with rune-names. Of these nine manuscripts nos. 3-7 form a group by them­
2. *Cotton MS. Otho B 10: one fuporc with rune-names. selves: their fuporcs are accompanied by a short treatise on
runic cryptography. Therefore they win be examined apart
3. *Cotton MS. Galba A 2 : one fuporc with rune-names, in Chapter II.
one with values only.
The remaining four versions seem to be independent; nor
4. Oxford MS. St. John's College 17 : one fuporc with rune­ does any of the Continentai fuporcs betray a special degree of
names, one with values only. relationship with any of the English fuporcs. On the whole
Unfortunately these four manuscripts do not represen~ dif­ the Continental fUporcs will be found to provide a more varied
ferentstrains in the tradition: nos. 3 and 1- are very closely information on English runic lore than the material preserved
related, and no. 2, which, like no. 3, is known only from a in England. Chronologically, too, they are more important:
,.~
facsimile in Hickes's ThesauTus, seems to have been influenced St. Gall MSS. 270 and 878 and Brussels MS. 93 11 -93 19 belong "!
(at least in the facsimile) by no. t. This notably restricts the to the ninth century, whil~t no English fuporc is older than the
.i..:
value of our material, though the evidence is perhaps not so tenth. This will compensate in some measure the damage
poor as Wrenn would have us admit. which their passing through the hands of more or less ignorant
(2) In the Continental tTadition we meet with problems of a scribes may have done.
different order; these have been discussed in a general way in
the Introduction. On the Continent we have an actual break I. London, BTitish Museum, Cotton MS. Domitian A 9
in the runic tradition. We no longer believe in 'German' (saec. XI).
runic lore to be found in manuscripts: all we find can be traced
to English sources, with some rare additions of Scandinavian The runic material in this manuscript was mentioned as early
material. ail 1696, 'viz. by T. Smith in his catalogue of the Cottonian
There are nine manuscripts with English fuporcs written on library (I). H. Wanley copied it for G. Hickes, who edited it
the' Continent; six of the nine fuporcsgive the names of the in his Thesaurus (2). Of the fuporc with which we are Con­
runes as wen as their values, two only the values (one of which
is incomplete), one neither names nor values: (1) T. SMITH, Catalogus, 134; " 5. Alphabetum Norwegicum siV'e Runicum.
6. Fragmentum Rt,micum ".
I. Vienna MS. 795 : fuporc with rune-names and values. (2) H. WANl.EY, Cata/ogus, 239.
2. Brussels MS. 9311-9319 : id. G. HICJa!S, ThelfltWUl, Grmnmatica Anglo-Sa:wnit:a. I3 6 ; Grammaticae
Idmrdicae Rudimenta, Tab. II, 3.
2:
3
 8r Chaldaeorum (ita appellantur) litene: sunt vera Runiae. Epistola
cerned here, Hickes gave two versions : one in the order of the
Dionysii ex.igui, ad Bonifacium prirnicerum Notariorurn. Sec.
manuscript, and the other arrange~ as an alphabet. More fere VIII. Scripta.
recently it has been examined by G. Hempl, C. L. Wrenn and 9" Fragmenta duo:
C. E. Wright (1). (I) &: quodam libro canonum ecclesiasticorum. Saxonice: et,
(::I) De S. Augustino, ex Beds, ut videtur, desumpturn. (A,*d:
Usual Cottonian binding, brown leather with the Cotton anna stamped­ Saxonice). .
in gold; on the back the inscription: TRAGrATUS 1 VARII 1 HISTO­ lOr, . I IV Alphabeta Runica bina (a Wanl~io D. Hickesii gratia descripts)
RICII ET I MISCELLANEII MUS. BRITIBIBL. COTTON. 1DOMITIAN cum explicatione Latina nornitium veterum Runarum.
A. IX, and the number 691/a s. Older press marks on the inside of the 11 v Voces Saxoruae Latine redditre.
binding: XXI. a (crossed out), 652 a (id.), and 691 a; on fo1. *1; XXIII A
l::1r • Versus quidam teclmici de calendario, cum historicis observationibus
(crossed out), XXII A. Since the runes are found on a single leaf, apparently
marginalibus.
the last of a manuscript now lost or "Unidentified, it is not necessary to
examine the composition of the codex in detail (2). A survey of the contents 13" Regullead inveniendum tenninum plischalem, literas dominicales,
bissexturn, et numerum aureum;
is given in Planta's catalogue (3) :
14" Catalogus regum Anglile, a R. Ina ad R. Edwardum I; et quot anneS
fol. 3' S. Aldhelmi, abbatis Mahnesburiensis, epistola ad Heahfridum;
singuli regharunt.
contra Scotorum artea, quos tanquam sciolos perstringit.
ISr &:perientia optima 'pro minutioIle sanguinis.
IsvVersus rhythmici de regibi.ls Anglorum.
I6r Versus prognostici de die S. Pauli, &c. et de indulgentia quam
concessit'Clemens P. V. pro recitatione quarundam precum: cum
(I) G. HEMPL, Hickel's Additiofts. aliis rhfthmicis rtlonachalibus.
C. L. WBBNN, Late Old English Rune-Names. 17r Revelatio B. Hildegardis.
C. E. WRIGHT, POttJcript.
ID., Robert Talbottmd Domitilm ..4.: IX. Medium JEvwn 6 ([937), ISr Prophetia, rhythmicis versibus exarata, cujusdam canoruci regularis ~,

170 f.
(Johannis de Bridlington) .. secundum opiruonem vuigi, qui febribus ,
infirmatus est," ut ipse auctor dicit in initio prophetile; cum prolixa
On the runes cf. also; expositione de rebus in ista prophetia, pnedictis sub RR. Edw. II.
W. GRIMM, Ueber deutsche Runen, 97 f. and Tab. Ill.
G. STEPHENS, Monuments I, [0::1, no. 9; II, 829 f' J no. 62. et III. hie in: Anglia et in Francia accidentibus. Is qui coi:mnentarios
T. VON GIUBNBBl!.GER, Die angels4c1uischen nmerrreihen, 3 ff.
scripsit, cos inscribit Humfredo de Bohun, comiti de Hereford, &C.
et constabulario AngliEe.
K. SISAM, Oytwwulf, 3 1 _6 .
B. DICKlNS, The Stindwich Runic Inscription, 83.
84' Nomina regis illius qui sanctam crucem inveniet, secundum diversas
H • .ARNTz, Rrmen und Rl/.'I'Ie!I.tumUt1I, 183 ff.
prophetias a u t h e n t i c a s . · ­
(::I) I owe the following information to Mr T. J. BROWN, Assistant Keeper 84" Nota de hora.et minuto, quando natus erat R. Edwardus IV.
of MSS. : " .. , in- Cotton MS. Domitian A 9 the juxtaposition of folios 8, 8S r . Litene congnitulatorim Pii IV. P. ad Edw. IV. R. Angl. de ipsius ad
9 and I I appears to be quite fortuitous. Their dates appear to be these : coronam successione. I I Kal.Apr. 1461.
f.8, 8th cent.; f.9, u;th cent.; f. II a , perhaps loth cent.; f. IJb, perhaps
8S v The names of the lordships, with the Bageons belonging to the
l::Ithcent. in part. The dating of the last two is a very doubtful question D. of York, .
indeed. Folin 10 was formerly attached to the bottom of f. 110,."
86r Enchiridion ad Laurentium; continens institutiones philosopll.ile et
(3) A CatalogtUl of tbi Manuscripts in the Cottonilm Library, 573 f.
. theoiogilie.
A few addenda :
(1) Fol. 2r (upper margin): Mitte nobis historiam bede de gestis The following items in this manuscript are of special interest:
Anglorum. . fol. 8: Lowe. has given an account of this interesting uncial
(2) In Aldhehn's EpiltoUJ ad HeoJifrid- there are two OE. glosses:
dqdrantium; eogOTa (4r ); typo: rednesse (7 r). leaf (1): Fo1.8r begins with part of a Hebrew alphabet
(3) Fol. 9": Extracts from a chronicle for 1113-11 14· and a note on the order of its letters: [.:.j UeTSUS nobis
(4) Fo!. I I ' : Bede, Hiltoria Eccluitutica. IV, 5: I, 27: II, 3·
(5) An .. Elenchus Contentorurn in hoc codice" was added on fol. 1 (I) E. A. LoWB, Codices Lat. Ant. II, nO.-I8s. Cf. also the Catalogue
by a modem hand. of Ancient Manuscripts in the British MWer.lm II : Latin Manuscripts, 68.

4
5
 inuerslJ. a dextris!namque ad sinistram partem eam finiunt/
(fig. 5 and pI. I): A first hand drew the runes, with below

propter et nos hoc eadem modo liUera[s) posuimus

each its value, in three lines: f s / teat j g. Below

Then: CHALDEORUM LITTERAE HAEC SUNT... the last five runes, however, this same hand wrote the names

EGYP[... j... UTTERAE GRECAE CUM 'NU­ in full, and the values above them. A second, somewhat later

MERO (?)..., with the three alphabets. Much of the hand added the names of the runes not given by the first scribe,

rest 'is hardly legible: Secundum numerom dierum writing them neatly above each rune in 11. I - 2; in I. 3 it gave

(oncep!tus est Dominus, etc.; a fragment of a trilingual

different names to the first three runes, variant values to the

glossary: [ ... j hebrei, Vmera grece, dies /anne. Finally: last two, and added a sixth rune (calc). Wrenn dates the first
;1
}
Epistola Dionysii exigui de ratione paschae(fragm.). Al­
hand within the first p~rt of the twelfth century, the second
though this page contains no runes, it is important as slightly later (I); Wright assigns both to the eleventh cen­
one of the t:arliest .collections of ' spurious alphabets'; tury (2), whilst Sisam dates the first about the middle of the
these alphabets play a part in the later history of the tenth and the second in the early twelfth century (3). The
. runes (p. 274 if.). second scribe seems to have intended to ' correct ' his prede­
fo1. 10 is an inserted paper leaf on which a strip of parchment cessor. A much later (sixteenth century) hand gave a few
(ca. 20 X 125 mro) has been pasted. This strip was more ' corrections' : the v:illues d and m for m and d re­
formerly pasted to the lower margin of the next folio. spectively, and (this time correctly) st above the stan-rune.
It contains a Scandinavian runic alphabet with some The rest of the page is filled by :
very late forms (13th century); to the left, in green ink: 1
( ) an explanation of the rune-names, in a sixteenth century

AI!pha!be/tvm! /Nar/wa/gi/cum; a sixteenth-century hand: f. feoh id est pecunia; g. gifu id est gratia; h. hegel

hand indicated the equivalents of a number, of ~nes

id est grando; e. ethel id est patria; d. deg id est dies;

(bighiloptu). tho thorn spina (vel sortes); m. man id est homo (vel mann);

fo1. II,ca. 208 X 130 mm,'written area 175 X 98 mm; one a. ar id est reuerentia; reo resc id est fraxinus; ea. gear id est

column, 31 lines. The recto side bears the old folio annus vel ear; st. stan id esnapis (vel z);.R. Rad id est

number 10 in ink (i. e. the Catalogue foliation) and consilium; B. berc id est cortex; S. sigel id est velum;

the modern pencil number I I. In the lower margin v. vr id ellt noster (4).
on the verso side the quire mark XIII. Wright te,nta­ Ker and Wright identified the handwriting as that of the

tively dates this mark in the thirteenth century. He antiquary ,Robert Talbot (r 50S? 1558) (5).

is probably right in supposing that this leaf was at one (2) To the lower right the ~ame h~nd added the following note:
time the last of thirteen gatherings (I); at any rate the ther ys souch an/other alphabet [...J / in the end off em] y f
verso side is much soiled. The outer margin has been old saxonice be [de] / de historia ecclesie / gentis Angloru~
repaired by tw'o patches of parchment (max. 35 X 10 that/with owt bordes in [...J flast leaf off that bo [kJ.
and 42 X 8 mm), b~ing pasted on to it; apparently the Whether the manuscript meant by Talbot has come down
manuscript to which the leaf belonged was at one time
or other given a binding with two clasps, and the latter (I) c. L. WRENN, Late Old English Rune-Names, 28.
(2) C. E. WRlGHI', Postscript, ISO.

damaged the last leaf. b) K. SISAM, CjI1teWU}f; 3 r6.

Fol. IlV contains the runic material arranged as follo~s (4) The glosses between brackets are in another hand; vel sortel has been
almost wiped out; after resc an h has been crossed out.
(I) C. E. WRIGHT, Postscript, 151. (5) C. E. WruGHI', Robert Talhot.

6
7
 to us remains an open question. Wanley's catalogue facsimile with the original, as this is the only way to value his
mentions five copies of the OE. translation of·the Historia facsimiles in those cases where the original no longer exists.
Ecclesiasticaj of these only Cambridge MS. Corpus Christi The two scribes responsible for the runic material are referred
College S. 2 (= no. 41) is said to contain .. Elementa "to as • A ' and • B '; unless otherwise stated, A wrote the runes
quinque priora Alphabeti Runici" (I). These five runes and their equivalents, B the names. • C' is the sixteenth
can hardly be .. souch another alphabet" as meant by century • corrector.'.
Talbot.
(3) Another reader had already left traces of his interest in feoh. ur. ~orn. or: J'6\. ".n·. ~.f... fen. * ~ r,,~el.I"'0r\. .
~ ..~. I..., !!>c'!-r. . ~r~
runes before Talbot's time. Under e. ethel etc. he scratched .r· n.~ .~ .~1'" .h· X· r: H'l' I . +.S· ~ ''1'' r·
the runes f u I» 0 (?) with a dry point; under d. deg F . ~ t5 .
lJ (. . "S . uu. h . n .
0 . ~t p. .x. r 1 • (10

etc. we can make out part of a runic alphabet, abc d e f

g h i kim n 0 p scratched in the same rather careless
~'r. berc.• j>el. lIe~ l~s"" .nc . mMO' 'pro. At : .er~ y" t'r

way. Finally, between th and thorn, there is another :i·t·M·~·~'*·~j-~·~·f·fYl T·

scratched g. These scratched additions are probal?ly .-r . b . e- . m. t 1n3. ~ . O~ "-. "-~ . y . eclr­
oren1;. cur. lola. ~ • el: cale.
based upon the ful>orc written on the same page; unfor­
10. q . }c: re. :5
tunately the crucial letters (q, x, y, z) are missing. Cf.
p. 263 f.
The glosses on the rune-names were at one time considered
*~ ~ . rh·~·)€· r
1"r'Cf'¥. ~~. F' 5~1'"
as authentic runic lore. Wright has proved, however, that they FIG. 5
only reflect a six,teenth century scholar'~ knowledge of OE. (2). . u: B writes the name u,r with short r; Hickes gives it a long ..:~
Of the fifteen glosses, eight are completely justified. Those on insular r. )

e and a are due to confusions in the ful>orc itself (the OE. .,: the value and the name are both written with crossed iI.
words are correctly translated, but ethel is not the name of the Wrenn ar.gues that porn. (the spelling of Cotton MS. Galba
e-rune, and ar is not the name of an OE. rune, but an errorfor A 2) is a better form than ilorn (1). As long as no full and
ac); one is due to the glossator's confusing two runes (gear and detailed study of the variant spellings p, ii, th (and even d
ear), and in the .four remaining cases
. he has not translated the and t) for Idl and 161 has been made, the spelling of the
rune-mu;nes, but homonyms or sirrular words (rmd • consilium ' : name with either iI or p cannot be used as an argument
rad • riding'; berk (1) • bark' : beorc • birch'; segl • velum' : for the superiority or inferiority of a manuscript (2).
sigel • sun'; ur • noster' : ur • bison '). Moreover, Cotton MS. Galba.A 2 is known only through
In his paper. on Hickes's additions to the Runic· Poem (MS. a facsimile in Hickes's Thesaurus and this does .not improve
Cotton Otho B 10) G. Hempl (3) has laid the basis for all further ..
Wrenri's position.
study of MS. Domitian A 9 by a detailed account of whllt the (I) C. L. WRBNN, Lau Old English Ru1l8-Names, z7.
manuscript actually offers. In the following survey I have (z) Neither W. KRu:.BR,. Angelst'ichsuche ·Pa/(uographie, 13, 34, nor other
authors on English palaeography supply detailed infonnation on the relative
indicated where my examination of the manuscript led me to
frequency of these variant spelling deviCes, which it would in fact be the
differ from Hempl's results. I have· also compared Hickes's grammarians' task to provide. .A good statement based on a limited material
is found in H. STROM'S Old English Personal Names (u8 ff.).
(I) H. WANLBY, Catalogw, 115. Cf. also J. BLOMFIELD, Runes, I84ff.
(z) C. E. WRIGHT, PO$tscript, 151.
A fuller survey of the use in Icelandic manuscripts is found in H. SPEHR,
(3) G. HBMPL, Hicku', Additions. Der Ursprutig der ulandiscMn Schrift, 8 ff., 3Z ff.
8
9
 between·~ and a; consequently he did not recognize the
h: this rune is preceded by two dots, the t-rune by three.
A consequently indicates the old subdivisions of the fuporc variant as h.
into groups of eight runes. There is no such indication i: B can have had only a vague idea of the names of the runes.
of a break in the last 17 runes, and so it is not possible to He seems to have used an alphabetical list of names.
decide whether A lumped all the additional runes Misled into al. by the acrostic principle, he assigned eight
together with the third group, or whether he grouped them names to the wrong runes. Although A had written ing-­
in a separate ;Btt, and in that case, where he began his in'the right place, and although B himself wrote the name
fourth group (d. infra). inc above the ing-rune, he also gave the same name inc to i.
Just above the name there are a number of faint strokes, Afterwards' the name was partially erased; therefore it is
which Hickes interpreted as a variant form of the h-rune difficult to decide whether B actually wrote inc or iuc.
consisting of a vertical stem with two slanting cross strokes. Hickes's facsimile reads eac, also adopted by Stephens and
Hempl rejected this interpretation and reconstructed a von Grienberger, but corrected by Hempl.
variant h-rune. (:j:I:) To the upperleft of this variant he dis­
j: below the rune A wrote g~; perhaps one more letter (d)
covered a faint a, to the left of the latter, and somewhat
was erased, and its place was taken by the dot which
lower (i. e. just! above the rune-name wen) the word vel,
separates two consecutive names. B first wrote geor or
and in the space between gifu and wen: &. This & vel a
geur, then corrected it to gear by underdotting the third
he interpreted as a gloss made by a scribe who had read
letter and writing a above it.
the additional rune as a. This glossa tor would then have
discovered that his interpretation was wrong, and in ,3: although A had given the rune the right value eo, B mistook
erasing & vel a' would also have erased the left vertical the rune for s (no doubt he connected it with Roman S) .",.;:.
stroke of the variant h-rune. The result was that at first and wrote above it the name of the latter, sigel:
sight only one vertical stroke of that rune was visible. p: the manuscript writes peoril with Carolingian rj Hickes's
On this, point I offer a different interpretation. So little facsimile has insular r.
is in fact visible that it is more a question of interpretation
x: below the rune A inserted the value usually found in OE.
than of reading. . First of all the name of the rune reads
documents : x. But B seems to have hesitated (on account
Hegel, not hegel. I admit there is no reason to have a
of the acrostic principle?) and finally left open the space
capital H here, but comparison with the h of leah shows
above the rune. He assigned the name iolx to k instead,
that the first letter of the name of h cannot be a minuscule.
and at the end of theJuporc p.e added a rounded type of x
Second, I accept Hickes's interpretation of the additional .
with the name calc (d. infrq.).
rune. Hempl's objection, that an h of this sort is unique,
is not valid. It is actually found in three other manu­ s: the s~rune shows the rare English type found on St. Cuth­
scripts. With a different value, 1), it occurs in four more bert's coffin, on the Thames sword and perhaps in one
manuscripts; in two runic alphabets it stands for x, and or two other inscriptions. In the manuscript material
in one for y. How these different values of the rune are there seems to be only one more instance of this form
to be reconciled will be examined elsewhere (p. 7 1 ); here (Oxford MS. St. John's College 17). Right above this
we are only concerned with its existence. Instead of rune A or:B added the normal type, to the right of which B
& vel a, which is rather awkward, I read ;B vel a : in his wrote the name. Of the latter very little is visible. Hickes
interpretation of the additional rune the glossator hesitated simply omitted the' variant' (= usual) rune and the ~me,
II
10
 and for the rune in the line he substituted R Stephens, d: cf. under m.. Stephens' (following Madden) read manis,

on the basis of a copy by Madden, conjectured the readings an error of the same sort as Hickes's lagir for lagu.

co or et for the name. Hempl read fe ~d supposed that B, <E: below the rune A wrote the value oe. Above it, there is

having taken the· rare s~rune for £, had begun to write the a variant rune rather like x; to the left of this rune B wrote

name feoh, but then became aware of his error and left pro, the r of which has a curious curving stroke going

off after fe. Mter having examined the manuscript my­ upwards and backwards (1). Hickes omitted the additional

self, I can only propose to read sig; whether B left off at rune; Stephens considered it as "an old Scandinavian'

this point, or whether the rest of the name has become type of the M" (i. e. a type of m found in the shorter

invisible, I cannot. decide, as the page is. much damaged Scandinavian fuPllrk (2». It is true that part of the rune

in this place. At any rate neither co, et nor even fe seem is ;russing (ink flaked off) so that it could easily be taken

to be possible. for a variant of x. But Hempl is probably right in identi­

t: the rune is preceded by three dots placed in one vertical fying it. with the type of re-rune found in the Thames

line; cf. under h. In the name tir Hickes's facsimile inscription; this type also' occurs in a couple of manu­
lengthens the r so that it looks rather like an insular s; scripts (O"ford MS. St. John's College 17 and Leyden MS.
the same happens in yr and orent. Actually B uses a very Voss. lat. F. 128). As to pro, Hempl interpreted it as
short Carolingian r. But cf. infra under ea. meaning that the upper rune may be used for the lower.
e: again misled by the acrostic principle, B transferred the B may have found this form in his exemplar; but then one
name epel from (E (*opil- > oepel > epel) to e, and omitted does not see why he omitted the name. With B's poor
the real name of the latter rune altogether. knowledge of the runes it is of course possible that he did
not really know what name to give to (E, having transferred , .
m.: B mixed up m. and d, although A had given them their >..~

correct Latin equivalents. Hence the name deg above the epel to e to conform to the acrostic principle.
m-rurie and mann above the d-rune. A much later reader, a: the reading of the name is doubtful. Either it was first ;;. ~ .
C, 'corrected' the values to make them agree with the written ac, which was then alter!'!d to aT, or vice versa. If
names by adding d above A's m, and m above A's d. we may judge from the colour of the ink, the former
1: the dot which separates the name lagu fro~ the folloWing explanation is the more probable. The 16th. century
name is placed high and rather close to the final u. .This glossator read ar which he translated by reverentia.
explain!! Hickes's reading lagir (accepted by Stephens and y: on the spelling of this name in Hickes's facsimile, cf. supra
von Grienberger). .under t.
IJ: Hempl states that A first wrote ing below t;Jte rune, and ea: here too A wrote the name ear in full (cf. ing), but gave
then corrected it to ng by erasing the i. This letter, no equivalent. For the remaining .five runes, however,
however, was not removed completely, but two dots A gave both the names and the values. Hempl supposed
remain. I doubt whether the partial disappearance of t4at A used two different sources, "( I) One' that had the
the initial i is due to any interference on the part of one .runes and below them their values, and that ended in ea.
of the scribes. Parts of sQme other letters are also missing,
the reason being that the ink flaked off from the rather (I) This curved stroke is probably aCCIdental; the whole can hardly mean
that • the variant rime stands for 0 (pro 0) " as such an etymological inter­
greasy parchment. B here repeats the name inc, already pretation lies entirely beyond the scope of D's runology.
given to the i-rune (q. v.). (z) L. JACOBSEN-E. MOLTKE, RJmeindskrifter (Text), 9 6 3.

12 13
 ."

This is the stage of development of the Old-English fuporc g: to the correct name and value given by A (gar, g), B added
that we find reflected on the Thames knife ... (2) One that
the name (?) et. This may perhaps indicate what sort
had runes for io and for the differentiated velars k and g, of source B was using. At an early time it seems to have
as well as for the spurious runes cweora and stan" (I).
become the custom to have the z of the Latin alphabet
The names ear and ing and perhaps also gil. (if = ger) show followed by a couple of abbreviations and extra letters :
that Hempl's reconstruction of A's sources is not absolutely
... (= est); 7 (= et, and); p, <1; 'fJ = (pret); p (= w) (I).
certain. B mistook ea for t and added the name tirabove This use seems to have influenced runic alphabets: in
it. This time tir seems to be written with a long r, which Oxford MS. St. John's College 17 and in Cotton MS. ,L
rather looks as an insular s. Closer inspection shows. Galba A 2 the z-rune is followed by a ligature of an e­
that this lengthening of the r was probably carried out am'l a t-rune, explained as "& ", and by the usual abbre­
afterwards; whether by B or another scribe I cannot viation mark 7. CoIning immediately after the z assigned

make out. to stan, B's et shows that he considered this a necessary

j: in his facsiInile Hickes makes the name and the value complement to the alphabet. This, and still more B's

(both written by A) change places. Above io, B added a blunders with inc (for is), sigel (for ih, eoh) and epel (for

name orent, a hapax not explained thus far (2). In eoh, eh) indicate that his exemplar did not present the

Hickes's facsimile the name reads osent, with insular :f names in the fuporc order, but rather in the order of the

(cf. under t and ea). alphabet.

q: the q-rune differs from the ea-rune by having the latera k: at the end B gave his idea of k (in fact rather = x), with

strokes more developed. The same rune is found for x above it the name calc. With this opinion he did not

or z in a number of alphabets, but the value q only occurs stand alone, as is shown by a couple of runic alphabets ,"'::..

in fuporcs. Hickes again has the name and the value of where the k closely resembles B's k.

this rune change places, and will do so for the remaining After -ea A left a space for one or even two more runes, and
runes too. B added a different but not unparallelled form began a new line with j. This, and the different arrangement
of the name (cf. cur in Trier MS. R. III. 13; qur in Mu­ of the values and the names beginning with j Inight indicate
nich MS. lat. 14436). Stephens reads the name written that A (or his exemplar) used two sources: one extending
by A cwreora; the manuscript does not support this reading. from f to ea, the other giving the remaining five runes. But
k: the name calc and the value k as written by A were can­ I rather believe that A (or his exemplar) was aware of the fact
celled by B, the first by underdotting, the second by a stroke that the last five runes were added. at a late date; they probably
across it. For calc B substituted iolx. He apparently were never so well established as' the first twenty-eight, and
mixed up the runes x and k. therefore more information (in casu the names) had to be given
about them. Perhaps this allows us to date the exemplar used
st: this rare rune, which is hardly ever found in inscriptions,
by A in a period when the last five runes were still felt as new,.
must already have caused some trouble to A : he gives the say the beginning of the ninth century.
name as stan, but the value as se. B added the value z,
It is not easy to settle the linguistic status of the two scribes,
and Cst.
or of their sources. The materi'al is too limited; moreover
(I) G. liHMn, Hickes's Additions, 138.
either of the t\~o scribes may have adapted the forms of his
(2) Could this orent go back to an insular CfMrt f
(I) Cf. A. C. PAUES, The Name' of the Letter 3, 441 ff.

14
IS
 i'Y!

exemplar to his own dialect and scribal habits and have retained
others, the whole giving contradictory evidence. On the basis few excerpts from Wanley's account are given between brack­
ets) (1) : .

of oe one may perhaps locate A or his exemplar in the North (I).

As to B, the clues are ambiguous : gear is West Saxon, but deg I. HomiUa de creatione, casu Dmmonum, & adventu Christi. F. I.
and hegel rather non-West Saxon (z), bere and peoril could go Z. In HexaemClQll. 8.

3· Depositio S. Basilii. 17.

together in Anglian surroundings (3), but wen js atypical 4· De S. Mauro Abbate. 27.

Kentishism (4.). S· Passio S. Juliani, & ejus sponsm B~. 3z b.

It is mainly on the basis of this one form that Sisam assigned

the whole to Kent (5). Perhaps the prehistory o( this fuporc
6.
7·
Passio S. Sebastiani martyris. 40. [39 b]

Passio S. Agnetis, virginia & martyris. 48.

1
8. Passio SS. Joannis, & Pauli. 53.

is somewhat parallel to that of a good deal of OE. poetry: a 9· Passio S. Eugenie virginia. 55.

Northern prototype may have been partly adapt.ed to its new 10. De S. Euphrosyna. 61 b.
II. De S. Cbristophoro martyre. 69.

Southern surroundings. We must keep in mind B's obvious

IZ. De S. Maria J.Egyptiaca. 77. [76 b]

deficiency as a runic scholar: of the twenty-eight rime-names 13· De SS. Septem Dormientibus. 96.

he knows, eight are assigned to the wrong runes. Of course 14· De inventione S. Crucis. 116 b.

the late date of the manuscript must be kept in mind when we IS· Passio Alexandri Papal, Eventii, ac TheodoU. 118. [117 b]
judge B's contribution. The evidence provided by A is pro!Ja­ 16. Sem;lones tres ad inatituendum populwn in religione, in quorum
altero introducitur diabolus, ostendens. cuidam Anachorebe omnes
bly of great value, but B's additions are poor indeed. With peenss inferni. I ZOo
this restriction Wrenn's judgment may be accepted. 17, In natalem S. Joannis Baptisbe. :136. [XVIII, fol. 136. Coru:lruio
Iwmiliile de S .. iEtheldrihtll1, en-ore BiiliDpqi hue tramlocata, ut
irrfra est 'fIitkre, etc.]
18. Ritus ordinandi Monachwn, cum precibus, Latine. 140 h.

2. Cotton MS. Otho BID. 19· Historia Holofemis &: Juditlue, ubi plura de captivitate Jud!eorum;
& ad finem, historia Malchi MonaclIi ex Hieronymo. 14:3. '1
,""

ZOo Ristoria libri cade,ntis de 00::10 coram porta Ephraim apud Hiero­ ;",.'.

This manuscript was almost completely destroyed by the solymam, in qua vil.ria aunt pnecepta de sanctificando sabbato,
&C. & in fine ait, tertium hoc scriptum use, quod Dew de c(1!Jlo mUmt,
fire of 1731. Our authorities for it are T. Smith and H. Wanley, neque post hoc aliud tlJCS'/!eCtandum. ISZ. [151 b)
who gave detailed analyses of the contents (6), and G. Hickes, 21. Canon!!S antiqui Synodalea, Latine. Inic reperitur confessio fidei

who edited the Rune Poem (7). This is Smith's. description Catholiae, quam Papa Damasus misit ad Paulinum, Antiocbenwn
Episcopum. -ISS. [tkl."aJw.)· .
of the contents (the folios are those added in the annotated u. Poenitentiale Saxonicwn. 161. [z60 b]

copy of Smith's catalogue in the Ghent University Library; a :.a3· Characteres Alphabeti peregrini, numero taDtuni decem. Aliqui

ex his videntur ~ Uteris RuniQS similes. 165 b.

(I) A's rare funn of s might perhaps seJ:'Ve to confum this view, as it also :.a4· Liber. Geneaeos, h. e. a 37. capite ad finem; SUonice. 166.

occurs on St. Cuthbert's coffin; but it is found on the Thames SCIIIIIlIISlIX as :.as· Depqsitio S. Swithini, & ejus miracula, de quorum silentio Auctor

well, and that seenis rather to be connected with Kent, cf. B. DICKINS, TM recentior increpat priora tempora. 18:.a. [181 b]
26. Natale S. EdmuncJi, Regis & martyrls. 186.

Sandwich Runie lmcriptimt, 83.

z7· Natale S. Georgii Martyria. 190.

(z) E. Srilvm:ts-K. BRUNNER, AltenglUche GrtJ1lfllftltik, §§ 5z, 91 b.

(3) E. SllM!liIS-K. BRUNNER, AJtenglis£he Grammatik, §§ 84, 120. :28. De S. Etheldrytha. 193. [Notandum vero, quod hujusce sennonu

(4) E. SIlMIRII-K. B.RUNNER, AltenglUche Grf1.fm!latik, § 31. paTS POSterior,low trampolitis, Iwhetur hujus in Cod. MS.lol. 136.

(5) K. SIIIAM, C~, 316. quam tanquam hi loci duiderata1nt ruutuit quidam neoterieru, lorte

Joannes JOIfelinwj. .

(6) T. SMITH, CatDlogw, 70 f.

H. WANL!"Y, CataIogus, 190 ff.
:.ag. Pa&sio s. Matgaretlll. 195.

(7) G. HICKJ!ll, Thuaunu, Grmmnatica A1IIJ1o-S~onica, 135.

(I) No. Hut. :.a36z in thst library. This copy was annotated shortly after
16
6
" 17
 In the copy of Smith's catalogue mentioned supra the number
of the manuscript is marked with red, which usually means
In his chapter De diale.cto Normanno-Saxonica rive A,nglo­
Normannia, & de dialecto Semi-Saxonica Hickes discusses the
total destruction. W. W. Skeat,.however, discovered forty-five
Scandinavian influence in England at the time of king Cnut
damaged leaves in the British Museum and Napier one more
in the Bodleian Library. All these leaves contained fragments
(1017~1035). He considers the runic material of Cotton MS.
Otho B 10 as a manifestation of this influence:
of 1£lfric's Lives of Saints (I); the folio with the runes seems
to have perished. For, bound up with the collection of 1£1­ Hoc ut credam faciunt runarum Danicarum, tam simpli­
fric's Lives of Saints, there was a single leaf which Wanley cium, quam duplicium, descriptio quaedampoetica,
describes as follows : ' Anglo-Saxonice explicata; quae in bibliotheca Cotto extat,
Otho B 10. p. 165, quarnque vix antea et ne vix observatam,
XXVIII. foi. 165. Folium quod olim ad alium quendam
nedum publici iuris factam, planam quasi ab omnibus
[mum pertinuit, nunc hujus pars, in quo continetur Alpha­
doctis spectatu dignam, hic cum runis rere incisis, operre
betum Runicum cum explicatione Poetica, Saxonice, quod
et sumptiis pretium exhibere judicamus, Latinis additis
non ita pridemdescripsi rogatu Cl. D. Hickesii, qui in Gram.
' ex adverso elementis, .ad ostendam runarum potestatem,
Anglo-Saxonicte, cap. zz. de Dialecto Normanno-Saxonica.
una· cum iis nominibus quibus appellantur ipsre runre (I).
p. 135. illud typis evulgamt (z).
In Hickes's rendering of the Rune Poem (2) we have to
Even if the manuscript had been saved, the situation would
,distinguish two parts: the text and the COpper plates. To the
not have been very different from that presented by Domitian
left of the page the runes f - k with their names and phonetic
A 9. Since Smith does not mention the Rune Poem in his equivalents were printed from a copper plate of 277 X 33 mm;
survey of the contents, we are led to suppose that it was inserted
below the text a second Copper plate of 18 X 128 mm gives the :>~.
between the time he saw the manuscript and the time Wimley
runes stan and gar and, separated from these runes, by a vertical '~
described it (3). Unfortunately, Hickes's edition, though line, this note :
partly a facsimile (or assuming the appearance of a facsimile),
is not a very trustworthy substitute for the manuscript evidence.
Has characteres old w n x fog (cf. P1. II) ad alia
sestinans (read festinans) studioso lectori interpretanda
the fire. Wanley's description contains a couple of slips: item V occurs (interpretandos?) relinquo.
twicej instead of fol. 160 b' and 164 b we find 360 b and 'a~ b.
(1) W. W. SKBAT, Ae/fric's Livel of Saints IV (Early English Text Society,
To the right the text itself was printed from type, in such a
Original Series. 114). London, 1900. XV fr.: "Once a valuable MS.. but way that, each rune is followed by the corresponding stanza.
now much burnt and partially destroyed • '•• Since the MS. was burnt. some There are twenty~nine stanzas, corresponding to the runes
of the leaves have been collected and bound uP. but are much out of order and
f - ea; there is no stanza on the last two runes of the first
partially destroyed j and moat of those that can be partly read are much charred
and blackened. The leaves have been renumbered in their present hap­ COpper plate (q, k) nor on the two runes to the left of the
hazard order, as this is the best that can be done •• , Not only have several second (horizontal) plate (stan, gar).
leaves of this MS, been destroyed, but some fragments went astraYj as Prof.
Napier has discovered that a leaf of it came into the possession of Thomas (a) The Text.
Hearn. the antiquary. who wisely deposited it in the Bodleian Library. where
its class-mark is ' MS. Eng. tho e. I.' ,.
(a) H. WANLI!Y. Catal6gw. 19::&.
(]) The fact that SMITH should have noticed ' ten runes' on the verso
The poem bas been printed' almost a dozen times since
side of foL 165 [cf. p. aI. note (I)] may be considered an objectionj but in this Hickes's editUi princeps, most recently by E. V. K. Dobbie in
CIISe W ANLEY. may have made a rqistake in assigning the Rune Poem to foL
(I) G. HICI<I.!S, Thesaurus. Gra;"matica Angu;.Smronica, IH.
165 [cf. p. 17. nQte (I)]. (2) G. HICKS!!, Thesaurus. Granunatica Anglo-Saxonica, 135.
18
19
the Anglo-Saxon Poetic 'Records (I). Dobbie's edition is
deciphered occur in the fuporc of the vertical plate and are

provided with a good introduction and full notes; therefore the

. explained in the Rune Poem. These runes seem in fact· to

discussion may here be reduced to a brief outline. ,

have nothing to. do with the Rune Poem, and their cursory style

The poem consists of twenty-nine stanzas, most of which

lays near the idea, that they were a scribble, perhaps in the ,

have three or four alliterating lines. Two stanzas consist of

margin of the leaf containing the poem. Read as a word or

two lines only, and one of five. The manuscript used by

words, the runes give no meaning. This is probably why __

Hickes cannot have been older than the late tenth century, as
Hickes left it to the reader to try, and make some sense out of

appears from the many late spellings: y for e in unstressed

them. I can see no more in them than a probatio pennae (I).

syllables, -un for· -urn. The original, however, must have been
The plate printed vertically to the left gives 'the fuporc

written, at an earlier date. Scholars are unanimous in com­

engraved in vertical columns; the left column lists the values

paring the poem with the older OE. poetry, and therefore it
of the runes but also three names : ing, ear, cweor4; that t-o the

may be assigned to the eighth or early ninth century.

right the runes and their names. The names are usually

Each stanza gives a paraphrase of one rune-name. The

written to the right of the runes. It seems obvious that the

runological value of the stanzas is quite unequal: whereas

,two runes engraved to the left on the horizontal plate below
some of the poetic definitions seem to, contain 'a good deal of
the text belong to the same. fuporc. A number of peculiarities

genuine runic lore, others must have been thoroughly moder­

must be examined in detail (cf. pI. II) :

nized.
w, h, n, 3, 9 : of these five runes variant forms are included
There are two related Norse poems, one Norwegian, the in the fuporc,
other Icelandic (cf. p. xxvi). Although they are of a much later
date than the OE. poem, and cover only the sixteen runes' of m, Q, d, J, ea : in five cases, too, variant names are given. In
the shorter Norse fUPllrk, they retain many archaic features. three of these cases, the names do not agree. This raises ~,

Moreover, there can be no doubt that all three poems are a number of questions : 'Was this fuporc written by one ,
related, whatever the degree of relationship between them may scribe? or were additions made by another hand (cf. Do:­ l'

be. Therefore I have reserved the text of the OE. Rune rnitian A 9)? Finally, how trustworthy is Hickes's repro­
Poem for a collective discussion of all three poems. \ duction ? In the absence of auy description of the original,
only a study of details may eventually allow us to 'solve
. these problems.
(b) The' Plates.
w: the tWo variant forms differ only in style : one is a ' pointed •

Let us ,begin with the right hand part of the horizontal plate rune, the other a ' rounded 'form. The value is indicated

below the poem, the transcription of which was give~ supra~ twice: by uu !n.the left hand column, by the usual cursive

All the runes which Hickes leaves to his studiosus lector to be form of w above the second rune.

(I) E. V. K. DOBBm, Thi A",Jo..S_ Minor Poems, xlvj ft., clxii, as ft.,
h: no less than three variants of this rune are found in the
IS3 ff. Cf. also: . second ~Iumn : (a) the normal type with two cross-strokes!
W. GlUlIfM, U~r dlWtsw &men, ;:u7 fi. (b) the rare type with one vertical shaft transected by two
J. M. KBMDLB, A",.1o-SCIX01l. Runes, 339 fi. horizontal (sometimes oblique) s~rokes, (c)' one like the
B. DICKINS, lUmi.c and Heroic Poems, 6, IZ ft.
T. VON G1UI!NBImGER, Das agl. Rwwrg«lieht. Anglia 4S (19Z1), ZOI ft.
(1) These runes may probably be identified with the 'ten chBl'8Ctel1l'
W. KELLER, Zum altt:ngIiIclum RU1fntIJ«lieht. Anglia 60 (1936), 141 fi.
Cf. further bibliography in Dobbie's edition. which SM,m fouitd on foJ: 16s", although the number of charactel1l does not
agree (HICKI!!! has only nine), and some difficulties remain [po 18 note (3)J.
1
20
21
 'V(

first but with only one cross-stroke (hence rather like H). ea: beljides the name ear in the first column, we find a variant
form" car to the right of the rune; to the left of the rune the
n: of the two variants of this rune, the second is a somewhat
obviously erroneous name tiro The resemblance between
carelessly made normal type, whilst the first is a completely ea and t may explain this mistake. Car may be a mistake
normal type to which a short additional cross-stroke has for ear or for cur; in the latter case it would rather belong
been added. This addition is probably accidental; it is to the next rune (cf. infra).
not found anywhere else (1).
q: both the name and the value are in the left hand column:'
j: the value is indicated by gtt., as in Domitian A 9.
k: for this rune the facsimile has neither value nor name.
3: the first rune is the normal type, the second looks rather
like a Z the horizontal strokes of which have been leng­
st, g: the name and values (.3', st) of st and the name of g are
given to the left of the respective runes.
thened.
After a thorough criticism of Hickes's reproduction, Hempl
p: the upper lateral stroke of this rune is missing; only <:oncluded that
something like an his left.
(1) the values in the left hand column are taken from Cotton
m: above the letter m indicating the value, a d has been added, MS. Domitian A 9. Hickes'slist of values actually repro­
and to the right of the latter the word deg. Whereas this duces all the peculiarities of its model : UU, gil, ing, ear,
word is in insular script, the additional d shows a modern the mixing up of m and d; even the ductus can often be
type (cf. Domitian A 9). To the right of the rune itself, recognized.
we find the letters an. Consequently the rune must be (2) the variant rune-names and runes are taken from the same
included to give the reading man. The similarity of the source, cf. h, 3, deg (= [m]an), mann (= kg), ear, tir
• <

",,'"
m- and d-runes must have led to confusion, cf. infra (= ear); also wen?
under d.
(3) "the extra runes below", too, would be borrowed from
1): thefirst variant shows the normal form, the second a more Domitian A 9. Hempl apparently means st and g, perhaps
sophisticated type also found in Cotton MS. Galba A 2 also q and k, not the group of nine runes mentioned on
and in Oxford MS. St. John's College 17. In the .column p. 21, which does not really belong to the Rune Poem.
of the values we read ing, in that of the names iug; the (4) the remaining names would also be Hickes's addition, but
latter is obviously a mistake for ing, cf. lug in Vienna here Hempl does not n~e the source: "The way that
. MS. 795. Hickes writes the names makes it appear that putting
d: above d indicating the value, the (erroneous) value m, and them in was an afterthoughr with him; indeed, I believe
to the right of d the name mann; to the right of the rune lean. trace them to their source.. but I refrain from saying
the correct name kg. more until the necessary material is in my hands. Of
course it is possible that the names stood above the runes
j: the value io is repeated to the right of the rune, above the
in the manuscript of the Runic Poem, having been inserted
name iar.
by some later scribe. but before Hickes's day" (I).
(I) One might perhaps connect this variant with a fonn of Q found in Unfortunate~y The Old-English Futhorcs and Alphabets, the
some manuscripts (*); or was it.originally intended as a variant for h, cf. type
(b) of that rune? Neither of these explanations is very likely, and theref9ce
(I) G. HBM:PL, Rickes', AdJitimu, 141. cr. the quotation on p. xlv,
foomote (l).
r have considered the additional stroke as spurious.
22 23
 ;.;

work in which Hempl apparently intended to prove his point. names for m (deg), d (mann), ea (tir), (d) 1Ihe variant values
was never published. So we are left guessing what evidence for m and d, and perhaps also (e) the runes q, k, st and g.
he had at his disposal: If Hickes took aU these names from The absence of a name for the k-rune may reflect Hickes's
the same list (this is by far the most likely hypothesis). it cannot hesitation before the conflicting evidence of Domitian A 9
have been any of those discussed in th.is study. Cotton MS. (d. p. 14 f.).
Domitian A 9 shows different forms in the rune-names giJu This does not solve all problems. One might e. g. ask why..
"i
hegel nead gear bere ior iolx (against gyfo htegl nyd ger beart Hickes added a po!nted variant of the w-rune, and not'ih~ ;
Ulr eolhx); moreover it lacks the names fod, 3 ande. Cotton MS. pointed forms for r, j, x, band (2 as well, which could also be
Galba A 2 dijfers in the names royn hegil ned gyr eth eob ior borrowed from Domitian A 9. (The omission of the variants
quertl (against wen htegl nyd ger eh iar cweortl), not to men­ for sand (2 is less striking, as Hickes also omitted them from
tion its imperfect forms peo;;" (?) and beor. .A comparison his facsimile of Domitian A 9). Yet no other solution is more
with Oxford MS. St. John's College 17 is still less satisfactory : satisfactory. ..
htegil Md geoJu ech ilx tyr oepel quar ( : htegl nyd gyJu eh eolhx After removing the intrusive material we obtain the following
tir epel cweortl); moreover wand j and their names are mis­ fuporc with the corresponding names of the runes:
sing in the Oxford manuscript, the order of the runes has been
upset, and some have been confused almost beyond recovery. ~
feoh
(I
UI"
~
porn
~R
OS
h X .P N i-­
rCld cen 9yfiJ. wen h'Zgl nyd
The Continental fuporcs need of course hardly be examined
in this eonnexion, as they seem not to have been known in
Hickes's days. I
is ger
<I> I h eolhx
eoh
r SIgel
peor~
'1 t g MN
HI" beorc. e h "man
There is then only one criterion by which to decide whe1lher
,~.
the rune-names given by Hickes actually occurred in the ori­
ginal : do they linguistically agree with the forms of the poem? . ~ .~ ~H ~ ~ ITt ~ 'f
There seems actually to be nothing contradictory there, .except fagu. ",n9 epel daeg ac yr iar' "ear
';esc
FIG. 6
wen vs. wyn 11. 37, 8S, roynna 1.94. But this. 'UJen is precisely
a name which may have been .borrowed from Domitian· A 9 The forms iug and ear actually found in the faC$imile may
(cf. Hempl's conclusion (2». This possibility does not solve either have been scribal errors, or they may have originated
the problem altogether, for if Hickes actually found. the form in the process of copying and engraving (I).
'UJen in Domitian A 9, we may ask : did the list which provided The linguistic status of the list of names is not entirely clear.
1Ihe other names not also present a name for the. w-rune, and There is not only the form wen, which points to Kent (2). but
if so, why did Hickes not borrow that name? also en, a case of levelling typical of Anglian (3). But by the
The most. plausible explanation seems to be the final hypo­ side of the latter we find both eoh and leoh, and apart from eh
thesis emitted by Hempl : Rickes found a set of rune-nameS and wen the language may be eharacterized as late West Saxon.
with ~he RUM Poem, but these names were probably not due Perhaps these abnormal forms are hardly surprising if we keep
to the scribe of the poem itself. This would do away both (I) Another posSibility should not be excluded: in Domitian A 9 the
with the contradiction wen: royn, and with the necessity to q-rune looks very much like an --~e, but was given the name cur by acribe B.
It is not impOssible that in the course of thepr:eparation of the engraved plate
have Hickes borr~w the former. from Domitian A 9. From
Cur of Domitian A 9 became car of Otbo B 10. Cf. p. 14.
the latter lIickes would only have taken: (a) the values of the (z) E. SlBVI!RS-K. BRUNNl!lt, AltenglUc~ Grammatik, § 31.
runes, (b) the v:ariant forms of w, h, D, 3, Q, (e) the variant (3)E. SlBVI!RS-K. BRUNNl!lt, AlunglUc~ Grammatik. § 119.

24 25
 -"

in mind the long evolution postulated by the text of the Rune

Black 'leather binding on heavy boards, stamped in gold, with five brass
Poem.
bosses on each cover, and remains of two clasps. Parchment of unequal
quality. The 177 folios are arranged into ao quires, numbered in a later
hand: ·1· bY), ·II· (ISv), III (23'), IIIl (31°), V (39'), VI (SI V ), VII (6I V ),
V
·VIII· (67 ), 'X' (83v ), ·XI· (9IV), ·XII· (ggv), ·XIII· ( 107'), 'XIIII. (IlSV),
3. Oxford, St. John's College, MS. I7 (saec. XI ex./XII in.) V
·XV· (la3 ), ·XVI· (13Iv), ·XVIII· (162 v ), ·XIX· (170V); 6 paper fiy-leaves
in front, 5 behind (I).
Although this manuscript has been studied in detail, espe­ Ca. 345 x 245 mm (written area ca. a3s/270 x 175/1gS mm); 1-3, but­
cially as a reflection of the progress of science in the late tenth mostly a. columns; 35-54 11. to the p~ge.
and in the eleventh century (I), no full account of it seems to In his survey of the contents Coxe distinguished the following items (a):
have been published since Coxe's Catalogue of 1852. The £O!. IV Regube de quatuor hwnoribus, de rninutione sanguinis; etc. Incip.
codex has been identified as one sent to J. Leland by the anti­ .. Hi quatuor humores dominantur in suis locis; sMguis dominatur
in dextro latere."
quary Robert Talbot (1505 (?) - 1558), but this identification V
a Nomina herbarum. Incip. " Sarrninum, id est, Cerfolium".
does not stand unchallenged (2). It was given to St. John's Prognostics e libra Galeni. De mensuris. Incip. "Calculus est
College by Hugh Wicksteed (3). If we may judge from mar­ ciC4j!ris grana ".
r
ginal notes, the manuscript once belonged to the Abbey of 3 Versus acrostichides, quorum acrostichis .. Surnme sacer, te summa
salus tueatur amicis, Gloria Dunstano devoto necne benigno".
Thorney; it was written b'efore IIII (4); later it was in the 3' Tabula numerorum et de· !nodis numerandi variis. cf. Bed&: Opp.
library of St. Augustine's 'Abbey, Canterbury (5). Basil. 15 63, I, p. 114. De temporum mtione ab Adam" usque ad
presens tempus".

(I) See e.g. C. H. HAsKINS, Studies in the History of MedioJnJa1 Science, r

4 Luna quibus diebus bona est, et quibus non.

83 f., 3ag. 4' S. Cypriani episcopi Coena.
L. THORNDIKE, A History oj Magic I, 680.
5· Alphabetum alireque tabul&: Runicm. Hickes, Thes. III, p. 4 figur. ii.
C. W. JONllS, BedmI Psetukpigrapha, la7 f.
In margine est versus ! Molis. on. erme. frimot. prici. si. pidis. osti.
A. V~.oo VIJVER, Les _es inJdites d'AbIHm de Fleury, 1# fr. ridirnot ".
H. ~L, Byrhtferth'~ he/ace :!tM Epilogue to IW MtlfIUal' Speculum 18
6' Tabula orbis universi geographies. De quota feria inquirenda in
(1943), aB8-3 0a. . . unaqusque die tabula.
(a) J. LBLAND, De Rebus Britmmicis Collectonea (1774), III, 97. According
6' Arbor consanguinitatis et affinitatis, secundum S. Isidorum, cum
to C. and D. SINGER (Byrhtferd's Diagram. Bodleian Quarterly Record a expositione partim metrica.
(1917-1919), 47 fr.) the volume seen by Leland cannot have been identical 7r Divisio Philosophire.
with the Oxford manuscript. These authors planned to edit the whole Oxford
7" Tabula exhibens concordiam mensium atque elementorum a
-codex, but this plan seems never to have been carried out.
Bryhtferd, sive Brihtfertho, IUOnaCbo Ramesiensi, edita.
(3) Fol. a' Liber CoUegij Sancti Ioannis Baptiste in I vniversitate Oxtltl. 8r Tabulre plures astronomicm.
-ex dono Hugonis I Wickateed Mercatoris Scissoris London I Pattis Ioannis V
I2 Procemium Brihtferthi, Ramesiensis monachi, super Bedam de
Wicksteed oHm pm:dicti I Collegij !locij. temporibus.
(4) Fol. :lI9' [J08S] Hoc anno fuit concilium apud Gloecestre in natale v
13 Compotus tam Greecorum quam Latinorum et lEgyptiorum cmtero­
Domini in quo FuIcardus uiceabbas Thomeiam perdidit ... Et abbas rumque.
GVNTERIVS directus a rege Willelmo seniore Thomeiam uenit •.. fol. agv
14' VenerabilisBed&: versus de temporibus anni duodecim. Ejusdem'
[lIn1 (in a somewhat later hand) Hoc anno duo f':ratres de Wintonia scilicet versus alii. (opera, t. I, col. 47 b).
domnus Ordmenm 7 Antonius detulerunt Thomeie [con. -iam] reliquias
sam:ti episcopi Atheluuoldi conditoris illius loci uidelicet spinam dorsi in mense
-octobri, and on fol. 3v : A nativitate Christi usque ad presens tempus [sunt abbonis Inonachl et in eadem libro I astrologia Nignn (Higini) philosophi 1
Mni)MCX. ' astrologia Marciani capelle 1 Compotus helperici doctissimi et 1 figuraciones
signorum abbonis.
N. R. KER, Mt:dieual Libraries, 105, dates the codex in IIQ9-HIO, but
H. lIENBL, Byrhtferth's Pre/ace, would mther assign it to the late eleventh (I) Five folios originally belonging to this manuscript are now bound up
-century. with Cotton MS. Nero C 2 (fols. 80-84), cf. N. R. KEI:I, Medieval Librar;I'S,
105, footnot~ a.
(5) M. R. JDms, Ancient Libraries, p. 3a9, no. 1157, and p. sao: Astrologia
(2) H. Con, Cataloglu II, 6, 5 fr.
26
Z7
 ,
Versua BedR: sacerdotill sex: super regulares feriarunt.

To this list tbe foUowing retnB.Iks must be added :

Item versus super regulares lunares duodecim.

Versus de amns p~, hoc est de septem sideribuB.

fol. I' A ~etrical fragment (with drawings), by a later hand.
14V Versus de septem diebus. Leetiones de septem embolismia et in SV For a full description of the page with tbe runes, d. infra.
Endecsdem, cum gIoasulis marginalibus. 6r In the lower margin a modem hand has inserted the solutions to
I~ Kalendarium, c:um e:xpoeitione de mensium IIDmimbus, etc. necnon the cryptograms on the preceding page (cf. infra).
diebus festill per Bingulos menses cannine natatis. S' One of the two rotae on this page has part of the text written with
22' Tabulle de feriis inveniendis, lume curBUS et &ignis zodiacl, cum notal! Sanc:" Btmi/at;;, in which the vowels are indicated by one (a)
sententUs ex Beda super easdem. to five (u) dots, e.g. c ••• TC :.: I . .... , u, t;· 'c 1m • ••• Tt • •• s _=,-'
29r Tabuhe de circulo decennovali,. de epactia, allieque. ciTculus uitM [i.e. uital!l ac mortis; etc. (i.e. the sphera Pythogqyae). j
35' Figura numerum exhibens infinitum, c:um Abbonis ratione super 16' 11. The comment added ~ the calendar comprises extracts from
eandem.
Hrabanus Maurus and Heriricus (I).
25' Compqt.'fJ$ 'IlVlgaris qui dicitur ephemerida Abbonis.
36 ' Argwnentum [BedalJ de luna et de mensibus secundum antiquOll,

c:um tabulis. 42 11. Computilltical fragments: comments on Gerben, extracts from

Herigerus. '
37v De pOllitione septem ste11arw:.O errantium.
39v Tabula exhibens sedem • Sanctle Trinitatill et individwe Unitatla'. 62' At this point the order of the quires seems to be disturbed; fol. 6zT
Ratio Macrobii de situ orbis: begins chapter XVI pf Bede's De 1IIltura TntItfI, and fol. 65' has a
42v Ratio regularum abaci: c:um taboos annexis. note; hic debet sequi hunc liber de tenIporibus qui est in anteriori
quaternione.
so" De ute arithrnetica ex Bwthio, etc.
I04r Cronic:um Bedll.

S6v Tabuhe inacriptle, a. Cribrum BoetU de multiplicatione.

b. De divisione absci. 144' 1£. are by several later handa.

c. Numeri Rit:hmimachiz. IS4v Coxe locates items 46 and 47 erroneously on this page; they are

SSv Bedal presbyteri de temporibus liber minor, in capitula viginti aero found on fol. IS7v•

distributu&, pnevia tabula. 175' 'In the margin,a charm wid blod rim of Mill.

62' De natura rerum libri fragrnentum.

65" Ejusdem Bedae de temporum ratione liber, pneviis pnefatione et The codex has been rightly described as " a scientific ency­
:>..!:~
capitulis. clopaedia ,. . ,Vande' Vijver has demonstrated that it reflects
123' Heril'ici monachi expositio COmpoti, in capitula triginta octo distri­ t
buta; c:um pnefatione. '
into al. the teaching of Abbo of Fleury at Ramsey (986- 8),
I3S v Dionyaii Exigui. ad Petronium episcopum epistola de Paschalis where Byrhtferth was one of his pupils (2). This is important r--<.'

festill ratione. because we shall find the runes in the neighbourhood of al­
137' "Item Epistola prefati Dionisii ad Bonifacium et Secundinum de phabets known' froin other sources to have been circulating
eadem. Petavii de Ratione Temporum App. edit. 16z7, tom. II,
p.874­
in France.
139' Tabulle ~bens Indictionem, Epactarn, diem Pasc:ha:, etc. ab anno The runic material first came to the notice of Wanley and
532 ad annum 960 inclusive. Hickes (3). It was edited in part by C. L. Wrenn in his study
144' Tabuhe argwnenti similis ab anno [422 ad an. 2612.
IS~ Argumenta titulorum paschalium. Bedal Opp. tom. I col. 206.
157" De signis secundum Virgilianos. Tabuhe numerorum faustorum vet
contra. (1) The QE. names of the months are also given in' the calendar. Other
IS8v S. Hieronymi de gradibus Romanorum tractatulus. .anglo$tUrmka in this manusCript include glosses to the table of degrees of
159' Ejusdem Prognostics temporis. consanguinity (6V ); Nomina dierum secundum Anglos: I. Sunnandmg, etc.
159" Tabuhe grarnmstiae, c:um glOssulis marginalibus. -on 7Iv; a series of glosses on 74v; explanations of the OE. names of the months
168" Quomodo, in compositione mutantur Iitene, et. in quibus et ubi -on 76v : Eastermonad: aprilis. e08tra emmpascha uoCatur apud'eos, etc.
manent. (2) A. VAN DE VIJVIlR, Le, IBU'QTN iniditllS d'Abbon de Fleury, 1# 11. The
[69' ReguLe prosodiaae. manuscript also contains later material, down· to Gerland (1081), cf.
170" De nominum deelinationibus, pnevio schemate; aliaque grammati­ C. H. HAsKINS, Stpdks in t/ae History of MI!Iiiaeval Sciftu, 83 C., 329, and
cali8. de adverbiis, conjunctionibus et verborum conjugationibus. L. THORNDIKE, A History of,l'4agic I, 680. •
175' Traetatulus medicus de parrlbus humani corporis, cum medicinis (3) G. H,ICKBS, Tlwaurw, Grammaticae Is/.andicae Rudimenta, Tab. II,
ad. moroos variOll. Manu aliquantum recentiore. nOR. 6-11.

z8 29
on the OE. fupo~cs (I). His edition is not entirely satisfactory; (3) A fictitious alphabet ending with T, values and letters,
besides being only fragmentary, it contains a number of mis­ almost identical with the littere caldeorum et siroru'm of
readings and rearranges the material in an arbitrary way. As Cotton MS. Titus D 18 (cf. p. 338). It is a precious
will be seen from plate III, fol. 5v contains, from left to right: indication in that it connects the Oxford alphabeta varia
(I) In the left margin, a cryptogram with similar Continental collections.
Molis . on . erme . fdmot . prici . si . pidis . osti . ridimot., (4) A cryptic alphabet, in which the vowels A, E, I, 0, Uare'
By changing the vowels according to the rule: e = a, expressed by one, two, etc. G's : A G, E GG,J GGG,
i = e, 0 i, u 0, a u it may be read : etc. (1).
Miles in arma fremit, prece se pedes iste redemit.
What is meant by that prece is not clear; charms are found (5) - (7) Three cryptic alphabets; the first substitutes one,
only in the final pages of the codex, and none of them two, three, etc.' dots for the vowels, the third the letters
seeIDS to fit the situation referred to in the cryptogram; immediately following after them in the alphabet (A = B,
was the cryptogram itself meant to work the magic? E ~ F, etc.); cf. notae sancti Boni/atii on fol. 8r, and the
cryptograms on this page (item (16». The second cryptic
(2) (This and the following items in vertical columns). alphabet represents each letter by that which comes imme­
A Norse fUPllrk of sixteen runes, giving also the values, diately after it in the alphabet (a IJ, b C, etc.); z
and the names. When compared with the manuscript, is ·renqeredby .AA.
Wrenn's rendering shows the following divergences:
(8) An English fuporc with forty-three runes, giving the
4~h rune] "Os (almost rubbed away)" : the manu­
names of the runes. It is obviously the result of an un­
, script' has only e for the name of this rune; 15th rune}
skillful compilation: the order of the runes is badly dis- "
" Liur ": MS. laur; last rune] "Reidr": although
turbed, names have bee~ changed, and a number of variant :
the final T is somewhat hidden by' an ornamental line~ .- ~
forms, some of which show only trifling differences, have' ~.'
there is no doubt about the reading reiiler.
b~n inserted, A detailed account follows on pp. 38 ff.
The knowledge of the Scandinavian runes displayed' in
this first fupllrk is poor: 4 has both the value and the (9) A Norse fuPllrk of a later type than (2); it gives the names
name e; n is called nou, and R reiiler instead of YT. More­ of the first sixteen runes, and the values of seven additional
over the order has been disturbed: instead of i a s t b runes, most of which are of the' punctuated' type (2),
we find i t b s a. Yet the runes theIDSelves show a fairly Cotton MS. Galba A 2 (or rather Hickes's facsimile of the
pure and archaic type, closely resembling the Gerlev runic material in ,his codex) presents a nearly identical
inscription {2}. This agrees with the name tiUT for the fuporc. Most ofthe rest Of fol. 5 v is also found in Cotton
t-rune; the same spelling is found in Leyden MS. Voss. MS.. Galba A 2. Therefor~ it will be necessary to compare
Lat. Q 83, which contains the oldest manuscript Scan­ the two manuscripts in detail before 'discussing the runes
dinavian fuPllrk. by theIDSelves (p. 37 f.).
(I) Cf. the Irish coll aT guta ' c. for the vowels', a fonn of secret ogbam:
(1) C. L; WlU!NN, lAte Old English Rune-NamIIl, 31 ff. Cf. R. A. S. MACALnlTBR, Secret Languagn, sz; and Chapter II, p. 149.
G. HtCKES, Thuaunu, Grammaticae rslandicae Rudimenta, Tab. II, nos. 6-II. (z) C. L. WlU!NN, Late Old English Rune-NamIII. 33, " ..• it [i.e. rune wJ
G. STEPHENS, MOOumenu I, loS f., nos. 31-36. is given withol1t Latin equivalent ~r name near the end of the second Norse
II. A:aNTz, Runen rmd Runennamen, 111. note I. etc. version"; this does not agree with the manuscript situation: there is a w
(2) O. VON FIuiumN, R:unoJ7Ul, 122 f. (]. BRIilNDUM-NmLSEN).
as used,in OE. writings, but it serves to indicate the value of a dotted u-rune
L. JACOBSEN-E. MOLTKE, Runeindskri/teT (Text), 292 ff., (Atlas)' Pl. 226 fI. in the ful>IIIk.

3° 31
...
 ----...
~

(10) Nemnivus's alphabet: the letters and their equivalents. each for 'c, J, g, h, k, m, ch, s, t, azathot, yrchoni and z.
This alphabet too is found in Cotton MS. Galba A 2; it MoreQver some names have been corrected: Clwritech
will be briefly discussed in Chapter II (p. 157 ff.). twice tc? Chorizech, Atathot also twice to Azathot, whilst
the puzzling pror which follows after this name is explained
(I I) An English fu~orc, giving the runes and their values.
twice by: uel pro R. This pror connects the Oxford
The first five runes of this fuporc are, found immediately
below the last Nemnivian letter, without any mark to manuscript with a Continental tradition again centeriIlg..,'
show that a new series begins; the remaining thirty-five in France (paris MS.' lat. 5239, Strasbourg MS~ ~26,
Florence MS. S. Marco 604, cf. p. 348).
runes fill the next column. This series corresponds again
to a fuporc in Cotton MS. Galba A 2, and will be discussed (IS) To the right of the last column there
' . are four rotae. The
together with the latter's version (p. 48 ff.). It seems first three are to be used in computing. the date of Easter,
to have been mixed up to some degree with Nemnivus's and are inscribed : (a) DIONISIVS; (b) VICTORIVS;
alphabet in the common ancestor of. the two versions: (c) laterculus secundum Antiochos. The fourth represents
three letters found in the oldest manuscript with this in a schematical way the different points at which the sun
fictitious alphabet (Bodley MS. Auct. F. 4.32, cf. p. 157), rises or sets in the course of the year: Ortus solis in equi­
viz. the 25th letter, et, the 31st, uuit, and the 32nd, oe, noctio; etc.
have found their way into the fuporc both in the Oxford (16) In the right bottom corner a contemporary hand added

and in the Cotton manuscript. This, and the addition eight Jines of cryptic writing :

of some variant:s, has increased the number of runes in the I. Muuilc xixilxisniiiixc ilc cxis uttximiiiixc xilidiiiixs.
fuporc to forty.
2. Cii uxiiiibixs lXpsixs mysturixi uri' nxiiiititixs.,~",
(12) A runic alphabet of a composite nature, mainly based 3· XIII XlIII XIII XVI XX IX III XX xm XVI:
upon the English fuporc. For a number of runes variants XX IX 1111 V XIX. XI V VII V XIX. V XIX
have been included; there are thirty-two runes for a - z, 4· XVIII IX 1111 XlIII III XIX IX XlIII XVII V
and five more after z. Since no values are indicated, it is XXI XIX V XIX
not always easy to decide which value a, given variant
5· : IVX IIIX XX III XX TIIX IIIX XIX X 1111 XI
may have. This alphabet is also found in Cotton MS. V IIVX ~ IIIVX V I1VX XIX XX I XIX
Galba A 2, and will be discussed together with the latter's
6. I mx IIV XI IIIVX XIX XI II V XIX I IX
version in the chapter on runic :lJ.phabets (p. 264 'ff.):
7· [IJkd nfskup· sfgfsft ,mk.cfs ektdsknkof hsbuft.
(13) A fictitious alphabet of twenty-two letters (z missing), 8. Da eralec maiu iuq iuartsnom ibit matut.
with the equivalents to the right: It is remotely related
As we shall find related or even identical cryptic systems in
to the' Egyptian' alphabet of Vatican MS. Regin. lat. 338,
other manuscripts too, those just given 'may be discussed
Munich MS. lat. 1#36 and Avranches MS. 107 (cf. p.
briefly. In the firsti six lines the principle is the same: some
274 f.). It may well belong to the same tradition as '(3).
or all the letters are indicated by Roman numerals which show
(14) The collection of alphabets ends with that of Aethicus their place in the alphab~t (0 '= 1, b = II, etc.). Lines 3-6
Ister, of which both the letters and their names are given. usethls system integrally; in 11. 1-2 only the vowels are cryptic,
It was compiled from at least two different sources : there but in the latter the difficulty is increased by writing the numerals'

are two variant letters for b, and two letters and two names backwards: in the former the whole sentence seems to be

32 7
33
 ~

written backwards. In St. Gall this cryptic device went under

Waniey gives the following description:
the name of clopfruna, at least if we may judge from St. Gall
MS. 176 (cf. p. 162 ff.). In 11. 5-6 the difficulty is again in­ Codex membranaceus & antiquus in Octavo, qui quondam fuit, ut dicitur,
peculium S. Anselmi, in quo continentur Sennones Latini numero 43.
creased by writing the numerals ,backwards, but apparently ex Augustino, Gregorio, &C. collecti. Post haec occurrunt
also by reversing the order of the words. In the seventh line I. Q\.IJEdam de Computo Ecclesiastico Saxonice.
each letter is represented by that immediately following in the II. Carmina quedam & Medicamenta Normanno-Gallice' & Latine,
alphabet (a b, etc., cf. (6», and in the last each word is litteris Saxonicis.
written backwards. A fairly contemporary hand has tried III. Medicaments contra varios _bas. Saxonice.
to unravel these cryptograms, but seems to have given up the IV. Alphabeta Runica diveraa, q\.IJE cum aliis ex hujusce Bibliothecre
Codd. MSS. descripta D.'Hickesio imprimenda dedi (I).
attempt in l. 5. On fo1. 6r (lower margin) a modem hand
(seventeenth century) has given the solutions, except for 11. 5-6. The nmic material need not have been contemporary with

It is not surprising that both attempts stranded at the same the bulk of the codex : it may well have been added at a later

point : there must be some or other mistake in the series of date. From what scanty information we have, saec. XlI 2

numerals of 1. 5. The cryptograms give the following readings: seems a fairly safe terminus ante quem. Some of the runic

I. Clauem consilii cia (?) sic committe sodali. items may of course-not to say must-have been copied from

2. cum uobis ipsis mysteria uestra notatis.

an exemplar (or from several) one or two centuries older. More­

over it is quite probable that the runic material represents

3. non quicumque uidet leget, et­

several traditions, which were brought together only at a

4. si doctior extet.
relatively late date. Nor should we forget that the runes had
5-6 OJ rei, res ut signata latebit
been inserted in three different places (2), and perhaps also at
7. [H]ic merito referes liber discrimine grates different moments. The ownership of St: Anselmus may at ~~

8. ad celare uiam qui monstraui tibi tutam. the very .most help to localize the manuscript; Wanley's cau­
A similar system of cryptic writing is used on fo1. 8r , cf. my tious " ut dicitur" shows that the manuscript itself provided
no clues (3). .
notes to the table of contents.
In order to enable us to examine the relationship between
A V and A VII were ' igne et madore carrupt(i) '. A IX· and A X 'much

the Oxford manuscript and Cotton MS. Galba A 2, an account mutilated " A XV ' only fragments '. It looks as if only a few remnants of

of the latter will be given next. Galba A II perished in 1865, the bulk having been destroyed as early as 1731.

In the Ghent copy of SMITH's Catalogus, too, the manuscript is marked in

red = 'destroyed'.
4. British Museum, Cotton MS. Galha A z (saec. XI/XII ?). (I) Catakgus. 2131. Cf. T. SMITH, Cata/ogus, 61 :
1. Diversi sennones ad populum in festis diebus habiti, & de variis
argutnentis, puta, ad Sacerdotes, de Episcopis, in dedicatione. ad
This important manuscript was destroyed by a fire at the uitaniam majorem, de nativitate DOmini, ad pamitentes, &C. F. I.
binders' in 1865; but the runic material which it contained z. Runica qUlEdam, praecipue in fine libri. 101, 1'1.7, IZ\I.

may have been lost at an earlier date (I). Here again we have. 3· Saxonicaq\.IJEdam de computo Ecclesiastico.

4· Exorcismi quidam & medicinalia; partim Latine, partiro Saxonice. 1216.

to rely upon Wanley's account and Hickes's facsimile. s· Nomina ventorum, Saxonice. 129.

(1) At any rate the Catalogtu oj the Manuscripts in the Cottonian LiiJT(uy. ('1.) This does not appear froin WANLEY's description, but the Ghent copy
Deposited in the British MtUeUm (1802), 242, seems to imply that not much of SMITH exp ressl l1 states that runic items were found on fols. 101, IZ7, 129.
can have been left of the codex: • Galba A. I.II.III. & IV. Desiderantur' ~ (3) For the origin of this manuscript it is important to note that MS.
Galba A VI, VIII, XI, XII, XIII. XIV, XVI, XVII were also missing, whilst Galba A 3, ,containing «Sermones ad populum, para secunda. Incipit a
sennone XLIV. & desinit in sennooe LXXXVI", etc. once belonged to
34
35
 In Hickes's facsimile (I) we find seven different items : the possibility that· Hickes • insulanzed' the original, as
he did for hand B in Cotton MS. Domitian A 9 (cf. p. 9).
(I) "Inscriptio Latine, litteris Runicis, ex ejusdem bibliothecae The fu~orc shows no such striking resemblance to any
Cod. MS. Galba. A. 2"; in transcription : of the other examples mentioned by Wanley and Hickes,
Ego dixi, Domini, convertere oc[u]li. Domine, con­ that we could suspect its _evidence. In this respect it is
fiteantur sacerdotes. Domine. salvum fac regem, salvum of special importance that it is quite different from th~ __
fac populum tuum, Domine. fuporc-with-names in Oxford MS. St.. John's College 17.
It is written in Norse runes, with some peculiarities show­ The latter manuscript has a number of items in common
ing that the scribe was not _completely familiar with this with Galba A 2, but on this point it must have drawn
type of writing (ON. dotted k c instead of == g ; uncial from a different source.
() = d; ON. h = g) (2). For a discussion of this fuporc, d. p. 45 ff.
Wanley does not mention this inscription in his catalogue,
(4-) _ A Norse fup~rk of twenty-two runes, giving the names

hut Hickes's statement seems to leave no place for doubt.

of the first sixteen (= the shorter alphabet), and the values

Immediately below this inscription the facsimile reads

of six additional runes. This fup~rk corresponds almost

., Alphabetum Runicum, ibidem", whilst a brace shows

perfectly to the se~ond· Scandinavian fup~rk of the Oxford

that -the indication "Alphabeta Runica ex eodem -C6i:lice

manuscript.

MS. GALBA. A. 2" (vertically to the left) applies to all

the alphabets except the last (this, a fully punctuated (5) An English fuporc, with above each rune its value, again
Scandinavian runic alphabet, was taken from -MS. Galba agreeing with the second English fuporc in the Oxford
codex. The two versions will be examined together on "~~
A. 3).
(2) A Norse fuPlirk; the usual sixteen runes, with one addi­ p. 4-8 ff,
tional, punctuated rune for g. The values are ind~cated (6) An English runic alphabet, probably based on (3), but -.
above the runes; of special interest is Norse R = Ie. For . with some additions. It is practically identical with

the first two runes the names are also given; fer (a hyper­ Oxford item (12). Discussion on p. 264- ff.

Scandinavianism due to an English scribe ?), ur._ (7) _Nemnivus's alphabet, again closely resembling the Oxford

(3) An English fuporc, with the names of the runes, in two version (item (10»._

lines: f - b, e -; z. Hickes's facsimile seems to render

-The problem of the relationship between 0 (Oxford MS.

some of the palaeographical features of the original :

St. John's College 17) and C (C.otton MS. Galba A 2) is not

insular r in porn, rad, gyr, tiT, beorc, yr, ear, qileril, ror, gar
a simple one. First of all, we must always bear in mind that

(but not in ur); fin gyfu (but not infeoh); gin gyju, hegil,
for the latter we hav.e only Hickes's testimony, and that his

gyr, sigel, -lagu, ing, dmg, gar; f.O in wyn. Yet therei,s also
facsimiles are no photographs; and, second, that the two manu­

Fontaine Abbey (T. SMITH, Catalogus, 61 : • Liber oUm S. Mat'ire de Fontibus •.)
scripts present unique material by the side of items they have

A Northern origin would make it easier to account for the wealth of Norse in common. The differences .between the versions of the

material. Cf. also N. R. KIm, Medieval Lwrariel, 142-; and p. 411· common stock are on the whole: trifling.'

(1) G.HICKES, Tlraaunu, (hammaticlU ItlmuliclU Jau;li:mmta, T.,b. VI.

The order 9f the items differs ori one point: Nemnivus's

G. STl!PHBNlI, M_mm" I, 103, nos. 12-15; no, nos. 40-42-. ­

or
(2) Thill rune for g may be an Engli8h type, ruuitely· j; else a Scandinavian alphabet stands last· in C, but between two runic items in 0 :

b was mistakenly used for g. the symbol for -the latter (i.e. ~otted k) being C (4) (5) '(6) (7) 0 (9) (II) (12) (10).

already used fot c. ­

37
36 -~
C (4): 0 (9)
Wrenn, who first edited the names of the runes, prints them' in
C has beosc (with insular s, for beorc), 0 beaT;' 0 has a
the order of the Norse runes; the manuscript presents them
<

final p-rune not in C, where it may have been illegible

in a much disturbed sequence (I). Moreover Wrenn gives
(the facsimile has three dots).
the impression that not only the names, but also the values of
C (5): 0 (II) the runes are indicated in the original (2); actually the scribe
where C has a well-made c, 0 has a form which is prac­ indicated the value of only one rune. Another hand, possibly .
tically identical with u; between the regular values of the seventeenth century, added the values of the first six'"
e and m C inserts another m, probably misled by the runes (I u th 0 T c) and corrected logu to lagu; these indications
form of the e-rune (M) ; C sund ( 1) : 0 sunt; above the have of course no value for our subject.
curious additional e (corresponding to Nemnivian elalt The fuporc consists of 43 runes, all but one having their
in Bodley MS. Auct. F. 4 . 32? Cf·. p. 158) C adds mei ds. names written to the right of them. The scribe himself seems
C (6): 0 (12) to have added the glosses uel cen and uel gylu, thus giving
C gives the value of the runes, but this is probably a late variants for the names coen and geolu. He may. also have
addition, at least if we may judge from the writing; corrected the first hmgel to .ha3gil. The handWriting shows
C also has three variants not in 0 : one for e the no special features, except that both Carolingian (geolu [3 Xl,
additional e from C (5) ?), one for m (the d-rune) and sigil, sigel [2 Xl, logu, ing [2 Xl, gylu) and insular g (ha3gel
one for s. [2 Xl, hg [2 Xl) are used (3).
C (7): 0 (10) The fuporc shows the following forms and names:
o adds a variant for s, and gives the value z omitted f.h
.f'c.en t,yfu

u.r ~OI'n'
by C; C writes the name of the 29th letter hunc, 0 hi.
The common ancestor of 'the two versions of this Nem­
nivian alphabet seems already to have been a compilation,
rnp 0'

F
'(9,,1 s(,.1 peo~ peo,.i pG'or\
"ad caen geot", geofu. geofl4.

R
h X:l8Cq,
i.
S19[1

~ ,,,"'"

r r ~
as it includes a variant for, b (no. 30, corresponding to the (Ix t)'r beorc. mech ech ec:h "';"
regular b of the Bodleian alphabet).
The above balance would place the' two manuscripts about
~ ~
rt
19914. h_.3!I hatJet ned. kale dal d-3 o+tl oe~,,1 <i.c
BMYM
de...
r~S~1hM~A;~~
on the same level, but this may be unfair to C. The form
beosc in C (4) might well correspond to beOTC of the original,
y,.

as the e~amination of Hickes's facsimile of Cotton MS. Do­ yr _, ...r Cfl.tt.r "luar- "ll&.t.r
tn, tn, z.
mitian A 9 shows. Similarly the additions in C (5) may have
crept in during the process of preparing the copy for printing. n I><: 't' rf rt 1:' ~ ~ 'f'':f
FlG·7
We have also to consider the forms of the runes, which in C
are drawn more firmly and look on the whole more true to (1) C. L. WRENN, Late OldEnglUh Rune-Names, 33. note1: " I have varied
the MS. order of the names only so far 118 to make the OR. forms stand opposite
style than those in O. Yet, in the' absence of one of the two to their NOI1le equivalents n.
parties it will perhaps be safest to assume that the two collections (2) C. L. WRENN, Late Old English Rune-Names, 32; "... I ha~e not
of runic material are partly derived from a common ancestor. reproduced the rune-symbols and their Latin equivalents, since these have,
118 it is shown above, already been given by Hickes and Stephens, and are not
A. The fuporc with the names of the runes in Oxford MS. of ·special interest In themselves". .
St. John's €ollege 17 (item (8» is one of the most puzzling. (3) In thc first Scandinavian fupllrk Iwgol is written with insular g, in the
5e(;()nd with Caiolingean g.
38
... 39
 At' first sight one is struck by an amazing wealth of variant as the seventh rune, but the next, w, was missing in his
runes: three variants are given for the names geo/a, sigel (-il), own fuporc, and the order of the rest badly upset. This
peorll, t[IUlT; two each for ech, halgel, daJg, oeiel, yr, ing. But may have discouraged him from continuing his effort.
closer inspection shows that this wealth is only apparent; it Yet one may ask why he did not copy the correct fuporc
covers some appaling lacunae. The three runes called geo/a all over, if his own proved so poor.
are actually g, g and j; one of the s-runes is probably a minus­ g: the second variant for g shows a unique, highly stylized--'
cule S; the three p-runes are actually hardly differentiated; forin. .
and more such flaws will appear in the discussion of the details.
s: the first s has the normal form; the second is the rarer
The order of the runes is much disturbed, although it is variant of St. Cuthbert's coffin and the Thames scramasax;
possible to distinguish sequences retaining the original order : the third is either a variant of the second or, more pro­
f u 1.J 0 reg... t b e m 1... d (£ a ~ yea.. But it is difficult bably, an insular type of minuscule s. The name appears
to account for the way the other runes have been shifted around. twice as sigel and 'once as,'ngil. The latter, which is the
That the runes g and j should have been inserted after g is older form, has a puzzling parallel in' the first hagel, cor­
only what we could expect after the names of the former two rected to hagil (q. v.).
(gar, gear) had been dropped in favour of geo/a. But elsewhere p: the first two runes differ only by their lower lateral stroke,
there are no traces of ~uch planning, only a couple of remains pointed in the first, rounded in the second; the third is
of the old order (.. p x ... h n .. ). I
identiCal with the second,., but has an obviously spurious
This fuporc calls then for discussion on a good many points : addition to the left of the vertical shaft.
.>t.~1-
c: the name coen is glossed cen. The former is either an old t: the name tyT is not found elsewhere with this spelling;
~
or a Northern form (umlaut of *ko'll.i- (1» whilst the latter it may be due to Scandinavian' influence (cf. tyr in the
is the regular Southern form. sel,':ond Scandinavian fUPllrk, item (9).
g: this rune shows the same opposition between geoju and e: the compiler was evidently misled by the form of e and
, gyju, though here the relation between. the 'two forms is altered the' name ech to 1IJ.ech. The second e-rune, which
not so clear. The gloss uel gyju is added only to the first looks somewhat like a Y, is found in only one more manu­
(i. e. the original) g-rune, but this need not imply a better script (in the 'Alanie' alphabet of Munich MS. LH36).
insight into runic matters' with the' glossator; . his gloss A connexion between these two concurrences is extremely
probably applied to the two other geoju-forms as well. doubtful; it is also hardly· possible to consider this form
The two glosses correspond exactly to the forms found in C. as runic.
The conclusion must obviously be that the compiler of 0 m : the tn-rune has been interpreted as a variant of the e-rune
tried to correct the fuporc-with-names which he had on account of the formal resemblance::
already completed, with the help of a version' closely
related to if not identical with that in C. ,He got as far 1: the scribe's loga was probably corrected to lagu by the
hand that added the values of the ·first six runes (seven­
teenth century or later). .
(1) A,t this pOint it does of cow:se not matter whether this is actually the i •

aw:n
original fonn of the name of c; we can only compare the fonn com with h: if we knew to whom the correction halgel > hagil is due,
proper names l1li Comred (H. STROM, Old E",1is1a Perstmal Nama, 144)' and
suppose that the scribe attached some similar meaning to the IUlle-name, if any. we might b~ able to throw some light on the structure of
·41
40
 B 10 and Galba A 2. The second is an x-rune; exactly
this fuporc. But a single i is too' little to decide who
the same form occurs as no. 17, with the name ilx. The
made the correction. A form in -il (which is the older
transfer x > n may be due to an intermediary misled by
of the two) is also found in Galba A 2, but there by the
the acrostic principle ('Y = i[lxJ ing).
side of sigel. not sigil. .z: the z-rune seems to be ea with two short strokes added

3: the second h-rune is of course no other than the rune to .the left of the vertical shaft. The same rune (without

which is normally called eoh or ih. additions) is also used for z in the is-runa fuporc and the

k: between the a and the I of the name kulc, a letter (c?) De inventione. alphabet.

seems to have been erased.

The evidence provided by the Oxford fuporc with rune-names

is not as valuable as Wrenn assumed (I); it is perhaps superior

d: the second d-rune is simply the normal type, in which the

to that provided by scribe B of Domitian A 9, but decidedly

transecting strokes have been lengthened left and right

inferior to that of A's part in that same manuscript. Its North­

of the vertical shafts. ern origin is shown by the rune-names feh, coen, ech

<2 : the second m-rune is that of the Thames scramasax, also

(unless eoh), oepel. The name coen may even be hyper­

found in other manuscripts.

Anglian : whereas the rune ..name cen is supposed either to go

y: the first y-rune is a slightly rounded u-rune without sub­ back to *kaun or to have a WGmc. e, here it must have been

mixed up with cen(e). ('tkOnja-) , cf. Coen- in proper names.

script i or y; the second is without parallel (1).

. Sievers-Brunner declare geofu a late WS. form (2), but both
ea: this is the only rune for which the scribe indicated the the status of the stem *gelj· and the relationship of the various
value, apparently without special reason. OE. forms are not quite clear (cf. e. g. p. 244). On account ,~

q: the first g-rune is derived from p by turning the lower of the -il suffixes, the first draft of this fuporc can hardly have
lateral stroke to the left instead of to the right. The been posterior to the eighth century (3). "
same type is found with the value p in the isruna fu}lorcs, ' The important Norse, ingredients of the collection agree of
but there the original p-rune has taken the place of g
(cf. p. 124). The second form seems to be a variant of· (1) C. L. WlUlNN, Late Old English Rune-Namel, 34: "Enough hill! now

the first, whilst the third looks like a combination of a been said to indicate that MS. 17 of St. John'~ College, Oxford. is deserving

of more consideril~on than it has hitherto received from those interested in

g-rune as found in Domitian A 9 with a calc-rune. The runes, whether Norse or English; and it may well prove that it should supplant

name guaT is unique. It may perhaps. be compared with MS. Cotton Domitimt. A. IX in its position of best and earliest exemplar

quaT in Munich MS. lat. 14436 (' Arabic' g). <>f the la~r OE. fupm now actually I'.xiant ". When WRENN (po 33) states

that, "except where otherwise noted in the above list, his [i.e. the scribe's]

( ?) : the 40th symbol is hardly runic, unless we interpret it names correspond accurately to his symbols", the following cases should be

+
as e t = et, cf. the symbol following z in the runic alphabet added to the exceptions noted by him : geofu as the name of i and j: meeh as the

name of e; ;ng as the name of x. The spurious runes for I, p, e, d, y and q

of this same manuscript (item 12, cf. p. 266). should also be kept in mind. We shall see, however, that the fupore without

names in 0 and C, and that with the names in C, provide as good evidence

1]: of the two runes with the name ing the first is the somewhat as any other version written in Engla,nd~

sophisticated type of 1J also found in Cotton MSS. Otho (a) E. SIBVI!RS-K. BRUNNER, Altengli.sche Grmnmatik, § 91. note 8; § 1 II,

note 8; geofu ~ repeatedly in the Durham Liber Vitae and may consequently

he claimed for Notthumbrian as well.

(I) In a few, runic alphabets we find a p-rune resembl~ this y; but this
(3) E. SmvERS-K. BRUNNM, Ait.e1rglische Grammatik, § ISa.
resemblance will be purely coincidental.
43
42
course well with a Northern origin. Wrenn would eVf:n go used sporadically at a late date (I); but it is hardly possible
so far as to ascribe the compilation to a Dane. His main to connect those· inscriptions with our alphabets. Altho)lgh
argument is the absence of the 'W-rune from the English fuporc. some of the Norse material is .older than the eleventh century,
Since this rune did not occur in the Norse fUPllrk, the compiler this revival of the interest in runes may well be connected
would have omitted it. But an Anglo-Saxon 'W (i. e. originally with the active part played by Englishmen in the Christian­
a w-rune) is used correctly to indicate the vaIue of the dotted ization of.the North; not only did many work in Denmark,
u-rune in the second Norse fUPllfk; so its omission from the Norway and Sweden, but a number of churchmen from these"
English fuporc will rather be accidental, the more so as we see countries studied in England and were ordained there (a source
that the compiler upsets the order of this fuporc precisely of conflict with the see of Bremen, to which the conversion
where the w-rune should have come in. Wrenn might rather of Northern Europe had been officially entrusted). Perhaps
have called attention to the Norse name of the t-rune, tyr (in Wanley's remark on the origin of Cotton MS~ Galba A 2 may
the first Norse fUPlirk tiur, in the second tyr). But this form be remembered here : St. Anselm, to· whom that manuscript
may have crept int.o the English list from a Norse fuPlirk. On is said to have belonged, was one of the foremost promotors
the whole the evidence for a Norse compiler is rather weak.. of the' Northern mission.
At any rate the Oxford collection of alphabets itself could The ancestry of the Oxford collection consists at least of
hardly be considered to provide such evidence : a number of three branches: (I) the runic collection (including Nemnivus's
mistakes prove that it has only second or third hand information alphabet) als() found·in Galba A 2; (2) the English fuporc with
on the Norse runes (nau; reiiler for the yr-rune; beor). Yet rune-names; (3) the Continental additions. Whether the notae
this Norse material shows clearly to what extent ' Danes' and sancti Bani/atii belonged to one of these branches, or should
, Saxons' had merged. The Northern origin of the English be set up as an independent fourth, I cannot decide. We may
fuporc with rune-names agrees well with this cultural frame. summarize this in a stemma :
It is not impossible that the interest in the English runes, of , ~-~

which the Oxford manuscript is proof, was an indirect result Runic Collection FuPoI'C with Continental collection
of the Scandinavian invasions, or of English missionary work
in Scandinavia (cf. infra).
/~ m~~m.
Equally important are the connexions of the Oxford manu­
script with the Continent. Three of the alphabets on fol. 5v .Galba' A 2 . - - - - - > Oxford St. John's < --Nome

(iteIris 3: 'Chaldaeo-Assyrian', i3: 'Egyptian', .and· 14: College 17. Bonifatii ?

Aethicus Ister's alphabet) seem to \have enjoyed considerable

popularity in some Continental (especially French) scriptoria. The fuporc without rune-names (item (II» will be discussed
There they were usually added to collections of, alphabets of together with the corresponding fuporc of Galba A 2 on p. 48 ff.
the three sacred languages, or at least of Hebrew and Greek.
They probably reached England as part of such a collection, B. We now come to ~he fuporc with rune.-names of Cotton
but the Oxford compiler left out the 'classical ' material, and MS. Galba A 2. It will be remembered that it is known only.
by various additions obtained a fine collection of 'native' from a facsimile in Hickes's. Thesaurus, and therefore the
alphabets. Whether all the alphabets were regarded as curiosa­ remarks on p: 9 fr. must be kept in mind. These are the runes
by the time the manuscript was written we cannot know for
sure; for, in the North of England at least, Norse runes were (1) E.g. Oil the Bridekitk font. (Cumberland, saec. XII).

44 45
and rune-names given by Hickes's facsimile of the fuporc in Otho B 10 (cf. p. 22). The name, too, must have been
partly illegible. In the facsimile the letters peo and hare
Cotton MS. Galba A 2 :
clear; but between peo and h there is some space, and to
feoh uri porn os rad cen 9yfu wyn. hegit the left Qf h, and quite close to that letter, a stroke resem­

~ Il ~ ~ Rh X~ ~
bling i. There can be no doubt that the engraving is an
attempt to render the state of the manuscript. As to the.
ned is 9yr et.h ~o Ih eolhx sigel tir beorc original reading, it was probably pear} with insulk"r:'

/' h if .11 t
Of that r only the short right hand vertical stroke remained;
l 4> &
of } the lower part was missing, leaving what looked

c.ok man lagu m9 dc£9 e~el e~el zsc like h to the copyist or the engraver.
(U:.
x: the form of the rune is stilisti~ally unique; cf. k.
M M t ~ H·.~ ? ~ ~ Q: the rune shows the rather sophisticated type also found

yr ear c.a.lc. 9uer~ lor stan 9ar z in Cotton MS. ·OthoB 10 and in the Oxford manu­

h1 'f rhJ, ~ * ~
FIG. 8
~ ~
The runes are well drawn, with neat, firm strokes; they
O!:
script. The name mg is of course a mistake for ing.
this is the only rune for which a variant form is given.
The variant is of the type known from the Thames SCra­
masax and from a couple of manuscripts.. In these in­
present no forms which point to a long manuscript ancestry. stances, however, the vertical shaft reaches at least as high
Only two runes, rand j, a,ppear with rounded forms. Similarly as the upper angle of the quadrangle.
the names give the impression of belonging to a trustworthy y: in the rune the subscript mark has been interpreted as y '~ "'~
tradition. The facsimile seems to follow the original closely : (hence the dot over the y-like subscript).
insular r is used everywhere except in uri only .3. occurs. A
k: shows~he same squarish style as x. To the right of the
couple of puzzles may be due to accidents in the process of
copying and engraving, perhaps also to a somewhat defective rune there is a spurious stroke.
manuscript. The following remarks apply to peculiarities in q: as usually, the symbol for q is derived from thep-rune;
this fuporc : the type found here only returns in the runic alphabet of
3: the rune is inclined to the right and a short stroke has been this manuscript and in its twin of the Oxford codex.
added .to the lower left of its (originally vertical) shaft. st: a st-rune made up of two overlapping triangles is also
This stroke must have been there in the original, for it found in the runic alphabet. It does not occur in English
turns up a,gain in the runic alphabet (item 6) and in the inscriptions, but is recorded from Westeremden (inscrip­
related alphabet. of the Oxford manuscript. The name tion B) (I).
eth is an error for ech or for eoh; in view of feoh the latter z: the place of % is filled by a fanciful Roman Z.
is probably the originid form. It is not impossible that
precisely at this point the manuscript presented some As far as the names provide linguistic criteria, this fuporc
difficulty, d. the next rune. . must be loc!J,ted somewhere. in the South: feoh, eoh, beare.
The evidenc~ is not altogether clear : a form querll· by. the side
p: part of' the upper lateral stroke of the rune is missing;
what is left reminds us somewhat of the p in Cotton MS. .(I) H. A.m4-rz-H. ZElSIl. RunentIenkmtikr, 394 f.

+6 +7
of peoTi! may perhaps reflect a special treatment of the vowel Q: apart from iu (if == iw), inc is the only rune-name written
following fJ) (I). The name gyr perhaps' allows to narrow in full (sunt! uult are no rune-names, cf. infTa).
down ~e area where the fuporc was written. In Kent y d: the values of two pairs of runes seem to indicate some sort
became e (cf. wyn ~ wen); but here, where we should expect of opposition: p = d, d = ddj g = g, j = gg. From the
ger, we find a form gyri this is probably an instance of reverse point of view of Old English, the double spellings are
spelling not unknown from Kentish manuscripts (2). Therefore quite obscure. It is of course tempting to call attenti,oft~ .
we may assign the fuporc to Kent or a neighbouring region. to Wimmer's explanation of the d-rune as being formed
The forms ned and hegil agree well with this supposition. The 9f two p-runes. But, first, we cannot suppose this genetic
latter may imply that we must date the fuporc _fairly early, feature to be reflected in a fuporc not older than the
perhaps before A.D. 800. eighth century; and, second, this certainly will not help
C. The survey of the fuporcs written in England will now
us, to explain the opposition g : gg. Unless we simply
be completed with the examination of the fuporc without rune­
explain the two double spellings as errors, we have to
names found both in Oxford MS. St. John's College 17 and
find some justification for this curious phenomenon. As
Cotton MS. Galba A 2.
far as I see, it is not possible to account for it within the
The differences between the two versions were discussed
frame of Anglo-Saxon spelling habits. _Double dd and
on p. 37 f.; they undoubtedly allow to derive the two versions
(occasionally) gg do occur, but only to render geminated
from a common ancestor. That ancestor probably had the
d and g(the latter is usually spelled cg); there seems to be
following runes with their equivalents :
no ground for interpreting. the runes d and j as geminates.
f u d 0 r c 13 uu h n i 99 x' p (u .$ t; b e m
The opposition between d and dd in our fuporc is that
fAr ~ R tl Xf ~ l' -I +j' ~; 't' '1 t .& MM between a dental spirant and a dental stop. The same
opposition is expressed in the same way in Old Irish
.~;

**
line dd oe 6.

~ H ~ r; F ft..1~Xf
ill

F~G.
a y 1;!6.

~ fh~~ 1< BSD{1~

<3 9 q sunt k oe ... uult
9 .
s e ... manuscripts (e. g. the St. Gall. glosses); these also use a
parallel device : g for a guttural spirant, gg for the corres­
ponding stop. This. however, does not seem t~ agree
with_ the use' of these spellings in our fuporc •. where the
A number of details require special mention :
situation seems to be exactly reversed: g for the stop.
p: the form of the rune is almost identical with that of the gg for the spirant. To be sure, the stop vs. spirant char­
w-rune; on the value d, see under d. acter of the OE. gutturals is still a matter of discussion.
3, x: the values of these two runes seem to have been inter­ As a result of his analytical studies Sievers believed that pal­
changed: the 3-rune is equated to .x, the ,x-rune to iU. atal g, as well as velar,' was an occlusive in the older texts.
This iu is either an old, 'esp. Northern spelling for w, On this Luick (I) based his hypothesis that initial palatal
corresponding to West Saxon eo, or else a Northern spelling g andj fell together in OE. lro/, at least in the older period ..
for ifIJ; in the latter case we have an extremely interesting If this were the case, the spelling gg for original Gmc. j
equivalent for usual eoh, th (3)· would become somewhat less surprising; but that would
j: on the equivalent ge, see under d. leave single g for theg-rune unexplained. There~ore,
it will stilI be safer to assume some amount of Irish in­
(1) Cf. E. S~.K. BR'IlNNER. AltenB1ucM Granmuztik, § II) A. 2.).
(a) E. S~-K. Bll'llNNER. AltenBwcM GrammatiJc, § 31 A. I.'
(J) E. SIBVBlIS-K. 'BtnlNmlR, AlterrgwcM GrammatiJc, U 2 34 A. 3, 25 0 A. 2.
(I) K. LufCK, Historische Gram'l'l'ltltik, § 633.

8 49
48 "
 fluence in the prototype of this fuporc. Double spellings
to the left of the vertical shaft, probably a spurious addition.
do occur fairly frequently in the English inscriptions
(Ruthwell: double 'tJ, t and i; Bewcastle: double S; g: this rune is represented by its epigraphical type (manu­
Bingley: double d; Hartlepool: double d; Brunswick . scripts usually have a square crossed by two transecting
casket : twice double. i; Mortain : double 0, II: and twice strokes). It is followed by a regular g in a slightly decadent
double i). If the opposition gg : g remains obscure, we form, perhaps meant to gloss the g.
must not forget that there may be a two or three centuries' q: as Usually, the q-rune shows a form derived from -til;
distance between the prototype and the copies that sur­ p-rune.
vived. Cf. also 3 : x.
st: the name sunt is probably due to the interpretation of It
(E: an OE. re-rune of this type is unique; the same form
appears with the value io in several fuporcs, and for g in as Ii sunt. In C the last letter looks rather like a d.
No. 38 is a variant form with the value s.
a number of alphabets. Formally it is identical with j.
This pseudo-re-rune, however, is followed immediately k: a short horizontal stroke (probably spuriou~) has been
by a symbol which is obviously a slightly modified (E. added to the vertical shaft.
The latter is given the value a. One might suppose that
(uult) : this letter does riot belong here; it is the 32nd letter
oe is a mistake for eo, cf. eor for the same rune in Munich of Nemnivus's alphabet.
MS. 14436 (' Arabic' e), but this leaves the equivalent
a for the real (E unexplained. To explain this a as a e?: this pseudo-rtlne is quite puzzling, and still more' so are
second mistake (for oe or e) is to add to the difficulty. It the words written above it inC: mei d'S- mei Deus?
is much more probable that the -equivalents have been The < rune' could perhaps go back to an m-rune l><J, to
.~~~
interchanged., or rather, that the two runes in question which the value e was given, as in the fuporc with rune­
~
changed places: thus the original would have had : 5C oe, names oLO.
* a. The latter is a Norse type of a and pointS to an
early date (not later than ca. 900) (I). We must not look
From this discussion of details we may conclude,:
very far for the reason why this a-rune was added: it (I) on account of the spellings uu w) and iu iw or io)
was probably a gloss to the ac-rune, the first rune after the fuporc will be much older than 0; this is corroborated
the re-rune. At first it may have been written by the side by th~ forms of the runes, which show hardly' any signs
or above this ac·rune, and one or other copyist will have of decadence, or fanciful distortions. To be sure, neither
inserted it into the fuporc. A correct (E is given as no. 35 o nor Hickes's facsimile show typically old features in
(value oe). the handwriting of the values; insular g does not even
occur. Yet, if the retention of old features can be an
a: the a-rune is practically identical in !\hape with the p-rune.
argument for an early date, a 'modern' handwriting
especially in O.
olE: the rune is poorly drawn, so that. it 10(lks rather like F.
A better olE follows farther down (no. 36), and there is a
If the explanation of *
cannot be an argument against early dating of the original.
as the Norse a-rune is right, the
fuporc with this addition cannot be later than the .ninth.
third symbol with this value at the end of the fuporc. century. !
This .last is a regular type with a slanting stroke added
(2) Several of the values are puzzling or simply misplaced.
(I) O. VON F'amsJIN, Runoma +5, 14 C., 140. I¢. It is nbt impossible that the values did not belong to the
50
51
 ~

fuporc from the very beginning, or that they Were crowded

in in such a way as to lead to errors. The Codex formerly belonged to the Cathedral Library in
(3) The fuporc seems to have reached its present state by Salzburg; in the catalogue of 1433 it was no. 160. At the
degree~ : beginning of last century it was removed to Vienna with other
(a) the fuporc till the calc-rune; Salzburg manuscripts; in the Hofbibliothek it was first known
*
(b) additions or ' glosses' : for the a-, X for the g-rune;
as Codex Salisburgensis LXXI, then id. 140, and also under .
no. 35 2 7. It is still often referred to as the ' Salzburg Alcuin::.:-­
(c) corrections; the second a!- and the second m-rune; manuscript " although there is no proof that it was written
(d) accidental intrusions from Nemnivus's alphabet. in Salzburg.
The runes (and the Gothic material) were first communicated

5. Vienna, Osterreichische Nationalbibliothek, MS. 795 to W. Grimm by the historian G. H. Peru and again by the

(saec. IX in., X). librarian B. Kopitar. The fullest account of the manu$Cript

is due to T. Sickel.
No manuscript with runes has received so much attention Plain white vellum binding (17th century). Parchment of varying quality,

as this Vienna codex, and for various reasons : it contains an well Preserved. The codex is made up of z05 folios = 30 gatherings,of which

important collection of Alcuin's letters, compiled during his the 5th to the z7th are marked 'il' to' z' (- fols. ZI- 19 1): these form the

centnl part of the codex; four quires were added in front and three at the end :

lifetime and perhaps by no one less than his pupil Am, then
Z I [1-4] +3 IV [54] + IV (-I) [Z9-35] + Z IV [3 6-5 1] +IV (':1)

ArchtJishop of Salzburg; in it is preserved some fascinating [5z-58] + Z IV [59-14-] + IV (+ I) [75-83] +.IV [84-91] + III [92-97] +

Gothic material (1). At one time this G1:lthic material' played II [gS-IOI] + 4 IV [loa- I 33] + III [134-139] + 5 IV [14D- 179] + II

a: very important part in runological studies, but nowadays

18
. [180- 3] + IV [ 184- 191]
[200-Z05].
+ III [19Z-I97] 4- I [198-199] + III
~,
scholars rather tend to adopt a more reserved attitude; we shall
Format 230 X 137 nun (written area 197 X 100) one column except on

see there are some grounds for scepticism. fuls, 5-:110: two. z3-35 lines to the page. Written by several hands, to which

Sickel assigned the following parts: (I: ZIT_I6:11r,176l'_I78r, 184r-19O" and,

(i) FROBENIUS FORSTER, Conspectw omniu",. opmDn Akuini. Ratisbowe, r

1.,00, 6, IS. . together With )',I"-4 ; fJ: 16z"-17I l' ; ,.: I7Z1'-176", 17S"-I83v, I!)ZT_ 99r and, .
1
togetherwith l¥, 11'-41'; 'I: 51'-zo"; C: :ll00'-Z03": 8 made additions and corrections

. ID., Opera Alt:rmri. Ratisborue, 1777 = J. P. MIGKII, Patrologia lat. CI,

901 C if. ' aU through the manuscript (I). Chroust further distinguished three

T l'
H. G. Pmm:, Archlv der Gesellschaft fUr iIltere deutsche Geschichtskunde 3 scribes (I: (II :-= ZI -58 , 83 r :lf.; ( I ' == 59T.75"; (I' =- 76T-8:11".;he does not

(I8zI),604. ' . , state which (I was at work in I V -41' (I). Several of these hands show a mlU'ked

. ID., ltalilmilche RBise, 46z.

W. GRIMM, Zur Litkratur d8r Runm, I if. = Kl. Schriften III, 85 if. In., Dk tlfIgW4t:/uischett~, 5.

H. MAs!ll\f.tWl!:!•. Gotthica Minora. Z. (. d. A,., ~ (1841), zg6 if. W; LUFr, Studien, esp. 76:1f.·

R. VON LILmNCRON, Zur R.u:tumIeIwe, 7 f. A. CHitoUST, . MomtmerJta pa/tuogro.phica I, vii,3 ("Ants Sammelhand_
A. KmCHHoJIF, DIU gothische .R.rmerwJphObet, %Oif. 8Chrift "). .
l. ZACHER, DIU gothiscM AlphObet, I if. S. BUGGE, Norg. Indskrijur -<Ide t:eldre IWner. Indiedning, 4Z if.
P. l~, M(I1IfUMIIto Alcuiniana (Bibliotheca Rerum Gennanicarum' VI). W. S1'Rlit1'BllRG, GotUches ElemtmtaTbuch, 36• ...0.

Berolini, r873, 137. '' , , ID., Dk gotUche Bibel. XXX.

T. SICtaa., AlcuinstTulien. Sitzungsberichte der philosophisch-historischen O. VON FRI:!!sEN-A. GRAPE, Om Code:!: Argmtew, dels tid, Iwm om 6den.
Classe der kaiserlichen Akademie der Wissenschaften. LXXIX. Wi.en, 1875, Med ett appendix av H. ANnimssoN (Skrifter utgivna av .8venska litteratur.
~a ' . silllskaPet :117). UppsaJa, 19:iS, 1:115:1f. • .
K. Fou, G.chichte tier SabilmrgtJr Bibliotheken. Wien, 1877, 8. G. BAi1sBcR:B, A1irogtms, 156 fl.
T. VON G~BRGER, Die gertna'llisc'- n.qremrmne1l1. Die gothischtm J. BLOMJI:i:m:.D,.Runes, esp. 209 if.

~~. (H T. SICRBI., AlcuirutiuJien. 471 fl.

(z) A. CmtOus'r, Mimurnenta fJobNographica I, vii, 3.

52
"" 53
 ~

insular inftuence, not only in the abbreviations (especially fl), but !llso in the s· (no title) Orthogmphical treatise, sometimes attributed to Alcuin (I).
ductus (<x'). 19'v FO'f1IJQe liturarum ftICUruium GrtlllCOS (a Greek alphabet).
19 A Greek syllabary; Roman numerals.
Sickel dated the bulk of the codex in 798 or sbortly after,
=' An English fu~orc; an incomplete Gothic alphabet.
and connected it with the name of Am: the collection of In the lower margin a cryptogram.
Alcuiniana would have been made in France at the order of :wY Gothic alphabets, etc.
Am, former Abbot of the Abbey of St. Amand, and Bishop, 2Ir Augustine, Quaestionu i1f EfJangeJiis Matthaei et L'UCQe.
59' AUgustine, Seruw in Epistola Pauli ad Roma.rws.
afterwards Archbishop of Salzburg (I). In the spring of that 84' ExCerpts from patristic woIb Uerome. Augustine, etc.).
year Arn, coming from Rome where he had received the pallium, 148r Alcuin, on the Epistle to the Corinthians.

stayed for some time in the North of France, as he ,had planned 149' AlcUin (?), on the Epistle to the Ephesians.

v
I49 Alcuin (?), on, the Epistle to Titus.

to meet Charlemagne before returning to his see. According 150v Alcuin, Letters. '
to Sickel fols. 1-20 and 150 ff. were written during or imme­ I~· Notitia ecclesianun twbis Romoe.
192" Alcuin, Letters.

diately after this journey. It appears, however, that Sickel

I97v Angilbert, Letters.

disregarded some chronological differences. When we compare r

199 A!cuin, Letter to Candidus.
the handwriting of th~ runic and Gothic material with e. g. zoor Charlemagne,. Letter to Alcuin.
that in the facsimiles given by Chroust, it becomes obvious
Fol. 2oJ'shows the following arrangement (cf. plate IVa);
that the former may well be a century younger. Therefore
I believe with von Grienberger (who does not seem to have (1) An English.fuporc of twenty-eight runes is written in two
known Sickel's study) (2) and Dr. F. Unterkircher; who kindly columns, f - s, t - y. Above each rune its value has
examined the manuscript for me, that the text on fols. 19 and been insc;lrted; to the right of each is its name. In one
20 is a later, tenth century addition (3). This is of the utmost case two names are given : f, feek, fe. .
'-{'l­
importance for the appreciation of the runes and the Gotkiea. (2) . Below the y-rune we find the key to one of the forms of "

Contents: (4) thenotae Bonifatii: a is followed by one dot, e by two, etc. .1- ~:
"
fol. Ir Table of contents (14th century) (5).
(J) To the right of the runes t - (2 part Of a Gothic alphabet
1· Aleum, Two letters to Am.
has been added (a to u, in the original order). To the
(1) On cultural relations between St. Amand and Salzburg see e.g.
right of the Gothic b and g there is an erasure; the j, too,
B. BtlICHOFF, Schreiblchulim I, 1#, 265. seems to be written over an erasure. In. the right top
(:a) Most scholars discuSsing either the OE. or the Gothic material do not corner,' above the number of the folio (' 20 '), there is
seem to have known Sickel's study; Miss BLOMF'IJtLD is a rare exception.
(3) G. BAliSECKIl, Abrogans, 156, considera the Orthographiabrevis originally
another Gothic j. The right half of the page is blank.
to have been a separate manuscript: .. [ ..•. Cod. Salisb. 795 'in Wien]. Er (4) In the lower margin, partly' cut away, a cryptogram:
enthilt an erater Stelle Briefe Alcuins an Erzmschof Am von Salzburg, den
Schiller ArbeoS. Ihnen folgt, urspriinglich eine selbstiindige Handscbrift,
die Aleum zugeschriebene Orthographia breuis... " (1) The-text of this treatise is no doubt closely related to that ascribed to
(4) Tabuloe I, 134. Alcuin, but it is not identical with the text printed by KEIL (Gram"ia1ici
(5) T. VON GRlENBBRGER, Die g~ nmenllillrUlll I, 185 f.: "Hic Latini VII, 295-312; KmL did not use the Vienna manuscript). FaoBl!NltJs
sunt rescripti libri in hoc volumine contenti I Primo Epistole quedam albini I ascribed the treatise in this codex to ALcmN: "Nos, cUm vetustiasimum
Secundo Orthogmuia breWs! Item questiones Ewangeliorum Augustini I Item codicem Ina. bibliothCCiE illustrissimi metropolitani capituli Salisburgensis,
Augustinus super EpistolaJ ad Romanos I Item questiones diuerse doctorum numero 71 subnot\ltum, et circa initiuIn sec. IX exaratum evolventes incideri.
super epistolas pauli lItem Excerpts de diuersis trsctatibua I asnctorum mus in opWlcul~ orthograPtricum. inter genWnas epistoIas Alcuini' ibi
primum super EpistolaJ ad Romanos I Item Explanacio G1adiorum qui descriptas comprebensum, opinari C(epimus idipsum ease opusculum beati
dicuntur I in passione domini I Item Epistole karoli ad Albinum," Alcuinihucusq~e desideratum ". (MIGNII, Patrologis lat. CI, col. 901 C).
54
S5
 ·V. indirectly from Alcuin or his circle (1). To be sure, the
XX. I. XI. V.1. XVIII. XX. VIllI. VII. I . XVIII. XV.XVII. 1. V . ,manuscript contains material which would be most readily
XX available in AIcuin's immediate neighbourhood; Sickel's recon­
XVIII. XI .I . XII. I. XVIIII. [V]. struction ·of the circumstances is on the whole convincing. But
i. e. ualeas uigeas praaul amate. This device is called chlophruna what is true for the texts written in 798 or soon after, does not
by Ekkehart IV of St. Gall (cf. p. 162 ff.). necessariiy apply to additions made a century or more lat!!_r.,
Fol. 20 V is wholly devoted to Gothic material; of main

Therefore it willbe safe to forget the hypotheticalconneiion

with Alcuin for the time being, and to look for internal eVidence.
interest to us are :
Although the Gothic material falls practically beyond the scope

(I) A Gothic alphabet with the names of the letters, arranged

of this work, a few words about its meaning can only help us

approximately in the order of the Latin alphabet (a b g d

..better to understand the fuporc. von Grienberger believed
e jj h i kim n up q T S t fJJ U :JC!41v pl. For several letters
that fo1. 20 V reflected the collaboration of a native Goth and a
variant forms are included (one for ad e f hilT S p, two
German whose native dialect showed the consonant shift;
for II and u). The formsof most letters in this alphabet
authors taking a less extreme position· still thought information
come very close to the 'classical' type of the Biblical
proVided by Goths (e. g. t,hose settled in the South of France)
manuscripts. had been incorporated here (2). As a matter of fact two sets
of data must be kept apart :.
(2) A cursive Gothic (?) alphabet in the original order (the
symbol for '90 ' has been omitted, and so has that for '(a) those derived from a Biblical manuscript i. e. the ex­
tracts with transliteration (or translation), the notes on
'900 '). ..\)1,,:­
pronunciation and part of the alphabets; "">

(3) Gothic excerpts with transliteration or translation into (b) ·those !ierivedfrom another source, perhaps from oral

OHG. tradition : the names of the letters, perhaps also the cursive

(+) Remarks on the pronunciation of Gothic (the Gothic words· alphabet.

in Gothic script): I do not doubt that a German scholar would have been able

ubi dicituT genuit j ponitur (I) to derive the information of the first order from a manuscript

ubi gabriel g ponunt & alia his similia such as the Codex Carolinus fragments in Wolfenbuttel

ubi aspirationem ut dicitur

jab libeda jab libaida
(I) T. VON GBmNmmGBR was aware .ofchronological differences in the

material of (ols. I-20: fola. 5-zo", including the IUrles, must be dated in the

diptongon ai pio e longa

tenth centurY; the Gothic material would be a still later addition. Yet

PTO ch q ponunt.
VON G~ not only attributed the runes to .ALcu:tN (Die 'genMmscJum

~ I, 187: " Da nun die vorgebundenen '" blltter briefe A1chwiries

This runic and Gothic material has always been interpreted enthalten unci die Orthographia brevis se1bst diesem ge1ehrten Angelsachsen

on the tacit or explicit assumption" that it came directly or zugeschrieben w:ird, 80 isi es ~ ~einlich, we auch die runemeihe

auf mitteUungen A1chwines berube, urn 80 l1lehr, als ibn-: namen, wie sich

(I) The words uhi tlidtur partially hide a Gothic word beginning with a, zeigen winl, northwnbrl8chen cbarakter.besitzen, etc. ")j the Gothic alphabets

most of which bas been wiped out. It looks 118 if the compiler fust intended too would go bad!: to him. '

to gO on with these Gothic extracts and interlinear esplanatiolll, but became (z) O. VON Fim!sEN-A. GRAPH, Om Codu ArgentllfU, IZ5 if.
aware of the inadequacy of the latter, 'and substituted a few' rules' he had G. 1hEsBCKlI, fibrogatU, 187. suggests the presence of Goths in the Salzburg

discoveIed thrOugh his comparative Gothic"'()HG. study. . ares.

S6 57
(Gothic-Latin bilingual) (1), especially if he knew the Greek·
. We can now proceed to the study of the fuporc. The forms
alphabet. Such an inductive study woUld explain the equation
-of the runes show on the whole a remarkable degree of ' runic
OHG. ch = Gothic q. and some awkward renderings in the
style'; they were obviously copied very carefully, though not
interlinear to the Gothic excerpts. The confusion of sand z is always with much understanding, as we shall see :
also significant. .
As to the letter-names, at one time they were considered as f f& . v cf o· r c 9' Uu

genuine Gothic forms of the rune-names. But this assumption rfec~. Aut' ~c!or" ~O$ R..:J.. ~cen X~.ofu tyn ~~h'K!tl"
requires so many explanations ad hoc. that Miss Blomfield was
n I n ~&h p I~x s t·8
quite right to question it (2). There can be no. doubt that
these names are somehow related to the names of the runes;
but that relationship is rather like that between the • cursive'
inet.ed Iis + OM" th ~ peord't'c$ ~$Ylifl if'tC ~ boer"
Gothic alphabet on this page and the • classical' Gothic al­ • m I nlll ~ oe ~... Go l
phabet. The following pairs, especially,'are not easily accounted Me",r1,non tl~,;u ~1\J1I Hb~e,;~oe~fl ~t.e ~U$ 'reor It\yr
for: j, gaaT; hi cho:nna; 0, utal; x, enguz; z, ezec. Equally FIG. 10

striking is the fact that in the alphabet u. takes the place of 0

A number· of peculiarities call for comment :
and vice versa. It looks almost as if the scholar who compiled
this material had confronted the information he derived from ( 1) Only one rune, f, has two names : fe and fech. The former
a Biblical manuscript with some information gathered in a is identical with the name of the Gothic f on fo1. 20 v.
region where rests of a Gothic population were on the point ( 2) The scribe Illay have had some trouble imitating the
of being absorbed by their neighbours. and where the Gothic features of the original. InluEgil ,he wrote something
language, as far as it survived, was increasingly showing the .,,"'-1
between a ligature :2 and two independent letters ae. It
't
influence of the surrounding dialects. Of any part of Alcuin is not impossible that he inserted a, after first having ,.
in this matter there is no trace (3). written hegil. What looks like :2 in the name of the rune J. ~.:

(1) A. DOLD, ~ Prooeniewl der altlaleinischen RIimer/JrWtexte in den r h3s a different origin. In all other cases he simply
gotisch-laleinischen Fragmente1l des Coda Carolinus fJ01I Wolfmbiittel. In: writes ae (naed. goer, daeg, aes). Especially his mistake
Au.r der Welt des Buches. Featgabe ••• G. Leyh (= Zentralblatt fUr Bibliotheks­ lug (instead of ing) shows that he was copying from an
wesen. Deibeft 75), 13-::&9. Compare also WALAHPlIID STIIABO'S remark in
De rl!!bw ecclesifI!JtU:U 7: .. Et (ut bistoriae teatsntur) postmodum studiosi older exemplar with long i; this exemplar may have been
illius gentis (i.e. Gothorum) diurnos Iibros in suse locutionis proprietatem somewhat damaged, which would explain the form aes
transtulerunt, quorum adhuc monumenta apud nonnullos habentur", quoted instead of aesc.
by G. BAlISECKB, AIITogans, 156 (note 2). Baesecke supposes that Walahfrid
refers to Am of Salzburg, because the latter is believed to be respoi1sible for (3) For three runes the value is indicated by two letters con­
the Gothica in Vienna MS. 795. nected by & : 3 i & h, (or i & ch), x I & x (or i & x).
(2) Although I do not agree with Miss BWMF1BLD on all points (e.g. I do
not consider lIARDER's treatment of rune-names or letter-names justified at
and -g = n & g. . Of these the last is the easiest to explain :
~: Introduction p. xlvii), I believe the study of this material can only profit phonetically, as well as graphically, IIJI is in some way a
by following the general tines laid down by her: I. Instead of regarding the combination of nand g. The formula can hardly mean
~ontents of fol. 20 of this document as p~rily Gothic, or (with still less
justification) as runic, we should seek to relate them to the activities of an
that sometimes this rune Iltands for n, sometimes for g.
eighth-century IIlphabet-fancier" (RtmIIS, u8; Miss BLOMFIJ!LD does not i
traversed by a vertical. stroke ") is typical of Anglou:x:on script, but it is also
accept von Griie\lberger's date, Runu, ::&10, note I).
found on the Continent (Cf. W. M. LINDSAY, Norae Latinae. Cambridge,
(3) The abbreviation mark in dicitvr (" the right-hand branch of the t
1915. 373 f.), and so it is too ambiguous to support ALCUIN'S authorship.
58
59
 The first instance, however, seems -to require some such which another hand added e 'above the a of rad, and a
explana#on : 3 is actually used both for h (e. g. RuthweU second a after the d; this change was obviously inspired
alme3ttig and i (Dover j3Slheard). The value i may by the form reda in the Gothic alphabet on fol. 20V.
well have its origin in the name ih, which did not agree
g: geofu will not be due to velar mutation of WS. giefu, but
with the acrostic principle. In the manuscript it looks
is rather a Northumbrian form to be connected directly
as if a c had been inserted between & and h, giving a ~th the stem *ge1J.. ( I ) . . .
reading eh for the second value. This may either be a
concession to OHG. scribal habits, Or else a • correction' w: both the value uu and the u- in uyn may be archaic spell­

ings (2).

'of the same type as Tad> raeda or Teda (d. infra). The
case of the x-rune is more obscure. First of all, it is not h:' on the ;e or 'ae, cf. supra. The ending -il (for original

clear whether we have to read I & ,x or i & x. The case syllabic I following g) represents an older stage (3).

of lug = ing shows -that the exemplar may have had either n: naed could perhaps be explained as a mistake for nead

I or long i.' Most ful>orcs transcribe this rune by x, but (and gae, for gear), but,such forms hardly fit into Anglian

Brussels MS. 9311-9319 has il. The scribe cannot have or even Northumbrian surroundings. Therefore ae will

meant that this rune could be used either for 1 or for x

rather have to be explained as rendering I~I (nid, ger), cf.

(the other possibility, an I-like x, or vice versa, need Aed- in the Moore MS. of Bede's Histona Ecclesiastica (4).

hardly be mentioned); but he may have implied that it

could be rendered by i, for in some runic alphabets this j: for gaer cf. under n and r. The equivalent gg is quite
rune takes the place of y, (e. g. Leyden MS. Voss. lat. puzzlinS' The double spelling of g may of course be
F. 12 8). Therefore the reading i & x is to be preferred. accidental : a reader may have written a second g by way
The value i may have been inferred from the name of the of a pTobatio pennae or for some other reason, cf. coscos
·1

rune, just as was the case for 3. (for cos, i. e. oos) in Vatican. MS. Regin. lat. 33 8. von
Grienberger believed it was imitated from uu W, but F':.,':

By examining the rune-names we may obtain the approximate he gives no reason why precisely g should be doubled.
date and -localization of the prototype : We met gg and dd in Oxford MS. St. John's College 17; the
f: feeh is the normal Anglian form; if ch is not an adapta­ explanation tentatively offered for those two cases, viz, Celtic
tion to Continental spelling habits, it may -point to an infl.~ence. may also hold here, although there is no such
early date (1). striking similarity between the Vienna ful>orc and that
],I: the name 10m retains the English device for spirant /ill, in the Oxford manuscript as to all{.lw one to connect the
which was not entirely unknown on the Continent (HiliJe­ two (s)· Yet, if the isolated gg of the Vienna manuscript
brandslied, Lex Salica). By the side of this one 4, d is is n9t to remain problematic, we have to account for it in
used twice for the dental spirant: peord, oedil. This , (I) E. SJ:I{VImS-K. BRIJNNBR, Altenglische Grmnmatik. §§ 91 A. 8, (I ( A. 8.
situation is also in favour of an early date (2). (2) E. S~-K. BRIJNNBR, Altenglische Gram'lll4tik, § 111 A. I.
(3) E. SIEVERS-K. BRIJNNBR, AltengliscM Grmnmatik, § 152.
r: At first sight the name looks liker;eda with a ligature ;e" -(4) H. STROM, Old EfllJlish Personal Names, Il3. The explanation of '&1
and - all commentators have read .so. Close inspection, {still less of ae) l1li 8 late variant for lao (E. SII!VBlIS-K.. BRUNNmt, Altenglische
however, shows that the scribe actually wrote Tad, after Grmmluztik, § 16/,-.) is for chronological .teIlllOns impossible. ,
(5) The two manuscripts also have Mtw saneti Bcmifatii in the neighbour­
(I) E. SIIMIRS-K. BltuNNBR, AltimtIliscM Grmmfllltik, § 'U3 A. I. hood of the-runes; but those cryptic devices are found 80 often that they can
(2) -H. sTRoM, Old EfllJlish Perronal Nan-; 129 f. hardly be considered conclusive. '

6,0
61
 the same way as for that in the Oxford codex: the Vienna: bare possibility. If more certainty could be gained concerning
the authorship of the Orthographia brevis and the relationship
fullorc probably goes back to a very old original, where
of the various items contained in fols. 1-20 of the Vienna
such archaic features may well have occurred (I).
manuscript, we could perhaps come to a safer attribution. But'
3: the name ih (without breaking) is an Anglian form corres­
for the time being we must not forget that there were many ,
ponding to WS. eoh (2). more. Englishmen of Northumbrian descent on the Continent'
p: cf. It· than the one Aleuin; some or other obscure scholar may have
x: the name ilcs seems to correspond to ib: in Oxford MS. had more time to dabble in such curiosa. Nor spould we lose
St. John's College 17, unless cs be interpreted as an error sight of the fact that the fullorc may even be older than Alcuin's
for ix; the latter explanation would at the same time arrival on the Continent.
account for the unique cs spelling.
s: for the i in sygil, d. under h.
6. Brussels, Koninklijke Bibliotheek, MS. 93II-93I9 (saec. IX).
b e: bere without breaking is' an Anglian form; the same

,
The runes of this manuscript were discovered by U. F. Kopp,
applies to the name of the e-rune (3)'

and communicated to W. Grimm, who edited them in his

d: cf. under r.
Deutsche Runen (I); they were edited again by F. Mone (2).
(£: the name oedil has the normal Anglian form; for -il d.
But the only full account, that by G. C. van Langenhove (3).
under h. seems to have reCeived very little attention.
ea: eor is an interesting variant for the more common name The origin of the codex is unknown. During the seven­
ear; it points to a Northern origin (4)· teenth and eighteenth centuries it belonged to the library of
the Jesuit community in Antwerp, where its pressmark was
There can be no doubt that the prototype of the Vienna W. 180: .. Bibliothecae PP. Societatis Iesu Domus Professre
I- ~.:

fullorc came from Northern England, and that it was con­

(I) U. F. Kopp, PalmJoirraphia critica. Mannheim, 1817-1821, I, 316.
siderably older than the tenth century copy that has come down
W. GRIMM, Ueber deuUI:hB Rtmm, 147 f. and Tab. II.
to us. Linguistically speaking there is nothing against dating (2) F. J. MONE, Qwdlen rmd Fanc/umgen, SS2 f., and facsimile no. I (P.ISS).
it in the eighth century. Consequently there is a possibility GlIIMM'S infomtation came from Psris, whilst MONE saw the manuscript in
that Alcuin acted as an intermediary, although I hardly believe Brussels. Owing to a couple of divergences between his facsiprile and
GRIMM'S, MONE did not venture to conclude that the • Paris MS.' and the
that something like i & x can be ascribed to him. Therefore • Brussel,l MS: were identical, although he wss aware of that possibility:
Alcuin's authorship should 1).ot be considered as more than a "Vergleicht man diese Runen mit den' Isidorischen aus einer Pariser HS.
bei Grimm ( ••• ), 80 springt ihre Gleichheit in die Augen und If;eigt eine
(I) Miss BwMFDlLD rightly calls attention to the inconsistencies in the
gemeinsame QueUe. UngliickJjcherweise ist die BrIlsseler Hs. an demselben
Gothic material! the note on the pronunciation of Gothic j (" genuit ") and
Punkte defekt, wie die Panser, im tlbrigen gibt sie .einige Ergllnzungen, die
g (H Gabriel ") seems to be contradicted by the letter-names gam (- j) and ich anzeige ; .... (footnote:) "Ich sollte glauben (nach Grimm S. 147). dies
geuua (_ g). Even if the Gothic material and the runes are not to be placed
sey dieselbe HS. mit der Pariser, denn die Gleicbbeit ist 2\U auffall.end, nur
on the same level, an analysis of the former may pemaps throw BODle light &timmt Kopp's Abzeicbnung nicht mit meinern facsimile, und nach Antwerpen'
on the latter. ist keine Hs. des lsidor von Paris zurllckgegeben worden. Die Brilsse1er
(2)' Hence there is no reason for providing this fonn with the ssterisk of
aber war wirklicb in Paris. Sind beide Hs. eins, 80 nehme man meine
reconstructed forms, as e.g. in E. SIlMDlS-K. BRUNNER, AltmgliscM Gram:nuU;ih Nachricht fUr eine Berichtigung dec Koppischen Abzeicbnung ".
§ 250 A.2. (]) G. C. VAN LANGliNHoVB,Bnuselsche ·R_. I. Ik lsidori R_.

(3) E. SlEVIDlll-K. .BRUNNIDl, AltengliscM Grammatik, § I 19 f. Acsd6mie Royide de Belgique. Bulletins de la Clssse des Letl:res et des Sciences

(4) K. LUICK, HistoriscM Grammatik I, § 119, specifies: 'South North­ Morales et Politiques 1923. 214-;11]8. .

umbria '.
63
62
Antuerpiensis Ludouicus Nonnius Medicinae Doctor obser­
IZsr, <Would point to a German origin (I). In some parte of the text we find
vantia: testanda: ergo L . M . D . D ." (on initial parchment traces of an Irish ancestor: fo1. 3&' posaitionem, 6?r DE NATURA RENUM,
fiy-leaf). Ludovicus Nonnius, or N unnez, 'physician, natural­ 88v possitione, posaicione, S9" Assis, and the confusion of tI IUld i, 0 and II
ist, Latin author and poet (born in Antwerp ca. 1553), may (Ociani, 'adepiscenda, Sufonias, Zorobabil, lohannis, zelotis, decim, precidente
columns: insignes aociwt, Erupa [= Europa,] geumetrica, etc.). .
have had the manuscript from his fa~her, Alvarez Nunnez.
professor in the University of Leuven. From a letter pasted 'Contents (a) :

, foIs.
V
on the initial fiy'-leaf, it appears that Nonnius jr. showed 'the Ir_a : A list of Tironian notes with their meaning (Partly not filled in)

manuscript to Heribertus Roswey. the famous Jesuit contro­ in 3 or 4 columns; originally ruled for '.1.6 n., but this ruling
WIllI disregarded, and there are '.1.7-3'.1. n. to the page. Dated by
versialist and precursor of Bollandus in the field of hagiography Schmitz (J) saec. IX[X but, judging from the writing, rather saec.
IX in.
(Utrecht IstMJ-Antwerp 1629) (cf. infra). During the French r
occupation of Belgium the manuscript was sent to Paris, but 3v Hymn to the Virgin Mary, written by a curious untrained hand.
3 A fuJ?orc, etc. (cf. itifra).

after Napoleon's fall it was returned and incorporated in the 4' Isidorus, De oJficiis.

r
Royal Library under no. ISS. When the inventary of this 37 Id., Libw proetnitmt.m de lihris 'IWfJi tu wteris tertamenti.

.Sr Id., Vita vtll obitm ranctonun qui in Domino pTtlcesrmmt.

library was made, it was listed as no. 93Il-9319. There seem 57' Id., A.lUgoriae i1l I_am. rcriptJlJ'am.

to be some indications that it was written in Germany (cf. infra). 67' Id., De Mtura TtITUrII.

90r Id., DijJlWentiat!.

Modem calf binding, stamped in gold; on the back: JSlDORl , OPERA Ioor Id., DtI sokstuw.

8£ c. 'X. SlEC., arms of Belgium f (shield with) 931I-I9' BIBL.

ROYALE. Strong, rather heavy pa.rebment,·well preserved but showing traces
10
7" Id., De prqprUtate II!N1I01Ifl1n vel rerum.

I I 7' Id., Doctrina tit ./ida tlCclesiastka.

of much handling on the tint and last leaves. An older binding with one IZ4' GltJrae SPiritaks .iu:da Ew:1urimn epUcoptmr.

clasp (1) left traces (l'WIt) on the tint three leaves. Ca. 335 X :IIS rnm laS" Sytu:m.yma CiclWortU.
(written area ca. '.1.60 X IssfI60 mm [marked double left and right: + a x
9 mm]). The codex is made up of several parts, though M(II'Chai's implica­
tion, viz. that it consists of nine fragments' (hence the no. 93U-93I9), is not
Fol. 3 v is arranged as follows : ,""
Originally the lower half of .the page contained a table,
bome out by the codex itself: fol8. 4~a7 form three quatemions mllJ;ked • a ' ­
, c '; fols. 37-II6 = ten quatemions marked' I ' • X', The composition being a square divided into 25 by 27 (?) columns; in the
of the manuscript may be summarized as follows ; small squares thus formed (about 3 mm square), letters
fly-leaf (formerly pasted onto the binding) + I (+ I) {I-3J + 3 IV [4­
I were written, but at a later date all this was erased; only
a7J + IV (+ 1) [ZS-36J + n IV [37-124J + II (+ a) [IaS-130J; 130' a few letters ([...] r q [ ...]) in the first line are still visible.
wu at one time puted onto the binding. The table may have had some computistical meaning, or
Part of the text is missing after fol. 36 (36v: " Desunt duo capita quae . may have been an acrostic poem (4). To the right of this
habentur in Parisiensi editione anni 1601 ''). Written by a great number
of different hands, all (except fol. I-Z?) fairly Contemporary. Roswey(I)
table, and going from the top to the bottom of the page
dated the .codex "antiquus •.. supra annos, ni failor, quingentos ", which (I) A gloss 'Nlt:iut (= fJlWle:it: mistaken for ceroi:t: 1) 'lIl fo1. uS- seems to

would mean sace. XI. At the other extreme Mone (followed e.g. by Stephens) confirm van· den Gheyn's assumption; siinilarly on £01. 130 ', between mlrnu

dated it tentatively sacc. VIII (1). In view of the undeniable archaic features, and IOspis: hell or hell[tllr, and between oIJS&e7IOIJU IUld impudicus: fu == frd I'

van den Gheyn's date : &aoo. IX ('.I.) may be adopted, with a preferenc~ for lXI, The handwriting of the rune-names reminded Prof. B. BISCHOFF of products

.No marks of origin seem to have survived; according to van den Gheyn the of the IICriptorium of St. Oyan Gura) •
musical accents accompanying a couplet found in the upper margin of £01. (a) Cf. J. VAN DEN GBlIYN, Catalogw II, '.1.7'.1. f.
(3) W. ScHMITZ, Cormtumtarii 1IOt.tuum tironiorum (Lips.ie, 1.S93), 9, tab. 13 1,

(I) In the letter pasted onto the initial fiy-leaf, 132; I'D., Studien - lateiniscJum Ttuliylfl'aphie. Fortaetzung. Programm

(a) J. VAN DEN GHBYN, CatologtU cln mlJmucrits ds la BibliothAque Royale des Kaiser Wilbelrri-Gymnaaiums zu ~ln, XIII. Schuljahr, ISSI.

de Belgique II .(I90Z), 27'.1. f. . (4) The fonner explanation is the more probable, cf. the table H Aetas
Cf. also H. BI!I!SON, lsidorsr.udien, '.1.6, 3'.1., 34. 35, sa, 71, I I 9. lunse in alphabeta di81incta" in BlmB, De /.emporum ratkme, ed. C. W. JONlS,
Bt!tIae OPlWa de ~, 2'.1.5.
6-+
9
6S
 (the page being turned 90° to the left) there is a fuporc.
full details of which are given infra. To the upper left written below them. The fourth rune was i; it was cut
of this fuporc a fourteenth century hand added a table of through half; the letter indicating its value (g) was preserved.
the contents, probably at the time when the volume was Closer inspection shows that two scribes may be responsible
(first?) bound together (I). for the runic material : the hand of the rune-names is firmer
The structure of the manuscript does not allow us to connect and heavier than that which wrote the values; it Uses only open q
the runes with Isidore's works, as their presence in the same (whilSt the latter has the normal minuscule a) and has on-the
codex is entirely accidental. The inclusion of the runes may whole a more archaic appearance than the other hand (I).
have been prompted rather by the preceding Tironian notes The runes, too, show differences: those on the first line are
than by any of Isidore's works found here (2). drawn in a leisurely way, the downstrokes being obtained by
The fuporc is not a careless addition of the probatio pennae passing twice or more over the same line; the runes in the

sort. The runes are carefully, if not always skillfully drawn. second line show a simpler and more hurried (though perhaps

Their size varies between 6 and 18 rom. The folio (single also a more skiIlful) technique. If we assume that two (and

leaf) on which they were written must originally ·have been not three) scribes were at work, A may have written the first

higher, but in the act of binding the lower margin was removed, line of runes their names, and B the additional runes all
+
the equivalents (2).
resulting in the loss of at least part of two runes. What is left
shows the following arrangement (plate IV b) : Since the. change in the technique of rune-drawing coincides
with the break in the fuporc (i. e. where the lower margin was
I) The runes f to re, with above each rune its name, and
trimmed), several questions arise : did the runes of the second
below its value. Of the last rune only the vertical shaft
is left, of the name only the first letter (e), and of the letter line belong to the fuporcfrom the very beginning, or did it .....':"'
originally end with the :e-rune? How many runes written
which once indicated the value (a) only part remains.
by A were lost? Did B copy his runes from the strip of parch- < ;~.
2) Below the last three runes of this first line there are three ment that was cut away? The order of the additional OE. runes,

. more and part of a fourth; to the left of the first there is which should normally help us to answer these questions, is

a small x-cross, possibly a mark of reference. No names not beyond doubt jtself. As far as the ea-rune, the order

are given for .these four run.es; only their equivalents are· corresponds to that of the Thames scramasax, of Cotton MSS.

(I) In hoc uolumine continentur hee. Ysidonu de I eecWiaaticis officijs. Domitian A 9 and Galba A 2 as well as to that of the imma

Idem de ordinibus eecWiasticir. I Item liber proemiorum. de librie nouj ac manuscripts. But all other fuporcs either have twenty-eight

ueteria I testamenti. Item de ujta uel obitu sanctorum. qui I in domino prl!CeS­ runes (i. e. they include a, a:, .y and ea), or add not only k

serunt. + [in tire margin, 17th century: + Allegorise I sacl'le scripwne] Item
liber ysidori de I natura rerum. Item differentie ysidori I episcopi • junioris and i, as the Brussels fuporc does, but also ;,st and q. Yet

[corr. from uin- 1] sparuensis. Item de 80lsticio. I Item de propriet:ate &er­ this does not necessarily mean that these three runes were once
monum uel rerum. Item I diffinitio eccluiasticorum docmotum [read: found in the Bruss'els manuscript: st is hardly used in English
dogmatumJ. Item I glose spiritales jUJ[ta eucherium epircopum. I Item
sinonima ciceronis. inscriptions, q never; the use of j too is rather rare. If the
(2) Consequently the Brussels manuscript cannot be compared with St. Gall fuporc followed epigraphical usage,it may never have had
MS. 878 (cf. p. 83), where the runes are found immedistely after an extract more than thirty runes (f - i)" Or 31 if we count j.
from Isidore's Et.ymologi(u, precisely the chapter De litterir (I, iii). ARNTz's
statement: .. Ebenso ist in Briissel das Ful;lork in die Isidoruberlieferung (I) E.g. the liiatures in pm and berc.

geraten" (R_ u1'lll Runennmrum. 190) can only lead to unwarranted (2). A's s!tare mey again be divided: Al mey have drawn the runes to which

general~tion8. A. added the names. But this distinction is not necessary, and dOes not lesd
us any funher, except perhaps in the case of .g.
66
67
 shows that the scribe at first skipped this e, writing g
It is possi.ble that all the runes were originally written in below the rune c, uu below g, h below w, n below h (and
one line; the page having bee!l trimmed so as to drop all runes also ibelow n ?); to the left and right of g there are traces
beyond e, B would then have added them below that first ofuu, and above uu a rest of h. The scribe must have
line, indicating by x that they were to come after e. ,A second noticed his mistake when he came to i, and then made
trimming then resulted in the loss of part of the two last runes, good his oversight.
re and g, and of the reference mark after e. It is not probable w: the name of the w-rune caused considerable trouble to van
that B copied the runes from the strip removed at the first Langenhove. He supposed that the scribe heard a final
trimming, as he would then have inclUded the names as well. palatal n (OE. wyn) ,as luI and therefore wrote ng. But
So either of two possibilities remain: since there is no other convincing example of a form being
(1) A only wrote the runes f - e; B added y - g; the folio based on dictation, we must leave the origin of uung
was curtailed only once; or . undecided (I).
(z) A wrote a fuporc of thirty or thirty-one runes, the last
h: the name of this rune has given rise to some discussion.
four or five :runetl, of which were lost; B reinserted the
At one time or other the name must have read hagal, but
missing runes, after which the folio was trimmed a second
the first letter was either partly· erased, or written over an
time. erasure (with the result that part of it was lost again). The
In judging the readings given in older editions. we have to first possibility is a priori the least probable. It is true
keep in mind that at one time the runic material cou14 be read that scribe A had some trouble with k (cf. the name of e,
only with difficulty. One of the first readers made ~n u~for': bee; and kine above re), but there seems to be no special , "',
tunate attempt to facilitate the' reading by the application of a reason why he should have erased the initial h of hagal. "
reagent. The result was that later readers found the writing There are, to be sure, examples of a name agal, or rather .' . ;'
much obscured. van Langenhove had the worst stains re­ agalc, in Paris MS. 5239 and Strasbourg MS. 326, but
moved (I) and now practically all. the readings are certain. our fuporc has no import;lnt features in common with the
The fuporc bas the following runes, names...'and values': runic alphabets where this form without h is found. There­

IV At" ~ '~ 'r x~~}li .~ f"'(Yi' fore I have kept the reading~gal (2). .
n: to the right of the first n-rune a second one has been added
f

,
1
u

t
th

$G: t bet ~M 'foh~eHtl ~

b
0

A~ rh ~
e
I"

m
C.

I
9

in
uu

dOlI!.'"
h n

r
f

.,..
'lit!"

Y
rh

eo
p

k ,
in a lighter ink. It is also possible that the latter rune
was written first but, being not well placed below its ,name,
was wiped out again. The 0 of the name is partly hidden
by 'a fold in the parchment~
'FlO. 11

j: unlike the preceding runes, this rune and the next have
Notes (z) :
}): for the name Kopp-Grimm read than, all later editors thorn: (1) It will hardly do to connect this form with OE. wong, wang, or even
c: the value e is written over an erasure, and so are the values with OHG.rmmg. •
(z) The erasing of this h may perhaps be connected with the corrections
of the next three or four runes. Careful examination in the l.ist of values. Perhaps B began to change the name of h to make it
fit the erroneous :value be had written below the rune, then realilred his mistake,
(1) G. C. VAN LANaBNHOVE, lndori Rurum. z14·
(z) c:
Cf. G, VAN LANOI!NHOVE, lndori Runerl, 2ZZ ft., liS well as GRIMM'S restored the reading Iwgal'and correCted the values for the runes c - h.
and Mom's editions., 69
68
 two names. van Langenhove supposes that the iar above
as h:ee or hat; van Langenhove accepted Grimm's reading.
the rune is a Continental rendering of DE. gear; ger would
Actually the name consists of II. followed by a ligature of e
be a Continental (OS.?) word serving as an f:;xample. I
. with another letter. In view of the preceding name bere
believe such a reconstruction is unnecessary: iar may
(where the last three letters form a ligature), I believe
either be a Continental Gmc. form of the word • year',
the ligature may rather be interpreted as ee than as et;
or, less probably, the iar we find in some English fuporcs
consequently we have to read the name of the rune hee.
(Cotton MS. Otho B 10; cf.ior in Cotton MS. Galba A :a, This form may go back to eeh.
ian in Vienna MS. 751); the form below the rune is simply
non-WS. ger (I). t: after the page had been subjected to the treatment men­
tioned by van Langenhove, it appeared clearly that the
3: the first scribe seems to have had quite some trouble in
name of this rune should not be read lag (Grimm, Mone),
assigning the 'right names to the right runes. Thus he but lago.
wrote ine above 3, kinc above <l'!, odil above Q (the latter
two runes also changed places). B corrected the first Q, <l'!:rune no. 22 is obviously an m-rune, as we may also
mistake, writing ik below the rune; but in the remaining infer from the Occurrence of the y-rune in the :a4th place.
two cases he was misled by A's error (cf. infra). The The name and value, however, are given as hine and in.
name ine was so read by Grimm, whilst Mone proposed This may suggest that scribe A was in fact not one, but
hie; van Langenhove doubted both readings, and proposed two different persons: Al drew the runes, and mistook
to interpret the word as reflecting DE. iw, a variant form Q for <l'!; As added the names, and overlooking A/s mistake,
for eok (2). When examining the manuscript myself, inserted hine and odil in the right places among the names
I could only read inc. This reading, of course, somehow (but consequently over the wrong runes); B agrees with As.
conflicts with the name kine above the m-rune, but I To explain the name kinc we may simply refer to the ''¥~'"
see no other possibility. scribe's hesitation about initial h (cf. hee; [h]agal ?), but

there may be more to this form. In the isruna group we

p: the name of the rune is much obscured by the old reagent,

shall find a variant y-rune, and that same rune will turn

but no doubt seems possible: we have to read pert.,

up with the value II. in runic'alphabets; we already met it

x: part of the name ilix, too, is obscured by old stains, but as a variant for h in Cotton MS. Domitian A 9. This

the reading is not doubtful. may perhaps explain Why the Brussels manuscript has

t: above the t of the name ti there is a short horizontal stroke, both inc and hinc.
which van Langenhove tentatively proposed to interpret No. 24 is the usual English Q. The name is obscured by

as an abbreviation mark, especially in view of the form a darker spot in the stains left by the reagent. It seems

tir found in several other fuporcs. I believe the stroke to ~ve caused some trouble as early as Mone's time.

.to be accidental, and so is the i-like mark after thet indi~ Kopp-Grimm had read odil; the form in Mone's facsimile

cating the value (neither does van Langenhove attach any was interpreted as odil by Kirchhoff (I), but Stephens

meaning to the latter). read othl. van Langenhove himself read othl in the manu­

script. The difficulty lies in the fact that 0 forms a ligature

e: Grimm read the name as het, Mone, followed by Stephens,

with the following letter. A comparison with the name

(1) E. SIIM!RS.-K. BRUNNER, AltetwlUcht! Grammatik. § 91 h. rad shows that the second letter is d, the loop of which

(2) E. SIIM!RS-K. BRUNNBR. AltenglUcht! Grammatik, § § 234 A. 3. 250 A. 2.

(I) A. KmcHHOFF. DIu gothische Runenaiphabet, 27.
70
71
 sets in rather high and is left open precisely to form the deniable Continental influence: not = DE. nied, nead, ned;

ligature with the p(eceding o. What Stephens and van hagal = h;egl, hegl; odil = (BpeI; dag d;eg, deg. The na,mes

Langenhove read as h is the vertical shaft of d followed by i. thorn and ken may reflect Continental orthography, but could

e: of the rune only the vertical shaft remains, of the n.ame be English as well; geuo is rather an adaptation of gefu (cf. infra).

only part of an e, of the value, part of an a. . Such translations as not, odil, dag seem to contradict van Langen­

y: Mone's facsimile led Stephens to consider the y below the hove's assumption that the names were written under dictation:

rune as part of the rune, the whole being <I doubtlessQ '~. they rather result from a conscious effort to substitute Con'::

tinental forms for. the English rune-names. The forms not,

ea: the value eo can hardly have been transferred. hither from odil, dag, and also thorn if the form of that word corresponded

the rune 3, since the latter has a correct ih below it; eo must to the scribe's dialect, may prove that the scribe was familiar

then bea variant of ea, cf. Vienna MS. 795, where the with Franconian, and, if we may.judge from geuo, with Middle

name of this rune reads eor. or even Low Franconian (1).

g: of the gar-rune only the left half remains,
with the value g Thus far we have made no distinction between the language
below it. The rune shows the epigraphical type, not the of A and that of B (or of their prototypes). As a rule B pre­
usual English manuscript type, i. e: a square'crossed by serves the DE. character of his material far better than A, so much
two intersecting strokes. . so that if his handwritting were no pure Carolingean minuscule,
On the basis of his discussion of the runic material. van one would not hesitate to declare him an Anglo-Saxon: Of
Langenhove concluded : course names could easily be translated, whilst· values wo~ld
rather be copied mechanically (odil : oe e. g.), but even the two
(1) the runes and the names are of English origin, although
names which B writes out retain the regular OE. form (ger, ih).. ..
the text affords too few data to decide to which dialect
The prehistory of this fuporc must be rather involved, at "'-'
the .prototype belonged.
least to judge from A's part. His exemplar may have been
(2) the names are the outcome of a Continental scribe's attempt poor in places; or else he wrote from memory. The latter
to render the English names dictated to him. assumption would explain such forms asuung, hine, hec, and
(3) this scribe was probably a German, but it is not possible the errors in the order of the runes more easi.1y than van Langen­
to decide whether his own dialect was Middle Franconian, hove's hypothetical dictation.
Low Franconian or even Saxon. On the whole the importance of this fuporc lies especially
in the adaptation to Continental phonology. Only the .evidence
About the first point, 1 believe, there-cannot be the slightest
which B provides appears to be genuine.
doubt. The fuporc includes such typically English features .. ,.,

as the runes 0, j, and the additional runes; such names as

os, rad, ken, inc, ligil, ti, /ago, QC can only be English. Perhaps 7· St. Gall, Stiftsbibliothek, MS. 878 (saec. IX med.).
it is not impossible to determine the basic dialect as well. The
forms fech, pert, herc and hec (if *ech) point to an Anglian With Vienna MS. 795 this is the best known manuscript

prototype {I); erst] may imply a restriction to Mercia, corrob­ with runes.. It has been studied mainly on account of the

orated by eo for ea (2). On the other hand there is the un.,. Abecedarium nord(mannicum), a· doggerel poem on the runes

of the shorter Norse fUPllrk, . written in a curious mixture of

(1) E. SIBVI!BS-K. BRUNNl!R, Altenglische Grommatik. §§ Il9, no.

(a) E. SIlMms-K. BRUNNER, Altenglische Gr_tik, § § sa, 35. (1) W. BRAUNE, Althoc1uIeuUt:/ul Grammatik, §§ 163. 134.

72
73
',.
Low and High German (I). As for the time being we are only ·24Z ff., 3ZZ If.• 3Sz ff. So p. 3ZI must for Some time have been the last page.
concerned with the English runes, this Abecedarium will not be Fonnat ca. ZIO X 137mm (written area 160 X loomm); 33 liQ.es to the
discussed in detail; our attention will in the first place go to the page. The manuscript has been assigned to various dates: the beginning of
English fuJ>orc which precedes it in the manuscript. ·saec. IX (v. An), XI/XU (Scherrer), IX ex. and later (Bruclmer), IX*
(Lehmann). whilst Mommsen and quite recently Bischoff called attention
MS. 878 does not belong to the old St. Gall stock: in I.457 to a note on an earthquake in 849 as a starting point for dating this codex.
it seems to have belonged to the Cathedral Library in Chur, A great many hands wrote more orless extel1llive parts of this codex. Amongst
afterwards to the Glarus historian, jurist and politician Aegidius .them Bisc~off distinguished four stadia of the same hand, which added material .,,' .
by and by, with intervals of several years. He believes this hand to be that'
Tschudi, from whose estate it came to the Abbey Library. Its ·of Walahfrid Sttaba, on the whole, I think, on convincing grounds. .1'0 this
importance has certainly not diminished since n. Bischoff hand would be due: WI: 194-Z40; W II: 5-43, 50-69: WIll: a54-257,
Z8 Z 8
identified the scribe of the major part of the codex with nobody 0 Z 1 324-3 , 3 9-335. 34 -350; W IV: 43-47. 77-95. 168- 1 71, 17z- I 76,
z60-z61.
24 - 4 , z43, z77-3 0 7, 3 15-321, 335-339, 366-391 (t).
less than Walahfrid Strabo : MS. 878 would in fact be a sort
of vademecum compiled by Strabo over a considerable length Contents (z) :
of time {2}. . P.
5 Donatus minor, and other grammatical extracts.
50 Donatus maior.
Modem binding, brown leather back and comers, greyish blue paper over
boards, with the inscription: Grammatica {AJrithmetica II_a. Bek. Gram­ 70 Latin poems; 7z De uoce; 79 De barbarismo; 80 De soloecismo;
mat. Prisciani. PhilDtophica. Medica. Historia. etc. 878. Parchment of unequal 8z
tropis.
De
XII uitiis; 83 De metspiasmo; 85 De schematibus; 86 De
quality, with many repairs and traces of much use; 197 fols. = 394 pages 91 Bede, Dill arte metrica.
numbered ' I ' to ' 394' (' 16' occurS twice, • 383 ' has been skipped) and 131 Bede, Dill schetnatibm et tropis.
arranged as follows :
148 Priscian, I,"titutio de nomine et. uerbo.
2 papers fly-leaves {I-4J + III {S-I61J + IV (+ I) [16*-33J + 18 IV 170 Grammatical extracts; Isidore, Etymologiae VI, ii; 176 De XII
{34-32IJ + IV (+ I) {322-339J + III {340-3SIJ + IV {3S2-36?J + III sigma (poem).
(+ I) [368-38IJ + III [38z-394J. 178 Hrabanus Maurua, De computo.
The old quire ms.rks show that one quire is missing between pp. 177 and 178: 24Z Bede, De natura rerum. '~"'
65 'IIII', 113 'VII', IZ9 • VIII', 177 • XI', 193 C XIII " Z09' XlIII " a6z Bede, Uber de temporibru.
2ZS C XV', Z4I 'XVI " Z73 'XVIII " z89 • XVIIII', 3ZI • XXI'. The z77 (belolf, tide :) Excerptum de libra Albini magistri (no text).

manuscript is made up of five parts, in this chronolo~ical order: 178 fr., 5 ff. z78 Adbreuiatio chronicae; Computistical extracts.

0
3 3 - Excerpts from Orosius, Cassiodorus's Historiae, and Jerome

(I) Cf. into al.: Epist.73.

W. GRIMM. Ueber MuUCM Runen, 138 fl.; Tab. II.
315 Excerpts from Isidore's Et;)l71iologiae: De accentibus; De ligura
ID., Zur Litteratur der Runen. z6 ff. - Kl. Schriften III, III ff.
accentuum; De posituria; De litteris .
. H. MAssMANN, Neue Runen, 32.

3Z0 A Hebrew alphabet: a Greek alphabet; • Anguliscum. = a fu.\>orc;

H. HATI'BMKR, Dmkmahle I, 417 C., Taf.1. Abecedarium nord(mannicum).
G. STBPHKNS, Monumenu I, 100 f., no. 6. Z2 z
3 - 3, 340 -44 Disputatio de uera philosophia Albini magistri
F. DlBTRICH, Fibif northumhrircM Ru.nerupriJche. 5. Die RunentJeTSe des Alcuin, Grantm4tica.
St. Galler Code:Jt 878. Z. f. d. A. 14 (1869), 119-123. Z
3 4 Calendar for Fulda or its neighbourhood; 3Z7 Epistola lpogratis
(K.) MOLLBNHOFF, (]ber tku Abecedarium Nordmmmicu.m. Ibid., IZ3-I33.
P. PJPI!R. Am Sand Galler Handschriften HI. Z. f. d. Ph. 13 (I88z), 445 ff. ad Antiochum et Antonium; 331 Recipes (with OHG. glosses), etc.
335 Epistola Karoli regis ad Albinum Magistrum.
K. MOLLENHOFF-W. ScmmmI, Dmkmllkr I, no. V; II, 55 fl. 344 Computistical items; 35 J blank.
J. H. GALLD, AlUaecJuisCM Sprachdenkmaekr, z63 fl.; Taf. XIIa-XIIb. .35Z Epistols Antimi medici ;.. ad Titum imperatorern et ad Teodericum
E. WADS1mN, Kleinere aluiicfuiscM Sprachdenkm4kr, ZO, 129 ff. regem Franconun.
G. BAl!SBCKl!, Dtu Abecedari#m Nordmonnicum, and most general works on
366 On bloodletting; 368 De mense nouembrio, de pomis; 370 De
rwiology.
(a) B. BISCHOFF, Bine Sammtllhandschri/t WalaJifrid StrabOI (Cod. Sangall. uino du~ce faciendo; 373 De nielancolia; 374 De observatione totius
878). In: Aw der Welt des Buche,. Festgabe ..• Georg Leyh (= Zentral­
blatt filr Bibliothekswesen, Beiheft 75) Leipzig, 1950, 30-48. {I) For full ..details cf. B. BISCHOFF, Hine Samm4haruischri/t, 34 fl.
(z) G. SCHIntJmR, Ver:reichniss, 30 7 f. .

74
75
 ~ ut inoolomitas custodiatur; 375 Hippocrates a:rchiater Antigono but in the long run caused more or less dark stains (from
regi. brown to bluish black); in some places the original writing
37S Excerptum ex atoria ecclesiastica; 380 Excerpt from Orosius.
39" Confectio antidotiqvod dicitur potio amara, and other recipes; can hardly be made out (I). At the top of p. 321 there is a
394 later additions (IZth century). Greek alphabet with above each letter its numerical value and
There can be no doubt that the runic material has some its name (11. 1 - 9); this alphabet is anno.l,lnced on the last line
connexion with the grammatical material of the preceding pages. of the preceding page: ALFABETUM GRECUM CUM ...
Grammar takes an important part in this vademecum, which NUMERO (1. 33). Ll. 10-11 show an erasure; nothing can-De
covers a wide range of other subjects as well (rhetoric, computus, made out of the original text. Then follow Greek diphthongs,
history, medicine, natural history). The series. of extracts. with their values shown above them': DIPTONG AI (e),
from Isidore ends with De. litteris, i. e. on the alphabets of the EI (?), OY (u), 01 (y),andthree letters (S CJ. t) with this
three • sacred languages', Hebrew, Greek and Latin (corres­ text: isti tres cara[cter~s] I ad numerum tantum' per/tinent.
ponding to Etymologiae I, iii, ... etc.) On p. 320 it is immediately The next line is again blankwith perhaps an erasure about the
followed by HEBRAICE. LITTERE (L 22) and a Hebrew middle. In L 15 the inscriptionANGULISCUM has been
alphabet with the names of the letters (11. 23-30). This al­ retraced with a darker sort o(ink, and so may have been the
phabet is remarkably genuine, especially if one considers that three runic words at the end- of the line. These words have
this period derived its knowledge of Hebrew mainly. from more also been treated by von Arx, but they can fortunately still be
or less trustworthy copies 'of Jerome's works. For a number read with fair certainty (except the last two runes) :
of. letters variant forms are given (haJ, mem, nun, pe, zadi). It
would not be surprising if this alphabet could be traced to one ear e a K K a [1 c]
of those learned Jews who sometimes assisted Christian theo­ ~,
logians in their exegetic work (I). The runes of the fuporc are written in three rows, f s I
Much of this alphabet is hard to read because 'the reagent t - jIg and tw.elve more runes. The· whole extends from ~. ;'

applied on the next page has penetrated through the parchment L 17 to 23. Another hand, using a lead pencil, copied the
and caused stains both in the preceding text and in the Hebrew y-rune twice between the lines.. A, recent reader numbered
alphabet. As was pointed out before, p. 321 was for some time the runes in the fuporc order; but after 19 (=e) he skipped the
the last page of the manuscript, and so the writing on it suffered m-rime, whilst after 21 (1J) we find the figures 19 (d), 4 (<2),
in places; -especially the text of the interesting Abecedarium 9 (a), 22 (y), 23 (ea). This reader obviously knew something
nord(mannicum) had become faded by the beginning of the about runes : his second ..., e. g., connects the rune <2, with 0
nineteenth century. Between 18.21 and 1828 the then keeper (= no....),~ probably because their values are related. But his
of the St. Gall manuscripts, L von Arx, treated it with a-reagent knowledge of the runes cannot have been very profound.
which for the time being made the faded letters legible (2), J;Jecause'the runes e and d are somewhat similar in form, 'they
(1) M. MANITrus, GeschkhU,l, ago, Z94· both received no. 19.. The 9 we find 'above a and z must be
(z) Q-. 8cBmmER, v.~, 308 ..... I. von An.: lieferte z Facaimi1e's
explained in the same way :- the ink of h (= no. 9) had partly
dieser Ste1le,eines i. J. ISZl ror W. Grimm, Ueberdeutsche Runen Qjtt.
18"1, Taf. II (p. 138 und 140-147) und daa zweite IS"S mit Anwendung von flaked off, and what was left looked rather like a. As a result
Reagentien ror W. Grimm's Abh. z. Litt.d. Runen in Wiener Jahr. Bd. 43 h, a an!! z received the same .number. '. von Arx applied his
und sepamt Wien IS"S p. 42 und ,,6-zS ". The recipe of the reagent used by
VON An is given by G. H: PmtTZ, ltaliiini.sche Reise, 51 I f. (I) A pootogmph made with infrared rays brought out some more of the
Cf. J. H. GALLD, AlhalCilmcM Sprachdtm1mweler, Plate XU-; photograph, text than was 'Visible on an ordinary photograph; but reading in bright daylight
and XIIb : VON An's copy, proved still' more successful.

76 77
r~agent to runes 2, 3 and 14 (u, )J, p); therefore these three beginning of the first line have variously been read as wreat,
appear surrounded by dark stains. wreaw (von Arx), )Jreal (Lachmann), )Jreat (I). Of these the
On _1. 24 begins the much discussed ABECEDARIUM_ first and the last are the most probable readings. One more
NORD[MANNICUM] (I). Much of it has become entirely rune may have preceded w (or )J), but if so it is now impossible
or partly illegible; at any rate the reading of those letters or to decide which. Neither wreat nor )Jreat are entirely satis­
words which caused difficulty to von Arx is still at least as hard factory. The former has been interpreted as a form of the
and uncertain as in his time, if not more. A number of English verbwritan • to write '; on account of the form ullman. in-the
runes were added between the lines and for this reason the text first line of the Abecedarium this reading is rather attractive,
must be given here. The contents, however, must be examined but the vocalism can hardly be explained. On the other hand
in connexion with Norse TUnica manuscripta. It is not possible one fails to see why a word peat • troop; violence; threat'
to decide for sure whether the additional runes are due to the should have been written there. So all we learn from the
first scribe : Abecedarium is the confirmation of the value of six OE. runes,

r . none of which seems ever to have been doubtful (2).

The three runic words which precede the fuporc may be read
rf(I,U forma.n flur after PthUdS thritten ~os ,sthemo ~ rat end ear eaK Kale. They call for several remarks. First of all,
[7] ~ R'ft ~ta.6u obaro os uurihn the runes are much smaller than those of the fuporc (only ca.
3 mm high); second, they may have been retraced in a darker
M i- ~

Yc.hhon thanne
cliuot
*h~9a.llna.uthabClt liS .far l.-t e.nJj $01

sort of ink, just like the word ANGULISCUM on the same

line-though the traces. of the reagent make it difficult to decide
this for sure. For obvious reasons it is equally difficult to-{,
A decide whether they were written by the same hand as the,

t~(LA &brfca. «.,nJi rnan I 1a.90 thtltohto ,l yr Ai brha.bQt (1) fuporc; but there are reasons to believe they were. Another':,
point of interest lies in the use of Roman Kin ellk and kale.
mIdi The fuporc actually has a symbol for k, but it was not used
FIG. 12
here. As to the meaning of these three words, the first and.the
Six English runesPare compared with their- Norse equivalents : last are in all probability identical with the rune-names ear
f, h, n, a, m anl y. Why the other· OE. equivalents, esp. and cale (or kale). I do not believe that the spelling ea in the
o and c, were omitted, I fail to see. The runes below the former may be explained as the phonetic rendering of the
(I) It looks 118 if even the application of the reagent did not allow von Arx English diphthong ea; it is more probable that these names ar~
to read more than the four letters NORD; now all the reatof the line is one simply English forms transcribed in runes. This leads to the
dark blot. In it I believe to have been able to read the letten leUr but· they conclusion that a Continental scholar was at work here: no­
seemed larger than the letten in the reat of the title and may have been added
later. Scholan are fairly unanimous to read nord[mannit:um).
where .in the English tradition do we find the rune-names written
(2) I do not give a critical text, but those readings which have been most in full with TUnes. The double occurrence of K may also betray
commonly accepted. This puzzling mixture of Low and High Gennan will a not very experienced rune-master. What eak means I fail
perhaps be more eaily undentood if it is examined from the angle ofWalahfrid
Strabo's part in this manuscript. A few remarks on the text·: above the u-rune
another hand wrote ie, but I oould not make out whether at one time there (J) K. MOLLENHOFF-W. SCHI!Rl!R, DmkmIi1er II, 56.
W8lI any text between thisie and the upper f at the beginning of the poemj
(2) On the other hand the fact that the ON. j'T-rune {i.e. R} is here equated
above I> there seems to be a i); the first three letters of the last word look mther to the English y~rune anticipates the later evolution in Denmark (after IJoo);
d. L. JAcoBSBN-E. MOLTKE, DtltI1tIi1:rks 1Umeindskrifter (Text), col. 980.
like bah.
79
78
to see. It could be the English conjunction eoe 'also', but The values of the runes are not indicated, but we may assume

in its context· one would rather expect another rune-name. that the runes simply retain their original order. . If so-and

Perhaps the original form was corrupted by the reader who there is no indication of the contrary-Qnly the more or less

retraced these runes. exceptional forms of a few runes require some comment.

The fuporc is firmly and on the whole carefully drawn. At 15: in the purely English tradition, this type of x-rune does

first sight, however, one sees another· hand at work fro1l1 rune not seem to occur. It is found in Munich MSS. 14+39 0 '

38 on: and 19410, Exeter MS'. 3507, Cotton MS. Vitellius A' 12

and Philipps MS. 3715, i. e. in runic alphabets\vritten

. 1 2. 3 4 5 6 7 8 ~ 10 11 12 13 1+ 15 16
h Xt~ l I +J' .~ I
on the Continent or probably originating there.

~ ~ ~ ~ ~ ~
29: the rune j is very rare on the Continent (I), but the related

j seems to have been more popular.. Only Munich MS.

17 18 19 20 21' 22. 23 2+ 25 26 27 28 29 14+36 has ear for e and ios for i in the same runic alphabet;

t&MP1~*H~~~~T*
at least the former goes back to the j-rune.

30: the g-rune is of the epigraphical variety.

50 31 32. 33 34·35 36 37.38 39 40 41 +2 3I : it
would be practically impossible to ascertain the value
Jt~X~ R~J\~l:){'f~* FiG. 13
of this rune if, in Munich MS. 14+36, it did not turn up
with the name cq,k (its value i in that manuscript must
be due to a mistake). Otherwise this variant form of the
Runes 38-42 are in fact unskillful repetitions of runes in the rune k is unknown. It probably arose from the type .'S:'
fuporc, a sort of pt'obatio penno.e, except no. 40, which is the found on. the Ruthwell cross (where it has straight lateral 1:

regular OE. form for no. 15. One rune is too little to decide strokes), through the addition of two ·more lateral strokes.
whether the author of these additions really wanted to record We have then a fully developed fuporcj only q and at are
this variant, or whether this fonn arose accidentally. missing, but these runes probably never reached the Continent
Runes 32-37, too, repeat runic fonns from the fuporc, except (at least not in the manuscript tradition; at is bown from
no. 35, which is a Roman K similar to· those in eak and kak. Frisian inscriptions). Whether the fonn hur implies that
Consequently nos. 32-37 may be read gar hur. The former some form of q-rune appeared in the prototype, we cannot
is the name of rune 30; the latter may be compared with cur, make out, unless the variant k-form were in fact a q-rune; the
the name of. the new q-nine, in' Cotton MS. Domitian A 9 third q variant in Oxford MS. ·St. John's College 17 may be
(hand B); we shall meet it again in the isruna-group and, as· a related form (quar). But the evidence of Munich MS. 14+36
([Ur, in Munich MS. 14+36. In the fuporc, however, there seems si.dficiently conclusive to settle the case : in the fuporc
seems to be no q-tune (that' found in other fuporcs is either an equivalent for k must have been meant.
derived from the p-rune, or it is identical with the eaN11l1e). Little can be said about the age of this fuporc. It may'be a
The names gar, hur, ear and kak may have been the only names century older than the St. Gall manuscript, but for that we
found by the compiler. Here we are reminded of Cotton MS. have n9 evidence. The sureness with which the runes were
i
Domitian A 9, where the first scribe only recorded the names (I) This relative rarity may have Bome connexion with the occurrence of

of the last six runes (besides ing, where the value could not the same fonn for :II; its resemblance to Roman X probably played a part in

very well be • abstracted' from the name). thia1lllltter.

80 10 81
....
 drawn proves that this copy cannot be far removed' from' the logical matters : even if the tract De inventione ·litterarum. may

English Erototype. be attributed to the former, its runic lorels so inferior to Strabo's

If these runes are really due to Stfli.bo's hand, and they ate fuporc (and to the Abecedarium, for· that matter), that one

certainly by the same scribe as the preceding pages, where would expect the relationship to be, the reverse' (I). Cf. infra.

could he have acquired this knowl~dge ? He was . born in 808 The' English runes added interlinearly in the Abecedarium

or 809 and entered the Abbey of Reichenau when still young (I). are the only instance of an actual comparison between the Eng:-..

From 827 to 829 he studied in Fulda under Hrabanus Maurus. lish and the Norse variety of runes. To be sure, they";re

Walahfrid had. a wide range of interests,· as appears e. g. from found side by side in other manuscripts (e. g. Cotton MS.

his botanical poem H ortulus. In 829 he was called to the Galba A 2, Oxford MS. St. John's College 17), but nowhere

imperial court as preceptor to .the young prince Charles; in 83 8 else do we find undeniable proof that the two systems were

he became Abbot of Reichenau; from which he was expelled compared in detail.

two years later, to .return in 842.-He died in 849· The fupor,c of the St. Gall manuscript is of special importance

It is not impossible t.hat Walahfrid had his runic lore from on account of its context. The extracts from Isidore's chapter

Fulda, that important centre of English culture founded under De litteris are precisely those which were used in compiling

. Boniface, or from' some outpost of that famous abbey. There a short tract on alphabets and secret writing going under

can be no doubt that the English fuporc was known in Fulda; the title De inventione linguarum. . Here the runes are really

some manuscriptS written there use runes as reference marks. ,integrated into the ~ystem of Mediaeval learning: they are on
The inclusion of the Fulda calendar in St. Gall MS. 878 shows one level .with the Hebrew and the Greek alphabets. They
that Walahfrid began collecting his vademecum there, if not have not yet received their own introductory notice,' but· the
earlier. The Abecedarium nordmannicum, too, points to, a necessity of the latter will be felt as soon as Isidore's text will .:-t'
region where Low and, High Getman met; like the Hilde­ have been condensed and cut up into paragraphs, one for each, -:
brandslied, it may have originated in Fulda. Its meaning as a alphabet. .Therefore we may probably consider this part of- :,."
whole has often been discussed, and is still not entirely settled; Strabo's vademecum as -a preliminary state of the De inventione
I believe, however, that more light could be gained if. it were text.-We shall see that there is a special reason for doing so :
connected with missionary plans for. Denmark or Sweden. the. runic ,alphabet in' /)e inventione is a mixture of English
Some knowledge of the runes would have been an important and Norse material, and such a mixture could only be composed
asset toa missionary setting out to workjn the 'North {2): I where both runic systeIOS were known. The possibility of
doubt, however, that Hrabanus was Strabo's mentor ·in runo­ . De inventione having originated from Strabo's c~llectioil. will

be examined in the chapter o'n that tract (p. 377 f.).

(1) M. MANrrrus, Geschichte I, 30Z fT.

B. BISCHOFF, Eine .8ammelJlandschri/t, ,41·
Cf. also F. VON BEZOLD, Kaisenn Judith und ihr Dichter Walahjrid Strabo.
Historische Zeitschrift 130 (19Z4), 377-439; . 8. Ghent, Universiteitsbibliotheek,' MS. 306 (saec. X).
O. HmwING, Zum Problem des Karolingischen " Humanismus ", mit besonderer
Riicksicht auf Walahjrid Strabo. Studium Generale I (1947/8), 389-397; As far as I know the fragmentary fuporc in this manuscript

'G. BAPl!ll.CKE, Die Karlisthe Renaissa1iCe, 15 z, 173, zO<}. .

(z) The early date of the manuscript need not be an obstacle : cf. B. NI!RMAN, has· neve,r been published. The manuscript originally belonged

En wten mission prJ Gotland WI tiden omkring dr 800 e. Kr.' Fomvllnnen 1941, to the Abbey of St. Maximin near Trier, where it was no. 183 ~

'3 0 -40 • The activity of Fulda in the missionary field is' still a matter of on fol. ~r there is an older'pressmark E. 7 (14th century). It

discussion : E. E. STENGEL, Zur Friihgeschichte tIn Reichsabtei Fulda. Deut­

aches Archlv flir Erforschung des Mitte1alteIS 9 (195 z ), 5 13 fT. (esp. 5Z0 ). (1) On Strabo's interest in Gothic- cf. p. 58, note (1).

82 83
...
came to Ghent about 1800 and first belonged to the private .127 PRAEFATIO BEDAEPRESBITERI= Bede, &positW ollegorica
collection of P. Lammens, librarian, who gave it to the Town de ,tnu:tu7a templi SalomonU.
and University Library, afterwards University Library, in ISI8. il9 A Latin elegy, inc. Plangamus cuculum Da/nin. dulciui11U!, nolmmr.
Modem binding (' ANNO 1749 '), pressed white leather on boardS, two
of
'1.2'1. Aletter sent by Peter, Doge Venice,·to the Emperor Henry and
one Hi(l)dibertua, Archbishop (later hand).
clasps missing; on the back the inscription BEDA./IN PARA-/ SALAMO.
Rather heavy and rough parchment, well preserved; a few margins (p. 57 f., On p. 126 the text of Alcuin's letter ends on 1. 16; 1. 17 is blank.
97 f., IU f., 221 f.) have been partly or entirely cut away, prohably to remove
owner marks. The I I I folioll are mal"ked ' I ' - • 221 • on the odd pages;
Between II. 17 and IS the values of the runes are writtetriD.
15 quires, mostly quatemions : uncials and haIfuncials. The runes theInselves follow.on t 19;
7 IV [I-U2) + IV (-I) [113-126) + 3 IV [127-1741 + II [175­ they ate 8-12 mm high. There can be no doubt that the runes
ISo) + 2 IV [I8I~212) -*- III (-I) [213-2'1.'1.). are contemporary with the manuscript : they are in the same
Old quire marks indicate that the codex is made up of .two parts (16 'I', dark ink as the preceding pages, and filled with the same yellow
32 • II " 48 I III " 64 • 1111', 80 • V' but 14'1. ' q. 11('). There can be no
doubt, however, that these two parts were meant to fOl"m a unit : the numbel" colour as that currently used for capitals. It is even probable
of lines, the ruling, etc. correspond exactly in the two parts; the yellow colour that they are by the same hand as Alcuin's letter. In the
used to fill the capitals in the fust part is also found on pp. 131 and 133 in following figure the respective positions of runes and equivalents
the second part.
FOl"mat 265 X '1.00 mm (wl"itten Bl"e8 ca. '1.10-157), 3'1. lines to the page.
have been kept as far- as possible : .
Written by sevel"al hands, which de Saint-Genois ascribes to the late tenth
centUl"y (I). Some hands, howevel", show decidedly mOl"e arehaic features, £"0 rC. t'.U~n 1k
YArfR}..X r N+J 1­
which require an earlier date. I do not believe that the writing can be dated
later than IX'. There are traces of insular influence, e.g. in the abbrevia­
tions. It is not known where this manuscript was written; at any mte it seems
to have been in Trier at an early date. FIG. 14
,-:{',
Contents :
P. 1-'1. Ornamental pages. It appears immediately that the letters indicating the values._
On 1 a croaa in a circle, with on it five smaller circles (in red, yellow, of the runes have not only been displaced, but that the equi- '"..
greyish blue and black);· in the smaller cireles the inscriptions RRR valent of }l has been skipped, whilst that of g is -a halfunciaI t.
(top), VVV (left), A..AA (right), FFF (below), LSP (middle), i. e.
rtlS R _ ruit, 'fJictor Vitalis (lelJiet, aurum a rwbiI rmfert, ferro It is not very likely that this t originally belonged to }l; there
jrigore fame, ? ('1.). Under an arcade on, '1. the inscription Si'lUis is nothing to indicate why this letter should have been shifted
Dbltukrit anatlurma sit. amen. to its present place; moreover this explanation leaves the g-rune
A later hand (14th centUl"y) added on I, by the side of the old press­
mark B. 7 : haec continmtur in hOt; uolrmiine / beds in parllbolas Solo­
without indication of its value. I rather believe that the t going
monU libri trtlS I beds de factura templi Salmnonis expositW ollegorica. with the g-rune is· the scribe's interpretation of an insular g
At the top; N. 183.
in his exemplar. The abseiic¢- of an equivalent for }l maybe
_ 3 (no title) Bede, In parDboim Salomonis libri treI.
explai~ed in. the same way: an uncial or insular d would be
I'1.3 Alcuin, Letter to Charlemagne (no. 164).

~'1.6 The fust u runes of a fulx>rc (3).

rather like the following 0 and may have been overlooked.

There can be no doubt that this fuporc is a copy. ,Only

(I) J. DB SAINT-GENom, Catalogue mitMdilJlU et raisormi des mo:nrucritl

de la bibliothJque de la fIilU et de l'uniwrsiu de. Gand. Cand, 1149-185'1., thus can we explain the shift of the values and the omission of

383. no. 548. one equivalent. The ductus of the runes leads to the same

(2) The first four groups are also found, with a full transcription, in Vienna conclusion. If the copyist rendered his exemplar carefully, it

MS. 751, fol. 39v : cE. p. '1.00 f. and. note (1). Around the cirele there is an
inscription, almost completely effaced; [...] SANCTI MAXIMINI. must have shQwn very archaic forIns: the r-rune is open,as

(3) To the right of the runes, in a later hand: Codex rnonarterii sancti that on St. -Cuthbert's coffin; the c looks rather like a rare

muimini prope treueriis. intermediary between the original Germanic type and the clas­
84 85
.....
~ical.English c-rune (I); this type is only known from one or fuporcs examined in the next chapter) have only twenty-eight
two English inscriptions (the scanm1wdu-solidus) (z),and that runes. There is a considerable amount of fluctuation as far as
of Hantum(Frisia)(3}.llut the scribe'!iJ being unfaIDiliar the runes beyond no. 28 are concerned. Their order varies,
with runes may have led to this form, !Itarting from t~e English and this no doubt indicates that their status was not the same
type with a straight lateral stroke. The value u for th~w-:rune as that of the first twenty-eight. One rune, known from the
also points to ~ rather early .'date. o~ .perhap~ toa Northern Ruthw:ell inscription (Dickins's k), does not occur at all in..our·
model. But only the runes which the scribe .'(for a reason manuscripts.
unknown to ~~) did 'n~t copy could I:).ave all~wed us' to date The fu porc material is quite heteroclitic as far as the rune­
this fuporc more accurately.' ., , llames go-they reflect ,chronological and dialectal differen­
By farthe most interesting feature is the value k assigned to ' ces-but strikingly uniform if we consider the runes. Runic
th~ i.,.run~. This 'is a question, h()wever,~hich can hardly be lore seems to' have been much the same all over England.
eJa!.mined adequately here, for it is an evolution in which the As to its cultural background, that we shall better be able to
name of the rune must have the most important part. There-· cirt!umscribe after we shall have examined other material as well.
fore full details will only be given in conn~xion with an occur­ As a rule the fupores show little connexion with their envi­
rence of a corresponding name beginning withk., Briefly ronment, except that of the Rune Poem; even those that became
stated this seems to be what happened : ,the OE.rune-name part of alphabet collections can hardly be tied down to a set of
gerIJ~rI was interpreted:as 0 H G. ger, her by ,a scribe more familiar circumstances. Perhaps this is not entirely a matter of coinci­
with High Ge~ma,n ;t~ with Ehglish phonology. ' From this dence; nor need the general scarcity of fupores be such (cf.
he derived the value 'k. Thus fuporc~ which provided no rune p. 426 ff.).

for k (i. e. those with only twenty-eight runes) could offer an . One more peculiarity: the few Continental items that have

equivalent for. .that letter when turned into an alphabet.; , come down to us all point to different s~rains of tradition .

.If it had reached, us complete, the fragmentary fuporc of This can only mean that several (not to say many) fuporcs"

the. Ghent ~anitss:ript would probably not have bee,n. inferior were imported to the Continent, a point which the coming

to that of the Vienna Codex. ' chapters' will amply corroborate.

. ,'." CONCLUSION •

The: fllJ?«>.rc ,xnateJ,'ial, whether English or Continenlill, falls

into two main types: some fupores have twenty-eight runes,
others more than thirty. This differentiation is probably based
upon chronological, perhaps also upon regional developments.
Early fUPorcs like that· of Vienna MS. 795 (and the isruna
(l).CE. the fourth and fifth forms of the k-nme in fig. 37(P. 41) of O. VON
Flm;:ImN~8RuP/OTftll.
(a) C. F. KEARy, Catalogue, Ixxxiv f., 1.
C. H. V.SunmtU.AND, Atiglo-SQII:01I Gold CoifWglJ in tlu! Light of the Cnmdall
Hoard. London, 1948, 40, 79.
(J) H. ARNTz-H. ZErss, Runmdenkmi.iln, ass.

86 87
....

 CHAPTER II

THE FUPORCS (II): THE ISRUNA TRACT

.As we saw at the end of the first chapter, the manuscript

fuporcs discussed there may be considered as independent
sources. Apart from the fuporc-without-names in Cotton MS.
Galba A 2 and Oxford MS. St. John's College 17, and apart
from the puzzling case of Cotton MS. Otho B 10, they show
no traces of relationship.
Besides these versions of the Old English fuporc, there is one
which occurs in no less than five manuscripts :
(I) St. Gall, Stiftsbibliothek, MS. 270 (SG);
(2) Brussels, Koninklijke Bibliotheek, MS. 9565-9566 (B);
(3) Trier~ Priesterseminar, MS. R. III. 13 (Tr); ,,,,,,
(4) Vatican Library, MS. Urbin. lat. 290 (D);
(5). Salzburg, Stift St. Peter, MS. a IX. 32 (S).
All five manuscripts follow up the fuporc with a text on runic
cryptography. The four varieties described in the text all
proceed from the same principle: the Germanic fupark was
divided into three sections or groups of eight runes (cf.p. xviii).
Each rune could be defined by two figures : one indicating the
group to which it belonged, the other its place in the group.
To us something like a fraction would be the simplest device
for writing With this system: 1/1 ~ f, 1/2 ,-: u ... 3/8 =0 (I).
The Germanic peoples developed various devices, each of
which uses a different way to indicate the two figures. . Four
such devices are described in the short text : isrUM, lagoruM,
haluzlruM and stopfruna; a fifth device, clop/rona, may be of
a different nature. From the .name of the first device I shall
call this text the imma tract.

I1
(1) Following most recent publications. and to simplify matters, I intend to
use this device to render the runic· formulae •

89
....
 At a glance we shall see that the' five manuscripts represent 19th pentury~and, which Bruckner (I) identified as that of the librarian
two different versions of this tract: the first four are derived K. Kolb, inserted a list of the contents: Conhnmtur in hoc CodiC6, etc. ..•
from one prototype. the last one from another. Closer inspection Quae pog. 511 d8 Tunis habmtur,. tkli1l6tinda curaui, et D. Carolo Grim (Carolo
crossed out) commtmicaui, qui,ea edid;t inli/wo [?] misso Uher die Teutsche Runen.
will show that, amongst the first four. nos. 1 and 2 again belong Gottingen z8n. In W. Grimm's work, however, we find in a footnote:
together, and so do nos. 3 and 4. Hr. Prof. Mane hat ne[i, e. zwei Runenalphabete] dart entdeckt 14M m.ir ""
freiem Gebrauch freundschaftJ.ich mitgetheilt (z). Strong, rather stiff parchment,
yellovrish to almost white, with a few holes. The codex has 33 folios, arn.mged: ,
THE MANUSCRIPTS. ' as follows (pagination between square brackets) :
III (3- 14] + 11(15-33] + IV (Z3-38] + II (+ 3) [39-53J + II (- I)
I. St. Gall, Stiftsbibliothek, MS. 270 (saec. IX). (SG) (I). [53-5 8J + tIl (- I) (59-68J; pp, I-Z, 69-70: paper fly-leaves.
Format z02/z05 x 141 rom (written surl'ace 140/150 x ca. 105 rom). One
column till p. 58, 1. 9; two on p; 51, and from p. 58, 1. 10 to the end; tltJ:ee
The runic material in this manuscript was known as early as on p. 57 (partim). The nUmber of lines varies: *1..11: 31; '*III-IV: 37;
W. Grimm's time, and has often been published since. It was *V: 2.9; *VI: z8. Written in one rather small but elegant hand. There
considered as the only source of the isruna tract until quite seems to be no mention of the manuscript in St. Gall catalogues before 1461 (3);
according to B. Bischoff, the handwriting dOes not point to St. Gall. There
recently. Although mllny works on general runology quote it Ilre traces of insular influences : IT = autem, :l = con, + = est.
or refer to it, no critical study of itS text has been offered thus Grimm and Steinmeyer date ,this codex ' Saec: X', Iiatteiner and Scherrer

far. The manuscript is usually supposed to have been written , saec. IX '; F. Steffens specifies' saec', IX exeunte " and this is'also the opinion

of Bruckner (4) : 860/70-890/900. In view of the many u-like a's (in the

at St. Gall, but it is not mentio,:led in the older catalogues of fullorc and the'alphabet ofily Z against la a's, bUt in the text 18 against 1 3 a's,

that library. At least one authority believes it originated in this seeming to indicate 'that for the 'scribe the u-like a was the more current)

another scriptorium. cf. p. 91. a date s,aec. X may be excluded; B. Bischoff would place it about the middle

of saec. IX. '

Modem binding, Iejlther back and IXIfllers, greyish bb~e paper over boards,
Contents (5) ,
,.!:,
with the inscription: DialscticalAlbini&1lfragmentumf d8 muska, and the no. z7o.
Paper fly-leaves in front (2~ one 'not counted) and back. (r.,.counted):' In iront a P. 3 [AJctualis. contemplatiua. recte credentivtn
, .

(1) G. PJmTz, I talilbri.sche Reise, 46z f.

[P]hisiu. Theoretica. orthodoxorum (other hand saec. IX). Right
top: D. n. 479 (no. of Kolb's catalogue); 1; 4 : Dialectica Albini lEt
W. GruMM; Ueber dinJ.tsche Rtmm, 106 if. and Tab. II. fragmentum de Musica. L 7' the' old mark CC: Below,' reversed,
J. M. KEMBLH,On Anglo-Saxon RUfU!Sj 338 and Plate XV, figs. 4, ,5. a fragment of a 'library 'catalbgue (saee; XV ?).
H. HA~ Dmkmahk 1,417 f. . 4 DISPUTATIOALBINI DE PARTIBiTS DIALECfI'ICA (Rubr.)
G. S'I1!PHI!NB, ,Monuments, I, 10Z f. (no. 10), IIfJ7 (no. ZI), z39; IV, I. [Aleuin's DiakcticaJ.
G. Scmnnum, Ver"eicJan;ill, IOI f. . ",
38 (Expl.) : Socrates non disputat. and a poem: Quid rogo ciuilils ctipiat
E.StlM!RS, Rtinm tmd Ruruminschrift.im.·, 259 n.
cognoscere mores, 1 Haec precepts legat,q'u\, liber iste tenet, etc.
E. STBlNMJm!R-E.' StBVERS, AltJwclukutsche GlolSen, IV, 446 f. (Rubr.) : Disputio de dialecti~ & de uirtutibus' sapienttissimi regis
T. VON GamNBllRGllR, Die angelIiicluischenn4ne11reihm, 6 if. caroU &'albini magister [sic],
F. S~s, Latemuche Pal4ographie", Plate 53 b •
P. LEHMANN, Bibliothtiluklrtaloge I, 114; I. 38. (I) A. BRUCKNER, Scriptoria III, 90.
R. A. B. 'MYNOBS, CtUriodori Senatom Imtitutianes, lCCCiy. (z) W. GRIMM, Ueber deutsche Runen, 107.
A. Q~UCKNllR, ScriptoriaIII, 39, 90, and Plate XXVI.
(J) P. LEHMANN, Bibliothelukattdoge I, no. 2.3, p. Il4, l. 38•
H. ARNTz, Handlnu:h 1, 2.75 f.; Hmu/Inu;h', 173. (4) W. GruMM,Ueber deutsche Rtmen,' 10 7. .
In., Eu- 1md WtU&emmen. .

E. STEINMEYER-E. SIIIVllRS, AlthocMeutsche Glossen IV, 446 f.

ID., Runen und Ru:nennamen, t72. if. '
H. HATTEMBR, Dmkmahk.I, 417 f:
E. RAucQ, Die Runm des Brimel8r Coda: N° 9565-9566,9 ff. G. SCHElUU!R, Ver"ek/miss, 101 f.
J. BLOIIfl'II!LO, RUfUlS, 2.ZZ if. F. STEF1J'HNS, Lateimche Palimgraphiel, text with Plate 53b.
R. DllRoLIIZ, Van Agambertw tot M~, 4B if.
A. BRUCKNER, Scriptoria 111, 90 •0

In., Dubthach'l Cryptogram, 374.

(5) G. 8cHmuum, Verflleiclmiss, 101 f.
90
... 91
 39 Upper IllIIl'gin: Omnium studiorum primam hane approbat quinti­
lianus. (Rubr.): DE MUSICA DISCIPLINA. and to the right its value, TUnes f to p on l. I,
[From ClUllidodorus's-lmtilutionu, De musica cap. V, with figures] • X to ea on 1. 3.
• 5 (ExpL): ... huius scienti~ uobis atria patefacit, and figure. Then, II. 4--' :
without title: a runic alphabet with above each rune its value;
A. Difinitionum prima est OY~DtJHC, latine substantialia... variant runes for a (2), d (2), g (3), 0(2), p (2),
B. Secunda ~es difinitionis est quae greu HNNOMATIKH , s (2), t (2), this giving a total of 30 characters,
dicilui, I latine notio nuncupatur.
Various excetp18, int. d. from Cassiodorus's Imtituti.l:nr# : .
written in two lines: a-p, q-z.
48 p[a]rsautem rethorica est sicut magistri trsdunt secu1arium littersrum The height of the runes varies between 5·and, 7 mm. ;
bene dicendi scientia, etc. Argumen.tstio dicta est q~i argut~ men­ they are drawn with evident care.
tis oratio.
49 Definitions of aritlmultica, teometrico., flItrtmMnia, ,apiMaia, mathematica. II. &- I I : the text on the iisruna, with the example comi.
SO DE DOCTRINA CHRISTIANA AUGUSTINI (Rubr.)
51 Quattuor a quadris consurgunt limite uenti, etc. (poem on the II. 12- 1 3 : the text on the lagoruna, with id.
winds). . - ­
52 A fuj;)orc, a runic alphabet aDd the isruna text (cf. Uifra); below, fonna II. 14-- 15: the text on the hahalruna,. with id.
-of the verb ypJ..po.: rPAtJlO rPAtJlHC rPAtJlH rPAtJlOJf1l
rPAtJlHTE rPAtJlOYCHN and an 'I,lIlintelligible scribble: facitaset­
16-18: the text on the stofruna, with id.
dileatioassidc(with I above c)esitintmcnosf. n. 19-20: the text on the clofruna,.
53 (No tide) Topicorum siue locorum ex quibus arguments ducuntur
triplex est diuisio ••• (on rhetoric), ending ;
54 A causis I argunumtum eat quando consuetudine communi rea qu~
In the fuporc there is a short vertical stroke to the lower left
trsctatur. of the fe-rune, and again of the s-rune and the o-rune, thus
55 Explanations of rhetorical tenna: [El!lergis, Leptologia, -Sindiasmos, dividing the fuporc into four groups of letters: 8: 8 : 8 : 4-.
Ypallage, Metriasmos ••, p. 57: Prathema .i. interrogatio. As the ink of these strokes is darker than the brownish sort of
57 Greek-Latin glosses (in three cola.), and further explanations. Gram­
matica est pulchra loquendi peritia a poetis illustribus auctoribusque I ink currently used in the manuscript, they may well have been '~ _.
collects ... Versus est copulatio legitimorum pedum cum certo articulo added by a later reader. Moreover there are pencil marks - :,,'
et cum cerro finel ... after r, n, X, m and a (the ac-rune), which mark off five groups
58 Definitions of figura, dijini&, genus.
Differentiae (uesper, ueapere, uesperum, uespera) I. 10 fl. : Biblical of five runes and one of three. A clue as to the origin of these
glosses, in z cols. : In matheo (rubr.), (59) in luca. marks is found in the list of the contents on the front fly~leaves :
59 Greek-Latin glosses.
II..: Eadem pagina 52 agit etiam de tribus scripturis secretis
6z b De catalogo ieronimi: Greek-Latin glosses.

6." Quibvs locis apo,stoli iaceant.

(Geheimschriften), quarum unalineis perpendicularibus I,
6.b Anticristos appeUatItT ... altern uncis, tertia punctis . : : . constat. Eae enuc1eantur hoc
Lat. glosses, with a few ORG. words interspersed.
modo : litterne alphabethi diuiduntur in quinque vel sex partes,

65" De regnia miliciaque uocabula (rubr.)

quae hic versus nominantur, ut cuique parti -quinque litterae

67 b Roc subter iaceant defuncta cadauera fratnun.

obtingant. Jam pro littera quacunque denotanda scribuntur

The arrangement of p. 52 is as follows : (I) duo signa ejusdem generis,» etc. FiD.aIly, small crosses have

11. 1-3: a fuporc of 28 runes, with above each rune its name, been pencilled to the lower right of h, and h. These indicate

the divisions of the fuporc as found in the Brussels manuscript

(1) Facsimiles in: G. STID'HBNB, Monumenti IV, I; (p. 100); they are probably due-to a modern reader who wanted

F. STl!FJ.IBN8, Lateinisthe PaliWgTaphie I, Plate 53 b;

to compare the tw() versions.

A. BRUCKNJ!R, Scriptoria III, Plate XXVI;

E. RAuCQ, DU RU1Ie/t des BrUsnler CatWe N° 9565-9566 (plate reproducing The fuporc wiJI be discussed with those in the related manu­

Steffens's). scripts (p. 122. ff.). The runic alphabet belongs to the matter

92.
... 93
of Chapter III and will.be examined there (p. 217); yet a few 2. Brussels, Koninklijke Bibliotheek, MS. 9565-9566
general remarks on this alphabet may not be superftuous here. (saec. IX ex.) (B).
The alphabet is not simply the alphabetization of the preceding
fuporc: it restores p in its original value (against the fuporc
value q), and has a new q. It has a variant for d which is n~ As far as our subject is concerned this manuscript is closely
other than w; the similarity between., and w no doubt led to related to the St. Gall codex. In many ways it helps us to
this confusion. , The question whether the alphabet may throw understand the latter. Its runic material was first edited some
ten years ago by Miss E. Raucq (I).
any light on the fuporc is not easily answered. From the
correction of the p/q-confusion, one might feel inclined to At one time this codex belonged to the Chapter Library of
the Abbey of Saint Laurent in Liege; it first had the press­ ,"
attach great' importance to the alphabet; but when w is listed
as a variant form of ,." we, wonder whether the correction was mark 6~6, afterwards 5~14; below the waved stroke: Sti
Laurentii in C. Leodii. Since it is not mentioned in the 12th
not the result of a coincidence rather than of runological skill.
As to the St. Gall text of the tract, it shows some anomalies and 13 th century library catalogues of that Abbey, it will
which point to a faulty transmission : probably havebee'n acquired after that date (2), although it
iisruna dicitur qu~ ... scribuntur; hahalruna dicuntur istc< may have reached the Liege area before. B. Bischoff formally
qu~ ... ostenditur (for ostendit or ostendunt); subtiliter
recognized the handwriting of ~ol. 8 as a typical St. Gall hand. '
(for subtus litera). Contacts between St. Gall and Liege must have been numerous
It. may consequently be several removes from the original. at one time: we have only to think of the St. Gall monk N otker
The manuscript context of this bit of runic lore is worth who became Bishop of Liege (972-1007). I have analysed
noting. This handy codex may well have been a teacher's background, and contents of the Brussels codex in another
study; therefore not all details will be given here (3). ~,
manual (I): It contains matters to be taught in the later stages
Modem red calf binding. P~t of good quality, some leaves palimp- _ .
t
of the trivium (dialectics, rhetoric), and also some for the qua­
sest (d. infra). The manuscript «:onsists of two parts: A = fols. I-U., B =
drivium (music). Of course we cannot infer. from this situation fols. 13-196. In A, the order of the fo~ is disturbed (pencil marts 5' 4,
that the cryptic systems ,explained in the tract were actually I x·' ~ 2, lZf 3, 13· = 5, then regularly 'on ttie first page of each quire:
r
taught in class. P. :52 being the last page of a quire, the text 2I v = 6, etc.); *1 = fols. 1-4; *n fols. '5-12, originally arranged as follows :
12V -., II -',Sf_ IOV. B consists of 23 IV and has 'an old quire numbering
V

on the runes may simply have been considered as a stopgap.

On the other hand, a magister with a sense of paedagogics may
v
=
20 v= I, 52v ~ V, 60 ' VI, 6S" = VII, 76v VIII, roo" = XI, loS' XII,
u6 XIII, 124'" = XlIII, I32v XV, and traces on fols. x56vand l72';
well have used it, to relieve a dull stretch in his course, e. g. at least Olle quire is 'missing at the end. ' - '
in connexion with the Greek numerals. A comparison with the Format 23 2 Xr75 mm; th~ written area varies in A; in B ISO X X35 nun;
A has' 19 lines to the page, B 2S. IHs, worth IlOting thai: some of the first
Brussels and Vatican manuscripts shows that we have probably pages of the matluscript; especially fols IT'v, 2f - .., 7", are palimpsest. The
to favour the latter possibility. older writing has been carefully remo~ed, but the roughness, of the surface
of the parchment imd some traces of an older lineation prove that some other
(1) We find similar textboolr;s in St. Gall MSS. 273 (together with s «:ollec­ text was first written on these pages. The date proposed by P. Thomss, viz.
tion of poetry),S55 (with gr&mmatica1 items), and especially 199. We can saec. X, may be somewhat late; E. Raucq and B.Bischoft are in favour of
only regret that these and innumerable similar codices in other libraries have
(I) E. RAuCQ, Die'Rutum des BrlUleler CodeJe N0 9565-9566.
only been examined for separate items of their «:ontents, and have never been
(z) J. GI!SSLI!R, L4 bibli6tMqru de 1'lIbbaye de Saint-Laurent d Liege au XlI"
studied as units lI8!Iembled for a very definite purpose. Only by !!Onsidering
et XIIIe sikle. :Bulletin de la Societe des Bibliophiles Liegeois 12 (19z 7). 9X ft.
them as units, and by taking into account small fragments of apparently: minor
(J) R. DmtOLllZ, Dubtlraeh'l Cryptogram. This paper was' printed before
import:a.nce; cah we obtain a more accurate picture of the teaching process in
Prof. B. BISCHOFF exiunined the facaimile in Miss E. RAuCQ's paper and kindly
Mediaeval schools. Cf. J. M. CI..llRK,TIW Abbey of St Gall, 97 ft. informed me of his findings.

94
'" 95
 the late ninth century (I). The two parts of the manuscript must be fairly 13' : The following enigmatic lines:
contemporary; B may be somewhat earlier than A,as the text of fol. IOV is Vt siculis omnibus

continued on I3 f• IVDIM. et cetera. q. NOM. M. R. F. V. T. C. S.

When we rearrange the first I2 folios as they were originally bound, the IUPA multo N. NV. ET. I. P. R. S. F. O. A. R.

contents shows this order: Patres hi quos uidi et cetera .

fol. If-3 f : Subdivisions of philosophy, u.-etica (II), practica (1"", 2f,) logica(2 V), A,I.L.P,M.A.Q.M.P.A,N.O.S.C.A.R.A.B. cons.
, spiritalis (3'), with diagrams (2). L. L. ex. q. nihil. A. O. nihil. ufo s. post. s. p. ex. I.
3"-4", 12', n V - ' : subdivisions, definitions and diagrams relating to Et cetera mortis
rlutorica. VBI N. G. S. Q. R. P. H. A. B. In qua urbe uiuimus
&pI. fol. 5v : HUCUSQUE RETHORICA. hic hic. S. NI. NUM. P. C. similes (I),
U V : seems to supplement the contents of fol. If; Porro sapientiam followed by a cryptogram (d. infra).
ueteres philosophiam uocauerunt id est I omnium rerum humanaIum fol.. 13": MARTIANI MilNEI FELICIS CAPELLAE I AFRI CARTAGI­
atque dininarum scientiam. Cuius I philosophiae tres esse partes NIENSIS I LIBER PRIMUS INCIPIT I DE NUPTIIS PHILO­
dixenmt, etc. Continued on LOGlAE ET MERCURII. At the end one quire seems to be
S': De phisica: subdivisions and definitions.
missing, for the text ends abruptly:
5v -6': Subdivisions of dialectica (with diagram)
196": '" et simplices quidem dicuntur qui tem[...] = ed. A. Dick,
6'-1"1: Subdivisions of Aethica; the four principal virtues and their sub­
p. 520,.1. 4.
divisions: prudent.ia (6V), itUtitia (7<), fortitudo (7"). temperantiD (7V ).
8': Runic material (d. infra). The cryptogram on fo1. 13 r consists of Greek numerals:
8"'-9' : Computwl table (3); in the right hand margin of fol.9 f a short I
text has been erased. IB E IZ IB E Ir. IZ E EA. r ./.d Ir r if Ir. IH A]
9", lor : Two computiatical tables. IA KI9 E IB
10" : Concordance of Greek and Latin numerals with the letters of
the Latin alphabet: I A a, II B b, III r c... XXIII K.r z, then
XXIIII 1Ld..• XXXI LA (i. e. AA) (4). Concordance of Greek IH K A .d B A IZ IH r IZ 9 IE IH (7 19
and Latin numerals, from A I to JfI DCCCC. , At the bottom of the page there is this note : ,.t,
Hic erras, Dupthace, in tuis notulis scribens H pro (7 uel
;'
(I) P. THoMAS, Catalogtu des mamucrits tk classlquu latins tk la Bibliot."~que r.;,.'

royale tU BJ"Weelles. (Universi~ de Gand, Recueil de travaux pubIi& par la (I) Professors ,B. BISCHOFF and P. VAN DE WOBIITIJNE kindly helped me to
Facuitt!: de philosophie et lettres, IS) Gand, 1896, 32. solve this puzzle. The first three lines are from Cicero, In Ven-em I (in A.
The date given by MARCHAL, viz. the first third part of the eleventh century, Caecilium I)
may be left unconsidered (F. J. F. Marchal, Catalogtu des manuscritl tU la [Cum qUaestOf in Sicilia Mesero, iudices itaque ex ea provincia
Bibliotheque Royale des Duu de Bourgogne. Inventaire N° I. Bl'UXelles- decessissem,l ut Siculis omnibus iu~dam diutumamque memoriam
Leipzig, IS42, 192). , quaesturae nominisque mei relinquetem, factum est, uti cum summum
(2) I wish to thank Prof; O. HOMBURGER (Berne) and Dr J. DuFr (St. Gall) in veteribus patronis multia, tum nonnullum etiam in me praeaidium
for having drawn my attention to a number of manuScripts with identical or suis fortunis constitutum esse arbitrarentur.
similar diagrams: Berne MS. uS, St. Gall MSS. 199, 273, S55 (on these three the second group from the same work, V, uS;
cr. also p. 94, note I). Such didactic devices-they look very much like Patreshi, quos uidetis, iacebant in limine, matresque m.iserae pernoc­
schoolboys' (or teachers') sununaries-occur probably in many more manu­ tBbant ad ostium carceris, ab extremo eonspectu liberum exc1ua&e;
, scripts and deserve more attention than they have' received thus far. ,They quae nihil a1iud orabant, nisi ut filiorum suorum postremum spiritwn
go back ultimately to Cassiodorus, but seem to have been l,Ised by other scholars ore exeipere ~ceret.
as well, e. g. AlcWn. For obvious reasons R. A. B. MYNORS, CmriodMi Sma­ and the last group from in .Catilimtm I, 9
torillnstitutious, xxxiv, could only refer to manuscripts with e:x.i:erpts of (0 di immortaleslJ ubinam gentium SumU8? quam rem publicam
some length. habemus? in qua urbe uiuimUlJ ? Hie, hie sunt, in nostro numero,
(3) Cf. BEDE, De temporum ratiou ed. C. W. JONBS, BetkuJ Opera de tempo­ patres c9nscripti •••
ribus, 356 : " Quot sit 111lll\ in kalendis per cyclum ". Only the first lettef (Of letters) of each word are written. A rew devistions
(4) Cf. BEDI!, o. c., 219 : "Pagina regularum n, but with Greek instead of from this rule rosy indicate that we have only a copy. Cf. RRR, etc. in
Roman numerals. Ghent MS. 306 and Vienna MS. 751, pp. 84, 200 f.

96 u
97
""
 pro e / uel pro aspirationis nota qu~ nec secundum brit­
Irishmen who had left· their homeland, possibly on account of
tan~cam linguam/in" i / pso nomine bene resonat.
the Viking raids, were living at the court of Mermen or
With the help of the concordance on fol. lOv, which, if we
Mennin(Merfyn Frych, 824-844) in Wales. One of them,
restore the original order, faces fo1. 13r , the cryptogram on the
called Dubthach, travelled on to' the Continent and thence he
latter may be read : . '
sent a cryptogram to his countrymen, possibly as a revenge
MERMEN REX CONCHN SALUTEM. SUADBAR
for some or other slight he had . suffered from them. With
SCRIPSIT.
the help,'of Greek annals (i. e. annals using' Greek numeralsr)
This same cryptogram is also found in Bamberg MS. Class~ they' decyphered the cryptogram and in the name afhis three
6 = H.J.IV. II, fOl.I09 v • There its two parts are imbedded companions Caunchobrach, Fergus and Dominnach, Suadbar
in a letter which enables us to reconstruct the circumstances passed on the cryptogram with the necessary explanations to
in which the cryptogram ~as: composed (1). A group of their master Colgu,' in order that he might teach· it to other
Irishmen intending to travel to'Mermin's court, so they would
(x) On the Bam~rg manuscript cf. the KatalDg tier Handschriftm tier not have to blush on account of their ignorance.
KiJn.iglic/umBibliothek RtU'Bamberg; Leipzig, 188, ff.,I,2, g f.; on
its text The Concen or Concin mentioned in the cryptogram (the
J. L. HBnmito, Et lilkBiJrag iiI Belymiilg qf Middelalderens Ktmdskilb til
Grtflsk. In: Oversigt over det Kongelige. Dimske Videnskabemes Selskabs
mistake Conchn is corrected in the note addressed to Dubthach)
Forhandlinger od dets Medlemmers Arbejder i Aaret 1889, xgB fl., and my was king of Powys (808 .- II) and died· a pilgrim in Rome
paper (p. 95, note 3). Full bibliography ill J. F. KENNEY, Sources I, 556, (854 or 855). Since Dubthach, Fergus and Suadbar may, be
§ 363 : The Bamberg Cryptogram. Including a couple of ObvioUli corrections,
the text of' the letter is as follows : .
identified with Irishmen known ftom other manuscripts, there
H\lC est inscriptio, quam Dubtach in srce Mermin Brittllnnorum regis denIisit may, be some hlstorical background to this letter. St. Gail is
ad probandos Scottorum sapientes, se ipsum excellentissimum omnium Scot~ precisely the place where we should expect such a cryptogram .....""­
torum Brittonumque opiruu::ui, scilicet putans ilullum' Seottigenaruln quanto to tum up, cf. p. 154 ff.
magis Brittonum doctorum in presentia Mermin regis istam scripturam perle­
gere atq'l:le intelligere potuisse. Sed nos Caunchobmch, Fergus et Dominnach . The environment of the isruna tract in the Brussels manu- ... "
et Suadbar opitulante Deo illa scriptum non Iatuit,· per annalem Grecorum script connects it closely with the St. Gail codex: there too we "
lillellum atque a1phabeti eandern inscriptionem inuestigantes. IB E IZ IB find at a few folios' distance (p. 45): Secunda species difini­
E lr. IZ E KA. r IA Ir r H lr. IH A IA K 18 E lB. IstiUli scriptu~
talis est, sensus: Mennen rex Conchen salutem. Si' 'ergo uolueris istain tionisest quae grece 'HNNOMATIKH dicitur, latine notio
scrlpturam dinoscere. perspicaci mente p~sciptBln 'Grecorum annaIis campoti nuncupatur, etc.; (p. 48)P[aJrs autem rethorica est sicut magistri
aerieiD'latinssque sequentes Iitems post ipsa greea e1ementa ordinatas animad­
IiClttito, lItque" cognoscito, !atiDU litems subsequentes grecis e1ementis predi­
centibus conuenire, sicut in pf:\lsCriptis lineis designsuimus. Cum ergo IB
grecas uicelicet literas in praefata Dubthachi scriptum aspiciendo uideris, Britonum regis illam inscriptionem non intelligentes erubescant. Nos autem

respice greclU'ilm titetlitum· aenem ante scriptam, atque in illa serie IB greca coram Deo testamur, quod nee causa' elationis aut tumid(l inBationis, quod

elemenfa tuilm>uirum (?)non latebunt; et quia XIImum e¢em Ii~ optinent absit, istam. uobis transmittimus expositiooem; sed istam latebram uestr&m

locUm, necesse' est; ut XlImamlatini a1phabeti m litersm designent. Item sanctitatCm Iatere fraterno amore non passi SUInus. Omnes in Christo fratres

gaudete. Valete.'
quia E in ipsa greci calculi eerie quinrum possidet locum, reete quintam latini
a1phabeti e Iiteram esse designatBln, atque ita' per c\ltem decurrens totum sensum IH K A A B A IZ. . IH r IZ 8 IE IH 8 18. Hie errss, Dubthsche, in

ipsius uel similis descriptionis intelliges. ', tWa notulis scribens H pro 8 uel pro e uel pro sspimtionis nota, qU(l nee se­

Norum sit tu(l prudenti(l, optime Colgu nosterque doctissime magister, cundum Britamucam linguam in ipSo termino bene sonat.

quod non quasi tibi ignoranti iatBln expositiunculam tranamittimus; sed The scribe signed his work by a cryptogram using the same device :

8uppliciter poscimus, at istam explanationem ignorantibus et' simplici<!ribus Ir A Ir 4 H A IZ 9 K IH IH A Z A KA B 14 Ir 14

!lOeb Scottigenis fmtribus trans Britannicum mare nauigare uolentibus per A"~mMrM"mrU~~mEUAffl8mmAm
tuam beniuolam caritatem insinueS, ne forte in presentia Mermin gloriOsi A IZ 8 18 H IB E 18 8 I A IB (= Narulliarius sagax oono animo
coriScripsel:at istam arlthmetikam).
98
: ... 99
 tradunt secularium litterarum bene dicendi scientia, etc. ; St. Gall manuscript. That the clophruna should not come at
(p. +9): definitions of arithmetica, geometrica, astronomia,
the beginning seems rather obvious.
sapientia, mathematica. This can hardly be a coincidence, as
As ii stands, the fuporc gives up the old division 8 : 8 : 8 4 +
. is also shown by features CQmmon to the two isruna versions for 9: 9 : 10, and this is of fundamental importance. There
themselves; .we shall see that in all probability the two manu­ was little chance for an example chosen at random to tally
scripts ultimately go back to the same original. well with both systems (all letters should be taken fromtbe
In the Brussels manuscript the runes are arranged as follows first group of eight). There is no doubt that our scribe (or the
(there is no definite lineation; the scribe seems to have followed compiler of his exemplar) had become aware of thisi:lifliculty.
the horizontal lines of the table on fo1. 8 v, but even those he This is shown by (2 (value 0) being shifted from the twenty- .
followed only freely; consequently the indications of lines in fourth place to the twenty-sixth. The reason was of course
the following description are only approximate, d. plate V) : that in the examples the 0 (= the w-rune : cepel has become
11. 1-6: the fuporc, in three lines: f - h (1. 2), n - b odil) of corui was indicated by the formula 3/8; by removing
(1. f), e -ea(l. 6), with above each rune its name two runes to the preceding group, 0 would have become 3/6.
(It. I, 3, 5) and to the right its value. There were two solutions : either to change 3/8 to 3/6 in all
1. 7: blank. the examples; or to shift the ce-rune to the eighth place in the
11. 8-9: Clophruna text. third group. Our manuscript shows the latter solution; the
11. 10-1+:
Isruna text with at the end the example (1. 1+). former will be found in the Vatican and the Trier manuscripts
11. 15- 16 : Lagoruna text with id. (1. 16).
(d. infra). But when shifting the 0 to a place where it would
11. 17- 1 9:
Hahalruna text, with the example inserted after fit the example, the compiler overlooked that there was another
isle (1. 17).
letter not belo~ging to the first group: i. Instead of 2/3, this~~<
1. 20: blank. had become 2/2, but neither the order nor the examples were
It 21-22 and 25-26: Stophruna text, with the example changed accordingly (I). . : ,.
in between the two parts (11. 23- 2 4). As to the text itself, it calls for a few remarks. The plural
The text is written in a sprawling, rather careless hand, dicuntur is used after all the -runa compounds, except the
which is also found on other pages in the first IZ folios, so there clophruna, implying that tuna is considered as a (neuter) plural
can be .no doubt about its being contemporary with the rest or a collective. But this use is not carried through everywhere :
of th emanuscript. The carelessness of the scribe appears into al. Hahalruna dicuntur ... que ... ostendit; on the other hand in
in the size of the runes, which varies betweell 6 and 13 mm., the stophruna paragraph we find the pronominal form illas
and in the runes of the example illustrating the lagoruna : the referring to these stophruna. ·And then there is of course the
, shorter' I-runes are simply inverted v's; the ' longer' ones enigmatic ciophruna paragraph (cf. p. 134 ff.).
too are strongly inclined to the right. For further details on the fuporc' see p. 122 :fl.; discussion
The scribe (or his exemplar) seems to have wanted to indicate of the text p. 131 ff.
the subdivisions of the fu)?orc': he tleliherately left enough A comparison of Band SG shows that their texts do not
space for two more runes after hagal and berg, and marked off depend one on the other, but that both proceed from a common
the three rettir by a positura. A positura is also found after source:
the clophruna paragraph but not after the other paragraphs :
it probably indicates the end of the treatise, and may imply
(I) The Z/3 of the manuscript can hardly be interpreted. as j (i. e. cOTUj) :

that the first paragraph originally came last,' as in the the j-rune is never given the equivalent i.

100 101
 <;:: = loa-us IV + 6 lQOse leaves;

B : ceri ,( ?), elox SG: cen, elux

D = 116-121 = 6 loose leaves;

tag ase ac odil tag odil ac asc

E = 122-137 = 2 IV.

Page ca. 265 X 188 mmj through trimming some fols. (int. al. fol. uS)

dicuntur (1. 10) dicitur (1. S)

are slightly smaller. Written surface; 210 X x3S mm. Fols. II6-121

subtus litera (1. 22) subtiliter (1. 16)

palimpsest. Ruled for; A: usually 28 D.; B: 33 n.; C: 28-33 11., but 50

qua (1. 26) qui (1. IS)

on fol. 114; D ~ 33 n., E: So-S2 D. These parts were written at different

ostendit (1. IS) ostenditur (1. 1+)

times. A, B, C, belong closely together (by one hand?) and are dated saec.

Xl/XII; Dbelongs to nec. Xlll, E to SIIeC. XIV.

These last two forms may go back to an abbreviated ostendl.

The two manuscripts also· interchange the names and/or the . In the vocabulary part there are traces of English influence :
places of C2 and a : under me we find an a-rune with the value e. g. radamus (: vespertilio, OE.. hreaaemus), uua.ndauuerpa

aa, and under DC an m-rune with value a; in B the me-group (: talpa, OE. wandew8orpe); abbreviations, e. g. autem;

N = enim; + = est, etc. But when Baesecke says :

comes. first, in SG that with DC. In either case the order has
« Dasz er (= Cod. R.I1I. 1 3 des Trierer Priesterseminars) ...
been disturbed, but whilst in B the runes and their values
insulare Ahnen hat, zeigen iiberdies insulare Abkiir­
have changed places, in SG it is the names.
One difference, finally, will be examined later : B has cloph­ zungen, ags. Sprachspuren und das Runenalphabet
von· HI. lIS v, etc. " (1),
and stophruna, SG elof- and stofruna.
I mUst take exception to his last argument, for this " Runen­
alphabet" had been transmitted for a long time on the CQn­
3. Trier, Priesterseminar, MS. R. III. I3 no. 6I).
tinent, and is not entirely comparable to the Engl~sh character
(saec. XIJXII) (Tr).
of the glosses or the insular character of the abbreviations.
The connexion of ·the imma tract with the' rest of the manu- ".
This manuscript has been examined especially for its ORG.
script (or eve'n with part C) is rather slight: Perhaps we may'''·
glosses, which were edited first by Steinmeyer and again by
point to the fact that it is found in the neighboUl:hood of me- . '
Gallee and by Katara (I). This last author's very careful
ieorological material in the Vatican manuscript as' well, whilst "
study provided the basis of the description following here.
the latter and the Brussels version are c~nnected· with com­
Brown leather binding over wooden boards (nec. XV), rests of two clasps.
137 leaves of parchment of various qualities. The codex is made up of 3 or
putistical items. Tr, however, hardly offers any due to explain
even 5 different man,uscripts (2), arranged as follows : its inclusion and still less to show for what purpose it may
A = 1-38 IV [t-8J + IV (- I) [I7-23} + IV [9-16J + VII (+ I) have been used. . .
[24:'3 8Ji The manuscript formerly belonged to St. MaximHl Monastery

B = 39-101 6 IV [39-86} + IV (- I) [87-93} + IV [9+- }; IOI

near Trier (fols. Ir; IV and 122":). Where it was written seems

(I) E. STElNMBYllR-E. SI;BV1!ltS, AltIwcluleuuche Glossm I, 314; n, 334. 590; to be unknown. For the OHG. glosses of C (which also

III, 432, 457 if., S70 if.; 'IV, 195 if:, 246 , 330 ; 620 f. (description). contains the isT'Una tract) Katara assumed a Middle Franconian

, J. GALLO, Uit BililiotheeluM mArchieven. Tijdschrift voor Nederlandsche original with some admixture of Low Franconian; this seems

Taal- en Letterkunde t3 (1894), 257-3 02 (267 if.).

ID., Vontudim, passim.
to indicate that the manuscript may be preserved not far from

P. J{ATARA, Dk GlorSf:!J des Coikx Seminarii 'r.,eui..,ensis R. III. I3. Text­ the place where it originated (perhaps somewhat more to the

ausgat>e mit Einleitung und Worterverzeichnissen•. Diss. Helsing£oIS, 19 12 ., North and East ?)., .

See also H. V. SAtJl!lll.,AND, Aus Handschriften der Trieret'Semi.nartnbliotlulk.

Gontenla :

Neues Archlv 17 (1892), 6t6.

fols. Ir_38v (= A): Seduliu8, Carmen paschale and Hymmu I, 1-5.
G. B~. Vocabulari'us, 73, 93, 95 if.
(2) As far as the 'composition 'of the manuscript is concerned, "the descrip­
(I) G. BAI!S1!lim. Votabularius, 97.
tions of STJlINMB'lI1I:tt and KATARA do not agree on all points.
103
102 ....
 39r-Ioo" (= B): Julius Solinus, 1>8 titu' orb;" terrarum et sl!ClllU
mirabilibm quae in mundo habentur. partly damaged. Other mistakes may result from the same
lOOV: Incipiunt presagia tonitruvm XII memium. cause: fed (for feh), doro (for dorn), tan (for ti?) (I). The
IOI V : Exordiuntur uero pre6gvrationea tonittvvm VI feriarum. omission of the clopf1'1lna paragraph may be explained in the
Inchoant predemomtrationea tonitruum horarum.
(all from Pseudo-Bede, 1>8 tcmitrais libellm). (1).
same way. A number of mistakes and divergences may be
lcaV , I. 18 - fol. 114r : Latin and OHG. vocabulary. due to other reasons, e. g. to the scribe's inaccuracy:
114 - fol. lISr: Life of St. Dionysius (unedited ?). Inc. Speciales I. S Iisruna .... i. litteram ... scribuntur (imtead of littera);
francorum protectores sunt dionisiuS lrllIrtyr domini preciosus.
115',1. 11 : Metrical Latin proverbs, printed by Katsra. Inc. Tantillis
l. 6 breworibus .I. ... longioribus .i.;
tantuM/nanciscens dogrrutta tantum. I. I I Hahalruna dicuntur qure ita scribuntur (Hahalruna ita
1IS": The imma tract, followed by Greek numemls, their value scribuntur in the other versions);
and nll1llea (d. below).
fob. ll6'-121" (= D): Marbod, Liller lapidum. Inc. Incipit Prologium
l. 14 qme supra in punctis quotw sit uersus subtw litteris
libri lapidum: Evax rex arabum legitur scripsisse neroru ... ostendunt;
122'-137" (= E) : (12Z" blank) Petrus de Riga, Aurora (incomplete). I. 16, 17 ut supra sint puncti qure Iitteram / et subtus ordo
Arrangement of fol. II SV : uersum.
II. I -3: fu~orc, with values to the left of the runes, and The arrangement of the fuporc as given in Tr has influenced
names above. The 27 runes are subdivided as the examples.· By adding the first two runes of ~he third group
follows: f - w / h - b / e - q. to the second, the scribe (or his exemplar) shifted ce from the
I. 4: blank. eighth to the sixth place. Hence the 3/8 which indicated the 0
II. S-8 :' lsruna paragraph, followed by example. e. originally ce) in the examples had to be changed into 3/6.
9-IO : Lagoruna text UnterIinear transcriptions in a thir­ In the hahalruna the scribe omitted a stroke to the left side of
11. 1I-13: Haha/runa text\teenth century hand (cf. infra). the last symbol, writing 1/8 instead of 2/8. On the other hand
II. 14-17: Stofruna text, with two examples. the text on the hahalruna shows the correct form ostendunt ..
The rest of the page was used by a fourteenth century hand (1..12; B has ostendit, SGostenditur, cf. p. 102). "
for inserting the Greek alphabet and numerals. The most striking difference lies in the examples. . Whilst

F 01. II Sv must for some time have been the last page of the the two versions mentioned before gave cryptographic renderings

manuscript: the writing is much worn, and a later hand retraced of corui, the Trier version transcribes coruw (1/6 3/6 I/S 1/2

a few words or letters and added an interlinear transcription 1/22/8). On the priority of one or the other reading, cf. p. 13 2.

to 11. 9 and I I : The repetition of the example in the stofrrma paragraph is

(Lagoruna) diciit q ita sc1biit2 (per I) Ira vt n n corui; unparallelled and may be due to the scribe of Tr; of the hahalruTla

(hahalruna) diit q ita sc1biit2 i sist8 pte q<>t9 vrsu. text, too, Tr offers a more expli~it version. It is hard to decide

From the appearance of this page in our manuscript, a few why the clopfnma paragraph is missing. Either it was omitted

clues on the history of this version and on the exemplar from by the scribe of Tr (or in his exemplar) because the device

which it derives may be gathered. The final letter of the name described in it was quite different from real runic crypto­

of the b-rune is missing, and ~o is the last rune (ea, %, aer, in graphy; or else it is an addition to the common ancestor of

SG and B.
the other versions). It is quite probable that these defects are
due to an exemplar which was hardly legible or had been
(I) The Rev. Krimer of the Trier Prieaterseminar read dola or dola rather
than dora; but-since the scribe uses a very long r (e. g. in rai) dora does not

(I) Cf. C. W. JONES, Bed4e Puudepigrapha, 45 if. seem to be entirely unlikely, .

IOf
.... lOS
 On the whole the Trier text seems rather to be a revised
IV [Bedae] compotus vulgaris, qui dicitur ephemerida (fragm.).
edition of the text found in the two other versions. 'Its relation V
2 Ratio spere Pitagorae quam Apuleiua descripsit.
to the rest of the manuscript seems to be quite accic,lental': it 3' Petosiria Micipso regi salutem [Pseudo-Bede, De divi1Ultkme mortis
et vitaeJ.
was simply added on a blank page, just like the Greek numerals
3" Remi Fanii (i. e. FaviniJ de ponderibus et mensuris.

below it (which were added two or three centuries later). ,. Tables of a paschal cycle, with annalistic notations refe.tring to

' 'Btaunweiler. "

4. Vatican Library, Codex Urbinas Latinus2C)O

Y
I2 Compotus Herimanni ri.e. Hermannus Contractus of Reichenau
: on the age of the moon]. •
(saec. Xex. {X1.) (U). 32' Tetrastichon authenticum de aingulis mensibus.
33" (A.usonius·sJ Monosticha de mensibus, and other computistical and
astronomical poems.
Well known to historians for its annals (annates Brunwil­
51" Libellus calculatoriae aetia Helprici [i. e. Helpericus of St. Gall,
larenses), this manuscript once belonged to the Abbey of St. ComputusJ.
Nicolas in Braunweiler near Cologne, founded in 1024- by 62'" Ciclus grsece, circulus latine etc. .Astronomical excerpts.
6
Count Palatine Ezzo. Formerly it was no. 288 in the Urbinate 4 Dionysiac cycles with annQlistic notes refeIring. to Braunweiler.
?1" .Alphabets. numerals, with into aL runes (c{. infra).
collection. How and, when thil? manuscript reached Italy
seems to be unknown. Its runes were first edited by H. Mass­ This lastpage of the manuscript is much crowded, 'so much
mann ( I ). For our study it is of special interest beca~se it so that it is hard even toasc~rtain in
what order the various
shows the coalescence of two different runic traditions. items were added. Moreover, as it is the last page. of the
Redbrown leather binding, stamped. with gold ornaments and the arms of manuscript, it has suffered, rather badly from rubbing and also
Pope Pius VI and Cardinal~Bibliothecary F. X. de Zelada (last quarter of the from dampness. Especially the right margin is much damaged:
18th century). Parchment of good quality, mOstly well preserved. r.rhe codex
consists of I +'?I folios. It was probably put together bit by bit as :the material
a strip of paper had to be pasted along the fold to hold the
could be collected. The ' composition, of the quires i,s irregular, and there folio. The result is that some of the runic material (irifra no. II)
are many single folios. ' is partly illegible. Moreover the writing of the preceding
Fonnat 312 X 250 mm; the dimensions of the writteri area, the number of . page (fo1. 7 Ir : TERMINUS PASCHALIS CVM REGV­

columns and of lineS'vary much. The handwriting reflects the composite

character of the manuscript. Yet the different parts may be dated within a
LARIB; AD FERIA' LVNAE QVARTA[corr. to -AE]

relatively short period:, ~aec. X. ex.-XI ,(2). The runes are contemporary DECIMA [id.J INVENIENDA) tends to become visible

with their surroundings (saec: XI). Folio I was added in the ISth century. through the parchment and this makes the reading still more

Contents: (3) ,• difficult..

fol. It Computistical and -astronomical excerpts,. from various authors.
Going from ieft to right and· from top to bottom we find in
(I) H. F; MAssMANN, Rrmen, 253 f. vertical columns :
'G.SrnPmmB, Monuments III,'I1,nos. ?6, n.

'T. VON GRIENBEROEil,' Die angeUdchsilcheti, nmenreihen, 6 f.

.(I) A Greek alphabet (to '1') with the names of the letters and
On the manuscript, cf. also their numerical values in' Roman figures.
C, W. JONl!S, Bedae Pseudepigrapha, 13, So, 54, 60, 68, 90, 138.
(2) A Greek alphabet (to '1') with Greek numerals (mia­
L. THORNDlKE, A Hist~ of Magic I; 6fJ3.

A; VAN DE VI)'VER, Les a1UVTe8 ineditesd'Abbon de Fleury, ISJ.

ebdacusie) and their numerical values in ROII],aI,l figures.
(2) On the basis of the Easter tables for the period 988-1082, MAssMANN (3) .-, (4) Part of a Greek alphaqet, A to 8, with the Greek
dated the manuscript in 1082; RBt:F'PERSCHEIDT,'however, concluded that. since
the tables were of 988, the manuscript cannot have been written much later
,numerals chile to ennpchie; then from A to 8 ==:; muria,s to
(footnote to MAssMANN's p. 25'1.). On the' difficulty of dating on the ~is ehdomta muriait, .continued in. the next column, to;1\=
of Eastertablea d. H. l:IENBL, Camputus, 23 f. ~. muriamuriait.
(]) For further details see the entry in C. STORNAJOLO, Codices I, 26,1.
,(5) Below this column, a. short note on Greek numerals and
106
.... 107
 diphthongs : ;i\ ennacussie DCCCC, etc. H~ tres not~ The blank space between these and the preceding isruna was

non ,litter" / sed pro numeris complendis / addende sunt obviously filIed by the /agoruna: faint traces of a series of

in alfabeto./ Diptongi grecorum / ai pro ae; etc. Items I-runes remain visible at the end of 1. 33.

completing the alphabets of columns (1) and (2). As was mentioned before, our manuscript is remarkable for

(6) The Latin numerals I-DCCCC correspc)nding to the Greek its combining two different tradi,tions: the isruna tradition

letters and numeric symbols A -;i\. and the Greek nu­ and the De inventione tradition. To the latter belong:

merals mia-niacu.sin. (a) the text in the right hand margin (8).
(7) Two items completing (3) and (4), a pseudo-Greek al­ (b) the runic alphabet (I I).

phabet with the names of the letters and their Latin To the former:

equivalents: 1\. alfa pro a, etc. (c) the fuporc (II).

(8) In the space remaining to the right of the page, a different (d) the fragments of the isruna text (10).
but apparently contemporary hand has added the' note on
Only items (c) and (d) will be discussed in this chapter; the

the Gothic alphabet printed on p. 317. other two in Chapter IV (p. 317 ff.).

(9) Mostly below (1) and (2): DE MENSIbus hebreorum./ The fuporc begins immediately below the runic alphabet~'
Nisan. I. Aprilis, etc. DE MENSIBVS [GRAECORUM]. the first two runes of the former have been crowded in after
I APELevs. i. December, etc. De mensibus egiptiorum. I the last rune of the latter. The remaining runes are written
Toth. I. 1111 kal. Septembris~ etc. in one vertical column, runes, values and names. Because of
(10) Below (3) to (7), parts of the muna tract, arranged as the poor state of this page, a detailed discussion of the fuporc
follows: is given here. correcting Massmann's readings and von ,Grien- ..t'
II. 28-31: the paragraph on the isruna, with example. berger's interpretations when necessary (I).
n. 32-33: erased. f : value f, name fue. The value f is clearly indicated, but

n. 34-36: text on the hahalruna, partly erased; the illustrating must have been skipped by Massmann. There is con­

example may never have been written. sequently no reason to read f, ue, as von Grienberger

l. 37: blank (erased?) proposes; the accent, moreover, is on u, not on e.

n. 38-41: a Greek alphabet, majuscule and minuscule forms. u: here again Massmann has overlooked the 'II. which in­

with the names and values of the letters in Mediaeval dicates the value; yet it is plainly visible between the rune

Greek : B uitta pro u '" ita, thita, lauda, etc. The order and the mime uor.

is disturbed, some letters are missing.

1'»: Massmann's rune looks much like n; actually the manu­

(I I) Below (8), in two vertical columns, a runic alphabet and script has a rounded type of 1'». the loop of which has much

a fuporc (cf. infra). faded. In the name dorn the accent is on the r.

It is hard to decide why part of the muna tract should have 0: In the lateral strokes of this rune the rjght hand parts are
been erased. The text may have extended originally as far vertical. To the right of the rune Massmann read ceOSj
as 1. 37; but' what one could take for traces of older writing in fact the value is 0, sepa,rated by a high dot from the
may as well be the text of the preceding page becoming faintly name eos;
visible through the parchment. At any rate it remains doubtful (I) H. F. ~, Runen, 355 f.; T. VON GRIBNBBRGER, Die fmgttlsiich­
,whether the muna text ever extended beyond the hahalruna. sisclum f'fINmreilum, 7.

108 109
""
r: the rune is very faint. The value is indicated by T. to the peta (?) and von Grienberger's correction to *pert are
left of which there are traces of another letter (an erased e? impossible~
or rather c, cf. the next rune).
:x: this rune comes ·much closer to the original OE. :x than
c: this rune too is barely visible; it has a straight lateral to that in the other versions (which have Roman X). It
stroke, not a curved one as in Massmann's drawing. may have been borrowed from the preceding ~phabet,
g: the slanting stroke j's much longer than in the other ver": but it is also found in the Salzburg fuporc. As against
sions. The u of the name gihu is somewhat worn, but Massmann's hix I read lux for the na~e, with faint traces
ther~ can be no doubt about the reading. of one more letter before I (e? with the x indicating the
w: the rune looks rather like a P open at the top. The value value one may practiCally read xelux). .
is uu, or perhaps iiu or nu; at any rat~ not h as with Mass­ s: th~ name sigi has an. accent on the first i.
mann. t: the letter indicating the val'l,le is about twice the size of the
h: whatis visible of the rune clearly points to h,.Il.!>t to the other letters and seems to have been insel1ed after the
sort of 0 given by Massmann; the value is clearly h; of the name Waf!' written; The name reads tu, or rather tii
name only haga is m,ore or less visible~ , (there is an accent over the last stroke).
n.: the rune is rather faint; the value is n (or. t(), the name no b: there can. be no doubt about the name being berh.
and one more letter, only the verti.;:al shaft of which e: the rune is hardly visible; what Massmann interprets as m
remains. Massmann interpreted it as I, .but d is equally might juSt as well be a regular e.The value is also quite
possible. faint, but the name is clearly echo
i : . Massmann seemi! tQ have- skipped a line. in, which only m: this rune too is very faint, and so is most of the m showing ...,
the rune can easily' be read; the rest was probably i, us. the value. Name: man. "
j: the vertical ~troke of the rune does not cross. the circle, 1: this and the following runes seem to have been retouched' :.'
as in Massmann's drawing. The value g is clearly visible, by a later hand, and not always in a rulppy way. The
and so is the first letter:. of the name. The third letter I-rune and the following lette! I have been merged into
Il)aY have been Tj (or b?)but for: the rest it is impossible some sort of M ..
to pecide whether Massmann's reading (or rather guess) IJ: of the name only 'lie is clearly visible'; at one time n was
is right : gar ( ?). preceded by another letter, as is shown by the accent.
3: this rune too has been overlooked by Massmann. Of the That letter was probably i.
rune itself only two short strokes are. visible. They may , d: .the value gassigned to this rune may go back to a faint t.

be completed tQ mean either 3'Qr ea., .butthe former is the 111 the name tag final g is only partly visible.

more probable interpretation (cf. 3 in the Salzburg manu­

C2: therune looks much like a minuscule g; in the other ver­

script). The value is k.1. the name is practically illegibler

sions too the lower ends of the downstrokes are turned

P': .of this rune two variant forms are given: 'One is the type inwards. The name odil is corrected from olil.

found in the other versions of this fuporc, the other is the

Z, a: the re-rune COmes first, with the value aand.the name

regular p-rune. The name looks like ppod or peod. The

ac; the a-rune looks rather like a somewhat carelessly

former reading c~uld perhaps be interpreted as due to the . . drawn p; its value is also a f its name asc. Both names

fusion of i1ie~atue and the name. At any rate Massmann's have an accent on a.
no
III
....
q: the rune is a somewhat distorted p, the value plainly q.
3/6 in the Vatican manuscript. The 1/8 for the last letter requires
The name reads ru, with traces of one more letter (r?). another explanation. As it stands, it would mean fO (or uu),
%: the value z is rather vague, the name still more. But which does not make sense. Again we must turn to the Trier
there is little doubt that we have to read aero What version to understand U. In Tr the example reads coruus,
Massmann took for a final h, h or I, is only the end of the not corui as we find in B and SG. Coruus may be transcribed
line which separates the runic material from the rest of 1/6 3/8 1/5 1/2 1/2 2/8. It is easy to see what happened in l..! :
the text on this page. the last two groups 1/2 2/8 were merged into one, 1/8, which
. (haplology' may have been laid near by 1/2 occurring twice.
The isruna text shows some interesting variants. First of
It is obvious, then, that the Vatican and Trier versions are
all, the word -runa is treated as a Lat. fem. a-stem, and hence
derived from a common ancestor; and there are other simi­
we find Isrun€ ... Hagalrun€ ... This eliminates the difficulty
larities which may point in the same direction, e. g. the correct
we find reflected in Band SG, where the scribes hesitated
ostendunt in the hahalruna paragraph (B ostendit, SG ostenditur).
between a Lat. neut. pI. and a Lat. fem. sg.; only Tr is consistent
in interpreting runa as a neut. plur. Then there is hagal- in
haga/~, which is probably due to the influence of the rune­ 5· Salzburg, Stijt St Peter, MS. a IX 32 (saec. X/Xl) (S).
name hagal. We find the same variant in the Salzburg version,
where it is said explicitly that this type of secret runes is based The contents of this manuscript mainly consists of canonical
on the rune· hagal, but that part of it is left off or changed texts. Owing to their importance the manuscript has received
(" partem eiusdem litter~ ablatam uel mutatam scito "). This much attention. But its runic~ material has hardly caught the
interpretation, also implicated in the Vatican manuscript, eye of runologists, although it was edited (with a facsimile) as
became rather obvious as soon as the meaning of hahal was no early as 1864 (I). Yet it shows an interesting development of
longer understood. Here we may perhaps look for the reason runic cryptographic lore, and gives us some important infor­
why the scribe of U did not finish his version: he must have mation not provided by the terse version found in the other'
been aware of the· fact that his hagalrum;, had nothing to do four manuscripts.
with his hagal, neither that in the alphabet nor that in the Binding of the 16th or 17th century: filletted leather over wooden boards,
clasps missing. Heavy parchment, well preserved. The 218 folios (2) form
fuporc. In fact the manuscript shows no trace of erased 28 quires, numbered I - XV and I - XIII, and arranged as follows:
hahal-runes. The Salzburg scribe, on the other hand, had no 2 IV [1-16] + III [17-22, Ph. 17-24J + 5 IV [23-6'1" Ph. Z5-64J + IV (_ '1,)
misgivings and added a (quite vague) explanation. [63-68 , Ph. 65-7'1,] + 15 IV [69-188, Ph. 73-19zJ + IV (- '1,) [I89- 194,
Ph. 193-Z00J + V (- Z)[19S-Z02,Ph. 201-'1,08] + 2IV [z03-zr8,Ph. zoc)-ZZ4J.
A third point of interest lies in the example illustrating the Format '1,78 X '1,'1,'1, rum, written a£e!l 2'1,'1, X 158 rum; one column, 30 lines
isruna, the only example given here. . It shows the following
formula: 1/6 3/6 1/5 1/2 1/8. Two' mistakes' must be (I) [G.] PHILLIPS, Der Codex Salisburgensis S. Petn.1X. 32. Ein Beitrag
zur GeSchichte der vorgratianischen RechtsqueDen. Sitzungsberichte der
exp,ained: 3/6 for <2, instead of 3/8, and 1/8 for the last letter. philosophisch~histori8chen Classe der kaiserlichen Akademie der Wissen­
To begin with, as we saw on p. 105, 3/6 for <2 is not schaften, Wien, 44 (1863-1864), 437-510, esp. 508 ff.
unparalleled: it is also found in the five cryptograms of the Trier ('1,) Originally the codex seems to have had '1,'1,6 folios. G. PHILLIPS, Der
Cotkz Salisburgemis, gives a different foliation, as he counted the folios he be­
manuscript. But there it is justified by the fact that the first two lieved to be missing. But since he qid not take account of all cases, I have
runes of the third mtt have been added to the second mtt, thus preferred to keep the foliation of the manuscript in this description, adding
shifting the (B-rune from the 8th to the 6th place. A fuporc show­ PHILLIPS'S numbering· between brackets. In the survey of the contents,
ing the same subdivisions 8: 10: 10 must be at the basis of the however, I have given PHILLIPS'S figures, as his account is usually followed
in the works referring to this manuscript.
112
12
II3
'"
to the page. The two parts: A = fola. 1-116 (ph. 1-120) and B = fols. 117-218
(Ph. I2I-2a4) are approximately of the same date; they may hBve ~n intended As to the reason why the Uruna text should have been included ,
&om the beginning to form a unit. The whole may be dated aaec. XlXI. here, Phillips gives a hint but rejects it immediately:

The origin of the manuscript is unknown. It shows some affinity with

Western Gennany: its collections of canons are, baaed on those of Regino, Man konnte wohl dafiir halten, dass es sich hier an vollig
Abbot of PrI.Im, or related to his (I). ungehoriger Stelle befinde, wenn nicht etwa es ebenfalls
The contents have been examined in detail by Phillips. Here follows his zum Gebrauche bei den Formaten empfohlen werden soli,
brief outline (2), in which I have included some additional data &om bis study:
was aber doch immer unwahrscheinlich sein mochi:e.(r).
fol. If : Summary of contents.
I" : Formata, sent by Bishop Ruodbert of Mea (883-90S) to Archbishop
It appears, however, that the regula formatarum (2) and runic
Wilibert of Cologne (870-889). material are, found together in two other manuscripts as well :
2-94 : Cresconiua, Conctn'dia Canonum. Leyden MS. Voss. Lat. fo1. I2 3 (regula formatarum + runic
95-1201: A collection of canons from Spanish and Gaulish councils.
120" : Bede, ~ remedw pw;atflrUm.
alphabet) and London B. M. Harley MS. 3017 (litterae for­
n.l w lS3f : Liber Canonum, from Hrabanus Maurua. matae + runic alphabet). Since these three manuscripts can
153": Canon of the council of Tribur. Letter of Pope John X. hardly be related, some explanation for this threefold occurrence
154-156 : Excerpts &om Augustine and Gregory. must be found. That runes should have been used to authen­
156 : Canons of the council of Tribur.

157-161": Canones Synodi Romanorum ad Gallo. episoopos.

ticate litterae formatae is, indeed, not very probable. But the
161"-170 : Divenae &ententiae Canonum.
regula formatarum may very well have reminded a scribe of a
170-171" : Praecepta S. Clementia Episcopi.
system of secret writing which was lingering on in some
171"-172" : Excerpts from Isidore.
172" : Ex decretis Vigilii' papae.
schools or scriptoria. The isruna tract, with its play with
172"-194" : Hincmar of Laon's PitMciolus. rank-numbers and group-numbers, in fact parallels the use of
194"-198: Ex dictis sanctorum Patrum, etc. the Greek alphabet and numerals as explained by the regula
198-204" : Collection of canQn8.
204"-208 : Conventua Ticinensis.
fairly well and may have served as an introduction. In the ...~~
,208 -212 : Capitula AngiIramni. other manuscripts only a runic alphabet has been added; its _
213 : Decretale of Hadrian II. inclusion will rather have been prompted by the preceding '.
213-2:13" : Collection of canons. Greek material. These explanations are, I believe, sufficient
2:13" ; Imma text, etc. (d. infra).
224" ; Regula formatarum. to account for the presence of the runes; no hypothetical use
in litterae forrilatae is' needed to justify it.
The isruna tract is well embedded in the material, of the Arrangement of fol. 2I7v (ph. 223v) :
manuscript. It is preceded by .. Concilio Heliberit. tit.
II. 1 -3: Introduction to the fuparc.
LXXV", (fol. 2I7f, Ph. 223 f ) and immediately followed by
.. EX CONCILIO TOLETANO. Tit. XI ", (fol. 217v, Ph. II. 4-9: Fuparc .
223 v); it is obviously written by, the same hand, which also II. 10-19: Text.
added the regula formatarum on fo1. 2r8 (ph. 224). The isrUna 11. 20-22: Examples.
tract is no hurned addition; the compiler seems to have inserted, 1. 23: blank.
-it with some definite purpose in mind.
(J) G. PHILLIPS, Der COth;JI SaJisburgmsis, 50 8.
(z) C. FABRICIUS, Die Litttrae jormatae im Friihmittelalter. Archiv fUr
(1) P. FOURNIER-G. LB BRAS, Histmre de, colkctiOflS cancmiqruls en occit/ent Urlrundenforaehung 9 (l9:16), 39-86, '168- 1 94.

depuislujaussesdkretalajusqu'audicrttde Gratim. Paris, 1931-193:1, vol. I, Dictimmaire d'archiologie c!rretimnt et de liturgie IX, lI, 1571-6.

:1116, 262, 263. 267, z69, 273, 30 5, 438. V. GARDTHAUSEN, Die griechisCM Schrijt des Mittelalttrs im Westen Europas.
(z) G. PHILLIPS, Der CodeJc Salislmrgmsis, 440 f. Byzantinisch-neugriechische JahrbGcher 8 (z931), 114-135 (U9 f.).
B. BISCHOFF, Das griechische Element, 32.
II4
'" 1I5
11. 24-28: EX CONCILIO TOLETANO. Titulus XI.
x: the rune has the same form as that in the Vatican manu­
In the Salzburg manuscript the text has been provided with script. It is the regular form of x in the De inventione
an introduction, explaining the fundamental principle of the alphabets and also in the' Syriac ' alphabet of Munich MS.
cryptic systems to be discussed afterwards: the fuporc is 1#36 (this alphabet is based on a fuporc as found with
<livided into four groups of letters, the first three having eight the imma tract, cf. p. 254 ff.).
runes each, the fourth four. In the other versions the first
"two groups must originally have had eight runes, whilst it is 9: the value of the rune is n (in in B, SG and Tr).
not clear whether the remaining runes formed one group of d: the intersecting strokes are placed so high, that the rune
-.twelve letters or were divided into 8 + 4. Only the St. Gall looks more like a perfect m-rune; the value t is also found
version maintains this situation; all other manuscriptll make or in the other versions.
imply changes in the subdivisions (cf. pp. 101, 105, etc.).
a: value aa; the second a is a pointed open u-form.
The Salzburg manuscript is also the only one to indicate
dearly the actual groups in the fuporc. On the other side it As to the text, it treats runa as a fern. sg. a-stem, using the
.omits the names of the runes in the fuporc; only the names of plural when the word is used alone (qu~ run~ dicuntur), but
the runes mentioned in the text are given. The fuporc calls a (collective?) singular with the compounds isruna, lagoruna.
hagalruna and strophruna. . .
for the following remarks :
The) Salzburg version is also remarkable for its explanation
p: the rune looks like a e-rune with a straight lateral stroke; of the hagalruna, mentioned before in the discussion of the
as the manuscript is plainly legible, this distortion must Vatican manuscript. Our author was aware of the difference
have occurred in the course of repeated copying. between the h-rune (name hagal) and the sort of symbols he
g: the rune shows the type commonly found with the isruna called haga1runa, and offered an explanation: to facilitate
tract, but the value reads s, In all probability ~]:lis is writing, part of that rune was suppressed or changed. This ..
ultimately a misreading for an insular g. which, if the top explanation can of course not be taken seriously : almost any '.
stroke was not very clear, could easily be interpreted as s, rune offers a closer resemblance to the hagalruna than pre­
cisely the hagal-rune.
especially when standing by itself.
Finally the examples illustrating the four types of crypto­
h: this rune has been so completely distorteci that it reminds graphy mentioned in the text call for comment. Three of
one only vaguely of h. them read 1/5 4/1 1/6 1/6 1/4 = RACCO; only the.hagalruna
gives an obviously mistaken 1/5 4/1 1/6 1/5 1/4 RACRO.
j: the value of this rune is rightly given as g.
The .word Raceo is in all probability the name of a person~
3: this and the following rune are inclined to the right, as if perhaps a .. Koseform " of a name beginning with Ragin- (I).
drawn in a careless way not found with the other runes. It may very well have been the name of the author of our
The value of the eoh-rune is h, whereas the; other manu­ version or of the scribe, but I think it is hardly worth· while
scripts have k. But h is also found in Oxford MS. St. trying to identify this Racco (if = the scribe) as long as the
John's College 17, and Vienna MS. 795 has i & h. history of the school of handw~ting to which our codex belongs.
has not been made.
11: 'the same type as found in the other manuscripts, with the
distortion mentioned under 3. (I) E. FOR8'I'llMANN, Namtmbuch II, col. 1200, classifies RaceQ under a stem:
Rae, but mentions also other possibilities.
:116
... 1I7
 On the other hand it is important to note that several of the of which, T= lis, had one small i-rune and five longer
collections of canons point to an origin in the Rhine area (I). ones.
This connects our manuscript, geographically at least, with We saw that it was hardly possible to identify the Racco

Tr and U. mentioned in the example. If we may depend on the evidence

For the rest, however, it treats its subject very freely, as the provided by the other manuscripts, which transmit the' exem­

text given here shows : plum' fairly faithfully, it looks rather probable that the scribe

Notum sit omni litterarum qu~ run~ dicuntuT1 scientiam found Racco in his exemplar. At any rate we must consider S

habere uolenti, I quia in IIl1 0r uersus uel ordines diuiduntuT. as a copy : only thus can the omission of is in 1. 8, and especially

Primus ordo continet litteras VIllo, Isecundus itidem VIII, the rather ' worn ' state of the runes 'be explained., The open

tertius similiter VIllo, quartus quattuor. a in aa, too, miIst be due to an older exemplar.

[Fuporc, with indication of subdivisions :] [P]rim02 'ordini Owing to the almost complete independence of S, it is hard
he deputantuT; Secundo uero he; Tertio uero he; Quarto he. to ascertain where it branches off from the other type (provided
De his litteris tres sequestrantur quibus singulis totum quod the latter pre;erves the more original type). If any store can
uelis scrifbere potes, id est [is] 3, lago & haga1. Quod be set on the absence of the clopfruna, one would suppose the
per i solam. scribitur, isruna uocatuT; I quod per lago, Salzburg version to have branched off from a 'type anterior to
lagoruna; quod peT hagal, hagalruna. Ergo si per i uel I the S,G-B version, or rather from that circulating in the Rhine
scrifbere uolueris; primum breuiori i veil ordinem notabis, area (BrilUnweiler, Trier?) reflected in Tr and U. This latter
10ngioribus"l uero litteram. Quod si per hagal scribere suggestion fits in well with the contents of S : its canonical
uolueris, in sinistra parte quotus I sit ordo, in dextera quota material is (in part)' closely related to the collection of canones ,I;,,
'sit littera notabis & ob facillitatem scri/bendi partem compiled by Regino of Priim (I). It may also be corroborated~'
eiusdem litter~ ablatam uel mutatam scito. Est & stroph by the substitution of hagal for hahal (cf. U).
'r.luna ! qu~ solis punctis constat, que' hac ratione , On p. 115 it was suggested that the isruna tract was perhaps .,
scribituT, vt superius or/do, inferius litter~ designentuT; included on account of its relationship with the regula fOTma~
& ut facillius intellegatur quod dicituT I pTomptum ponamv's taTum. In fact the two texts show a somewhat similar approach:
exemplum. AmeN. fo1. II7 v (ph. 223 v ) :
Then follow the four examples, each with its name : isruna; Notum sit omni litterarum qu~ run~ diCuntUT scientia~
lagoruna; hagalruna; strophruna. • habere uolenti quia in 1111 uersus .uel ordines· diuiduntuT.
Rem. 1 Ph.]dicantur; fo1. 118r (ph. 224.r) : '
II When Phillips saw ~he manuscript the P of Primo must Gr~ca elementa' litterarum numeris (read: -os) etiant

have been clearly visible, as is shown by his facsimile; exprimere nullus qui uel tenuiter greci sennonis notitiam
it has suffered some damage sirice, becoming almost habet ignorat. .
illegible. There is, however, a striking difference: whereaS the isruna
3 the name of the i-rune is omitted; Ph.] is et?
text is fairly correct, the text of
the Tegula shows an appalling
number of blunders and omissions, some of which make it
"breuiori ... longioribus: the difference in number may practically useless.
be due to the influence of the example, the first letter
(I) Cf. p. 114. note I.
(I) Cf. P.1I4 and note I.

JI8 119
....
 THE ISRUNA TRACT. U : Isrun~j litteris (I. l)i [ita) the Ms. reath rather like n&; [sit] om.;
(passim); longioribus I [scribatur) om.; c[or)ui : the second and third letter
have been damaged when the following parograph was erased; at firSt sight
From the combined evidence of the first four manuscripts, one might read cerui, but the strolul through tlrs 0 is accidental.
the following fuporc and text may be reconstructed : Edd.: Iismna) Steffens read UsmfIQ; autem] Steffens alone explained the SG
abbreviation C:OrIectly; Grimm read • K (C?)', Hattemer haec, AIntz
feh ur dorn oos .... a.t c.en 'lebo huun .. St. Galhm hat h mit Verweisungszeichen, von una (1935) ala haec

rf t1 u ~ d ~ 00 kr kc. ~9 tu u
. 8ufgel6st und vom Scbreiber wahrscheinlich auch so gemeint .. j Raucq
tskes no decision.

Lagoruna dicuntur quae ita scribuntur per l litteram ut

hagaJ nod I{s gel'" rh perel c.lux SI91 nomen corui.

Mh In I. ~k Ip Xx '1s SG. : qu~

+9 B. : literam

Tr. : qwe

~c a.sc. .ur aer U. : om. (erased).

r..a. ~ i. ~'1 't. V~(ant$:fn (B)~(U,5)

FIG. IS
Edd.: AIntz remarks: .. Beide Handschriften [i.e. SG and B] geben daa Iago­
Zeichen an dieser Stelle ala .L wiederj es musz aber ~ l sein, wie aucb
? das Beispiel ergibt"; such a change, however, it not necessary, lIS
I litteramay mean • the rune 1 '.

The text given here is based upon all four manuscripts, but Hahalruna dicuntur istae, quae in sinistra parte quotus uersus
mainly on SG and B, these being the oldest versions. All sit ostendunt, et in dextera quota littera ipsius uersus sit.
variants are given below each paragraph, and also the readings SG. : ist~ qu~ j [sit) first time om.; ostendit with abbreviation mark over
proposed by Grimm (I), Hatteiner (2), Steffens (3), Raucq (4) final t, -tur.
and Arntz (5). All other editors depend on one. of these B. : iste que; ostenditj dextra.
editions. Tr. : qwe ita scribuntur: in sinistra parte quotIU uersus sit ostendunt, in
Iisruna dicuntut quae i littera per totum scribuntur, ita ut dextera, etc.;, final [sit] om.
U. Hagalrun~ dicunt[ur);. [istae] om.; qU~j quotus sit ueraus; in dex{•.•]
quotus uersus sit primum breuioribus i, quae autem littera sit l

quota sit littera [•••] rest erased.

in uersu longioribus I scribatur, ita ut nomen corui scribatur
Edd.: SG ostendit] Grimm, Hattemer: ostenditur; Steffens: ostendit, " fUr
his litteris ita. ostendunt "; AIntz· would propose an alternative reading allowing to
SG : di' = dicitu.rj qu~ (parsim); II' ~ autemj litters. keep' ostenditur ( ... istae : in siDistra parte •.• ostenditur)j but since it
B : litera, etc. pasnm. obliges him to drop quae, it is ·better to emend osterulitur to ostendunt,
following the text of Tr.
Tr ! liuera (1. 1): [his litteriS] om. i the i followhtg breuioribua is a capital,
firs one following longioribua a mintucule i. Stopfruna dicuntur quae supra in punctis quotus sit versus,
(I) W. GRIMM, Ueber thtltsche Rrmea, 110 if.; c:orrecUng remarks are given subtus litteram ostendunt; sed aliquando minim illa faciunt,
by E. G. GRAFF, Spracluchatllt II, 523 f. ut supra sint puncti qui litteram significant, et subtus ordo
(2) H. HATTEMER, Dmkmah1e I, 418.
(3) F. STBFFI!NS, LateiRische PoliiograpJtiel, tranac:ription facing Plate 53 b • uersus.
(4) E. RAUCQ, DU Rrmm da BriiSseler CotWc N° 9565-9566, tranac:ription SG. : Stofmnaj q~j aubtus littera] subtiliterj illas.
on folding leaf opposite p. :z6. B. : Stophrunaj literam; illas.

(5) H. ARNTz, Eis- WIll Wassernmen, 7.71 f., an attempt to reconstruct the
original (U Die urspriingliche Form Uszt sich also mit einiger Sicherheit Tr. : Stofruna; qwej I. 2 litteris; [significant] om.; ueraum.

bestimmen "). U. : om.

12I
120
...
Edd.: SG stofrona] Hattemer read 6oofru'lUl; this ill probably· the form to
which Am~ alludes in 11 footnote: "Von W. Grimm Ilk ICDOfnma rune, or perhaps rather its selection, seems to be connected
(St. GaJUm) gelesen "; Grimm has the correct reading dofruna; with the introduction of X for x. Formally this new x
significant] SG has rigii which Hattemer read signant, all other edd" coincided entirely with the old g-rune, and this may have
Jigrrificont.
led to another type being substituted for the original g.
Clopfruna dicitur quae pulsu efficitur distinctis personis et As to the origin of the new g, Arntz proposes two solutions:
litteris, ita ut primum incipiatur a personis, postea a litteris. . either it is the original g, altered in such a way that itJ~ad .
SG. : q,,~. a (more or less) vertical shaft; or else it is the old j-rune,
B. : This por06"aph comes at the begirrtrin& of the tTO£1; it ends, h _ . with
a positura, which shams that it really bi/mIlls at thiI end; Clophrona. an s-like form of which is known from Scandinavian! and
Not in Tr and U, German inscriptions, with a vertical shaft added' (I).
Since the type of j which Arntz chooses as a starting.:.point
THE RUNFS (I).. in his second proposal is not known from England, I believe
. We have to accept the first possibility. This new g seems
The runic forms of S having been discussed separately to have been known only in a restricted area : Exeter
(p. 116 f.). only the more important divergences will be men­ MS. 350 7 and the related ~anuscripts point to W, and SW
tioned here; for the evidence of U, I also refer to the special England, and so does. Nemnivus's alphabet (2). This may
.discussion on p. 109 ff. The runes ·fl1 0 r w h i j 3 s t be throw some light on the origin of. the isruna fuporc .
mid (2 a & call for no comment as far as their forms are n: in B the rune is reversed, but this can .hardly be due to
concerned. Norse influence. In Tr it is not possible to decide which
u: in Band S this rune tends to become rather like a !ninus­ of the two types WBS actually meant, as thenme.18 ..
cule n. a regular X. ...;'

c: Band SG have a rounded h-like type, whilst the other p: this type of p-rune is not found elsewhere;. two of the. :,
three manuscripts give the rune with a straight lateral manuscripts, U and SG (in the alphabet), also have the
stroke. normal type of p. A similar rune Occurs twice in Oxford
g ! the same type occurs in a number of runic alphabets MS. St. John's College 17, but each time with the value q;
(Exeter MS. 3507 and the related Cotton MS. Vitellius it is obviously derived from the p-rune. The same
A 12 and Phillipps MS. 3715; also iIi Leyden MS. Voss. expianation probably holds for that in the isruna ful>orc.
Lat. F. 12 8). as well as in Nemnivus's alphabet (2). Its The letter which takes the place ofp must originally have
occurrence in the ' Syriac' (= runic) alphabet of Munich been q. We shall see that very often the form of q is a
MS. 14436 is easily explained: this alphabet is derived more or less fanciful distortion of the regular p-rune.
from a fuporc of the imma type. .The creation of this x: cf. g. The difference between SG, Band Tr on the one

(I) E. RAucQ. 1M Runm. da BriU,eier COtlu N° 9565-9566, 14 f.

hand, and U. and S on the other, is of special importance.

H. ARNTz, &rrm rmd Runennamm, 174 fr. The latter two borrow their x from the De in'Oentione

Although the values are discussed together with the rune-names (p. J2.4 fr.). alphabet.
some reference to them ill unavoidable in this section.
(2) It ill not found in the runic slphabets of Cotton MS.Galba A 2 and Ox­
ford MS. St. John's College 17 (G. STEPHI!NS, Momt.ments I, Dos. 13 and 31), (1) H. ARNTz, ~ rmd~, 177­
as Miss RAuCQholds (1M Rurrmda BrlUseler COtlu NO 9565-9566, J4, foot­ (2) Yet we should not forget that the runic alphabet of Exeter MS. 3507,

note), but in the Nemnivian alphabets of those manuscripts, cr. #ifra, P.IS7 fr. etc., was probably imported into England from the Continent.

122
.... 123
 not so easily accounted for. One might feel tempted to
IJ: the same type of IJ is found with various values in a number
of fuporcs and alphabets: for x in Paris MS. Arsenal n69 correct it into feu, the form found in some Scandinavian
(or == y f); for y in Munich MS. 14436 (' Syriac alphabet '); name-lists, in the • Arabic' alphabet of Munich MS.
for h in Cotton MSS. Domitian A 9 and Otho B 10 (f), 14436 (showing a strong Scandinavian element) and in
as a variant of the normal h; also in Exeter MS. 35 0 '], Phillipps MS. 3715, the evidence of which cannot be
checked. In the absence of clear evidence for Scandinavian
Cotton MS. Vitellius A 12 and Phillipps MS. 37 1 5.
We shall examine later on how these different values can inRuence in U (apart from that normally found in 'the
be reconciled. In U the two short strokes transecting De infJentione alphabet), the origin of Jue remains obscute.
the stem are not parallel, but they meet to the left of the At any rate von Grienberger's proposal fue = *f, ue may
stem. Much the same form is found in Nemnivus's al­ safely be rejected.
phabet (with the name kine), but it is doubtful whether u: the value u is shown to indicate long u by the spelling of
any special conclusions can be based on this similarity. the name uur in SG,B and Tr. U has uor, hardly a mark
q: at this place of the fuporc we expect y; in the'longer of diphthongization of u, more probably a mistake (I).
fuporcs q is no. 26 or 27. Our fuporc had originaUy y p: the five versions are unanimous on the value d, which is
for no. 27, as may safely be concluded from the name yur supported by the name dorn (SG,. B, U). Tr with its
in Band SG. The reading in U may also go back to such difficult exemplar writes doro. As Gmc. p became 4
a form. This, however, leaves the form of the rune, a and. then d everywhere in Continental Germanic, but at
regular p, unexplained. It looks as if two traditions had different moments, dorn may be a criterion for a rough
collided: one with y and the name yur, the other with p localization of our text. We must not forget, however,
and q (the latter two having changed their values to q and p that d for p is also found in early OE. manuscripts (2).
respectively); q perhaps had the name cur. Since y and q ,0: B and SG stress the length of this rune by spelling.
(i. e, original p) show some resemblance" confusion may' the value 00; the other manuscripts have o. The length"
have arisen at· this point. Cf. also p. 'The fact that U is also indicated by the form oos in SG and B; this form
has both forms for p is not so significant, as this may be must also have occurred in the prototype of U and Tr,
due to the inRuence of the De in'Oentione tradition. but the former misread the first 0 as e (/!los), the latter as c
%: the ea-rune appears to have been interpreted as z in a (cos).
number of Continental alphabets as weU, but it is also r: The name rat, compared with OE. rad, shows an adap­
found for x, y, k, and q, all letters to which no rune of the tation to Continental Germanic phonology. The shift
original fupark corresponded. may imply a conscious or unconscious change of meaning:
OE. rad • riding' : OHG. rat' consilium '. It may also
THE VALUES AND RUNE-NAMES (1). help to localize the isruna tract, as the shift of final d to t
did not occur farther North than the Rhine-Franconian
f: the value is J in all five manuscripts, but the names show area.
some divergence. Band SG have feh, a form reRecting
the OE. name. Tr has a mistaken fed. The Jue of U is (1) W. BRAUNE, Althocluleuuche Grammatik, § 41 : ii. > uo is very rare
in OHG.
(1) E. RAucQ, 1M R~;Us Bnus. CoUx N° 95 65-9566 , IS ff.
(a) E. SIIM!RS-K. BIn1NNE:R, Alttmglische Grmnmatik, § 199 A. Y.
H. ARNTz, Ru_ und ~, 192 'ff.
12 5
i24 ....
c: both the value c and the name cen faithfully reproduce· German', is not clear (I). The form nod, with final d,
the OE. model in all versions. is very rare as compared with not, but it is at least found
g: the value g is found in all five versions, but the names. in the Heliand and the Ludwigslied.The preservation
differ: SG, Band Tr have gebo, U has gibu. The latter of -d may help to localize the fuporc. One might also
may either be a variant form of gebo (i being due to the compare nod with OE. nop , boldness' , but since the latter
influence of following u, or e retained on account of fol­ is not found as a rune-name (OE. only nead, ned, nied) ,
lowing 0), or it may be a form borrowed from the preceding our nod is rather a translation of the OE. word.
De infJentione alphabet. Neither is the OHG. (nor the i: both on the value (i) and the name (iis) there is complete

OLG.) form corresponding to OE. giefu, gifu ; we should' agreement. The double i in the name indicate~ the

rather expect gwa. I fail to see what Arntz means by length of the vowel.

"gebo, womit die· germanische Grundform gefJo " Gabe " J: all five versions render the value of this rune by g, the

wieder gewonnen ist" (I); hardly that our gebo is the normal OE. spelling for I j I i this' spelling is also known

OGmc. form. Forms with geb- are not unknown from from OHG. manuscripts (2). l.n Germany the name ger,

English texts (2), and therefore it is difficult to decide which is found in three manuscripts, may have been

. where the form gebo arose. interpreted as ger '·spear::'.

,w: allfive manuscripts spell the value of this rune uu. On 3: four versions (SG, B, Tr and V) have the puzzling value k,

the name they agree with initial h, but U has hun against to which S opposes a more correct h. The value is ih'

hrpm in SG, Band Tr. In huun the first u may stand for (in U it is illegible). Arntz explains k by the influence

/ w /, the second for <OE. y. The latter sound often of the Latin alphabet. He seems to imply that the. runic

caused some trouble to Continental scribes: OE.j·r is alphabet (found in SG only) led the scribe to assign values "':-1,
spelled ir, uir, uyr, and also huyri, huyry, huuri, hyri. not in the fuporc. e..g., k, z, to runes which otherwise,.
Here we haye also aD. explanation for initial h:. it is a would not have fitted into an alphabet (3, ihi ea, ear). In -.
meaningless addition found with a good number of rune­ the case of p : q I have admitted t~at a runic alphabet
names having an initial vowel: his, hac, hur. Initial h may .have played a part, but that is precisely a point where
is often added in OHG. sources; also sporadically in the runic alphabet in SG corrects the fuporc (cf. p. 94).
OE. (3). I rather tend to believe thatk is a mistake, since S has h.
h: value h and name. hagal (U haga[.]) without exception. The latter agrees with the name ih, a regular OE. form (3).
Th~ name is an adaptation to Con,tinental Germanic. OE. p: apart from V,' ~here the reading is doubtful, the manu­

had either hfBgl (also hegel, ~gil, ~gel, hegil, hdl) or else scripts agree on the value p and the name perd. The latter

hagol; the latter, however, is never found as a rune-name. is '8.n adaptation of OE peonl, per4.

n: all manuscripts agree 'on the value n and the name nod. x: the value of 'this rune is x, and the name elux (SG, Tr,

Why Arntz should call the latter' neither English nor probably also in U) or elox (B), but the runes differ:

S and V have a type!borrowed from the De inventione

(I) H. A:RNTZ, Rwum und Runm1Iamm. 2.Z9; on the same page gebo is said alphabet (cf. supra).
to be " eingedeutscht ".
(2.) H. STRoM, Old English PerS01U.ll N_, IJ2. f. (I) H. ARNTZ, Runm und Runennmnen, ::&28.
(3) W. BRAUNE, Altkoduieuuche Grammatik, § Isza. (z) W. BRAUN!!, Althoc1uleuuc1!e Grammtttik, § lIS fi.

E. SIBVERS-K. BRUNNER, Aliengwclut Gr_tik. § :l.I7 A. t. b) E. SIImma • K. BRUNNER, Altengwclut GrammGtik, § 2.50 A. 2.

126
....
7
12
s: the name sigi seems to imply that the common ancestor
of SG, B, Tr and U was not free from mistakes: in the dering, unless one took to such complicated (and ambig­
common prototype the final letter of sigil must have been uous) formulae as n & g in Vienna MS. 795. The
omitted, d. sigel in English manuscripts, sygil in Vienna spelling 1U: in the name is a current OE. spelling for final
ng (I).
MS. 795, etc.
t: the value is ti everywhere, but the name seems to have d: the name and the value of this rune have been transcribed
caused some difficulty. SG and B have ti, U tii or tu, Tr a into OHG. : value t (U has g by mistake), name tag-(cf.
puzzling tan. This last form is clearly a mistake, perhaps OE. da1g, deg). The change in value is obviously due to
influenced by tag (the name of the originald-rune). If the translation of the name. The form tag may help to
U may really be read tii, this spelling indicates the length decide where the fuporc received its .Continental garb.
of the vowel; cf. Tiig in early English texts (and tu in ~ : the DE. name UJpel, epel, too, has been translated into Ger­
Vatican MS. Regin. 338). man; odil may again be a criterion for the localization.
b: here too there is some disagreement over the name : SG a,ce: these runes led to no small degree of confusion.
and B write berg, U berh, Tr ber (this last form is probably All five manuscripts, with the possible exception of S
due to an imperfect prototype). In berg ana berh we (which does not give the names), have mixed up runes,
have to see adaptations of OE. bearc, berc, probablyaccom­ values and names. The original situation may be restored
as follows:
panied by a change of meaning ('mountain' instead of
'birch', which should have become OHG. biric, birich). a, value aa, name ac (OE. ac);
The reading berg is supported by Munich MS. 14436. ce, value a, name alc (OE. a?sc).
The name a?sc has been adapted to OHG. phonology, but
e: the name is eh in SG, Band Tr, ech in U. The latter is
not ac, which should have given OHG. em. The actual
only a scribal variant, perhaps influenced by ech in the situation is as follows :
preceding runic alphabet.
B : a, value aa, name asc; ce, value a, name ac;
m : value m, name man without exception. SG: z, value a, name ac ; a, value aa, name asc;
1: the only point of interest is the final vowel : B has lagu, Tr : ce, value a, name hac; a, value aa, name asc;
against !ago in the other versions. The coexistence U : ce, value a, name ac ; a, value a, name asc;
of final 0 and u is well known both in OE. and OHG. S: z, value a; a, value aa.
grammar; the transition u > 0 may have occurred in Consequently the runes and the values are given in the
either territory (I). right order in B, the nam~s in SG, Tr and U; S reverses
l): SG, Band Tr give the value in, U and S n; in SG, Band the order of runes, but does not give the names. This
U the name is inc, Tr omits it. The divergence in the seems to imply that the COmmon prototype arranged this
indication of the value is due to the fact that the acrostic material in a way which could be interpreted differently.
principle could not play in this case : the rune stood for The double aa of course serves to indicate the length of
the final sound of its name, not for the first. Moreover the vowel in OE. ac (cf. J';pinal-Erfurt glosses z35, Corpus
that sound was hard to render; n was about the best ren­ 535 calor: aac; and Vati~n MS. Regin.33 8 : aac). U
simplifies it to a. On initial h in hac, d. w.
(1) E. SIlMnIS - K. BRVNNBR, Alteng1isc1ul Grammatik, § ++ A. 7. q: the value of this rune is the same in all five versions, but
W. BRAUNE, AltJwchdeutsclul Grammatik, § 58 A. ;a.
(t) E. SIIIVImS - K. BRtlNNBR, Alttmglisclut Grammatik, § 5.
u8 21
13
'" 129
 the name readsyur in SG and B, cur in Tr.ru(rJ in U.
The latter mey go back to yur with a shQrt y. The Tr may also have been borrowed from the English original. On
form may be a concession to the acrostic principle (cur = the other hand huun and yur must have arisen on the Continent,
tjUr). The reading tjUr also occurs in Munich MS. 14436. and so probably did aero We should keep in mind that ger
In the section on the runes I pointed out that in this place and man may be OHG. forms as well. But even so there is no
we should expect a y-rune, and that yur in SG and B (and question of a systematic "Eindeutschung": the compiler
also TU[r] in U?) may reflect the original situation (1). only transposed the easier words; he even shrank fromarela­
A q-rune borrowed from some or other runic alphabet tively easy case such as OE. ac : OHG. ei.h. His knowledge of
took the place' of y; that q-rune itself was in fact the ori­ OE. was very limited, and in a number of cases he was led by
ginal p. There is another explanation for yur, but it is the sound rather than by the meaning (rat. sigi, berg). More­
less probable. The name yur might go back to an original over two runes changed places (p, q) and the acrostic prin­
qur, either through a misreading or through a conscious ciple was abandoned in three instances (huun, yur : q; aer : z).
. change : in the Latin alphabet y was the last letter but one, Whether some Qf these alterations Occurred in the English
prototype we Cilnnot decide.
and this may have played a part in this change. On the
whole themst solution with its reference to the original The adapted names with dentals may help to localize this
order of the fuporc is to be preferred. German edition of the fuporc. The combined evidence of
dorn, rat, nod, tag and odil points to a centre rather far North
%;: the rune which receives the value z is the ea we met in the

English fuporcs. The name of ea survives in the name (Rhine or East Franconian ?), although at the early p~riod in
aer, which obviously goes back to ear. The question how which we have to date the archetype (saec. IXl) the interpre­
tation of such data is. rather difficult, cf. the many d's in the
. this rune came to be used for z is not so easily solved.
The same transfer is found in the De inventione group and' Keronian glossary (I). From feh, eh, ger, e/ux, and probably
also from perd and herg (i. e. OE. here) we may infer that the '
in Berne MS~ 207, but in Exeter MS. 3507 the same rune
stands for k, in Leyden MS. Voss. lat. F. 128 and Oxford original came from Anglian territory, and was written down:,'
at an early date (saec. VIII) (2). The fact that the g-rune is
MS. St. John's College 17 for x, in Cotton MS. Domitian
foqnd in other sources pointing to a Western region (Exeter
A 9 for q. It looks almost as if this rune had been used
to fill gaps which appeared in: the process of alphabetizing. MS. 350 7, Cotton MS. VitelIius A 12, Nemnivus's alphabet)
may help to narrow down the area. Cf. infra p. r 57 ff.
Munich MS. 14436 still preserves the form ear (with the
value e). but aer must have appeared in the common
ancestor of SG, B and U (in Tr this rune is missing). THE EXAMPLES AND THE. TEXTS.
The evidence p.rovided by the fuporc is not unambiguous;
yet it does give us some valuable information on the background We saw that, as far as the examples are concerned, the
of these runes. Twelve rune-names have retained their OE. manuscripts of the isTUna tract· fall into three groups :
form: ur, oos, em, iis, ger, ih, elwc, ti, man, /ago, inc, ac; nine (r) eOM: SG, B.

have been adapted to OHG. phonology,· or simply translated: . (2) eOTUUS: Tr, U.

dorn, rat, hagal, nod, sigi, berg, tag, odil, ase. To the first (3) raceo; S.
category we may probably add feh and eh. The name of g
(I) This' iriterpretation was adopted by Miss E. RAvCQ (Die Rtmm des (I) W. BRAtlNI!, A.lthoclukutsche Grammatik, § I63 A. s.
Briluekr Coda N° 9565-9566, 18 f.), but she did not know the other versions. (z) E. SUM!R8 - K. BRUNNER, Altmglische Grammatik, §§ 9 1 • b; 108, 6;
Ir9 f.

130
.... 13 1
This last example has been discussed on p. II7. It goes in all 'shall see, however, that on this point 8 probably follows

probability with the recasting of the text : Racco may be the an old tradition (p, 14. 1 ).

name of the scholar responsible for it. The groups or sections were known to Latin scholars as

It is not so easy to establish the priority of corui or COTUUf. versus (all five manuscripts) or ordines (8).

The archetype of U and Tr must have taken nomen corui to (b) Within ':this system each rune was defined by
mean 'the word coruUf', whilst 8G and B seem to imply an two figures, one indicating the group to which
interpretation such as 'the name Corui'. Although I can the rune belonged, the other its place in that
offer no explanation for cOTui, I believe it to be the original group. The figure indicating the group always
form, rather than coruUf, which looks like a rationalizing modern­ comes first.
ization (I). (c) The first cryptographic device indicates the
The version in 8G, B, U and Tr shows a simple and straight­ group by a number of short i-runes, the place in
forward parallelism, at least in the first four paragraphs. Each the group by long i-runes. From the name of
.device is. carefully described, and immediately illustrated by the i-rune, OE. iis or is, this device is called
'an example. 8, on the contrary, tries to systematize by treating iisruna (isruna).
the first three varieties of secret writing together. The fourth
(d) The l-rune (name lago) maybe used in a similar
variety is then discussed by itself, and all four examples are way: lagoruna.
given at the end. Only two manuscripts, B aJ;ld 8G, also
include a fifth variety. We shall see that this may well be a (e) The third device consists in marking the group
later addition based on a local tradition. by one, two or three short strokes to the left of
a vertical shaft. and the place by one to eight
The five manuscripts provide the following runological
strokes to the right of that shaft. In three
information :
versions (8G, B, 'U) this device is called hahal-:. .
(a) The OE. fu~orc of twenty-eight runes was divided into Tuna, in the two others hagalTuna (U, 8). As.;'
sections or groups, specified by 8 as three groups of eight far as I know, this name has not yet been ex­
and one of four runes. The combined evidence of the plained (1). The reason is, that scholars have
other four manuscripts also points to a system with groups mostly been misled by the variant with hagal-.
of eight runes, but we cannot make out for sure whether There can be no doubt that there is no con­
the last four runes formed a group by themselves, or nexion between the hahalTuna and the rune
whether they were simply added to the third section. hagal. I believe the solution lies in' another
One may even doubt whether the evidence of 8 can be direction. The first dement of the name is
considered conclusive by itself. By choosing an example OHG. hahal' cremacula " i. e. a pothanger with
containing one of the additional runes (a), the author of FIG. 16 a rack (2).' The compound is a good descrip­
the original of 8 was obliged to settle the status of these tion indeed of this type ofsecret runes, as fig. 16
last four runes. All runes in the example of the other
(1) H. ARNTz, Eis- mul Wasst7nlnm,273 : " Was mag er [i. e. der Schreiber]
version were taken from the first twenty-four, and so the sich aber bei der hoJuzlruna gedacht haben? Die Zeichen sind keine Runen,
question of the additional runes could be left open. We. sondem fonnal am einfachsten als /Qgorunos zu erkliren, wobei die Runen, die
das Geschlecht angeben, noch dazu Wenderunen sind. Was hohal bedeudet,
conn,
(x) For M. OLSEN'S interpretation of see p. Isz. I doubt whether weisz ich nicht. JedenfaUs sind es keine ' Hagelrunen .,. schon ihrer Fol'lTl
an equation coru.s =< ORG. hrahan == Hrabanna brings na any Closer to the risch nicht, und deren Name heiszt zudem immer haaal oder allen Calls hagl ".
solution of this riddle. (z) E. G. GRAFF, SprachscMtz IV, 772 s. v. halwla.

:13 2
.... 133
 shows (1). Consequently SG, B and Tr here retain the
unexplained. Fortunately one 'example of clop/runa, with
original reading. The explanation proposed by S is only.
that name added, has come down to us, paradoxical as this
a half-hearted attempt to justify the name.
may seem. By definition a cryptocheironomic system
(f) A fourth system indiCates the group and the place of the cannot be • written " but we have only to think of the way
rune by dots placed in horizontal lines. Either those in which a Morse message may be represented by a series
showing the group ~re on the upper line, and tho~e showing of stops and dashes, to understand how clopjruna couldbe
the. place on the lower line, or vice versa. From the represented in writing. The one example of clopj~na
OHG. word for «dot', stup/ or stop/, this device is called that has been preserved is a signature of Ekkehart IV of
stop/Tuna (sto/runa, stoph:runa) (2). . St. Gall in St. Gall MS. 176 (cf. Appendix I). But this
instance has nothing in common with runes: it is based
(g) The fifth device does not seem to have been used in writing. on the Latin alphabet. Each letter receives a number
Both the name clophruna, clojruna, from the OHG. verb according to its place in the alphabet, and these numbers
klop/on ' to knock, to tap, to rap', and the word pulsu(s) are then indicated by one, two, three, etc. dots: a = I = .,
in the text show that it must have been a sort of crypto- . b 2 = ,., c 3 = ... , etc. Ekkehart himself calls this
cheironomy (3). It seems to have been something like device chlophruna~ The' formula distinctU peTsonis et
a Morse code, but the text is rather obscure at this /ittms may probably be explained. as follows: this sort
point. The first four paragraphs distinguish veTSUS, i. e. . of signalling could be used· during the hours of silence
group, section, and /ittera, i. e. rune. A parallel system imposed by the monastic rule; first the name of the person
of tapping would probably require two different tones or for whom the message was meant was signalled, after that
sounds, one for the groups and one for the runes.' One the message itself. If this is the right interpretation of .,.t'
does not see very well how this may have been done in the last paragraph in Band SG, the name clop/runa is)
practice, and even so the mention of personae remains not parallel with muna, etc., as this device has nothing:,"
to do with runes. The other examples of the use of
clop/runa are also briefly discussed in Appendix I to this
(1) After fig. 53b in W:BoMANJoi, Biiuerliches Hawwesen und Tagewerk im chapter..
alten Niedersl1£/uen. Weimar, 19z7;cf. also figs. 45, 47, 49-521, 54. 55, and :
P. DE KEYsmI., Vie den lmientam {Jon IuIt Folklore·Museum te Gent. De Hangel. From Iceland we have some evidence that a similar
Oostvlaamsehe Zanten 17 (19421), 140-145. device, klapPrdniT, was praCtised on the basis of the Norse
(21) E. G. GlW'F, SprachschatsJ VI,. 659 s. v. stuph. The reading $oofruna
led S. BUGGB, 'DI!T Runenstein von RDk, 198, to the following conclusion:
fupqrk: But the text shows that this fupqrk is not the
" AlB ziemlich sieher darf man es femer bezeichnen, dasli die in den altdeut­ shorter sixteen-rune type, but one with new 'dotted'
schen HandschrUten aufgezeicluiete Geheimschrift eine nor dis c h e (nieht runes added; it even contains a p which had been ab~m­
eine angelslichaische) gewesen at. Hier fUr sprechen die 'folgenden Griinde :
•••21) Die Geheimschrift, welehe aus runden Punkten gebildet at, heisst in der
doned in the North before the end of the eighth century .
St. Gallener Handschrift s;oofnma. Oiesen Namen muss man .wohl mit Since the treatise also betrays some knowledge of the
M. Olsen aus altn. soppr , Ball' (dem kein angelsiichsisches Wort derselben notae Caesam and the notae sancti Boni/atii, it is not very
. Bedeutung lautlich entspricht) erklliren..... But soojrU'/'Ul is a misreading likely that these klapprdniT go back directly to ON. cryp­
going back to HATll!M.ER'S edition.
(3) R. A. S. MACALlSTl!R, Secret Languages, 37. tology; it rather looks as. if they were a late adaptation of
On klopj01l see W. WISSMANN, Nomina postfJl/IT'balia in den altgermanischen the Continental clop/runa. Further details in Appendix II.
Sprl1£hen. Gottingen, 193z (Ergiinzungsheft z. Z. f. vgl. Sprachf. II), 175, . There is of course the name clop/runa, the second
183. 193. Cf. also the klapprUnir in Appendix II.
element of which seems to point to a runic origin. But
13-4
.... 135
 . in OHG. the word TUna glosses Lat. summo. mysteria (I).
On the other hand S has a completely recast edition. S itself
and consequently it is no decisive argument in favour of
can only be a copy of that new version. Thus we obtain the
a runic origin. The name may of course have been following stemma .
created on the model of Uru7ltl. etc., but this does not
[Clophruna]
affect the interpretation of the device. X
The names of the first t~o devices may have been brought
over from England. Those of the last three were invented Xa Xb me inventioneJ
on the Continent. Sto/- and clo/- in SG (the former also '" /
in Tr) are more extreme Alemannic forms as against stoph­
and cloph- in B (and stroph- in S). At any rate the names ' "' ,,- /
Xb'
of the last two devices point definitely to High German territory.
This does not altogether agree with the conclusions founded
on the rune-names. The prehistory of the isTUna tract seems Xc
to be a complex one. It rather looks as if the tract had not
come from England in the form we know; on the contrary,
there is a possibility that it was composed on the Continent.
A further analysis of its background will show how this may B SG Tr U S
have happened. Xa may perhaps be claimed for St. Gall. SG was written in a
Before we study that background, one point remains to be centre with insular influence, B probably in St. Gall itself.
examined.: the relationship of the five versions. Occasional There the Clop/TUna tradition lived on as late as Ekkehart IV's .... ....
references have been made to this problem before. We saw days, cf. Appendix I. Xb may rather be claimed for the,.
that B and SG are closely related, but that neither can be the Rhine. area, and it is not impossible that Xc also originated "
exemplar of the other. There may well be intermediate links there. As to the way in which the tract was disseminated,
between each of them and their common ancestor (Xa); the B may contain some information, see p. 154 f.
CIoP/TU7Itl may have been an addition to the latter not found in
* **
the archetype (X). U and Tr have in common the example
coruus instead of corui and also misreadings for the' name of
the CB-rune (U eos, Tr cos), which may point to a common RUNIC CRYPTOGRAPHY.
prototype. Tr was copied from a poor exemplar but has a
• corrected' text, whilst U shows traces of in8uence from a The, problem of the origin of these cryptographic systems

De in'Ventio'118 alphabet. A similar influence may be detected may be approached from two angles ;

in S (the x-rune). The latter also shares hagalTuna with U, (a) from the runic evidence itself, i. e. especially froIn the

and consequently they may be derived from the same ancestor. examples found in the inscriptions;

• (1) G. EHlWIMANN. Guchichu tier deutsclum Literatur I. 46.

(b) from related cryptic systems. There is one form of secret

E. G. GIWIF. Sprawdrattt II, 523 8. v. rllna. writing !to which we shall have to pay attention in the first

A different interpretation is offered by E. WEBllB, Zu tkm Wort RmuI. Ar­ place: the. Old Irish ogham. This bears in fact so many

cbiv 178 (1941), 1-6 (not convincing).

similarities to runic Cryptography, that some scholars

136
137
""
 believe the two may have a common origin, or even proceed Each rune may e.g. be indicated by that immediately
one from the other. preceding or following in the fupark.
Before we examine these two fields, however, a few words
must be said about the ON. manuscript evidence. In Iceland (f) In Sweden we find a curious instance where runes of the
runic cryptography seems to have developed numerous new old fupark (or of the English fuporc) are carved instead
devices, as is shown by the copious list in Liljegren's Run­ of those of the new fuplP'k (the Rokinscription, cf. infra).
LiiTa ( 1). Liljegren had his material from late Icelandic (g) The most important type, hoth numerically and for-our
manuscripts; part of it may go back to the ON. period, but subject, is based upon the division of the fupark into
much is no doubt of later invention. On the whole this material groups of runes. .
appears so sophisticated, that it is safer to rely only on the It is hardly possible to decide when this last type of crypto­
epigraphical evidence. Cf. Appendix II. graphy originated. The possibility of such cryptography
existed as soon as the fupark had been divided into groups of
The Epigraphical EfJidence. runes. Our. earliest evidence does not seem to reach farther
back than the sixth century : the fuparks on the bracteates of
In .runi~ inscriptions we. find various types of cryptogra­ Grumpan and of Vadstena are divided into three sections of
phy: (2) eight runes (I). There are a few symbols in older inscriptions
(a) Runes may be left out, words may be shortened, con­ which may perhaps be interpreted as early i~tances of the
tracted. As a matter of fact, it is usually quite hard to hahalruna:
decide whether a given inscription presents this type of (1) on several arrows found at Nydam (S. Jutland, ca. 400 )
cryptography, or simply reflects the engraver's ignorance, we find a syt.nbol which may be read III = for t. On .~
clumsiness or carelessness. . one, 1/2 seems to be carved (2).
(b) The order of the runes in a· word or formula may be (2) On the Kylver stone (Gotland, earJy fifth century): a
changed. Instead of the word (or formula?) alu, which vertical stem with six downstrokes to the left and seven
seems to mean I protection " one also find'! lua or lau. or eight to the right (3).
The remark on the first type applies of course also to this
device. (I) See e. g. W. KRAUSS, Runenmschri/ten, figs. 5-7.
. (c) The inscription may be written backwards, entirely or (2) L. JACOBSl!N-E. MOLTKl!, Runeindskrifter (Text), 37 f. and Atlas,
nOs. 47, 49, 51; no. 46 may perhaps be read lIz. The autholll consider these
in part. , runes ' simply as magical symbols. ,!hilst W. KRAUSE, Runenins£hri/ten, 448 f.
(d) Special rune-like signs may be substituted for certain explains no. 5 I as an apotropaeic symbol, which later on was connected with
runes, e. g. for the runes indicating vowels on one panel the 15th rune; no. 46 he explains as a ligature ai, a solution already proposed
by WIMMI!R.
of the Franks casket. (3) Much depends on the interpretations implicated by retouched photo­
(e) Runes may be substituted for the runes actually meant, grsphs. That of H. AruITz, Runenkundet, Pl. V has six downstrokes to, the
according to a key known only to the prospective reader. left side and eight to the right; moreover it seems to indicate a, deliberate
grouping of the lateral strokes: [, z, 3, 4. 5, 6 to the left correspond to I, 3,
4, 6, 7, 8 to the right. According. to E. MOLTKE, however. there are only
(I) J. LILJ1IGRBN, Run-Lara, 53 f. seven strokes.to the right (Arkiv 56 (1941-194Z), [(8). W. KRAUSI!, Runm­
(2) O. VON FIm!sEN, R_a, 157 f.; inschri/ten. 431 f. and H. ARNTZ, Handhuch1 • 142. propose to interpret the
H. AruITz, Handlnu:h 1, 272 £1'.; L. JACOBSllN-E. MOLTKE, Runeindskri/ttr ~haracter as a multiple t-rune; but then the difference in the number of strokes
(Text), 847 f. on both sides of the shaft remains a difficulty.

13 8
... 139
(3) the inscription on bracteate no. 61 (Zealand no. 2 The cryptic runes have been called 'twig-' or 'tree-runes '
Stephens no. 57; sixth century) ends in 313 i' or e? (I) through a mistaken identification with the type called kvistronar
(4) the wooden 'sword' of Arum (Frisia; 550-650): 2/4 = in Icelandic. But in the kvistronar the lateral strokes go
j? (2) .

'~'~1 It
I
(5) the Kerlin ring (pomerania) : 2/1 or 1/2 = n, u or b. (3). 1

For most of these inscriptions other interpretations have been ,•t

I

~1t+'t+
offered (cf. the footnotes), and perhaps more plausible ones.
On the whole the instance from the Arum inscription is the
most convincing; but it seems a narrow basis on which to

~I) I~t~'~
found far-reaching conclusions. The possibility of devices
such as are described in the ist'una tract having existed before
ca. 800 A. D. should not be excluded; but how far they may
go back we cannot ascertain.

Hac'kness (Yorkshire) (4).

11'11*l . H~ckness
The face of the stone which interests us bears an inscription FIG. 17

in three parts (a) two lines of runes, (b) three and a half lines
upwards, whilst here they slope down, precisely as in the
of hahalruna, and (c) the Latin word ORA (in Roman).
. hahalruna of the manuscripts. Unfortunately the inscription
The first two lines have been read embdwre gnwguire. i has suffered much, and for ~everal symbols the number of ,~,
Thus far this part of the inscription has not been explained (5).
side-strokes cannot be made out (I have placed a dot where .
(I) L. JACOBSEN-E. MOLTKE, Runttiruhkrifter (Text), 535 f.; Atlas, 419, the number of side-strokes could not be counted by the scholars" : ,. ,"
Br. 61. The symbol has been interpreted as a threefold t-rune, d. W. KRAUSll, who examined the inscription) (I) ;
RfUIlmituchriften, 477 f. On bracteatea nos. :ax and aa (Overhomblek, N. Jut­
land, nos. a !IIld 3 : L. JACOBSEN-E. MOLTKll, o. c., (Text) 507 if., Atlas, 414) 4/8 2/4 3/44/ 1 3/· '/.3/3 ./ . . / . . /.
we also find combinations which could be read I/a or ala; but these inscriptions
4/2 2/4'3/8 1/1 3/5 3/1 1/1 .1 . ./ . .1.
are partly made up of pseudo-runes, and therefore cannot be considered as
evidence. 3/2 III .f. ./. 4/6 3/8 3/3 3/5 4/4 ./.
(a) H. ARNTz - H. Zmss, Runendenkmiiler, 106 if.; P. C. J. A. BollLBS III III 2/3 ./. 3/8
Fmsltmtl, 347 if. T. VON GlUENBllRGllR, Neue BeitTiige, a99, .was the first to
compare this symbol with the St. Gall hahalTU1lG. When ARNTz identifies But even if too little remains legible to allow of a plausible

a/4 with 3, this must be a slip; #4 j. reading, some information may be derived from this part of

(J) W. KRAUSB, RfUIlmituchrijten, 446 f. and fig. lB. The rest of the inscrip­ the inscription. As Macalister pointed out, it is founded OIl

tion reads ula or aiu, the well-known apotropaeic fonnula.

(4) G. B. B,ROWN, Arts VI, i, sa if. (esp. 67 if. and Pl. XIX, 1). a fuporc divided into four sections. 1 believe we can even go

G. SnwHI!NS, M _ t s I, 467 f. (photograph). further. Of the first group or section, only one rune ()Ccurs

(5) R. A. S. MACAI..IIITllR has proposed several explanations: it might be for sure : the first; of the second only the third and fourth

.. (a) a subatitutiary cypher; (b) a transpositional cypher; (c) a list of names

or other 'Words, written straightforwardly, but abbreviated so as to be unin­
(x) "Of the thirty-Jive letters, ten .re entirely effaced, and others are muti­

telligible to those to whom the names are unknown; (d) a mnemonic clue, lated so that the number ofside-etrokes - a vital question - is uncertain. For

explaining to the initiated the principle on which the following inscription in these reasons, and for reasons stated above, it is probably quite impossible

tree-runes is constructed. As a matter of fact, I know of no plausible inter­ to decypher the inscription" (R. A. S, MACALlSTBR in : G. B. BROWN, Arts VI,

pretation of these runes" (in: G. B. BROWN, Arts VI, i, 68). i, 68 f.).

140
141
....
runes; of the remaining two sections five or six different runes, the eighth century. With one or two exceptions, the order of
including 3/8 and 4/8. From this we may conclude that the the rettir or sections has been inverted : that beginning with t
fuporc consisted of more than twenty-four runes: probably comes first, then that with h, and last of all that with f. The
twenty-eight or more. Since only one tune of the 'first' following survey does not claim to be complete. The inscrip­
section is used, it follows that the sections were probably tions are transcribed in fraction-like formulae; the devices
numbered in reverse order, the one with the additional runes actually used are shown in fig. 18.
coming first. The high frequency of 1/1 (5 out of 23 legible
characters) agrees well with the fact that this symbol stands
for a vowel (a). The Norse inscriptions also favour the ~e­
versed numbering of the sections, as 'we shall see (I).
ROk (a)~~t t~t t t t1 (b) ~~II)(Q"I
The Hackn~ss inscription could be the foundation of the
history .of the' isruna devices, if its status and meaning were
better established. For the same stone also bears an inscription
«) l11JJJl11JJ 111JJJJJ
(d)XX~(')XXX

carved in a secret ogham, which has not yet been deciphered.

Unfortunately the date of the whole is uncertain: Collingwood
would place it towards the end of the ninth century, but Brown
has apparently good reasons for believing it to be a century
older. If so, we may assume that this inscription with hahalruna
is as old as the earliest appearance of these secret runes in
manuscripts, and probably older. Brawn suggests that there
K(n9r3tor~o&k . ~l1~SR~iY b Lom

is a connexion 'between the cryptic runes and the secret oghams,

and even that the former owe their origin to Irish influence : lund. ~~~ rfi fff=r=r. M"lrfell
"... it would be much more probable, indeed quite a
matter of certainty, that the pseudo-Oghams and the
M...howeVnl rrrrrrNo,um \\\1.2'
M~uhoweXVIl1 rTrrr11Jflf
twig-runes would be devised and executed by the inmates,
among whom Irish monks or nuns would no doubt be
found" (2).
All Norse inscriptions with comparable cryptic runes are
founded on the later sixteen-rune fup¥k. Therefore the
system as used there cannot be older than the second half of R.,brunn, " IIII " III " /11111/ ,III " , IIIIII " /"111
(I) The Hackness inscription,however, 'hardly allows of an interpretation;
transcribed on the basis of the above assumptions, it reads : R_dven 1,1 "'1.,,,./t: ~ . /' ' {~:
~ .r ~ ~ J J J -r
w m j f
u. m 8 a 3 h a
n a C 8 .j 3 0
VLI.t. .
a a e s
(z) G. B. BROWN, Arts VI, j, 1Z; cf. also R. A. S. MACALI'.STER, Buret Lon­
1.1
guagu, 61.

142
,5' (ZV ,'Vvt·l(}(f'Uct ( n
FIG. 18
(j<{- se Cr ;t1O-I0c;,\
. /
CV-'~43
;\t.. flv",,: ~ IIv ~ ( (' /?.i((iIt.!
....
Riik (Ostergotland) (I). Narum (Bohusliin) (I).
This, one of the most interesting and puzzling of all runic 3/'1. m. In this ins~ance the original order of the
inscriptions, offers a regular catalogue of cryptic systems. sections has been retained.
Rotbrunna (Uppland) (2).
(a) [.] 2/4 3/6 3/2 1/3 3/2 3/6 1/3 2/3 2/2 2/3 = [s]akumuk­
mini. '1./4 2/3 3/5 2/3 3/6' 3/5 = airikr.
RiHlV6n (Romsdal) (3).
(b) oossoosss = 2/2 2/3 = ni.
8 s t '1./4 '1./3 '1./'1.: 3/5 2/4 i '1.15 1/1 = 8stain: ridst.
(c) 3/3 3/2 = lJu; 3/5 = r. Valsta(Sodermanland) (4).
(d) 3/5 3/2 2/2 '1./3 1/'1. 3/4 lJR = runimolJR 3/5 '1./4 3/'1. 1/5 . 3/2 '1./4 3/5 liz rauR . uarb.
(e) '1./4 3/'1. 2/3 3/2 2/3 '1./5 auiuis. For the late Icelandic manuscript evidence, cf. Appendix II.
As this inscription is fairly early (ca. 850), it is an important Of the illustrations given in figs. 17 and 18, Rok (a) and (d),
element for the solution of the isruna problem. All similar ON. Lorn, Lunda, Maeshowe VIII and XVIII, Ma:lifell, Rotbrunna, .
inscriptions are of a later date : Rodven and Vllsta remind one of the devices described in
KingigtOrsoak (Greenland) (2). the isruna tract: Rotbrunna of the isruna, the others of the
'1./3 1/. = i [.] hahalruna. But not One of these instances. except perhaps
Lam VII. (Gudbrandsdal) (3)· Lunda, are identical with our manuscript cryptic devices.
['1.] /4 a. S. Bugge and M. Olsen interpret the short strokes in ihe Rot­
Lunda Rectory (SOdermanland) (4)· brunna example as s-runes (of the Halsinge type); M. Olsen
3/3 2/3 '1./4 k 1/'1. [should be 2/'1.] a = lJiakna. supposes that similar short strokes which he reconstructs from
MlBlifeli (Iceland) (5)· the Sigtuna amulet inscription were R-runes (name iR) (5).
1/1 3/4' 1/3 2/4 '1./5 = tomas. This of course destroys the parallelism with the isruna cOm­
Maeshowe VIII (Orkney) (6). pletely, for then we have to compare Rotbrunna with Rok (b).'
2/4 3/5 1/4 '1./3 3/6 3/5 ~rlikr.
Maeshowe XVIII.
3/3 '1./3 '1./5 '1./4 3/5 3/5 3/'1. '1.1'1. '1./4 3/5 = }Jisar runar. (I) O. VON ~,_ Runorna,. [58.
1/3 a lJ r = malJr. (2) O. VON F'RI.BsBN. Upplmuh nmlt#1Im', 46: Runorna, 158;
M. OLSEN, Sigtuna-mmdetten, 20.

(I) S. BUGGE, Tollmingen; Om lUmeindskri/terne paD. R6k-Sunen; Der b) M. OLSEN, RUnflr'i RtuhJen. Romsdalsmuseeu srbek, 1938-1941, 1 ff.

R1menstein 'VOn Riik. O. VON FRII!SEN, Runoma, [07 f. .

F. LXFFLIIII, Om R61utenen; Bidrag; (4) E. BRA'll! - E. WllSSEN, S6de:nnanlandsrunimkri/ter • .36 no. 47 and
E. BRATB, O'terg6tlmuh rvninIkrifter, 231 ft.; pI. 29. ~ead .rauR firt or rauR fart or rauR uart. Cf. O. VON FRIESEN,
O. VON F'RI.BsBN, R61utenen. Runorna, [58; M. OLSEN, Sigtuna-amu14tten, 21.
(2) M. OLSEN, KingigtUl'loak-lfmen; Sigtuna-amuletten, :23· S. BUGGS, Der RU1IenItein _ R6k, 201 f.
(3) M. OlSEN, En.fut/u.l:rk-innskrift iLom kiTh. M.o.M. 1943,88-<)6; Norgtll (s) M. OLSEN. Sigtuna-amu1etten, 20.
imIIkri/termea de yngrtI nmer I, 83 f., no. 36. To the above list the S&ierkOping inscription might perhaps be added, but
(4) S. B. F. JAmSON, S6nnllinthka runltemfynd· Fomvinnen 1 948 • 293 H. the interpretation of the portion which interests us here is not sure : kk. kiiii
(s) A. BJEKIITBD. lslmuh lUmeindskrifter, 151. kkk [.1]. cr. A. NORDiN. SOderkfJpingssunen. En ""ju.nnen rnmtm med magisk
(6) B. DICKIN8, The lUmic lmmptimu of Maulurwe. Proceedings of the syfu frdn .. overgangstidm '~. Fomvllnnen 1937. 129 ff.; NORDiN offers a
Orkney Antiquarian Society 8 (1929-1930), 29· different interpretation.
M. OlSEN, RJi.nar er '(istu rynamr menn, J77 ff. M. OLSEN, Sigtuna-mnukttm. also connects part of the inscription on a
O. VON F'RI.BsBN, Runorna. 102 (M. Olsen). copper plate found at Sigtuna. with the imm4 tnIdi~on, cf. p. 152..

14 1 45
144
'"
 The Ogham (I). letters of the first group consisted of one to five vertical notches
below (or horizontal notches to the right of) the stem-line;
On the whole few cryptographical devices used in classical those of the second were made by one to five notches above
Antiquity seem to be comparable with the isruna system. (or to the left of) the stem-line; the next five letters were similarly
Polybius's torch signal system is very similar: the (Greek) indicated by one to five strokes transecting the stem-line
alphabet is divided into five groups of five letters each; the obliquely (or cut at an angle on the two faces of the stone which
number of th~ group to which a given letter belongs, and the met at the edge used as a stem-line); the last five were made
place of that letter in its group are indicated by one to five in the same way, but here the strokes were at right angles
torches; e. g. 2/5 = K (2). But who will undertake to prove with the stem-line. Sometimes dots on the stem-line were
that there is a connexion bet'reen Polybius's signals and the substituted for the strokes of this last group. The letters
isruna? formed a unique and hitherto unexplained sequence. Later a
There is, however, a closer parallel both to the isruna and to fifth, obviously secondary group o( symbols was added to render
Polybius's torch signals in the Western world. As a matter ·diphthongs. The whole looks as follows:
of fact, the Old Irish alphabet or ogham shows such a striking
resemblance with the fu~ark, that some scholars have earnestly I II 11/ 1111 11111·I " 1/1 - " " I " " TI. 1/1/.
/ If~ fl/I//II
~
considered the possibility of both being derived from a common b I ~ ..$.n~... ~.1 .. t. c. q m 9 "9 z r
ancestor (e. g. Marstrander). H. Arntz has even tried to LII III 1111 11111 V A II
derive the ogham from the sort of cryptic runes described in r nllT nrrnlJlI\~1P
~ 0 Ur ea.
It Ll.r 01 10 U
the isruna tract. These attempts have met with little approval :
FIG. 19 .~;~
although the two systems are alike in the main lines, there are
such important differences as cannot be explained if we accept The inscriptions in this script (mostly funeral) are usually' ~.
a common origin or dependence of one on the other (3)· Cf. dated from the fourth to the seventh century A.D. (I). From'
infra. later manuscript sources we know the names of the letters.
The ogham alphabet seems to have consisted originally of Just like the rune-names, these names were actually words
four groups (aicme) of five letters each. In epigraphical usage used in the language, but they were chosen from one semantic
the script followed a line carved on the face of a stone, or else group: apparently they were all names of trees (2). Each
represented by a more or less straight edge of the stone. The (I) R. A. S. MACALISTER, The Archaeology of Ireland. London (1949),
328 ff. Ogham inscriptions are occasionally found in manuscripts, e. g. in
(I) H. AllNTz, ])as Ogom, corrected by St. Gall MS. 904 : C. NIGRA, Reliquie celtiche. I. II manoscritto di S. Gallo_
R. 'rHuBNEYSEN, Zum Ogam. Firenze, etc. 1872; W. M. LINDSAY, Early Irish Minu,cule Script, 46. On
R. A. S. MACALISTBR. Secret Language" 1 ff. the regular treatise on the ogham cf. infra.
J. VENDRYl!S, L'ecrituTe ogamique et ,//$ origines. Etudes celtiques 4 (1941), (2) G. CALDER, Auraicept, 275 f. : " Secundum alio, it is from the trees or
the forest that names were given to the Ogham letters metaphorically. More­
83-116.

Cf. AllNTz's bibliography in his Hantlbu£h" 277-2 79.

over beilhe, b, is from the birch of the forest for the first letter on· the path
(2) W. SOss, Ueber antike Geheinuchr~thoden, 148 f. of the Ogham alphabet. Luis, I, that is, elm in the forests. Fern, f, that is,_
(3) H. AllNTz, Dos Ogam, 394 ff.; Hantlbu£h" 294 ff. alder in the forest. Sail,s, of the Ogham, that is, willow, again, in the forest_
R. 'rHuBNEYSEN, Zum Ogam, 199 ff.
Nin, n, of the Ogham, to wit, maW' of spear, or nettles in the woods. Uath,_
In connexion with AIINTz's theory cf. also
h, of the Ogham, that is, test-tree or whitethorn, 011 account of its thorniness_
W. KELU!R, Die Efluuhung du Ogam. Beitrige 62 (193 8), 121 ff. Dur, d, of the Ogham is oak, again, in the forest. Tinrre, t, of the Ogham,.
H. jAENICICEN', Zum Ogam. Runenberichte I (1939- 194:i), II2 ff. to wit, holly or elderberry in the forest. Coll, c, of the Ogham, to wit, hazeL
For MARsTRANDER'S hypothesis see especially his Om nmene, 125 ff. in the forest. Quert, q, of the Ogham is holly in the forest .. or quicken tree,.

147
146 ...
 ;k i<¥ ¥ )< >< )(

group was known by the name of its· first letter, e. g. aicme

hethi • group, tribe of h ' (cf. in later Norse runic lore FreyslBtt
13 etc..
, group, family of f' c).
The ogham is built upon the same fundamental idea as the
b,· If" n 'j 'Y
isruna : each letter may be written as a fraction, the numerator
indicating the group to which it belongs, the denominator its 89~ /1\ ~~~ hd
etC'..

place in the group. But whereas runic. cryptography writes

both the numerator and the denominator, in ogham crypto­ 12 ( ( .~ ( ~«) ~. e.tc.
graphy the numerator is shown 'only by the varying shape of m 9 ng $,. r a. o.
the denoniinator. It looks as if the oghamists had not been a. 0 . u e

able to feel the numerator as a figure in the same way as the 4-8 II II III! II II III! 1111 IIII 1111 II II III! 1111
etc.
denominator. They seem to have been so spell-bound by their
fundamental system (which used a different type of notches
for each group) that they never reached the degree of abstraction 1+ .I '. II
.. II!. ,...
... /111 1/111
..... 111111 1111111' etc.
....... .......

~x:emplified by the isruna system. This does not mean that b I· fs n. h d

they were unable to invent cryptograms quite similar to: those FIG. 20
found in Norse inscriptions, and some even more complicated.
Therefore the cryptic systems derived. from the ogham must first three classes show a close resemblance to runic devices.
be briefly examined. . Calder's no. 13 (ebadach Itaind (ebad-fashioned ogham of
Our main source on ogham cryptography is a treatise on the Ilann ') reminds one of the X-shaped secret runes on theRok
invention and peculiarities of the Old Irish alphabet, edited stone, no. 89 of the 'twig-runes' of Maeshowe. In no. 12. '~':
by G. Calder (I). In his analysis of the part devoted to cryp­ the m-group is formed of scores looking like C, and the a-group;
tography, R. A. S. Macalister distinguishes the following forms of Scores like :::>, rather like in the Hackness ogham. No. 48~
of cipher: 1. Anagrams; II. Confusion by means of insertions; uses a somewhat similar device, as its name coll ar gttta ' c for
III. Tampering with the form of the stem-line or of the letters; the vowels ~ implies : the consonants are written as in the normal
IV. Substitutions (2). Although only devices classified under alphabet, but for the vowels ogham c, cc, etc. has been sub­
IV are directly comparable to the isruna type, some in the stituted. A cryptic alphabet in Oxford MS. St. John's College
17 in the same way substitutes C, CC, etc. for the vowels, but
or aspen. Mum. vine, m, to wit. mead [from it]. Gort, cornfield, I, to wit,
keeps the consonants (cf. p. 31) (I). The ogam ar ahairtar
fir. Geta1, ng, to wit, broom. Strm!. str, willowbrake in the forest. Otm, 0, to
wit, furze or ash. Ur, u,to wit, thorn. &Iad,e, to wit, yew. Ida, i, to wit, cethrur ' ogham called" four-man" , (no. 34) uses the letters
service tree. EOOd, ea, to wit. elecamplllle. air, oi, to wit, spindle tree. Ui1kmm, d t II to represent the four groups; each letter is written as
ui, to wit, ivy. Pin, io, of tbe Ogham, pine, apin, in the forest. Hence
are named caera ~ gooseberries; ifin. again secutUium alios is the name of
many times as the intended character has scores, e. g. dddd = s.
that letter. EmancoU, witch hazel. ae, again, to wit, c doubled according to
fact or aCcording to form, to wit, c across c in illl form ". (I) R. A. S. MACALISTER, SecretLa1lg'Ul1gu,52, adds the following comment:
H. MERoNEY. Early Irish Lettu-Nmnu.Speculum 24 (1949), 19-43· .. A cypher of this kind, in ordinary Irish letters, was a favourite plaything
(I) G. CALDER, Avraicept, 272 fl., frmil the Book of Ballymote. The text among the scribes of eighteenth-qmtury MSS. I have a fragment of a MS.
also occurs, with slight variations; in British Museum MS. Addit. 4783; d. on the Ogham alphabet, with directions for writing Ogllam CoU (' C Ogham ~)
R. THuttNlmnlN, Zum Ogom, I9O, and R. DBROLBZ, O,am•• Egyptian " • AJrican' IlIld· Ogham consoine (' Consonant Ogham '). In the one, the vowels are
and • Gothic' Alphabets. represented by C's, just as in the Ballymote alphabet : the diphthongs a~ also
(2) R. A. S. MACALIllTER, Secret Languages, 38 ff., esp. 46 if. represented by C's in various positions... "

148 1 49
...
The • ogham of Bricriu' (no. 14) indicates the place of the ponding OE. word. The / in clo/- and soo/- as opposed
letter in the alphabet by a number of strokes or dots, but it to klapp- and sopp- is due to the scribe's Alemannic dialect.
differs from the isruna and stop/runa in that it abandons the
(c) The St. Gall version has 0 as the eighth rune of the third
distinction in aicme's or groups: the sixth letter is rendered by
group. But in the OE. fuporc the OGmc. 0 had the
six strokes, etc. (1). No secret ogham expresses the numerators
value re or e, whilst 0 was indicated by the new 0, i. e. a
in another way than by different shapes of the denominators;
form derived from the old a-rune.
here is a fundamental difference with runic cryptography.
None of these arguments is convincing. The klapprunir
These are the data on the basis of which we. should examine mentioned in Icelandic manuscripts are found in the neigh­
the isruna devices. There can be no doubt that, there is a .bourhood of notae sancti Boni/atii and other devices imported
connexion between the OE.' and the ON. material. The ties from England or from the Continent (cf. Appendix II). Bugge's
between the Gmc. material and the ogham are of a more impon­ soofruna is not in the manuscript : there we find sto/runa, a
derable nature. S. Bugge decided that the 'group-number­ good OHG. compound. Finally, the value 0 of 'rune no. 24
script'. was originally founded on the twenty-four rune fupark. cannot be construed as an argument against an English
He ascribed the Hackness inscription to a Swedish carver, as origin. A German who found repel or epel in his prototype
he believed it to show traces of Norse influence (2). The could very well substitute his native odil for it; moreover one
~rinciple of the • ~rou~-number-script ~ would have bee~' should not forget that early English manuscripts have such
mvented by a ScandmaVlan. Yet he beheved that the wryole ' spellings as oidil or oedil (I).
system rested ultimately on Irish models, i. e. on the ogh~m : We saw in the Introduction that C. J. S.Marstrander assumed
at Hackness Irish and English culture met (3). The Rok runes and ogham to be derived from a common ancestor
inscription in turn shows English influence, and this expk,ins (p. xxix). In his theory on the origin of the ogham H. Arntz
the presence of the cryptic systems. Bugge did not go so far, chooses a different course. He derives the ogham from the ;
however, as to suppose that the carve,r of Rok had been in"'. " Gmc. hahalruna (2). The system exemplified in the Hackness, ",
England. As to the isruna tract, it would go back to an ON. Rok, etc. inscriptions would be much older, and the ogham
model because (4) : would owe its existence to relations between Ireland and the
(a) The clofruna of the St. Gall manuscript correspond to Germanic world in the third century A.D. It would take too
Liljegren's klapprUnir. long to, examine how Arntz derives each single ogham letter
(b) The St. Gall term soo/runa can only be explained by from a cryptic rune, the more so as R. Thurneysen has shown
connecting it with ON. soppr • ball '; there is no corres­ that Arntz's hypothesis is quite unsatisfactory from the chrono­
logical and from the formal point of view (3). G. B. Brown
(1) Cf, on the ogam BricremJ R. A. S. MAcAuSTBR, SlICTet Languogu, 56
and R, 1\. S. Macalister accept Irish influence at least for the
and H. MmwNEY, A DnlitJic LituTgy in Ogam BricTemJ? M.L.N. 6a, (1947),
187- 189.
Hackness inscription (4).
(a) S. BuGGE, Der Runemtein 'Von Rok, 194 fi.: emundro I onzsboa M. Olsen mainly examined the Germanic material (5). He
"Emund hat Rube (ruht) in JEsbo": .. Ein schwedischer Mann Emund fundamentally agrees with Bugge : the isruna tract is no more
ist in JEsbo (Eseby) gestorben und begrsben worden. Eine mit ibm verwandte
Frau (wohl seine Tochter oder seine Witwe) hat in Hackness· zu seinem (1) H. S'J'lll)M, Old Engl:Uh Pmonpl Names, 30 •
Andenken die genannte Inschrift des Grabkreuzes schreiben llll!sen" (196). (a) H. Ann;z. [)as Ogom., 396 fi.
Cf. also ibid., :200. (3) R. THURNBYSBN, Zum Ogom, 198 if.
(3) S.BUGGB, Dtn' Runemtein 'Von R6k, 199 f. (4) G. B. BROWN, Arts VI, i, 7a; R. A. S. MACALISTER, SecretLanguogu. 61.
(4) S. BUGGE, Der Rrmemtein 'Von Rbk, 198 f. (5) M. OLSEN, Sigtima-amuhtt4n; Grimhi1th 06 Gudru'IU nmeimukrifter. .
15 0 lSI
....
than an artificial elaboration based on an ON. model. In his (a) what is the relationship between the OE. and the ON.
paper on the Sigtuna amulet (ca. 1050) he interpretS the section material?
iii isiR lJisisiR auk is as concealing isTuna : is would be the. (b) is there any relationship between runic and ogham
name of the i:-rune (I), iR that of theR-rune (I in the Hiilsinge cryptography?
fupqrk) (I). Thus he obtains a sequence
p III 'and' (a) The evidence for ON. influence in the Hackness inscription
which he reads Ipbt or opuf. The latter, reversed, gives fupo, and in the isruna tract, given by Bugge, may safely' be
i. e. the firs~ four runes of the fupqrk. Olsen believes that the rejected; Olsen has only worked on the assumption that
numbers of runes in various parts of the inscription have a such an influence existed, without proving it. The
magical meaning. Similarly he points out that corui, the example absence of an uncontested example ofgroup-number-script
in the St. Gall manuscript, has a total of twenty-four long and in the Scandinavian area prior to the Rok inscription
eight short strokes when written in isTuna. But corui would should not be overlooked. The Rok inscription itself
only be a substitute for ON. hrafn, runic hrabn;. written in displays a striking amount of learning, not only in the
isruna. hrabn would have a total of twenty-four strokes, which cryptograxns quoted before, but also in the use of the older
number may again play a part in magic (2). Olsen!tas even fupark (or the English fuporc) for cryptic purposes. The
tried to reconstruct the runic inscriptions of a-rfffihiid (Guil­ blundering use of the older runes proves that there is a
rdnarkvi4a II, 22) and Gudrun (Atlamdl 4-tJ) on the basis of break in. the tradition: the carver substituted old runes
the Eddic texts, assuming that they wer~ conceived as cryptic for the new ones he was used to as well as he could, with
inscriptions of the isTUna type. F. G:enzmer attempted the the result that he obtained hypercorrect archaisIns (I).
same for a stanza in the Egilssaga (3). ' ,':I'hese interpretations Therefore some form or other of foreign influence cannot
will easily be challenged (4); Olsen's general view of the isTuna be rejected a priori. Judging only by the dates, this '.;:'
7

tract does not stand close inspection either. seems to be the most logical explanation (Hackness ca. .
800 (.?); isruna' tract: . ca. 850 or somewhat earlier; Rok'
There is nothing in the isru1f.a tract which poin~ to an ON.
model, except perhaps the n-rune .of the Brussels version. But ca. 850; the other Norse inscriptions are all much younger).
what value has this one variant against the combined\~vidence On the whole the comparison with the Norse material throws
of the other four texts? One should not forget that the, runes . little -light on the isruna problem. There is at any rate no
of the Brussels version are on the whole drawn rather \are­ clear indication that the devices described in the isTUna tract
lessly. The runes are English, the rune-names .are. partly were borrowed from Scandinavia, although there is obviously
English, partly German, and so are the names of the cryptic some relationship between. the .cryptograppy based on the
English fuporc and that using the Norse fupllrk. We shall
devices.
As a matter of fact two aspects of the problem should be find at least one instance of a borrowing in the opposite diTection,
cf. Appendix II.
kept apart:
(b) The material with which we are' supposed to find an answer
to the second question is quite heteroclitic, and hardly
(1) M. OLSBN, Sigtlma-anaikltttm, 16 ft. allows of a definite answer. But perhaps we can come
(z) M. OLSBN, Sigtvtuz-amrdetttm, 20 f.; Gritti/Ji.lIh og Gudru:tu nmemn­
slrrift.er, Z3 (and note).
(3) F. GI!NZMBR, DitI ~ thr Egiluoga. Arkiv67 (19SZ). 39-47· (I) E. BKATE, Z'IIT. Deutung der R4ker Imchrift, z8s, in S. Bugge, Der Rune:n­
(4) A. BAuttrnID, Mdln_. 87 if. stm. 0071 RAil.
15 2 IS3
....

 closer to a solution if we again divide the question into :a few dependents who spoke his own language. Mter a time
two subquestions : the inner school was given to Marcellus with Notker, later
(I) was the runic group-number-script borrowed from called Balbulus, and the other boys who wore the monastic
ogham cryptography? habit; and the outer one to Iso with Salomo and his contem­
(2) were some devices borrowed from one system into poraries. It is pleasing to remember how much the monastery
the other? ·of St. Gall began to grow under these auspices " (I).
A couple of years befor~, one of the most famous of all Iiish
If in the first question we maintain the term • borrowed "
] am in favour of a negative answer, but] have no other argument peregrini, Sedulius Scottus, had first appeared on the Continent.
Liege seems to have been his favourite resort; there he found
than this : from the very beginning runic cryptography attains
benevolent protectors in the bishops Hartgar and Franco.
a degree of complication (consequendy of perfection) never
L. Traube has given a fascinating account of the varied activities
reached by ogham cryptography (cf. p. 148). If, however, we
displayed by Sedulius and his companions (2). On the basis
substitute • inspired' for • borrowed " I see no, reason for
-of their works and of marginal notes in a number of manu­
denying that possibility. But I fear we cannot come to ia more
scripts, he was able to reconstruct an important aspect of the
definite answer unless new material be found. i
In the answer to the second question th~poss~~y.)of such
cultural life in the ninth century. Among the names mentioned
a borrowing cannot be denied : we find ;ham
and runes on
in this connexion, there is first Sedulius himself, whose name
is found in St. Gall MS. 48 (together with that of Dpbthach)
the Hackness stone, perhaps also in Oxfofd MS~ St. John's
and in Berne MS. 363 (where both the names Fergus and
College 17 (p. 31). The borrowing may have been both ways :
Dubthach occur). Then there is Fergus, whom we know
if the • ogham of Bricriu' may have s~rved is
a model for the
already from the Bamberg letter (cf. p. 98 f.), and who is also,..,
stopfruna, . the lwistntnor may have inspired \he inventor of
found in St. Gall MS. 904, in the Dresden codex Boernerianus ~ ..,.
ogham no. 89 in the ogham tract (I). Theioe can be litde
and in Berne MS. 363; he was a companion of Marcus, as.
doubt that Irishmen, who seem often to have Md a keen sense
appears from Sedulius's welcome poem to the latter. Dubthach"
for abstruse and cryptic lore, would show grdt interest in
we met as the inventor of the cryptogram in'the Brussels and
secret runes. Perhaps there is some eviden~ of th~ presence
Bamberg manuscripts; he himself wrote Leyden MS. Voss.
in the neighbourhood of our isruna text.
lat. 67, and is also known from the codex Boernerianus, from
]n the Brussels manuscript the isruna tract is followed by
other materials for cryptic writing. . The most interesting St. Gall MS. 48 and from Berne MS. 363. Other scholars
belonging to the same circles were Dongus, Comgan or Congan,
amongst these is no doubt the cryptogram briefly discussed on
Beuchel, Blandus, Dermoth" Maelchomber, etc. (3).
p. 97 1£. If the Brussels manuscript was written at St. Gall,
we have probably an indication of how the cryptogram got (1) J. M. CLARK, The Abbey oj St Gall, 33 (translated from EKKKa4RT'S
there. Under Abbot Grimald (842-872), presumably in 8so, Casm S. Galli c. 2).
(2) L. TRAUBE, 0 Roma Nobilis, 347 if.;
there arrived at St. Gall ... .. Marcus, a bishop of the Scots ", J. F. KENNEY, Sources I, 557 if.
corning from Rome. .. He was accompanied by his sister's W. M. LINDSAY, Early Irish Minu.rctl1e Script, 36 if.
son Moengal, later called Marcellus by our brethren after his (a) A dI:matio Beati abbatis, ecc1eriaeHonaugiaeJacta, edited by J. MABILLON,
A1I'IIO!a OTtiinis S. Benedicti, T. Il (paris MDCCIV), 699 f., ends:
uncle Marcus. . The latter was most learned in divine and' Ego Wellimannus rogatus scripsi & notavi diem & tempus & locum.
human matters... The Bishop remained with his nephew and Haec charta in Msguntia civitate scripta XI. kaL Julias, anno X. regni
domini nostri Caroli regis & imperatoris. t Signum Beau abbatis, qui
(I) G. CALDER, AuraU:ept, 311, 313. and fig. 20 mpra. hane ehartam fieri rogavit. t S. Conigani episcopi. t. S. Echoch'
IS4
....
1 55
 The connexions between Sedulius's immediate sum~undings Excursion: Nemni'VUs's alphabet.
and the St. Gall group cannot be doubted. Other important
On fol. 2-ot' of Oxford MS. Bodl. Auct. F. 4.32 (1) we find
centres of the activity of these Scotti seem to have been Cologne
the following curious note in a ninth century insular ha~d :
(where Sedulius found a patron in Bishop G~nthar, also men­
tioned in the Dresden Boemerianus and in St. Gall MS. 904). Nemniuus istas reperit literas uituperante quidam sco­
Milan, and Salzburg, which was probably one of their inter­ lastico Saxonici generis quia Brittones non haherent
mediate stations. rudimentum; at ipse subito ex machinatione mentissuae
Still in connexion with the Brussels cryptogram, it is important formauit eas ut uituperationem et hebitudinem deieceret
to note that Sedulius wrote a poem on an altar founded by gentis suae; de figuris et de no minibus dicens :
Ruadri, Mermin's son and successor. The name Ruad,ri is then comes an alphabet with for each letter its Latin equivalent
also found in St. Gall MS. 904, w~ch seems to have come from and its name, first the letters corresponding to Lat. a - !¥, then
Kildare by way of Liege. The \presence of Fergus and his ten additional letters indicating diphthongs. .The same al­
companions at Mermin's court ~U hardly be a fiction. We phabet is also foUnd, with only trifling differences, in Cotton
may safely suppose that many of t\te Irish scholars driven from MS. Titus D 18 (cf. p. 338), and, with some divergent forms
their country by the Viking raids sbught and found a first safe and without the names, in Oxford MS. St. John's College 17
refuge in Wales, which was comparatively free from the inroads and in ~otton MS. Galba A 2 (cf. pp. 3 2 , 37).
of the Norsemen, and at times succe~fully resisted them. Last Nemruvus may have been identical with the historian Nennius,
but not least, it is not impossible that Marcus himself was a who began his Historia Brittonum with these words : I
Welshman, but had been educated in Ireland-J.I): _Ego Nennius, Elvodugi discipulus. aliqua excerpta scribere
It is rather tempting to connect the isrrJ,na tnict--ia..!IOme w~ ;"curavi quae hebitudo gentis Brittanniae' deieceret ... (2). ";~,

or other with the activities of these peTegrini. The only-iDs~ce

."

of runic cryptography of this type known froni England is the I. Williams re;w.arks : " It is difficult to believe that there were,
Hackness inscription, which was found to show traces of Irish two Welshmen fiving in the early years of the ninth century, ',.
influence, There is perhaps one further argument in favour both equally sensitive. to the charge of hebitudo laid against the
of connecting our. fuporc with Wales : its g-rune is also found Britons, one called Nennius and the other Nemnius or Nem­
in' Nemnivus's alphabet, which I propose to examine briefly niuu8'" (3). The argument seems rather cogent, but need
now. hardly concern us here, as only the alphabet calls for our
attention. One glance at Nemnivus's rudi71Ulnta shows that
episcopi. t. Signum Suathar epi. t Signum Maucumgib epi. t. Sig­ the expression subito ex machinatione mentis suae should be taken
num Canicomrihc epi. t. Signum DoilguS80 epi. t. SignUm Erdom- cum grano salis: there can be' no doubt that all Nemnivus did
Nch epi. t, Signum Hemeni presbyteri. '
was to adapt an alphabet of his • Saxon' enemies; an OE.
Hohenaugia. Honau WIllI an tlcclesia Scottmum on a small island in the Rhine
near Strasbourg. The original of the text printed by' MABILLON se.ems to be· fuporc. Here follows the alphabet in its two varieties (Nl
Io,t; he used a copy of 1079. whilst the original seems to be of81o (or. according
to E. MC'HLBACHI!R, of 786). One might feel like to connect Coniganus.with (I) G. HICKIlS, Thesaurus, Grammatica A:nglo-Saxonica. 168 (facsimile);
Comgan or Congan, Suathar with Suadhar, Maucumgib with Maelchomber, Caudogu 01 Western Manwcripu II, I, Z43 fl'.
Canicomrihc (an Olriah genitive?) with Caunchobrach, Doilgua perhaps. (2.) F, LoT. Nenniw tit l'Historia lJrittommr (Bibliothhque de l'Eoole des
with Dongus. Striking though these parallels may look. the chronological hautes. etudes :z(3).· raris, 1934. 147. NENNIUS'S preface is followed by
difficulties. should not be underrated. Versus NennJni [or Nenniui?]·Qd S~. '
(I) J. M .. CLARK, 1'he Ablury 01 St Gall, 34. following a suggestion of (J) I. WILLIAMS, NottiI·on Nenniw. Bulletin ofthe Board of Celtic Studies 7
L. ThA'UlIE, 0 Roma Nobilis, 370. {193S), 380 f. .

I5 6 157
....
 Bodley MS. F. 4.32 and Cotton MS. Titus D 18; Nil = Oxford n : in Nt doubtlessly n.

*
MS. St. John's College 17 and Cotton MS. Galba A 2) : 0: probably derived from (2; influenced by g ?

n,K~
N.Bts t:.*.r 419
J!.
•...
*~ t
~. . J "- -
~
~
:3'::;;
~
.",
U'>'~
~
..r:. !::
~"'-:s
,!,:!
-
::t
...~
..c
~
E
.cs
...>t
..c
:J
i
~
c.
'-
::J
q: the Nl variety could be adapted from k.
r: a reversed ' r with a stroke at the back '.
t: a t put upside down? or rather a ditto ea ?
u: the • reversed V ' type of u.
x: probably from the x-rune.

"'0 ...s: v E

*
.6 -0 . ., 6) ..... ::J -9 ae: derived from f2?

C1'l .~
eu: a y put upside down (Nl )?

If 1 <: 1\
n, 'I- ~ Z ..I, ,.1.. )I<J elau (ego) : connected with the stan-rune (in N \I ecce)?

11 96
Itt ) a~.It, ~ The names (I) are sometimes obscure, but of those that are

'-
s::
....s::::
s::
s..
o
.J::.
-et6
Q..
..c
.p
,-
;J
c:r
"';d
So.
en
s..
:s
II»
:3\
~'-
..::1
+>.
L
s.. -
,­
.! -
L
......
0
L
tJ
'­00·
....
clear some tend to show that Nemnivus may have translated
them from Old English:
alar 'ennui' cf. OE. nead;

:; u U \
~
:J cusil ' conseil ' misunderstood OE. rad?

.S1\1 -~
.u
<If
'-
a.,
U
L.
;:)
cd
U
c
,-
'­
..c
C

U
~
0
E::
:
<J :;
......
c
b
tjexu (?) : cf. derw ' oak' : OE. oc.

guichr 'impetueux, colere' : OE. em?

~
"U
.",,* ::I :s
..6
III
'- ,~
....s::::
tr.I v
(101 :::!
f::i
0 huil ' voile' : OE. sigil.

" ~
eo)

'fJ/ 4" r 1- It\ ~ 'f( y

louber ' lumiere ' : OE. df£g.

N,{:
t\t' { >K ~ 1- ,~ ~ ~ XX
t6
* Ell :s
"
:s
.of
" .:J::
I~ -'l
0
OJ
61 U
Q)
U
c:\ol
muin 'bienfait, present' : OE. giefu.

nihn ( ?) cf. OIrish nin • ash ' : OE. ~sc?

rat • grace' : OE. 'lUytIn?

arm ' acme ' : OE. gar?
" ..

FIG. 21 Others may have been chosen because formally they resembled
rune-names, cf.lich and OE. leah, feh; parthand OE. pear},
First let us consider the forms of the letters : piT}; rat and OE. rad; uir and OE. ur, oyr and OE. yr, aUT
a: probably derived from a. and OE. ear, cinc or hinc and OE. inc.
b : in Nl most like a pointed h. Perhaps a closer investigation may discover other similar­
c : p on account of resemblance to Roman C ? ities; at any rate there can be no doubt that Nemnivus knew
d: a slightly distorted.,. the OE. fuporc and derived his' Welsh alphabet' from it.
e: in N \I identical with (2 in one of the Oxford fuporcs.
I; in Nl practically identical with f.
g: identical with the g of the isnma-fuporc. * **
(I) J. WILLIAMS AD ITHKL, Dwporth Ed,eym Davod AUF; 01· the Ancient
h; the OE. h-rune obviously served as a model. Welsh Grammar. Llandoyery 1856, 10 f.
k: may be derived from a 'double calc-rune' or from J. LoTH, Vocabuloire vieu;r:../weton, 88. VV.

an x as found in various manuscripts ( but cf. x.

l: an 1 turned left and with a second lateral stroke added?
*); C, Zrross, Gro:nl'llJatica Celtiea, 1059 f.;

H. ZIMMlIR, NmrrJm Vindicatw. tlber EntstehJmg, Geschichte und QueUen

der: Historio. Brittonum. Berlin, 1893, 131 1£.

15 8 159
'"
 CONCLUSION.
(and Irish) influence there. As long as the ON. evidence has
not been sifted, no definite solution can be given to the problem,
It is dangerous to found conclusions on one. rune. Therefore but the results of the investigation may with some degree of
I do not stress this point, but I do hope that further research certainty be summarized as follows :
may help to clear it. For the time being I only want to call The isTUna tract is asystematic account of runic cryptography
attention to the environment of the isruna tract. In four of founded on a peculiarity of the OGmc. fupark, viz. the division
our manuscripts it is found in the neighbourhood of Greek of that alphabet into three groups of eight ~une8. In the later
material. In SG it is followed by a paradigm of the verb English tradition the new runes seem to have formed a group
yp&q,w and by Greek rhetorical terminology borrowed from
by themselves. Secret writing ,on this basis may have existed
Cassiodorus. In B it occurs at the end of extracts from Cassiodorus at an early date, but it was probably developed on the model
copiously sprinkled with Greek words, and not far from a of the Old Irish ogham and its cryptographic variants. The
cryptogram based upon the Greek numerical system. In U it systems described in the tract may have been invented in
is found on a page mainly devoted to Greek alphabets and England (for the haluzlTUna we have convincing evidence in the
numerals, whilst its inclusion in S was probably due to the Rackness inscription), but the names of the last three are at
presence of the regula formatarum. Even if we no longer least translated into ORG., and there can be no doubt that the
believe that • who says Greek, says Scotti ~, it remains true that text in its present form was written on the Continent. St. Gall
the Irish diaspora played an important part in the diffusion of seems to have played an important part in the diffusion, but
Greek lore (I). At the same time .these Irishmen' displayed it is not possible to decide whether the tract was first written
great interest in all sorts of esoteric lore and especially in there, as the internal evidence is contradictory. At any rate
cryptography. Their native system of writing itself had a Irishmen may have' had a hand in circulating the text, and
there is some evidence that the tract reached the Continent via "'~'
marked cryptographic character. ".
One might now ask. whether the English themselves had no Wales. At an early date a somewhat modernized version:'
share in the elaboration and the drculation of runic crypto­ reached the lower Rhine area (Braunweiler, Trier ?), and from
graphy. The Rackness inscription proves that they probably there a new edition came to Salzburg in the tenth century.
played a part, but our tract contains few traces of their activity. The text describes four types of cryptography and one of secret
They no doubt furnished the fuporc., The<form in which it signalling. This last is found only in the St. Gall and the
came down to us, however, points to' a Continental centre, and Brussels versions; it remained in use at St. Gall for at least
one where direct English influence was not very strong, or two centuries. It does not seem to be based on the fuporc, but
where it was on the decrease. The man to whom we owe the rather on the Latin alphabet. The four other devices first
runes in their present form had no doubt Jittle contact with a indicate the group to which the rune belongs, then its place

living tradition. Whether some systematicaccoUJ;lt of runic

cIJP.tography existed in England we do not know. The names
.
in that group .

isruna and lagoruna go back to an English'source, but those of

ApPENDIX I.
the three other devices were at least translated into ORG.
The appearance of similar systems in Icelandic manuscripts is
A survey of Mediaeval cryptographyIies entirely beyond the

somewhat obscure, but there are clear indications of English

scope of this work, but a few examples which have come to my

(I) W. 'I'RAUSB, 0 Ronuz Nobilis, 341. Cf. especially B. BJSCHOPF, Das notice must be mentioned here, as they show a close connexion

griechisclu Elemtm:t, passim.

with St. GaIl. I do not doubt that a careful search might

160 15
... 161
unearth other instances, and perhaps show a wider radiation able, if above it he had not written Chlophruna. This is, as
of thes.e nugae. far as I know, the only instance of the term clopfruna not attached
(a) St. Gall MS. 176 presents probably the most interesting to the isruna tract. 'But what Ekkehart wrote were actually
case. It is dated saec. IX by Bruckner (I). The fly-leaf at stopjruna • dot-runes', not clopjruna. The latter were by
the back of the volume has been pasted onto the binding. It was definition not a form of cryptography, but of cryptocheironomy.
used by later hands for various inscriptions; it has much suffered The distance between the two was of course not very grt<at :
from dampness and worms. The upper half contains a text the simplest way to ' write' a tap was a dot, and vice versa a
beginning. : Qu~ est latitudo? Bonorum operum usque ad dot, or a number of dots, could best be ' sounded' by tapping.
inimicos dilectos dilatatio, etc. Below the text, a drawing The second part of Ekkehart's term, -Tuna, calls also for a
representing Christ carrying the Cross (latitudo, longitudo, remark. His cryptology is based upon the Latin alphabet;
profunditas (?) inscribed on it), and a boy carrying a sig~ with· it has nothing to do with the runes. Hence, by the time when
the trilingual inscription; and short notes. Somewhat lower
Ekkehart wrote his jibes against Crimalt (probably after 1034 (I»
than half-way down the page a waved line marks off the lower
only the name clop/runa remained. From this we may probably
half, which has been filled with taunting verse on one Crimalt
infer that the development of some forms of non-runic crypto­
or Crimait, e. g. graphy was at least partly inspired by the example of the runes.
Hauserit hoc si quem Crimalt (2) ex uase liquorem Other instances (using only Roma,n figures) are found in
Peruigilem tussim suscit[e]t atque sitim. St. Gall MS. 899, p. 21 (Simi XI. VIllI. XVIIII. V. XVII. V.
The author has given his. name in a cryptogram : XVIIII. XVII. XlIII. XVII. XVIII. XX. XII. XI. V. VII. V. =

Crimalto (3) fratTUm facetiori, similiter retrorsum lege), Oxford MS. St. John's College 17
,.L't
followed by groups of dots arranged on a horizontal line. One (cf. p. 33), etc. Of course not all instances are due to imitation ....
need not even take the trouble· to count these dots, as the of cryptic runes : the device was too obvious not to arise in
writer has indicated the number in each group by suprascript different places and at different times.
Roman figures : Some influence of the isruna system is, however, probably
V X X V VIII I XVII XVIIII needed to explain the following two examples.
If we take the 5th, loth, etc. letters of the Roman alphabet, (b) At the end of the De inventiQne text in Vienna MS. 1761
we obtain the name Ekkehart. The handwriting has been (fol. 105 r ) there are several devices for cryptic writing (int. al.
identified as that of Ekkehart IV of St. Gall (ca. 980-1057/60) (4)· one with Roman figures instead of letters, of the type
This cryptogram concealing his name would not be so remark­ just mentioned). In one of these the alphabet is divided into
three groups of six letters each, and one of. five. The groups
are indicated by short vertical strokes, the place of the letter
(I) A. BRUCKNi!R, Scriptoria III, 79·
(2) Corrected from ' Crimolt '. in its group by longer vertical strokes (II = a, 111111111 = z).
(3) Corrected from 'Crimolto'. In fact this device is simply an adaptation of the isruna.. Since
(4) FiIst by H. HATl'BMl!Il, Denkmahle I, 412. Cf. E. ScHuLz. tIber die
Vienna MS. 1761 is also to be connected with the isTUna
Dkhtungm Ekke/w.Tti.ts IV. von St. Gallen, in: Corona quertrea. Festgabe
Karl Strecker zum 80. Geburtstage dargebracht (Schdften des Reicbsinstituts group for other reasons (cf. p. 302). and is closely related with
fUr liltere deutsche Gescbi.chtskunde). Leipzig, 194 1, 199-235. esp. 226. Vienna MS. 1609, which in its turn has a St. Gall element in
On EKK!!HART'S autograms cf. P. :LE:HMJINN, Autographe UM Originale 1Ul7tf.­
htifter lateinischer Schriftsteller. Zs. des Deutschen Vereins fUr Buchwesen (t) In that year Poppa became Abbot of St. Gall. . In the margin of MS.
und Schrifttum 3 (1920), 6-16, - ID., Erjorscht.mg des Mittelalurs, 359-3 81 , 176, p. Z98. Ekkebart scribbled a bitter remark on the nouitas PoppcrnU. i. e.
esp. 369 J. (with bibliography). Pappa's reform.

162 163
....
its ancestry, there can be no doubt about the origin of this reform was aimed into at. to suppress the leisurely play with
cryptic device. such schoolboy devices?
(c) Vatican MS. Regin. lat. 421 consists of a number of
fragments, some of which, according to Dom A. Wilmart, ApPENDIX II.
originated in 'St. Gall (I). Amongst the latter there is fo!' 26,
with the following contents : The manuscript evidence for runic cryptography in the
North has not yet been examined as a whole. Some instances.
(I) a' concordance of Greek letters A -'K with Roman nu­
e. g. Bodley MS. 572, have received but little attention. By
merals I XX.
far the most interesting text edited thus far is the following
(2) a majuscule alphabet, with above each letter a Roman extract from J6n Olafsson's Runologia (I);
numeral (A I, Z = XXIII [corrected from XXIII!]).
Eigi skal Runer rista, nema vel clda kynne. var hann klokr
(3) Exempla literarum. An alphabet a i, then ik ia ib ic pvi er hier ~u fyrst Stafrof huar mei! peir villa med sUmer
id ie is iz ih i8, k ka kb he, with above t,hese pairs of letters iraletur. Stafkarla letur. Pera letur. Punckta letur
the rest of the alphabet (k - z). The whole is of course oc d-Ietur. Eru pat ecki utan tim Stafer.sem er .a. e. i. O.U.
based on the Greek numeral system. 5 .Pessum StQfum skytr maar inn i Ordenn epter pvi sem
(4) Sic scribitur, followed by an example using Greek nu­ hentar. '
merals as implied in (3); the transcribed example reads : .Paer Mal-rona Stafer mei! sinum pydingum. huar af
adnexique globum zephyri freta kamma' secamat. aller Runa Stafer taka pyding af & raadningar. & huertt
amen. (2). ad audru.
(5) Item aliud genus scripture: from a II to z = 1111 11111 10 .Petta er ira letur :
-">,"<
(6) Example; transcribed: abc d e f g h i kim n 0 p q r stu x Y z p. riett

ferunt ophyr connexa kimba per liquida gazas. . x deb c zt k I h i n m p q p s r g y auf o. iral:

(7) Item genus scripture secundum numerum literarum, the Stafkarla letur: b b c d f f g h k kIm n p p q r s

same as (2), but with a minuscule alphabet. txxyzp

(8) An example using this system, with an interlinear trans-' a e i 0 u
cription: 15 Pera letur:[p with I, 2, 3. 4, 5 strokes]
lazare surge f uenite morti tfollo rapae!. [d with I, z, 3, 4. 5 strokes] a e IOU
This folio seems to have been written at St. Gall in the 11th
century; it mayh~ve passed through Goldast's hands.
There can be no doubt that the basic idea of the in'una
remained ,vivid in St. Gall for at least two centuries after. the
tract itself was written down in MS. 270. Perhaps Poppo's
(I) A. WUJIlART, Codices II, SI4 f. According to E. DtlM:MLBR, Zrn ,Tier­
.1 ~ ~ 11-> f
~ ~ ~ ~ A-:*) ~ ~1
.r r- r' * I r
f1 X t *
~
'.~
r
job,l, Z.f.d.A. 20 (1876),213. fols. 16-20 and 27-28 formerly belonged to St. FIG. 22
Gall MS. 899.
(z) This is 8 distorted version of the meani.ogless sentence used in teaching [* (preceded by:) Riett er oc so giort..
writing because it contained 'all the letters in the alphabet, cr. J. M. CLARK, ** (preceded by:) .Pi giora sumer so.] Aar fyrer a.
Tlul Abbq oj St GaU, gil. This line also occurs in Retcbenau manuscripts,
together with.,the example in no. 6, which obviously served similar purposeS. (I) I. LINDQUIST, En skitmlumdskrift fran Flatey, rr6 if,

164 .... 165

20 Biarkan fyrer b. Hnes6l fyrer C. Sttingit Tyr fyrer d. Stun­ lEtt, pa tvler sem bagall er) skalltt telia fyrst naucl i hagals
ginn is fyrer e. fe fyrer f. stungit Kaun fyrer g. Hagall lett. Pa hagal. kaun. reid. os. pus. ur. fe. fiQlgaztt so
fyrir h. 6-stunginn is fyrer i. (j:.stungit Kaun fyrir k. 50 nauder framm fra naud i Stafrofe unzt par kemr sem fe
LQgur fyrir 1. Madr fyrer m. N aud fyrer n. Os fyrir o. er. par skulu vera nauder atta. is skal koma fyrir sig i
Plastr biarkan fyrer p. rang-snuit kalUl fyrix q. reid fyrer r. peine rett. skal su rett oc $0 fara riett sem naudar rett
25 Sol fyrir s. Tjr fyrir t. Dr fyrer u. Yr fyrir y. Puss aufugtt (Eg meina rett framm) til pess par kemr sem Ur
fyrer p. Hier eru pri-deilur. er, par skulu vera isar .viij..
55 hialm-rUner Spelld-runer lin-rlUler kvist-runir: Vreng­
rUner Knapp-runer hals-rlUler belg-runir: Sol-runer. SeSl­

~'II'"
runer: Ein-hver-fingar vinstre Einhverfingar. Ad-hverfingar
Stungv.-runer: hemlur: Mid-hemlur. Fymsku runer. Ensku

~Jlr?' 60
runer: Gra:n-lendsku runer: Vardar-runir :
Hier epter ern IQtun-villur tvennar. er en fyrste Stafr

wI'fl" FIG. 23
fyrer f. oc fe. Ur. Puss. Os etc. Meire ..... Minne .....
pa er enn eitt letur. oc eru kallaclar hnack-villur.
prer ern so ristnar, ok ern prjar rettir fes rett hagals rett oc
Tyrs rett : fes lett... Hagals rett... Tyrs rett.. ..
{Fig. 23 (a)] pesse rett heiter fes rett.

65 'Hier eptir eru KlapRunir: Prer ern ok teknar af pd­

fyrst er Fe. ur. puss. Os. reid. Kaun.

deilum. Prer ern til gamans. ef tveir Menn talazt vid, ok

Aunnur heiter hagals rett [Fig. 23 (b)]

vitia eigi aclrer skilie. I>11 skal klaPP!l til hvepss orclz. eru~\
30 hagall. naud. is. aar. S61. pat prjar retter : fes rett. hagals ok tyrs-rett. Til fes rettar
pridia heiter tyrs rett [Fig. 23 (c)] , skal klappa iij. hQgg aull jafn fliot. enn til Stafa seima, so:,
hier er fyrst tyro biarkan. maclr. lQgr. yr. 70 sa skilie sem po talar vicl. enn varazt at taka ecki utan
Petta heita prideilur. hier ern Hester Stafer i Stai-rofe. einn Staf ur hveriu orcle en fyrsta, Po at so astandizt at
fes rett. hagals rett & tyrs rett. Ein skal K visl til Stais Heire mege hafa. & gior so orcl, & giQr pjer vel kunnigar
35 huers { huerre rett. enn so fiQlgazt [al: enn so fiolga Kvisler aliar retter & Qll orcl, so pu villizt ecki. & klappa alltid til
scm. Stafer, duo Eumplaria] Stafer sem Kvisler i rett hverre. hverrar rettar, sem pu parftt Staf ur-taka. fe ur puss Os
So [al; pa em 24 Stafrof. pau em QII af prfdeilwn. duo &empll.J 75 reid Kaun II hagall Naud is ar Sol II tyr biarkan. madr.
ern & Qll Stairof i fiolne [FjQlnir er Oitins heite, enn mun bier lQgr yr. Til hagals rettar skal klappa ij fljot hQgg. AlIt
merkia vil1uletur mQrs e(!r Rwna ,book einhverial. ok ern aull annat sem fyr seiger. Til Tyrs lettar skal sla .j. hogg fliott
40 tekinn af prideilum. Skulu i Qllum pessum Runum sva semhierergiortt [3/12/1 I/IJ: fe. ha. t. SofiolgaztStafersem
margar K visler i fyrer [deest in duobus Exemplaribus' Vocula Kvister. Pat ern Kvistereclr Ord sem til hregre handarveit.
i fyrir] sem i Pd-deilum. l!!; so prjar rettir. 80 Nu parfftu at haia .e klapp til hagals rettar, &: huxa
Half-deilur ern petta. oc eru tvrer rettir Naudar rett ok Ordin, partil at kemmr at is. nefn is. ok sting vid skiott. pa
,,'f,."":
is-rett. Naudar = : :~ ~ ........ " .... : I- " ~ .. J- ~ ~: ~ f,. er E stunginn is j Malrnnum fyrer e. Nu villtu hafa .d.
45 "',,1-"'''' i "l-"k': ,,1-- ... ; ~I- " '·11:111: 1111: 11111: klappa til tyrs rettar, & nefn tyr & still vid. pa er d.
: 11111111: Naudar lett fer aufugt, enn is-rett riett. stunginn tyr fyrer .d. Brestr pjer. nefn Orclin i fes rett, par
pat er so at skilia : Pu (Naud skal kama fyrir sig i hagals 85 til pu finnr Kaun, sting vid. pa er .g. stungit Kaun fyrer G.
167
.166 ....
 Ecki hefi ee sied fieire letur, enn po skil ee einginn to eight i-runes. There is no epigraphical evidence of this
minn gode lesare. system.
The 'Irish alphabet' (1. 10 if.) has the letters of the Latin The numerous devices enumerated next (55 if.) are more
alphabet in a completely disturbed order; yet some pairs are or less sophisticated ways of writing based on the. pri-deilttr
still found in the original or in reversed order: de, be, kI, hi, . system (x).
nm, qp, ST. I can see no connexion with the ogham order, The klappni.nir (65 if.) are perhaps the most interesting of all.
except perhaps that a, 'It and 0 are found at the end of the They ate also founded on the Prl-deiluT system. They' are·
. alphabet. used by pe~ple who want to converse without being understood
The stafkarla let'ltT (13 f.). the pera letuT (IS), the punekta by a third party. The groups are indicated by quick tapping,
letUT and the d-letur (16; cf. the names in 11. 3 and 4) are all the runes by slower tapping. To render the dotted runes,
. variants of the notae sancti Bonifatii, i. e. the cryptic system such as e (= dotted i) the signaller is supposed to 'prick
in which only the vowels are at first sight unintelligible. The quickly' (make a pricking or stabbing motion with the hand f),
'Stafkarla letuT and the punckta letuT are well known from after having tapped the corresponding undotted rune.
English and Continental manuscripts (cf.' p. 389); the other Consequently the klappniniT are doubtlessly a form of runic
two systems are only further developments. The pera letUT cryptography, which was not the case with the elopfruna. But
starts from the Latin abbreviation mark for per: p. This at the same time there is a strong suspicion that the whole
symbol is made to represent a, and for the other vowels one to extract quoted above is influenced by some text .like the isruna
four Strokes are added. The d-letuT may have been inspired .. tract, perhaps. an expanded version of it. The stafkarla let'ltT
by the crossed d: d. and the punckta let'ltT have their roots in insular traditions, and
The runie alphabet (17 if.) is of a late type, but curiously so have the pera letuT and the d-letuT. The occurrence of p . . "".
enough its p reminds one vividly of the original p, its q is in the runic alphabet is rather suspect, too; and cou.ld not the,""
simply such a. p turned to the left. This p is explained as half-deiluT be deri~ed from a model with groups of eight runes?
pwIT biaTkan, i. e. an opened b-rune, a term which applies The klappni.niT themselves show a degree of sophistication
to the open b which is used for p in later Norse inscriptions (I), which is no doubt late. Finally there are the iTa letUT or Irish
but not to the regular p of the treatise. letters; even if they show no connexion with the ogham system,
The system of cryptography which is described next, the their name probably points to the source of this and other
so-called prl-deil'ltT (25 if.) is not based on this alphabet nor on devices as well. Therefore I believe that the klappnmiT des­
an extended fup¥k, but on the sixteen-rune ON; fup¥k. The cribed here are a late adaptation of the system mentioned in
system is called pri-deiluT because it is based on the fuPllrk the isruna tract (or in a siInilar collection of cryptological de­
being divided into three groups of runes. The device used vices, not necessarily arranged into a formal treatise) to the
to illustrate this. system reminds one of Maeshowe XVIII, NorSe fuPllrk.
Mrelifell and Bodley MS. 572; cf. also a couple of symbols in
the Rodven inscription. By the side of the pri-deilUT we find
the half-deilur (43 if.) : here the fUPllrk is divided into two
groups of eight. runes each; the'runes of the former group are
represented by eight to one n-runes, those of the latter by one

(1) O. VON F'RJIlSEN. R _. .104. 141, ;141 (fig. 74). (1) cr. J. LILJBGRKN, Rufl-Llira, 52 f.

~68 169
....

 CHAPTER III.

RUNIC ALPHABETS (I)

By way of introduction a few words must be devoted to the

difficulties which scholars who wanted to transfer a fupark to
the order of the Latin alphabet had to face. At first sight the
problem they had to solve may appear to be quite simple. In
the cases which interest us here they set out from a fuporc of
twenty-eight or more runes, so their difficulties can hardly
have arisen from a lack of symbols (I). But we must not forget
that the two alphabets were not parallel. Whilst for some Latin
letters more than one runic equivalent could be found,. there
was none for two or three other letters; on the other hand a.
number of runes had no corresponding letter in the Latin
alphabet. The following comparative survey will' 'helpto .'!-.
discover the points where difficulties may have occurred (for, ..,
the sake of clearness I distinguish between vowels (includinlL.
diphthongs) and consonants, although transitions from one
group to another will occasionally have taken place) :
(a) Vowels. A twenty-eight rune fuporc had symbols for the
following vow~ls and diphthongs: a, re, e, ~a, i, 0, ce (e),
u, y; occasionally the runes 3, IJ and j were also interpreted
as vowels. This waS amply sufficient to render the six
vowel symbols of Latin (a, e, i, O,U, y). In fact the alpha­
betizer was obliged to make his choice between two signs
for a (a, 2,), two or three for·e (e, ea, re [> e]). two for 0
(0, and re whenever the etymological value. of the latter
was present in the alphabetizer's mind), two or' perhaps
three for i (i, j, IJ). Several circumstances (the basic
(I) There;is no evidence that any of the runic alphabets found in Continental
manuscripts goes back to native runes; the latter died out aome time before
the importation of English runes began, or at any rnte they never found their
way into scriptoria.

I7 I
...

 fuporc. the degree of adaptation to the alphabetizer's i:onsequently, subject to changes occurring in the language. If
own dialect, etc.) may have influenced his choice, and thus now the initial sound of a rune-name was affected by a linguistic
we are able to distinguish a good many independent change, the value of the rune itself changed :. 0 = *oilil- > e
attempts simply by checking on these critical points. epel in OK through i-umlaut. In High German territory a
(b) Consonants. Here the problem was less complicated. whole series of consonants, were affected by such changes
Hardly any doubt was possible for b, f, h, i. m; n. p, T, S. (d > t; t > z, zz; k :> ch, kh; g > k; b > pl. When the
But here again we can predict where we shall find diver­ English rune-name hg became tac, the value of the'rune
gences: at those points where the fuporc had more or would at the same time tend to become t; tier) in translation
fewer symbols than the alphabet. A twenty-four rune gave ziu, porn gave dvrn. We have already met traces of such
fupark had only c to take the place of c, k .and q. When adaptations in the chapters on the fuporcs. But when a scholar
the alphabetizer started from a fully developed fuporc, arranged such an adapted fuporc into an alphabet, he had to
the problem was much simpler: there he found symbols solve a dilemma: either to change the values of the runes
for k (calc) and q (cweoTil). In many cases, ho~ever, these together with the names, and thus to alter profoundly the whole
latest .additions had not yet been incorporated into the , structure; or to transfer only the names, tac e. g. becoming the
fuporc prototypes, as is shown by the almost desperate name of the t-rune., SiiIce the affectation of the OHG. con­
attempts to find runic substitutes for k and q. On the sonants varied from one region to another, and all attempts
other hand there were two or even three possibilities for g to translate the names were not carried through to the last
(g, g, perhaps also j, cf. the English spelling g). The consequences, a great variety of' alphabetizations was bound
dental group also had a surplus: three runes (d, t, lJ) to to arise.
fill two places (d, t). But there was no rune corresponding Another element in the problem were those instances where
to z (at least not in the English fuporcs)nor to x although the two alphabets did not agree at all (Lat. x, z, sometimes~
some fuporcs assign the latter value to the fifteenth rune; also y; runes w, j, g and the extra vowels). Some alphabetize~';
on the other hand there was no use for the runes w, j and g. will use the runes which were left to fill the gaps in the alphabet.
Thus we shall find ea for x, g for y, g for z, etc.
As a rule a few runes will have to be dropped in the process
As to the actual procedure f9110wed in shifting the runes
of alphabetization; some alphabets end with the frank admission:
from the fupark order to the alphabetic order, it is too early to
supeTsunt mae... give more than a few general indications. An important
Not all alphabetizers had before them an unaltered English preliminary step was, that each rune was provided with its
fuporc. On the Continent they may have started from proto­ value in Roman script;. for this must have. given rise to the idea,
types which had been adapted to the non-English language of especially' with people no longer acquainted with the meaning
their surroundings, and this leads us to a special type of dif­ of the fupark order, that this apparently meaningless sequence
ficulties encountered by Continental scholars. should be converted to the well-known alphabetic order. This
The runes were' in all probability known, and learned, by procedure may again have led to differentiations : the values
their names. Just I1S the letters of the Greek alphabet were of the runes were not aiways easily rendered in Roman script.
called alpha beta gamma etc., the runes Were called *fehu *UT­ In some cases one had to be content with approximations, and
*porn- etc. The main difference was that, while the Greek sometimes differences of opinion become apparent, especially
letter~names had become meaningless words, the rune-names in cases where the acrostic principle did not apply. Thus the
were mostly if not aU actual nouns used in the language and, rune x is found with the values x, 1 & x, ii, even y. Here
172 173
'"
again it was up to the alphabetizer to select what he believed Fleury manuscripts it came into the possession of the humanist
to be the nearest equivalents of the letters in the alphabet. Pierre Daniel (many notes in his hand),. and afterwards in that
We have also to keep in mind that the alphabetizers were not of Jacob Bongars, the French diplomat, historian and philologist.
all equally skillful in performing their task. It is even doubtful At his death his library came into the hands of the Strasbourg
whether they all attached equal importance to this S9rt of work : banker and jeweller Rene Gravisset, whose son Jacob gave it
much has been made of the careful and longwinded attempts to the municipality of Berne in 1628. The major part of this
of men such as· Alcuin and Hrabanus Maueus to adapt runic codex is still kept in Berne, but during its peregrinations part
lore brought over from England for German audiences. But of it was lost; Miss B. Boyer discovered a number of the missing
why should they-the highest representatives of Christian . folios in Paris (Bibliotheque Nationale, MS. lat. 7520, fo1s. 1-24).
The Berne part is bound in white vellum (1935). The parchment is of
education of their period-have wasted so much of their time 'uneven quality and thickness. Composition: fol. I, 264, 2S7, then 2-195
on such nugae? Should we not first think of minor scholars, (+ Paris MS. 7S20, fols. 1-24 = formerly 212-235) (I), formjng 26 gatherings :
pupils and copyists, who devoted a few moments of leisure in IV [2-9} + III [Io-IS} + 3 IV [16-39} + III [40-4S} + 10 IV [46-12SJ
classroom or scriptorium to these interesting if not very prof­ + II [I26~I29} + 7 IV [I30-18S} + III [186-I91} + II [I92-19SJ.
Traces of an old quire numbering are found on foi. 23 v 'C " 77 v ' K " 8Sv
itable trifles ?Thete may have been circumstances .where
• L " 93 v (1), H7 v (1), 14S v • R " IS3 v ' S " 16I v ' T " 169v • V', I77v
people of greater authority and science took the trouble to • X " 18Sv 'Y'. Fob. 257 and 264 were probably ·at one time the outer leaf
examine briefly runological questions (cf. Chapter IV), but this of a quatemion; in the binding their order was reversed (2).
will have been exceptional. Format ca. 300 X 178 min, written area ca. 215 x 135/140mmj one.column,
27 or 28 lines to die page (Js on fols. 257 and 264). The manuscript is written
If to this we add that many alphabetizers can have had only in a rare' Irish Continental' acript, dated variously from saec. VIII to IX/X (3).
incomplete and vague information on the essence of the runes, Allowing a safety margin on account of the rarity of the script, I believe saee.
it is not surprising that some alphabetizations will look utterly VIII/IX is as good a date as any, with perhaps the stress on VIII in view of ,0;.
the decennoVennal table for A. D. 779-797; there is a red dot by the side of 792, ",
clumsy. but it may be accidental. At least two scribes worked at the manuscript; ..
. there are some later additions. E. K. Rand and others have considered the '
THE MANUSCRIPTS AND THEIR RUNES.
W. M. LINDSAY, Beme :107 (Palaeographia latina II = St. Andrews Univer­
sity Publications XVI) Oxford, 1923, 61. if.
I. Berne, Stadt- und HQChschulbibliothek, MS. 207 G. MICHELI, L'enlvminure au: kaut mo:J/en-dge tit les injl.uences irlandaises.
59 f., 66, figs. 79, 81.
(saec.VIII/1X) (1).
B. B. BoYER, A Paris Fragment oj Codex Benrensis z07. Classi.ca1 Philology
32 (1937), 113-120; In., Insular Contributions to Medieval Literary Traditio ..
The oldest owner of this manuscript we can trace is the on the Continent. Ibid. 43 (1948), 32. •
Benedictine Abbey of Fleury (St. Benoit-sur-Loire): Hic est O. HOMllURGRR, Die ilI.wtriertert. Handschri/ten der Stadt- und Hochschul­
biblWthek Bern. I. Die oorkarolingisclum. rmd karolingischm Handschriftrm,
libeT sancti Benedicti floriacensis (fol. 138v). With many other fol. 19 if. (not yet published).
(I) From the very extensive bibliography on this manuscript only the items R. DRRoLBZ, Ogorn, ' Egyptian', • African' and ' Gothic' Alphabets.
which are most important for our subject CILIl be· mentioned here : (I) I follow the old ink foliation. There is a more recent pencil foliation
J. R. SlNNBN, Catalogru Codicum MSS. BihlWthecae Bernensis. Bemae. • i '.' 197', but no author seema to have accepted it. Hagen proposed to
MDCCLX, 424 if. . number as follows: I, II, l.i9S; Homburger: I, A, B, 2-19S.
H. HAGEN, Anecdota Helwtica. (G1'IIIIlllI. Lat. VII), XV-XXXI, 39 if. (2) W. M. LINDSAY, Beme 207,63, supposed that the title folio (i. e. fol. I)
H. HAGEN, Catalogus CodielUll Bernensimn (Bibliotheca Bcmgarsiana). Bernae. was misplaced by the binder: he seema to'imply that originally fols. 257 and
MDCCCL1!Omn, 2SS. 264 belonged to the initial folios, although they may also have belonged to
F. LoscH, Die Bemer Rwwralphabete. Germania 30 (I88S), 287 if. a lost part of the codex.
E. K. RAND, A Vade Mecum oj Liberal Culture in a Mamucript oj Fleury. (3) H. HAGEN, Catalogus, 2SS: co s. IX-X"; W. M. LINDSAY: "IX" (?);
Philological Quarterly I (1922), 268 if. O. HOMBURGER: "VIII ex. ".

174 ..., 175

 manuscript as a product of the Fleury scriptorium, and, as far as I know, no De barbarismis; I77v De Boloecismis; 178r De uitiis; 17SV
objection has been raised against this BSsumption. D«:: metaplasmis; DE PROSA: 179' DE METRIS; 180V De
E. K. Rand has very aptly .described the codex BS a " vade mecum of liberal fabuJa; 18 I r De historia, etc.).
culture" (x). In the following survey I supplement Hagen's enumeration of
the contents as far as needed for the discuasion of the runic material: Hagen's "excerpta grammatica" does not describe the
fol. I f - . : two ornamental pages, the second of which gives a list of the contents of fol. I6S t ff. very accurately. There we have in fact
contents. a transition towards the computistical items origin:Uly found at
:a6..r : in the right top the name ' Hangars '. the end of the manuscript (d. 257v, 264t-v) : lSI v De rethoTtca
Part of Ch. I of Bede's De f'atiO'l'ltl t8mponI.m : De loque1a digitorum,
ending on z6.j.V. et dialectica; IS7v DE PERIERMENIIS; I89t De mathematica;
26.+': a Greek alphabet; a Hebrew alphabet; three runic and pseudo­ De astronomia; IS9v De mundo; De forma mundi; De cae10
runic alphabets, groups of variant runes, a lilit of rune-names; eiusque nomine; I92v Item de nominibu$ stellarum quibus ex
Aethicus Ister's alphabet (further details d. itifra).
257': an ogham alphabet, followed by a syllabary in the same script
causis nomina acciperunt; 194 v CVRSVS LVNAE PER
(6a be bi 60 bu, ca ce ci co cu, etc.). Between and partly over these DVODECIM SIGNA.
lines a later hand wrote a Latin grammatical fragment in question­ The above survey shows clearly why a later binder shifted
and-answer form. Several mostly fragmentary alphabets, amongst
which one may recogni2e runes, • Egyptian' and 'Mrican'
fols. 257 and 264 to the beginning of the codex: in tills way
letters, etc. (2) most of the alphabetic material was made to form some sort
The rest of the page contains a number of scribbles. of a unit. A glance at fols. 264 v and 21 shows that the runes
:aS7': four lines from Optstianus Porphyrius (also occurring on fol. 8r'),
ornamenting the latter were chosen from the three alphabets
in a later hand.
IN CHRISTI NOMINE INCIPIT CYCLUS paschalis XVIIII on the former. Therefore our study has to start with fo1. 26 4 v.
SanctiapudBITHINIAM IN SINODO NICENA TEMPORIBUSI Fol. 264 must for some time have been the last leaf of the
estque exordium huius suptld scribti cycli DCCLXXVIIII annal manuscript, as it is much worn. At the bottom a loose strip
ro[m.••] anni sunt. In five columns, from A. D. 779 to 797.
INCIPIUNT de titu1is pll8calis aegyptiorum, etc. of parchment has carefully been replaced with paper, and the ,,;,
2': title page, in capitals and runes (d. iflfra). lower part of fol. 264t has been covered with transparent paper . .
2' - IIv , 13' - 17' Donati Br8 prima, de litteris, etc. de barbarismo. to prevent further damage. Fol. 264 v too has suffered. The'
1 I' - 12v : Semi qui dicitur tractstU8 de finalibus litteris (end missing).
17': An quaedam anonyma.
somewhat greasy parchment has not held the ink well : some
18': Commenturn in Donati artem minorem, de Iitteris, etc. de bar-' letters, especially the heavily drawn letters of the alphabets, have
barisrno etc. Juliani Toletsni, ut videtur, grammatici sive illius flaked off; the black ink is as a rule better kept than the red.
simillimi.
77v : Servii centimetrurn.

Moreover the lines were cut so deep that the parchment actually
80': Excerpts de panegyrico Porphyrii Optatiani.
broke off, which necessitated the repairs just mentioned. Of
81': Commenturn in Donati artem maiorem vel Juliani grammatici vel
course all this does not make for an easy reading (I).
eius simillimi.
101'; Donati ars maior de octo partibus orationis.
Fo1. 264 v shows the following arrangement:
112r : An grammatica anonymi. 11. 1- II : the end of the loquela digitorum.
127': TractstU8 de nominibus mobilibus. 11. 12-15: a Greek alphabet with the numerical values of the
127': An anonymi conpendaria.
letters.
130'; .Asperi arsgrammatica.
140': Sergii in Donati artem explanatio. ll. 16-20: a Hebrew alphabet with the names of the letters.
14Br : An grammatica Petri grammatici. 1. 21 : blank.
1681 : Isidori excerpts grammatica (177' De differentiis; De gloais;

" (I) See his paper with this title. (I) The excellent photographs which were made for mc at the Stadt- und
(2) I have examined this page in detail in my paper Ogam, ' EllJIPtimt " Hochschulbibliothek brought out a number of details hardly visible to the
, Africon • ami ' GothU • Alphobets. Cf. also infra. naked eye in the manuscript. .

176 r6
177
...
11. 22-27: three runic (or pseudo-runic) alphabets. As to the characters of these alphabets, those of the first
II. 28- 2 9: seven groups of three runes (or rune-like symbols). are certainly runes. The difficulties of the alphabetizing
30 -3 1 : a list of rune-names belonging to a runic alphabet, process have mostly been solved in a clever way. For a the
ll. 32 -34 : . the letters of Aethicus Ister's alphabet, with their compiler chose a, for k the calc-rune (k), with which a K in­
names. . dicating the value seems to have coalesced in the prototype
On the line immediately preceding the runic alphabets there (cf. infra); for ~ he took the ~-rune, for y, y, for z, ea. This
seem to be traces of two or three words; but if so, their irregular last choice occurs in almost a score of alphabets; a tentative
appearance rather points to probationes pennae. Apparently explanation has been offered on p. 130, and we shall return
the runes, like the other alphabets, did not receive any inscrip­ to this point after the whole ~aterial has been examined. The
tion. A couple of letters (XA ?) were scratched with a dry point, form of the q is obviously non-runic; one may best describe it
but seem to have no relation to the following runes. as a heart with a figure 4 written below it. This character
The runes are written in a firm hand, which gives them a may have been supplied by one of the fictitious alphabets
quite 'runic' appearance.' If some readings are uncertain, referred to before; or else it may be a transformed Q. As a
this is due to the ink having flaked off or to the parchment matter of fact the q's of the two following alphabets seem to be
having been damaged. (Yet for some runes the true form may capital Q's with some fanciful strokes added below.
be supplied from the ornamental page (2r ), where they were Consequently the fuporc used by the alphabetizer comprised
also used.) They are divided over four lines, each of which the additional runes a ea y and k (also j and perhaps st, cf.
fits in between two lines of the ruling. There are three al­ infra), but no q-rune. The absence of this last character in the
phabets, separated by a series of three to six dots placed in a prototype is hardly surprising, since even in complete fuporcs
vertical line. The first alphabet, written in red, comprises the it seems to have been invented ad hoc. The forms of the rune;:'
characters of the first line and the first two of the second line; are very well rendered; the distance between the underlyi~g
the second, in black, the rest of the second line and the first fuporc and the alphabet cannot be very great. Yet the following
two of the third line; the third, again in red, the rest of the alphabets (II, III) show that to the compiler the runic characters
third line. The runes of the first line have consequently were a category of style, a set of types interesting from the
received ample space, whilst those of the next two lines are artistic point of view, rather than a sort of writing connected
rather crowded in places. The fourth line of runes consists with his native tongue and. with the mythical pagan past of his
of seven groups of three runes each, also marked off with sets people (I). The values of the characters in the first three
of dots (five or six); they are alternately red (I, 3, 5, 7) and lines are not indicated, but there is a minuscule showing the
black (2, 4, 6). value above each character in IV.
The characters in II-IV show even more damage than those in
I~&hHM~XHI~IM+~~~~~inr~~
the first alphabet. In II the equivalents of r stu seem to have

~%1~M~f~H~{§~Xf~{~1T~~

been partly retraced at a later date. In III the ink of all letters
has flaked off mote or less completely; c, m and u have become

."tMl +mH t~*~( II~~~ >1~ ~:r1~

i i i ooo999Vvvtttiiiaa.?l.
almost illegible, i, k, l, r, t and z are also badly damaged. In
IV the last letter has practically disappeared, and besides parts

:m:~~~~t9 .tbkY~ ~~¢ri~ =t:~rsx~~~~tt1

(I) On fol. 81" th.enI is an inscription in AetbiCUll's letters (de partibus
orationU). In Valenciennes MS. S9· runes and AethiCUll's letters are used
FIG. 24 together in one inscription, d. p. 406.

17~ 179
of letters, the minuscule equivalents above the letters have be explained as runes. F. Losch (I) made an attempt to account
mostly become invisible. Yet, as in I, the parchment is usually for each single character, but even comparisons with the whole
less soiled where the ink came off, and so the strokes look corpus of TUnica manuscripta known in his days could not make
white on a darker ground. the runic nature of all these letters probable. There is no
doubt that a number of characters look perfectly like runes,
A number of characters in IV are certainly runes, viz. the but the values do not correspond, in II d = p, f = x or ea ( f),
first character in each triad : P = ea; in III d = j, e perhaps = st, m 1. Other characters
i: the first i is the English 1,). The reason is rather obvious; seem to be only slightly modified runes, e. g. a, b, n, t, u in III,
Judging by the acrostic principle, the compiler was fatally and also a number of letters in IV. The most plausible expla­
led to consider a rune, the name of which was ing, as a nation for these lists of pseudo-runes is, that the compiler
variant for i. ' needed a number of characters for the ornamental script on
0: the first character for 0 is the (2 variant we also find in fo1. 2r. Mter having written out the first alphabet, he simply
went on inventing new types which would allow him to compose
Oxford MS. ~t. John's College [7 and in Leyden MS.
Voss. lat. F 12 8; it also occurs in the Thames inscription. a cryptogram offering more difficulties than mere runes.
F. Losch already pointed out that on fot 2 r the scribe used
g: the first g is j, a type known especially from Continental especially characters from II; those of III .and IV come next,
manuscripts, but also found in the Westeremden inscrip­ whilst he seems to have avoided using those in I as much as
tions, etc. possible (2). But perhaps his inventive skill was of the same
t{: w found its place among the variants for U; since its value kind as Nemnivus's (cf. p. 157): what he apparently created
outside of England could only be expressed by u (or uu), "subito ex machinatione mentis suae" was often, consciously
Il.,

its function as a u-variant is what we should expect. or unconsciously, inspired by other alphabets. The runes ".
obviously played the main part in this respect, but other al­
t: for!J there were two possibilities: d or t. Our compiler
phabets too may have provided a number of characters. Thus'
chose the latter, which is not found in any other alphabet.
a number of letters in II and III remind one of the • Gothic',
i : in the second i-group the first character is probably the , Egyptian' and ' Mrican ' letters which were in all probability
English 3. A value i for this rune is not unparalleled : known in the immediate surroundings of our scribe. The
Vienna MS. 795 describes it as " i & h" and the name ih first three characters in II, e. g., show an extraordinary resem­
is found in all Continental fuporcs. The second letter blance to the first three letters of the 'Gothic' alphabet in
of this triad may be the usual g, which came to be classified Munich MS. 14436 (d. p. 254).
amongst the i-variants because in OE. g was used for /j/. It is quite likely that the scribe first wrote out these alphabets
a: the other candidate for the place of a 'in the alphabet, re, on some spare space at the end of the codex before using them
has found its place in the group of substitutes. for the ornamental title page to Donatus's Ars minor (fol. 2r).
The artistic skill displayed on that page has rightly been praised.
In alphabets II and III runic elements are very few. In III To the left there is a large initial I (214 mm high), with knot
the h is the variant type found in Cotton MS. Domiti,an A 9
and in a few alphabets. The last character of III may be a (I) F. LoscH, Die. Berner Ru7Ullflllphabek, 295: "Ich glaube, dasz unser
Schreiber nur Zeichen verschoben und vermischt, nicbt dasz er Formen
slightly modified form of ea, which took the place of:;: in the geindert habe ".
first alphabet as well. But the remaining letters can hardly (2) F. IAlSCH, Die Berner Ru7UIIflllphabek, 293.

180 18[
....
and plaitwork ornaments, and animal heads at the ends, a fine eighteen are taken from II, thirteen from III and IV each, and
example of insular workmanship. At the top it has an animal only five from' the purely runic alphabet. The total number
ornament made up of four birds. The text is written in large of runes (from II and IV) is. twelve. The inscription differs
ornamental capitals, partly filled with green, red or sepia. only in very few points from the models set down on fol. 264 v;
Under each of the first three text lines there is a line of runes yet these differences may be important. First of all, there is
(or pseudo-runes) repeating the text. These runes are in red the use of j for i in nomine and donatio As we saw, IV classifies
ink, except the last two lines, which are alternately red and it amongst the variants for g. In III, however, the i may well
black. The first and the second line of runes are separated be a modification of the same rune; it must therefore be added
from the following line of capitals by an ornamental strip to the fuporc used by the compiler. The u in urbis is probably
(animal knots). For the fourth line of the title, however,this a variant of the third u in IV: the right lateral stroke has
arrangement could not be followed, as the artist seems to have been omitted. Losch preferred another explanation: he
miscalculated his space. Therefore he shifted the capitals to believed that the last letter of gramm(a)tic(i) was mistaken,
the left side of the page, and wrote the corresponding runes by and that the form of u in urbis was that intended by the inventor
the side of the text instead of below. A glance at the set-up to render u. He was forced to take this position because he
shows that there can be no doubt about the artist's intention: read the fourth word of the inscription semi (= sanctissimi)
he plan'ned the page so as to include 'the runes from the very instead of sUmi summi); but the latter reading leaves no
beginning. The capital text of the title reads as follows : doubt, and so Losch's. proposal may safely be dropped. The
last letter of urbis is again problematic. It may bea variant
N NOMINE DEI SUMMI

I
of the s-rune, s.uch as we also find on the Bewcastle Cross; but
NCIPIT ARS why did the compiler not list it in II or III ? This and a couple
DONATI GRAMMATICI of trifling differences (the confusion of c and u just mentioned;
URBIS ROMAE the slightly modified d in donati, cf. that in II, and n in in, d._
the n in III) can hardly mean that the compiler of fol. 264 v
and the ' runes' :
was not responsible for fol. 2f as well. Small variations were

*tlX~*~l+~+r6S=~:=

bound to occur, even if the artist created the extra letters

himself. Therefore the appearance of a new s-rune in ftrhis
has hardly any importance from the runological point of view.

:rti~~~~1\{~ iIicipit ars =

The r in III consists of two strokes; there is no trace whatso­
ever of the ink of a third stroke having flaked off; but the r

h~~~~ *Y~'r~ ~ X'1-g:a:=~ici + of gramm(a)tic(i) has an additional vertical stroke and forms
a triangle.
Even if we must admit the possibility of his having used
non-runic models, the compiler's skill in creating new' runes'
Rt~~T
= urbis ro­
'if1Xh FlG. 25
mae. is remarkable; but it was probably not unique. No profound
analysis was required to establish the fundamental principles
of runic script, and on this basis new, runic characters could be
By checking this inscription with the alphabets and groups of formed if needed. A number of runes found in alphabets may
letters given above, one will find that of its forty-eight letters owe their origin to such a procedure. This possibility should
182 183
....
'.~"----
 J
!
I
i
I
also be kept in mind when we discuss the origin of the addi­ I
to High German phonology. The same probably
tional runes. holds for
The last line of runes on fol. 264 v is followed by a list of chen : German influence may be postulated the more readily
twenty-three rune-names. There can be no doubt that these since there are other and undeniable instances of such
are the names of the runes in alphabet I. Unfortunately the a phenomenon (cf. quello; net, ret).
reading is rather difficult and uncertain. Not only has the ink
come off in places, but in the act of ruling, the line on which egch: . one might perhaps try to analyse this complex as
the names are written was cut so deeply, that the parchment eg + eh, but the accumulation of gutturals is probably
afterwards came apart. A couple of wrinkles in the parchment a scribal affectation. OE. eh would normally appear
and its being soiled along the margin add to the difficulties. as ech in the orthography of Germanic words as prac­
The writing is obviously contemporary with the bulk of the tised in France (I). Our scribe seems to have had
codex, and from the same hand as the rest of fol. 264 : many little notion of the value of h; cf. ish.
ligatures, insular g (except in hog;/). The names read as gobo one of several German substantives derived from the
follows (cf. the runes on p. 178) : stem *gib- or *geb- may be at the. origin of this form
ach. berc. chen. dei. egch. fegc. gobo. hagil. ish. kalod. (gebo, geba, .giba) (2). We may also start from an
logo. man. net. os. per. quello. ret. sigH. ti. u[r] [..]uch. archaic OE. gebu : in the oldest OE. texts b also stands
Ulr. ear. for the bilabial spirant, which is spelled f after­
Several of these names correspond exactly· to English rune­ wards (3).
names: bere (i. e. the Northern equivalent of WS. beore), dei hogil : No such form is known from OE. name-lists: these
(a form usually claimed as specifically Kentish, but found in have hegl, ha3gl, ha3gel or, still closer to the Berne;'
other dialects as well (I», man, os, ii, ear (these are common form, hegil and Juegil (4); a form hogol also exists, bJt
OE. forms). In sigil the first i causes no difficulty: before is never found as a rune-name. The name hagit is
palatal g the vowel y became i at an early date (2). The i of an adaptation of hegilor ha3gil to Continental Germanic
the second syllable is probably an indication of an early borrow­ phonology or orthography (5). OHG. has normally
ing (3). The name of y, uir, may also preserve an archaic hogal.
spelling, cf. the oldest Bede manuscripts, the Corpus glossary, the final h can hardly have a phonetic value. In the
ish :
etc. (4). For the remaining names some measure of adaptation prototype it was perhaps meant to indicate the value
to Continental Germanic phonology and orthography, or, of the preceding A-rune, and in the course of copying
alternately, corruptions of various degrees, must be postulated:
found its way into the list of names.
ach : though final eh instead of e is a not unknown early OE.
spelling (5), I rather believe we have here an adaptation (I) F. KAUFFMANN, Obu althochdeutsche Orthographie. Gennania 37 ( 189 2),
243-264.
(I) E. KRUISINGA, (Review of R. MtILLI!R, Ober die Namen des 1Iordhllm­ In., Das keronische Glos$aT Ilnd seine SteUufig in der Geschichte der alt1wch­
brUcJum LibeT Vitae. 1901), Anglia Beiblatt 16 (19°5), 145 ff. tkutschen OrtJwgraphie. Z. f. d. Ph. 32 (1900), 145-174.
(2) E. SmvERS-K. BRUNNER, Altenglitche GromrtUltik. § 31 A. 2. (2) W. BRAUNE. Althochdeu.tsche Grammatik. §§ 30, a07. 2'1.'1..
(3) E. SmvERS-K. BRUNNER, Altenglische Grammotik, § 44 A. 6. (3) E. SI1M!RS-K. BRUNNER, Altmglische Grammatik, § 19:1 A. 4·
(4) E. SmvERS-K. BRUNNER, Altenglitche Grammotik. § 94 A. Cf. the dis­ (4) Vienna MS. 795: hmgil.; Cotton MS. Galba A a: hegil.
cussion of various views in H. Sm6M. Old English Personal Nomes. 145. (5) The form hagil is given by E. G. GRAFF, 8procMchatfl IV, 797 as occurring
(5) E. SII!VERS-K. BRUNNER. Altmglitche Grammotik, § 206 A. 9. in a Trier MS. of Heinrici summanrmt.

184 185
...

kalod: for this name too some sort of coalescence must be he who chose the former probably did not even know
assumed. We saw that the rune to which this name the latter, just as he had to invent a rune-like symbol
belongs had absorbed the letter. k which indicated to take the place of q in the alphabet.
its value. As far as kalo- goes, this might be a mistake u[r] : the final r is hardly visible, but there is no reason for

for kale; but then final d reqtains unexplained. There­ supposing that there could· have been another letter

fore I propose to start from an arrangement like this instead.

kalc id [el]uch : only the last" three letters of .this name are plainly

.th K legible. To the left of u there came a long letter;

judging from the space it occupied, this must have

where id would be an abbreviation for idem (or result been lor, less probably, d. It was preceded by one

from: id + = id est). In this connexion I must refer more letter, only the upper part of which is more or

the reader to the fupore in St. Gall MS. 878, whiclt less visible. In agreement with other sources I pro­

also has Roman K for k. The letters id would then pose to read eluch, cf. elux in the isruna fuporc, helueh

have been appended to the name, giving kalod, whilst in Vatican MS. Urbin. 290, helach in the De in'Ventione

the variant (rather than: iridicator of the value) alphabets. .

K became part of the rune. Such a syncretism is far ear : if any further proof is needed to convince us that the

from being unique; it will of course occur most easily alphabetizer used the ea-rune for z, this name may

in alphabets which occasionally abandon the· order be considered conclusive.

of the Latin alphabet and which are sufficiently strange ~~

to escape being corrected, e. g. Aethicus Ister's The alphabetization· reflected by this list completely agrees
alphabet (I). with that of the first alphabetic series (1) : a for a, k for k,
o for 0, x for x, ea for z. There can be no doubt that the names

logo: obviously a mistake for /ago.

belong to that alphabet :, only the insertion of the two ornamental

net, ret: although final t instead of d is not unknown in early OE. pseudo-runic alphabets and of the groups of variant runes

spelling (2), these forms are rather to be interpreted caused them to be separated from the corresponding runes.

as adaptations of ned and red (corresponding to WS. This probably indicates that the alphabetizer and the creator

nead/nied and ra'!d) to High German phonology. of the pseudo-runes were two different persons: the latter

per : this name has lost its final consonant, d or t; in view started from a ready-made runic alphabet and may hardly have

of net, ret, the original form was probably pert. . realized what sort of letters he was copying and imitating. This

also eXplains why amongst the rune-names there appear such

f[ludlo h~re we have a certain proof of German influence.

strange forms as egch and gobo: the scribe of Berne MS~ 207

The compiler needed a word with initial q and chose

probably had no direct knowledge of the runes nor of their

the first that occurred to him. For obvious reasons

names. There are unambiguous traces of Continental Ger­

quello cannot help, to elucidate the OE. name cweoril :

manic, and specifically High German influence. On the other

(r) The 2200 letter of this alphabet in Vienna MSS. r609. r76r and in a hand there seems to be no indication of Irish influence in the

number of other manuscripts dearly t:onsists of two letters which were originally
runic material.

independent; similarly the third letter in Niimoorg MS. r966.

(2) E. SIEVEIIB-K. BRUNNER, AltenglUche Grammatik. § 224 and A. 2. The origin of the seven groups of three characters in IV now
186 187
...
becomes clear too: the seven runes not used in the alphabet, indicated above most of them, it would hardly be possible to
i. e. Jj, <2, j, w, }l, 3, z, were probably given at the end of the identify more than half a dozen. As far as we can judge from
alphabet. The creator of the two fanciful alphabets also the letters indicating the values, these 5 1/2 lines of characters
shaped two new characters to act as variants for each superfluous must be contemporary with the rest of the manuscript. There
rune. The fuporc from which he started may be reconstructed are all in all 101 characters, for 93 or 94 of which Latin equi­
as follows: valents are given (I) :

(' t?·n ~ ~ RhX~ Nt I *s:~

for:f u tor c· 9 u h n i 9 i P
(I)
(2)
(3)
I
a
2 3 4 5
a dee n 0 p s rae
u x a f g him x s .
k m nos q q r 5 t x 5
6 7 8 9 10 II 12 13 14 IS 16 17 18 19 20
h kim n 0 p s
abc e f g
a b d f k q
(4) abc d e . m n 0 q r 5 t

't'~ l' ~ MP1~* H<Pr~~1rh

for: x s t b e m lid 0 a. d y Z k
(5)
(6)
u ( ?) z ( ?) e
p q nun
abc d e f
tau mes n samech
g h kim n 0

As it is impossible to divide them up into alphabets, the char­

FIG. 26 acters in each line are numbered from left to right.
On the whole this fuporc has a very 'epigraphical' appear­ There can be no doubt about the runic. origin of a number
ance; only two characters, the u- and the y-rune, show some of characters: (I) 2, 3, 5, 6, 7, 8, 13, IS, 16, 17, 18, 19; (5)
concession to manuscript habits. Notwithstanding the strange 10, 14, 17, 18. Moreover there are a few for which such an
forms of some rune-names, which were more subject to mal­ origin may be postulated with more or less certainty: (I) 4,
formations than the runes themselves, I believe this part of the 9, 14; (2) 1,2,4; (5) 12, 13, IS, 16, 19·
.~~
Berne material deserves to be trusted more than runologists 2 5 4 5 6 7 II 9 1.5 14 15 16 17 18 19
have felt justified to do. Their hesitation resulted from two
facts : they examined the names too much by themselves, and (1)ta:: ~qHe.ff,+n,~,rr, 5s,Flh,Xk.n,~m:tn,Xq~p
as far as the runes were considered, they made no clear dis­ 1 'l. 4 10 11 15 ,4 'S" 17 1& 19
tinction between runes and pseudo-runes.
As to the origin of the underlying fuporc, the names of the ( 2 )~, nu,Aa, (5) ~d, ~f, 'f9, Fih, Xk, fl.iX1m, Xn, Xo
runes offer very few clues. The original was probably quite FIG. 27

old, but it is hard to localize it even approximately. Kent may

In the first line no. 2 is doubtless a, and no. 3 d. No. 4
have some claims in view' of dei; the vowels in net and perhaps
might of course be interpreted as a Q-reek H, but it could as
also ret point to a non-West-Saxon origin; but on the whole
well be a slightly modified e. The next e-symbol is probably C2;
the orthography is too shaky to allow of definite conclusions.
its lateral strokes seeming to be curved upward, one might
The contrast between the runes on fols. 264 v and 2r, and the
interpret it as an 0, but both the value e and the fact that 0
material on fol. 2S7 r is very striking indeed. Whilst the first
actually follows as no. 7, are in favour of the former solution.
of these pages gives a well-ordered catalogue of letters, and
The value e calls fora rune-name esc (cf. Erfurt gl. 772 and
the second an example of their artistic use, fol. 2S7 r contains
Brussels MS. 93II-93I9). The form for n has nothing surpris­
an...unsightly jumble of characters. If the equivalents were not
ing: the . regular n could easily develop into the more
(*) In one point I have anticipated the conclusiona of this study, viz. in
identifying j and j; (I) Cf. ~e plate in my Ogam, • Egyptian " • African' and' Gothic' Alphabets.
188 189
""
 +
~M~Mr~9 NX~f4~~~r::<n~
• natural' or elementary forms or x. NO.7 is a somewhat
clumsily drawn 0, no. 8 a rather badly distorted p. I have not
hesitated to connect no. 9 with s, especially because similar h kim n
for: a .d e e f p s u )' 0 0
forms of s are found in a number of other manuscripts (e. g. FIG. 28
Cotton MS. Titus D 18). No. 13 is a hardly modified form of These runes can hardly have belonged to one alphabet, as
h, no. IS of I, no. 17 of n, no. 19 of p. The other runes in there are two symbols for e and 0, but none for a whole .series
this line show more important alterations: in no. 16 (m) the of other letters. Perhaps this fragmentary list may be con­
strokes between the vertical shafts .do not intersect as in a nor­ sidered as a runic alphabet in nucleo : the outcome of an attempt
mal m; in no. 18 (0) the upper ends do not meet to form the to arrange the runes in an alphabetic order,' which did not
closed loop of reo The value 0 for this rune must be due to the succeed because the alphabetizer had too scanty a kno,wledge
adaptation of the English name to Continental Germanic pho­ of the runes.. At any rate these runes seem to have nothing
nology (CEpel > odil). The form of the second p (no. 19) to do with those on fols. ~64 v and 2rj they can hardly be due
is easily explained if we remember that the two hook-like lateral to the same compiler.
strokes sometimes met to form x (Niirnberg MS. 1966, Paris Before pronouncing a final judgment on the whole collection,
MS. 5239); a second vertical stroke was added in one other al­ a few words must be said about its non-runic elements. I have
phabet: the 'Alanic' of Munich MS. 1#36. As to the made an attempt elsewhere to trace these elements, and I believe
character k (no. 14), the most plausible explanation would be to have succeeded in identifying many of them as belonging
to connect it with the g-rune. The use of a rune of the voiced to spurious 'Egyptian' and 'Mrican' alphabets (1). Thus
guttural type for k is found in a number of alphabets. Usually our manuscript is connected with a group of alphabet collections
it is g or j .. The g-rune may have been transferred in the which seems to have been compiled at .least as early as the~'
same way. This transfer probably implies High German eighth century, and which was certainly known in Northeni
influence. France. This fits ,in very well with the supposed origin of the
The first character of the second line may be a rune (x ?), Berne manuscript. Of the remaining letters, a few are taken
but since its value has become illegible its origin can hardly be from a Hebrew alphabet, and one is an ogharn symbol.
ascertained. No.2 is probably u, no. 4 y (y is the value we The collection on fol. 257r , however, is a disorderlyaccumu­
should expect in that position, not a).. lation, whilst the related collections give the letters in neatly
In (5) no. 10 is the same d as that in (I). No. 14, h, is again arranged columns, usually with a more or less appropriate
clear, and so are nos. 17 (m) and IS (n). The form of the inscription. In fact not one alphabet in the Berne collection
f-symbol (no. 12) may be derived f~om thef-rune; less probable, is complete, only one approaching this ideal. Bits of alphabet
but not to be rejected a priori, is the identification of no. 13 sequences may be disentangled from the rest, but they hardly
with the j-rune, as its English name (ge(a)r) would lay near its ever comprise more than. six or seven characters. The whole
use for g. For k we find the same character as in line (I). seems to be copied at random from an exemplar which may have
For no. 16 I see no other explanation than to connect it with 1, b.een in poor condition. The scribe had apparently less under­
even though the shape of the letter rather reminds one of a standing for these alphabets than for the ogham material in the
minuscule s. The last letter of this line may be a distorted reo upper third part of the page. The first, partly erased line
Consequently the following runes· (in normalized forms) contains the ogham symbols arranged in the order of the al­
may be gleaned with more or less certainty from the collection
of characters : (J) Ogam, ' Egyptian', • Africtm' tmtl ' GothK' Alplwbets.

190 19 1
"
phabet, with the diphthongs following after z. The next six here: I (1-2) + III (3-8). A paper leaf has been inserted between I and 2,
where at least one folio is missing. Of fol. 7 the outer margin and most of
lines seem to be' an ogham syllabary, perhaps a key to crypto­ the outer column has been cut away; of fol. 8 only a strip about one third ot
graphy. It consists of groups of five vowel symbols, each the original width remains; it was formerly pasted onto the binding. The order
preceded by a consonant symbol: ba be bi bo bu, ca ce ci co cu, of the folios is probably disturbed, as the contents of 2v-3' and 5v-6' seems
etc. Only an Irish scribe could have possessed such an accurate to belong together.
The format is max. 320 X 250 mm (written area ca. 250 X 192 mm); two
knowledge of ogham script, and his nationality probably explains columns (each 75-100 mm wide, with 15-18 mm space in between); 3:1. or
at the same time the poor state of the runic material. 33 lines to the page. In this short fragment a great many hands appear (seven
It is hardly possible, then, that the runes on fol. 257r should or eight ?). one or two of which still show uncial features (a, j, m). The hand
responsible for the runes uses open g, and a consisting of two c's.
be due to the same compiler as the well-ordered and on the
Contents (I) :
whole very trustworthy alphabet of fo1. 264v. In the latter fo1. I r Fragmentum Synodi Romanae III sub Symmacho papa. an.501.
the tie has the equivalent a, in the former e; the latter uses the Fragmentum epistulae Iustini imp. sd Hormisdam papsm.
2 r Fragmentum commentarii !sidori Hisp. in Test. Vet.: in Levit.
k-rune for k. the former the g-runej and they have different
c. 21. Variorum condliorum decreta de accusatione episcoporum,
types of j and re. But if the runological value of fol. 257 r is presbyterorum vel diaconorum. .
small, it allows us to get a glimpse of an alphabet collector's 2 v-3 r, 5T-6r Frsgmentum Edicti s. Donationis Constantini imp.
workshop. It also shows how broad a basis is needed for 3v Concilii Nicaeni decretum de epistulis formatis una cum Iitterarur:n
computo et litteris nwneros exprimentibus [with a runic alphabet].
disentangling such alphabetic material. 4 r Civitates metropolitanae in provinciis Gallicanis.
4 v Nomina omnium provinciarum Romanorum.

6v
Divisio orbis terrarum Theodosiana.
7' [Fragment of a poem in a clumsy hand:
2. Leyden, Universiteiubibliotheek, Cod. Vossianus lat. F. I2 8 ... in .mense carsmen lire resonemus
(saec. IX). odens uirum inclitue cot de uoce ci modulemus almen, etc. ~~
On fol. 8' the same hand wrote ( maim.bertuse scric' and (maim~ ,
This manuscript is a membrum disjectum presumably written bert '. Fol. 7·-8 f were used by several other hands for probationes
pennae].
in the Abbey of Fleury (St. Benoit-sur-Loire) (1). So little
The first column of fo1. 3 v contains the text of the Nicaean
is left of the codex that it is hard to characterize it. It may have
decree on the litterae jormatae printed by L. Muller (2). This
been a collection of canones, as this could very well include the
regula was normally followed by a Greek alphabet with the
regula jormatarum (fo1. 3v ; cf. Salzburg MS. a IX 32 , p. IIS)·
numerical value of the letters. In the Leyden manuscript the
The manuscript now consists of eight folios in a modem binding with
vellum back and comers. The parclunent is rather rough and heavy; the Greek alphabet is written in the middle of the second column
lower margin has suffered from moisture. The eight folios are bound in of fot 3v; it is preceded, at the top of the column, by the runic
two unmarked quires and are numbered 35-42; the table of contents inserted alphabet. The runes are distributed over four lines : (I) a-g.
at the beginning follows a new numbering, I to 8, and this we have also adopted
(2)h n, (3) re - t,. (4) U 1;. The runes are apparently
(I) According to the catalogue of 1716 the manuscript (or one of the texts
formerly bound up with it?) once belonged to Pierre Daniel, the Orleans (I) As given on a paper folio inserted at the beginning, with additions of my
humanist, who acquired many manuscripts from the Abbey of Fleury. An own. The contents of the codex with which this fragment was bound up will
old ex-libris on fol. 7' has been partly erased : LlBER SANCTI (...]IS COE­ be found in the catalogue of 1716 (p. 368).
NOBII. SI QVIS EVM I FVRATVS FVERIT. DAMPNATIONEM (a) L. MOLl.J!R, VI!TSUS Scoti cuiusdam de alphabeto. Rheinisches Museum
ACCIPIAT CVM EIS QVI DOMINO DEO RECEDE A NOBIS. FIAT. 20 (1865); 363 f. His text is based on our manuscript and on MS. Vossianus
A )_( H )-. An attempt to read or photograph the missing word(s) under lat. Q. 33. Cf. also
ultra-violet light proved unsuccessful. About 16-20 letters may have been C. FABRICIUS, Die LittI!TOi! Formatae im FriJhmittelalter. Archiv fUr Ur­
erased, so [BENEDICTI FLORIACENS]IS may have been the original kundenforschung 9 (1926), 39-86, 168-194 (text on p. 39), and Dictionnaire
reading. d'archtlologU chretimne et de liturgie IX, :I.. 1571-6 (text col. 1574 f.).

19 2 17 ~93
 drawn by the same elegant, regular, somewhat sophisticated that which affected the shapes of c and d. The explanation
hand which wrote the regula and aIm the texts on fol. 4r -v and of e, however, is also closely connected with that of the following
part of Sr. They are written on what seems to be an erasure. rune. In the place of f we find a symbol which may be a
There probably is a connexion between this erasure and the Greek .p, or else the English j. If the values have really been
. fact that the Greek alphabet (which we expect in the place shifted one place to the right (by e being qropped), then this
taken by the runes) is found only lower on the page. Above rune actually stands for g. This is hardly surprizing, sinceOE.
each rune its value is indicated by a letter of the alphabet. spelling usual1y rendered I j I by g (ge, gil. Thus a non-English
The runes are carefully drawn; so carefully indeed that' one scholar could easily be led to interpret this rune as a variant
is surprised to find a number of puzzling or distorted forms for g. This explanation is more probable than that which
amongst them. Here follow the forms as given by the manu­ starts from Greek .p : St. Gall MS. 270 also gives j as a variant
script: for g, and the same is implied by the alphabet in Arsenal MS.
abc. d eo f 9 h I' ~ 1m II69·

~BJ ~y¢gH I"'¥t!><l

The symbol under g goes back to the type of g found in the

iSTUna fuporc. The broken line which there reminded of an
s-rune has become a regular capital S. 'The occurrence of
no p. q r stu 'x .Y 1­
this type in a manuscript coming from a centre with obvious

i-+ )f:WR. ~f1VW'fN FIG. 29

Irish connexions may be a precious hint: it supports the con­

clusions reached at the end of Chapter II. Instead of the
h-rune we find a Roman capital H; a symbol which looks very
much like h will tum up under :t, cf. infra. The monstrous ",
formation which takes the place of k seems to be derived from
...
The runes a, b, i, 1, nand r require no lengthy comment. a form which is either the English x-rune, or a k turned upside
The use of the ac-rune for a was obvious; it is also found in down; the latter is the more probable, but there is an additional
other alphabets. As is the case with the other five runes of difficulty : the same form turns up at the end of the alphabet
this group, its form is very well preserved. The forms of the with the value y. To start from an ON. k is hardly to be
runes for c and d seem to imply that at some previous stage recommended, as there are no other traces of ON. influence.
of this tradition the lower ends of these two runes had become There is a third solution: in some alphabets we find a symbol
invisible. The d of our alphabet is of course the rune lJ. In for q (e. g. in the De inventione alphabet) consisting of a vertical
fact several other anomalies may also be due to a defective stem with a lateral stroke curving up to the left. This too
prototype. The rune which takes the place of e might be the could. be at the origin of our k. On the whole, however, it'
English x (eolhx), perhaps even ea. But further inspection will be safest to start from a k turned upside down, even if that
shows that the original f is not found elsewhere in the alphabet. same symbol is used for y, cf. tnfra.
Therefore it will be safer to explain the e of our alphabet as The runes for m and d are often hard to distinguish, and in
an I-rune (the vertical stem of which has been broken); it was many instances we may be sure that the scribe did not know
then shifted' one place forward. We can only guess. why this how' to differentiate the two. In the Leyden alphabet they
change was introduced. It may have been because the alpha­ have changed. places : .the d-rune is found in the place of m,
betizer interpreted the real e as an m-rune; or else the e of the whilst the m-rune is found in that of t (cf. infra). For 0 we
exemplar had become illegible owing to the same damage as find the rare variety known from Oxford MS. St. John's Col­
194 ... 19S
 ea > x), one or two accidental transfers (£ > e; also j (g) >
lege 17 and from the Thames scramasax. The form of p is
/1); two more shifts are probably based on linguistic consider­
again puzzling: the vertical stroke is probably spurious, the
ations (lJ > d, d > t). It is quite possible that the alphabet
rest may go back to a regular p-rune (cf. some forms of p on
in its present fOr:n:l has a long history behind it; that would
p. 360). The q is not unique in runic alphabets (cf. Vatican
explain the insertion of H and V, and the transfer of the h-rune
MS. Regin. 338), but it is certainly of non-runic origin. In to the last place, changes apparently due to a • corrector'. With
all probability it was borrowed from the • Chaldaeo-Assyrian' its many additions and shifts the Leyden alphabet may be
alphabet found in several alphabet collections. Such collections
termed a poor alphabetization. The symbol for q indicates
were circulating in France; the occurrence of a Chaldaeo­
that the alphabet once belonged to a collection such as that of
Assyrian q in our alphabet has nothing surprising, if the Leyden Vatican MS. Regin. 338. .There are, however, no indications
manuscript actually originated in Fleury. The s is the reverse of a closer relationship between the Leyden and the Vatican
of the normal form, but such a change is rather obvious. For t runic alphabets; therefore their choice of the same q may
we find the English d, with a spurious h9rizontal stroke at the
be a mere coincidence. It only indicates that the two alpha­
top. It implies that the alphabet passed through the hands
betizers had a similar collection of strange alphabets at hand,
of a High German intermediary. This cannot be an argument
and that they borrowed the same type of q from it. As the
against the alphabet originating in Fleury, as Old High German two manuscripts are supposed to have originated in France.
may not have been unknown there (I). The symbol for u may such a coincidence is not altogether surprising.
be a Roman capital V, or, less probably, a reversed u-rune.
The use of ea for x is also found in the Oxford manuscript 3. Vienna, Nationalbibliothek, MS. TSI (saec. IX med.).
and in Vienna MS. I761, but far more often this rune takes the
place of 11 (in the latter case a variant form of the x-rune takes This codex consists of several fairly contemporary manu....,
the place of x). The choice of the regular x for y was probably scripts or parts of manuscripts. In the first of these we find:~'
influenced by the similarity of that rune to a capital Y. It list of rune-names, but not the characters themselves. Only
is found especially in those alphabets which have left their this first part will be examined in some detail; for the others
fuporc prototype a long way behind them. As to 11, the only the briefest mention will have to do, as their connexion with
plausible explanation I can see is, that the h-rune was used to the first part is purely accidental (I).
fill this gap; h had in fact become superfluous when it was (I) Only the more important bibliographical items on this much studied
manuscript can be given here :
forced out by a Roman capital H. W. GRIMM, Ueber deuUche Runen, 106; ZUT Litteratur Jer Rllnen, 1 Kl.
Consequently the Leyden alphabet contains at least two and Schriften III, 85.
probably three letters of non-runic origin (h, q, u); two are J. A. GII.J!S, Sancti Aldhelmi Opera. Oxford, 1f44, 103 ff. J. P. Migne,
of uncertain origin (k, p); two show a radical transfer (h > z, Patr. lat. LXXXIX, Z99 ff.
P. J!\FPti, Monumenta Mogunti1Ul (Bibliotheca Rerum Gennanicarum III).
Berlin, 1866, I I ff., Z44.
(1) W. BI!TZ, Deutsch und Latei"uch, 100; " Es mag sein, dasz Fleury, das G. STEPHENS, Mtmuments III, 14, no. 84.
Tochterkloster Clunys, das fUr die ags. Benediktinerrefonn 80 viel bedeutete, H. HAHN, Forschungen lI"UT deutschen Geschichte IS (18'5),9' :If.; ZI (1881).
such fUr die tlbernrittlung der ahd. Wolter ins Ags. eine wichtige Rolle 385 :If.
gespielt hat. Denn Althocbdeutsches gab es in Fleury sieher : Pa (= Pariser E. DtlMMLER, S. BoniJatii et LuUi Epistolae, in ; Epistolai! MerOfl1ingici et
Glossen) war vielleicht eine Zeidang dort, die eine Handscbrift des St. Emme­ Karolini aevi. I (= Mon. Gerni. rust., Epist. III), 383 :If.
ramer Gebets stammt aus Fleury, und die ahd. Gespriche kommen ebenfaUs W. DIEKAMP, Dk Wiener HandschriJt deY BoniJatilu-Brieje. Neues Archiv 9
aus Fleury, wie mir Dietrich Gerhardt mitteilte n. Contacts may also have (1883), 11-z8.
existed at an earlier date; at any rate Fleury was much older than Cluny (which , J. H. GALLID!, Altsaechmche Sprachdenkmaeler, z05 ff. and pI. VI.
was founded in 9l0).
197
196
 Parchment binding of the 18th century, with the imperial eagle stamped
in gold on front and back; inacriptiona • E. A. B. C. V. ' (= Ex Augustissima
Ji. Hahn, E. Ehwald and M. Tangl. According to this last
Bibliotheca Caeaarea Vindobonenai) and ' 17 • G. L. B. V. S. B. SS ' (= Ge­ author it was compiled in the following successive steps:
raIdus Liber Barn Van Swieten Bibliothecarius 17SS). The codex oonaists (I) Collectio pontificia, the correspondence of Boniface and
of 188 folios; fols. 1-77 ate well preserved, 78-188 have suffered from moisture.
E. Steinmeyer (I) distinguished six parts: A = fols. 1-77; B = 78- 128; Popes Gregory II, Gregory III and Zacharias. The
C = 129-162; D = 163-166; E = 167-172; F = 173-188. A consists collection consists practically only of letters addressed to
of ten unmarked gatherings; S IV [I uncounted folio + 1-39} + IV (- I) Boniface.
[40-46} + 3IV [47-70} + III (- I) [7I-7S} + 2 single leaves.
Format 300 X 19S rom; written area ca. 230 X 140 mm; one column, (2) Collectio communis, a first collection of letters not contained
31 lines to the page (B: 29 lines to the page, C : 3Si the format also varies in the papal collection. After the death of Lull, Boniface's
in the later parts: DEF = 270 X 197 mm, written area 203 X 142 mm; successor to the see of Mainz, many items were added to
B has two columns to the page).
this collection. The rune-names are found immediately
Contents (2) :
after the last item of this section.
fo1. It Letten of Boniface and othen (cf. infra).
39V into al. a list of rune-names. (3) Collectio Lulli. About the middle of the ninth century,
78r Acts of the Apostles; James; Peter I. the compilator of the Vienna manuscript added 43 more
129r Explanations of Biblical passages from Kings I. 23,3 to Acts 27, 17,
with OHG. glosses.
letters, mostly relating to Lull, some also to Boniface.
145" Latemthu noUlTum, related to the notal! Papianae et Eituidlemes He probably copied these from originals or concepts in
(cf. p. 288 f.). the Mainz archives (1).
162r .12th, 13th and 14th apocryphal letten of St. Paul and Seneca.
163' Augustine, Homilies. Since the originals of a number of letters could only be found
1661' A note.on Willibert's election to the Archbishopric of Cologoe (870). in Mainz, part A of our codex was probably written there,
167' Theodulphus, Capitula ad presbyteros.
173' Brevis adnotatio capitulorum in· qnibus constitutiones conciliorum
possibly at the order of Hrabanus Maurus (archbishop of '.
mogontiacense & remis & cabillione & turonis & ateiato gesto~ Mainz 847-856). This agrees well with the date of the hand.. .,
concordant. writing which is placed in the ninth century by G. H. Pertz,
1871' A fragment of Bede's Martyrology. W. Diekamp, etc. Other authors have dated the codex in the
1881' OHG. charms. .
tenth century, but judging from the handwriting, this must be
Part A is one of three manuscripts containing important collec­ considered too late.
tions of Boniface's letters (the others : MUnich MS. lat. 8Il2 The last item of the collectio communis had for some· time
and Karlsruhe MS. Rastatt 22). This collection has been been known as "Epistola anonymi ad sororem anonymam ",
examined at length by P. Jaffe, E. Diimmler, W. Diekamp, when Hahn and Tang1 showed that it was addressed by Lull to an
abbess and a nun whose names are not given; it must be dated
E. ST.I!INMEYBR - E. SIlMllIS. AlthocluleutscM Glossm IV, 636.
ca. 738 (2). The letter closes on fo1. 39 v (11. 19-22)
E. EHwALD, AldMl:mi Epistoloe. (Mon. Genn. hist., Script. Ant. XV), S19·
M. TANGL, Die Briefe des h. Btmifatius tmd Ltdlus (Epistolae aelectae in memo/res perenniter estote meique laboris sudori
usum scholarum ex· Monumentis Gennanise hiatoricis separatim editae I). puris/sirois orationum uestrarum precatibus fauere /
Berlin, 1916, XI fl. suppliciter posco dignemini.
M. TANOL, Studien rmr NtrUIlUSga1J6 tIer Bonijatitu-Brie/s. (I. Tell). Neues
Archiv 40 (19 16), 639-790·
On 1. 23 f. there follows a curious cryptogram: the palindrome
H. J. H:ImMANN, Frilhmittelalterliche Hanilschriftm, 122 fr.
G. BABSIICKB, Vocabulorius, 97. (1) M. TANGL, StJulien ;mr Neumugabe der Btmifatius-Bril!fe, 662 fr.
(I) E. STBINMBYI!R-E. SIJmllIS, AlthochtkutscM Glossm IV, 636. (2) J. A. GILES, Sandi. Aldhelmi Opera, 103 fl., included this letter amongst
(2) E. S~"l!R-E. SIJmllIS, AlthocJuleutscM Glossm IV, 636. A1dhehri's works; hence the title of O. B. SCHLUTT1!R'S paper; Aldhelm's
H. J. HBRMANN, FrilkmittaaIterliche Handschriftm, 122. Runic Alphabet, etc.
1 98 1 99
""
METRO HOC ANGIS SITIS SIGNA COHORTEM is It is not possible to decide whether the exemplar copied by
written from left to right and immediately below again from the compiler of Vienna MS. 751 contained these nugae, or
right to left and upside down; moreover the letters of the first whether he added them from some other source. The latter
line have been connected with those of the second. explanation is rather tempting, since fol. 39 v is the last page of a
The names of the runes are given next (1. 25 f.) On account quire i the few lines of space which were left blank after the
of a hole in the parchment (there are four in this one folio) the letter was completed may have been thought fit for this sort of
name hagal had to be written ha-gal; there is a larger space stopgap. On the other hand there is some proof that Boniface
between ilc (= x) and ian y) for the same reason. Ll. 27-32 an,d his circle had a taste for cryptography. As we shall see,
(the last in the lower margin) contain an example of a curious the tract De inventione litterorum states that Boniface himself
play with initials: introduced two types of secret writing into Germany (cf. p. 289).
R R R rex romanorum ruit (above rex: t regnum).
The Vienna manuscript contains several letters by Boniface
p p p Pater patriae profectus est.
and Lull with titles or subscriptions in a peculiar cryptic script
f f f ferro frigore fame
(fols. 2 v, ..V 'v, 39 v) (I). Since some authors have discovered
m m m monitum monumentum mortuus est.
runic elements in this script, a brief discussion must be
u u u uictor uitalis ueniet.
devoted to it. Out of the various cryptograms the following
a a a aurum a nobis aufert (I).
alphabet may be reconstructed : '

o~ ~ ~ rj~ ~ y~ !~~
(I) This play reminds us somewhat of the devices invented by the grsmmarisn
Virgilius Maro. cf. G. CALDIlR, Auraicept, xl ff. CALDIlR, o. c., xlii compares '
Auraicept 350I-3 with the device explained by Virgilius, but those three lines
of the Auraicept are obviously an example of lIotae Bamfatii. In St. Gall
;~
MS. 899 (p. 46) I came across the following instance, 'which STmNMIM!R also
found in Fulda MSS. Aa 2 and C t t, and in Munich MS. lat. 14737 (Altlwch­
deutsche Glossen IV, 436,6; 440,8; 457,6;549,26) :
Tres habuit turris scriptas in fronte figuras
Quas modo diuerso vir famulusque legunt
DMS.SSS.DDD.
\f' 1)0)p -DS-C;\-
m 0 n q r stu X y1.
Dominus dixit Domu$ mortui sepultu FlO. 30
Seruus dixit Domus magna senatorom
Dominus dixit Serous Malus Damnetilr
Serous dixit Dominus Malus Sepelietur The runic character of m and u was not questioned by Tangl,
Dominus dixit Serous Serpens Satanas who also referred to Diekamp's explanation of t as a " Rune
Serous dixit Dominus Demon Damnum
In Valenciennes MS. 411 we find a series closely related to that in the Vienna
manuscript:
V. V . V . Venit Victor Vitalis Lucius Beda. L. Quid spectas AngIe boB? B. Specto ruinam urbis
V. V . V . V . Vicit Viros Vesttre Vrbis vestre. L. Vides, sed non intelligis. B. Utrom intelligam veni et audio
V . V .V . Victor Venit Validus Et est. P.P.P.S.S.S.V.V.V.V.V.V.V.F.F.F.R.R.R.R.R.
A.A.A. Auferre Aurum Aroma (= a Roms) Interpretatio. Pater patrie profectus est. Secum salus sublate est.
R.R.R. Regnum Ruit Romanorom Venit victor validus vincens vires urbis vestre. Ferro fuga fame. Regale
F. F. F . Fame Ferro Frigore regnum ruit. Romanorum .Roma.
P . P .P ; = Pater Patrie Profectus Cf. J. MANcBART, Cato.logus, 390, no. 393, and 619, no. 625. On a similar
R.R.R.R. = Regale Regnwn Romanorum Ruit. cryptogram in Ghent MS. 306, see p. 84.
In MS. 843 of the Valenciennes Library there is a similar version worked into (x) M. TANcL. Studien. zur NtflflfJSgabe tier Boni/atius-Briefe, 723 fr. (with
a regular dialogue: facsimiles).

200 20I
...

• tyr' "(I). Several of the other letters are more or less cer­ this study (p. 6, 181) and we shall meet them again later on.
tainly derived from a Greek alphabet: c = k, e, I, r, s; in the The cryptic alphabet used by Boniface's circle seems to belong
name Erkenchinde (fo1. 39Y) ch is rendered by Greek X. For d to the same category : abc d f g q r are quite similar to the
an English uncial may have served as a model. No satisfactory corresponding letters of the • Mrican' alphabet in the Book
explanation has been offered for a, b, n (unless = Greek v); of Ballymote and in Brit. Mus. Addit. MS. 4783; m and n
p may be P without vertical stroke, f looks rather like an insular closely resemble the ' Egyptian' m and n in the same manu­
y and g like A written upside down. scripts. The' Gothic' alphabets of Cotton MS. Titus D 18
Let us now examine the characters which have been con­ and Munich MS. 1#36 also go back to the same prototype.
sidered as runic. If, as is commonly assumed, this type of What was said about similarities between Boniface's alphabet
cryptography was used as early as 738, this would imply that and Greek does not contradict· these findings: some of the
the m and t were borrowed before that date. But the Norse spurious alphabets may wen contain Greek elements or even
rune m was not created until the second haif or rather the end go back to varieties of the Greek alphabet.
of the eighth century (2), whilst the new t (Diekamp's rune The relationship with fictitious alphabets reduces the chances
" tyr ") is still younger. Therefore it will be safer to derive t of runic ingredients to almost niL At any rate we may safely
from some manuscript form of t, and m from Greek lP. examine the rune-names without taking account of the cryptic
On the other hand u could be the u-rune. But in view of alphabet (I).
the non-runic character of the rest of the alphabet, there is at The rune-names are written in a hand contemporary with
least as much chance that it is simply an inverted V, just as g that of the preceding letter and very similar, if not identical (2);
is an iriverted A. More light on these questions could probably their reading offers no difficulties :
be obtained by the stitdy of related scripts. As far as I know, asc. berc. can. donr. ehu. feli. gip. hal/gal. is. ker. lagu._,
the relationship between Boniface's cryptograms and an alphabet man. not. os. pert. quirun. rat. suigi1. taco ur. ilc. ian.'"
in Vatican MS. lat. 266 (fo1. Ir) has never been pointed out (3). zar. (3).
Yet, for all the differences between the Vienna script and that Two mistakes can be immediately corrected: donr is in all
in the Vatican alphabet, there can be no doubt that they are probability an error for dorn (4), and feli for feh. Seven of the
derived from a common prototype (a c d e f g kin r are prac­ rune-names have kept their OE. form and require no further
tically identical). The Vatican alphabet has clearly non-runic comment = is lagu man Os ur, and probably also here and feh.
characters for m, t and u, and t~ may be another reason not
to' explain the Vienna varieties as runeS'. (I) The Vatican matll.I8Cript gives the names of the csyptic letters : .alma,
Spurious alphabets going under the names 'Chaldaean', bina, calda, dotta, -a, J_ta, [.}g[.]s, lulled, (i missing), ka cana, tida,
mis, 1r000,ota, ponto., qui_, rOTot, risso, tOllta, .,..011, (~ missing; the name
, Mrican " • Egyptian', etc. haye; been mentioned before in began with s), YT, 2&da, d. These names have obviously nothing to do with
(I) M. TANGL, Studien .1M' Neuausgaie dsr Boni/ati:us-Brie./e, 726 and twie-names. To be sure, the ,name of.)l is YT OE. YT, the name of Yi but
footpote ::to the «Egyptian' .)I was called yph and the ' ChaJdaean ' .)I, .)In;
W. DU!ltAMP, Die WimIeT Hmubclrrift dsr Bonijatifu..Bri.eJe, IS 11. (on t see (:t) It is rather difficult to compare the two hands : in copying the letter
p. 17)· ' the scribe proceeded cursively, using numerous ligatures and separating the
(::t) L. JACOBSEN - E. MOLTD, RvneindskriJtsr (Text), 001. IOZO f. words only imperfectly; the rune-names, on the other hand, seem almost to
(3) On the Vatican manuscript see J. H. GALLER, Altstm:ksische Sprach­ have been written letter by letter.
~ler, ::tS6 f., and pL XIbi d. aJao (3) Cf. O. B. ScmUTI'llR, AMhelm's R:u:nic A~haiet.
F. DumuCH, Em _t/Iilisduls R_~habet, mit N _ dsr Budutabtm, (4) O. B. SCHLUT'I'BR, A/Jh8lm's R:u:nic Alphabet, reads dour docur dogur
erkliirl. Germania 13 (1868), 77-91 i (i. e. OE. dogOT ?). Such a reconstruction complicates things unnecessarily:
G. STl!PHBNS, Monumentl I, 114 (nos. 68, 69) and III, 9. tlogor never occw:s 118 a rune-name.

202 203
The first three and the fifth may have had the same form in the
MS. 207 with quello (p. 186) (I). The name of c may be a
scribe's Continental dialect, but this we cannot ascertain.. The
mistake for een, rather than be connected with the ON. name
names asc and hagal are adapted to Continental Germanic
kaun (d. caon in Munich MS. 1++36, con in Oxford MS. St.
phonology. In a n]lmber of cases the adaptation results in
John's College 17), as there are no other traces of Norse
more important changes. The name of e, ehu, looks rather like
influence. On account of the early date of this list, ui in the
the OS. equivalent of OE. eon, en (Heliand 388: ehusealcos).
name suigil can hardly be explained as umlauted u, at least not .
A number of names show decidedly OHG. features: gip is
from the point of view of OHG.; it may, however, be an OE. spell­
identical with the 2nd pers. sg. imperative of the verb giban,
ing for /y/. If so, it points to a very early date (firsthalf of the
although one ·does not see why precisely this form should have
eighth century) (2).
been selected (I); the dentals in *dorn pert not rat and tac (2),
In our general survey of the problem of alphabetization we
the guttural in tac, perhaps also the labial in gip point to a
saw that the last two or three letters of the alphabet will often
High German origin. The same may be the true of ker. Two
have caused hesitations, as the fuporc did not offer equivalents.
explanations are possible in this case : either the alphabetizer
The name ilc of Vienna MS. 751 becomes clear when we compare
started from OE. gar, which, translated into OHG., gave ger,
it with the form ilcs in Vienna MS. 795 and ilix in Brussels
ker; or from OE. (WS.) gear, (non-WS.) ger; the latter form,
MS. 93II-9319' The rune with this name had already received
which in OE. had initial /j/, was interpreted as beginning with
the value x in English fuporcs. The loss of final s will be due
a guttural stop, which then led to the German form ker. In
to an accident in the transmission. The form ian is rather
abstracto the two solutio.ns are equ.ally probable. But since
perplexing. Two explanations are possible. We may either
there is not one other instance where g is used for k, whilst there
start from the name of y, OE. yr, archaic *uir (d. suigil); this
are at least two more with similar names, one of which has the
could be misread as iur, and, in a context using open a, as iar; <.,
rune j (ker and j in Munich MS. 194-10; cer in Phillipps ~.

insular r could easily be interpreted as n, hence ian. Or else

MS. 3715), the second solution is by far the more probable.
ian may be derived from OE. iar, the name of j. If we assunie
The name of q looks at first quite puzzling. A name cweorn
again that iar had a final insular r, a form ian is easily accounted
has sometimes been postulated for OE. (3), but that word is
for. In the absence of the corresponding runes, there is no
never found as a rune-name. As a matter of fact the OE.
way of deciding which of the two solutione is right. Con­
rune-name cweortI, if it ever reached the Continent, did not
sidering the early date, the former may perhaps be favoured.
mean more there than it meant to English scribes : by the side
As to zar, the explanation is not so difficult as it looks at first
of such desperate solutions as qur (isruna fuporc) and quor
sight, if we think of her. The OE. rune-name gar appears as caar
(Munich MS. 1++36)'we find, more often, the name of the
in Munich MSS. 1++36 and 194-10, and in such a way that it
c-rune repeated, usually with another vowel. Only two alpha­
seems to belong at the end of the alphabet; the gar-rune is
betizers tried to give a rational form : the compiler of Vienna
actually found in the place of z and with that value in Vatican
MS. 751 with his quirun • hand mill, quem', and that of Berne
MS. Regin. 338 (there the name is gaar). There can be
no doubt then that the • superfluous' gar-rune was used for z
in some runic alphabets. This use may have been facilitated
(I) The fonn gip could perhaps go back to gif written with insular j; but
no such fonn is recorded in other name-lists. by linguistic considerations·: OE. gar may first have been
(2) O. B. 8cm.uTTER, AldJuIlm's Runic Alplulbet, emends tat: to tile to tiitl?;
a quite super;Buous. detour. (1) Therefore it is not necessary to reconstruct quirUlI. < quiruh < quiruth <
(J) E. SI1M!QS-K. BRUNNER, Alterrglische Grammatik, § 205, note I. queorth 88 O. B. ScHI..VTTIlR, Aldhdm's Rume Alphabet, proposes.
(a) E. SllNI!RS-K. BRUNNER, Altenglischll Grammatik, § 9+ A.
204­
20 5
...
 in Klosterschulen bestimmtes Compendium" (1). Its contents
changed to pseudo-HG; CaT (or. cam), which could then be
is important from the point of view of cultural history, and
pronounced with initial /ts/= written z (1). Unless the form
therefore we must regret that it has not been examined as a
zar is based upon OE. gar with insular g, which might perhaps
whole more recently. The runes which it contains were
have been interpreted as z, it supports the evolution gar >
known as early as W. Grimm's time, but seem never to have
car > Itsar/. As far as the Vienna manuscript is concerned,
been studied in detail since (2).
no such detour is necessary. The OK g may simply have been
The 34 folios are paged • I ' - • 67 '; the verso-side of p. 67 being blank,
chosen because it was the last rune (or one of the last) in the "it has not been counted. The first folio was at one time pasted onto the binding.
fuporc, its name then becoming zar through application of the Fonnat 220 X 137 mm (written area I80/zoo X lOS/IIO mm); 20-27 lines
acrostic principle. As the Vienna alphabet shows no direct to the page. The manuscript is mostly written by one Bavarian hand (pp. 1­
60; 62 I. 6 fl.; 65 1.7-67) about the middle of the 9th century. It may have
connexion with the alphabets mentioned supra (and conse­ originated in Tegemsce (3).
quently with the names gaar and cam) it is safer to accept the
Contents (4) :
latter explanation. p. I : Interrogationes in theologia, mixed with other matters (e. g. 22 : on
The monophthongs in the rune-natnes berc, feh and pert the division of the letters into vowels and consonants, etc.) and

seem to point to a non-West-Saxon origin, whilst Boniface, 24, 32, 33, 36-38, 58-60: Latin-9HG. glosses.

39: Carmen ad Deum, with OHG. translation.

Lull and most of their companions came from West-Saxon

41 ; Formularum colkctio.

territory. This need not imply that this runic material cannot 51: Latin poems.
be connected with the missionary movement led by Boniface. 58: A runic alphabet; Greek alphabets (d. ill/Ta).
For one thing, the early date of the prototype of this list may 61: Arno, Encyclica ad :rynodale colloquiwn in Rispah.
63: Ephmem Syrus, Homily.
account for the absence of breaking. German influence may
also have played a part. Therefore there can be no objection (I) L. ROCKINGER, Dra FOTI'I'Ielsammlungen aIlS der Zeit der Karolinger. Am .,4,
against the original of the list going back to Boniface's or Lull's Handschriften mitgdheilt von... (Quellen und Erilrterungen zur bayerischen
und deutschen Gesehichte VII), 1858, 24·
times.!liDiekamp has pointed out that the additions to Lull's (2) W. GRIMM, Ueber tkutsche Runen, III fl.

letter do not disagree with his character. On the. other hand Id., ZUT LitteratUT der Runell, 24 fl. = Kl. Schriften III, 110 f.

no Englishman can be held responsible for such forms as donr, H. F. MAssMANN, Neue Rutten. Anzeiger fiir Kunde des deutsehen Mittel­
feli, ian, or gip, ker, taco Some of the alterations are no doubt .alters 1 (1832), 27 f.
F. J. LAUTH, Runen-fudark, 40 fl.
due to a scribe of the Mainz archiepiscopal scriptorium or to J. B. SILVBSTRI!, PaliographiIJ ImifJeTselle, IV, 87, Pl 23 1 •
the copyist. This and the absence of the runes should warn K. Mou.itNHoFF-W. SCHERER, Denkmaler II, 353 it.
us not to value this material too high. E. STBINMEYl!R-E. SIJM!RS, AlthocluJeutsche Gionen IV, 567 f.
G. BAllSECKB, AbrogallS, 36.

Id., Bischof ArImJ von Frttising. BeitrAge 68 (1945), 75-134 (esp. I I 3).

B. BISCHOFF, SChTeibschu1en, ISS f., 159. 163 f., 270 f. and plate Ve.
4. Munich, Bayeriscke Staatshihliotkek, Lat. MS. I94Io w. SCHRODER, Di:e Verwaadtschajt der aIthoclukutschen Glossen n Gregors

(saec. IX). Homilien. Beitrige 65 (1942), 1-105 (esp. 93 fl.).

(3) B. BISCHOFF, SchTeilm:hulen, 163. The Munich catalogue has "Teg.

1410", but cf. E. STEINMl!YER-E. SllM!RS, AlthocluJeutsche G10lsen IV, 567:

L. Rockinger defined this codex as " ein fur den Unterricht .. Die alte Tegemseer signatur ist nieht mehr vorhanden It.

(4) L. ROCKINGER, Ueber bei FtmtUlls4mmlungen aIlS dem _nten und :t:ehllten

Jahrhundert in Handschriften der Stoatsbibliothek _ Mii:nchen. Gelehrte

(I) Unfortunately OHG. grammlUS seem to consider the spelling c = /ts/

Anzeigen der k. bayerisehen Akademie der Wissenschsften. MOOchen 1857,

as a • mistake' except before, and i, and thus we have no accurate data as to

its actual extension. Cf. W. BRAUNB, AltiweluJeutsche Grammatik, § 159, 461-477.

Catalogus II, 3, 242."

note z (p. 141).

206
207
 65: Eio's epitaph.
66: Epitaph for the scribe Hrotrohc. arrangement of the names on p. 58 shows clearly that the

67: Hartwig, bishop of Passau, Letter to Croso. second name is an intruder; but from p. 59 it does not appear

On pp. 58-59 the runes occupy a peculiar position. On p. 58 whether the scribe was aware of the discrepancy at the end of

there are, first, four lines of a text apparently intended for the the alphabet.

study of vocabulary (Nimirum ut frabri fer'ra'rii moris est, etc.). a: the rune is the a-rune; the form 'of the name, with final g

The runic alphabet, which follows on II. 5-6, is continued on for JkJ, points to a region where voiced and voiceless

the corresponding lines of p. 59 : a - I on p. 58, m Y on guttural stops could be confused, i. e. probably High

p. 59; the names are written on I. 5, the runes on I. 6. Next German tertitory (1). The same form occurs in Phillipps

come three Greek alphabets, the first of which gives the values MS. 3715.

of the letters, the last the names. They are also written across (z): the name must probably be read caar, an adaptation of

the two pages, but without having beef!. carefully planned. OE. gar (2). I have proposed to explain caar as the name

Mer the A of the first Greek alphabet come five forms of B of z in the scribe's exemplar. There are further argu­

then M. N, 0, etc. The final symbol t ('

1000 ') is followed ments for doing so. The g-rune, with the name gaar and

by M, i. e. the Roman nu:m:eral ' 1000 " and B. On p. 59 the the value z, takes the place of z in the runic alphabet of

scribe tried to crowd in the second alphabet on one line, but Vat. MS. Regin. 338. Holthausen and Schlutter (3) imply

with Xhe ran out of space and therefore wrote the remaining that OE. gaar was misread as zaar (the insular g is in fact

three symbols on p. 58 (1). The runic alphabet, too, contains not very different from some types of minuscule z). which

a striking anomaly : would explain this use. But in the Vatican alphabet gaar

e.g .c~r.6e.rlc.cen.doJ .eh.feh.geuo.hefh. IS. ker.I'b.go is written with the same g as daeg. geos (i. e. geof), hegil,
~.
".
/ago, sigil, a g which is quite different from the z used in
~ B· h· M·M·Y· X. N·1· t . ~
the same manuscript. This is also the place to remember
zar in Vienna MS. 751, which might be quoted as the
mlkn .n&.os.perd.cen.rat.srl.tfr ur.elcd.uyr missing link. As a matter of fact two explanations are
possible: either OE. gar became OHG. car, which was
f><l'~' ~.~ .. l\.l\: \i·t f\*'A~
then adapted. according to the acrostic principle and became
zar. Or else OHG. car (whatever meaning the scribe may
FIG•. 31
have given to this word, if any) may have been pro­
A name ca(a)r has been inserted after· a, where it has apparently nounced JtsarJ. as c may. have the value of an affricate
nothing to do; on the other hand the name of z seems to be (normally only before e and i, but occasionally before other
IIllssmg. The scribe must have misunderstood his exemplar. vowels too). With the more current spelling z, this
In the latter ca(a)r was probably the name of z, the supra­ would have become zar, and thus the name would have
script a being a correction (car> coor). The scribe of the fitted the acrostic principle. The former solution is
Munich manuscript mistook the suprascript a for an indicator­
(I) E. g. J. SCHATZ, Altbairische Grammatik, § 73. F. J. LAUTH, RU1ll11l­

of the value, and therefore inserted the name between those judark, 40 if., points out that g is found for c in other instances, e. g. legem

of a and b. This gave him twelve names for the eleven runes saUgam (p. 50).

(2) W. GRIMM, Ueb/17 deutsche Rumm, IIZ, read agcar as one word; in Zur

on p. 58; on p. 59 the situation was exactly reversed. The

Litteratur der Runen, 26 = Kl. Schriften III, I [0 he proposed to read ag, ear.

(I) In the Greek material we find a HG. interpretation: K: ' pro c et h '. (3) F. HOLTHAU!iIlN, Altenglische Runmnamen. Arcruv 99 (1897), 425.
o. B. ScHLUTTER, Aldhelm's Runic Alphabet.
208
... 18 209
 preferable because it is simpler.. In that' case the rune g fact rather. a complex one. OE. /jerI was spelled ger,
will have taken the place of z. because it was not needed which a German must have ul)derstood as OHG. ger, Upper
elsewhere, perhaps also because it came at the end of the German her 'spear' (which in fact corresponds to OE.
fu pore. But here we come to another difficulty: among the gar). Once the name of the rune had received this Upper
runic forms of Munich MS. 19410 there is no g; k takes German garb, the rune could take the place of k in agreement
the place of z. So we must assume that to fill that place with the acrostic principle. .
the form of one rune was combined with the name of n: the name net is a partial adaptation of OE. ned, Upper
another. Actually this will be found to have happened German having no final -d; cf. also rat: OE. rad.
in other alphabets as well, especially in the De inventione p: the adaptation of the names to Continental phonology
alphabet (p. 372). was not carried through uniformly: instead of perd we
b: the name .berie is an OE. form with secondary -i-, cf. should expect pert, a form actually found in other alphabets.
-berig for berg (WS. bearg) on the Franks casket (I). q: to fill the place of this letter the alphabetizer simply
d: in the name of the rune final -g has been palatalized to Ij/· repeated c and the name ern.· .
This change is typical of Kent, but there -ei seems to be r: cf. n.
the usual spelling of the resulting diphthong. The spelling s: the name sil results from the palatalization of g in OE.
ai is found in later Northumbrian (maiden etc. in the sigil or sigel, with subsequent contraction (I).
Lindisfarne Gospels) (2). Our alphabet has ei in another
x: this form of x is known only from Continental manuscripts
instance (cf. under h).
(St. Gall MS. 878, Munich MS. 1#36, etc.). The name
e, /: the forms eh and /eh with ' levelling' point to .an Anglian probably contains an error: Munich MS. 1#36, which, as :­
prototype. we shall see, is related with this manuscript, has elx. But
g: in the name, -u- (i. e. Iv/) was probably substituted on the -x would not easily become -cd. Perhaps we may recon­
Continent for the OE. spelling with -/-; in OE. texts u or v struct the. evolution as follows': for OE. ealh(s) the OHG.
with, this value occur only from the 11th century on, form elak was substituted; written with open a, final ah
whilst in OHG. it is a current spelling (3)· was then mlsread as -ed (for -d, cf. the confusion of -h and
h: the name of the h-rune presents an obvious mistake: the -I in heih).
final h stands for l, Mil being a palatalized form of heg(i)l. (4) y: uyr seems to be one of the numerous Continental attempts
Cf. under d and s. to render OE. yr; neit~er OE. nor OHG. spelling habits
k: the alphabetizer here depended on a form which had satisfactorily explain this form.
already been' translated' into OHG. The transfer is in z: on the name of this rune cf. supra (z). But if we agree to
accept ea(a)r as the name of this rune, a new difficulty
(1) E. SIlM!RS-K. BRuNNER, Altmglische Grammatik, § 164· arises : the rune which takes the place of z is not g but k.
(2) E. SmvERS-K. BRUNNER, Altenglische Grammati/l, § 126 A. 3 declore
earlyei-spellings in Northern texts (the Durham LiberVitae) , Latinizations '; Only the basic fuporc would enable us to find out how the
is there any proof that such spellings were not possible in contemporary OE. ? alphabetizer proceeded. As we saw under k and x, some
cr. E. KRuISINGA'S review mentioned on p. 184­ confusion seems to .have crep~ in. Perhaps the j-rune
(3) W. BRAUNB, Althochde:utsche Grammatik, § 137·
(4) The Munich manuscript being very early. this phenomenon can hardly
(I) Cf. Silhearwtm • Aethiopians • for Sigelh8arwan, E. SIlM!RS-K. BRUNNER
be explained as OHG.; the Alemsnnic transition of -egi- to -ei- seems to be Altenflische GrammaUk, § 214.4.
of a later date. d. W. BRAUNB, Althochde:utsche Grammatik, § 149 A. 5"'·
211
210
 coarse and greasy. The codex was gathered over some length of time; it
with lengthened cross-strokes was chosen for x on account
consists of 190 folios, mainly arranged in quaternions :
of its resemblance to Roman X. This left the regular 3 IV {1-24] + IV (+ I) {"5-33] + 19 IV {34-185] + 5 single fols.
x-rune unused. It must then have been mixed up with mounted on parchment guards {I8f>..190]
the calc-rune, the result being that the form which corres­ The quires are marked I I ' to • XXIIII '; thete is an error in the numbering
after • VillI', or else one quite is missing.
ponds to the latter took the place of z. Fonnat :115 X 160 mIn (written area ca. 145 X lIZ rom); ca. "0 lines to
The other runes require no comment : the c-rune takes the the page; two columns on fol. I. The manuscript was probably written in
Fleury. A gfeat many hands contributed to it, mOlt of them belonging to
place both of c and q (in spite of the acrostic principle); is, /ago, the ninth century. Laistner dated the codex A.D. 861-4. but Jones claims
man, os, tir, ur are taken over without any change. an earlier date: 817 (I). The quire which contains the runes is made up of
two ruled leaves (58/65. 59/64) and two unruled leaves (60/63, 61/6z); these
have only 16-17 II. of text. Part of the quire must have remained blank for
some time : whilst the hand which wrote the runes cannot be much younger
There can be no doubt that this alphabet originated in the than the rest of the codex, some formulae on fol. 61 11. point to the year 9Z0 (z).
Upper German area; this agrees well with the location of the Contents: (J)
scriptorium in which the manuscript is supposed to have been fol. If: Omnium faustorum & infaustorum catalogi fragmentum.
Zf: "Incipit computatio Grecorum seu Latinorum; primitus vero
written: Tegernsee. In fact the adaption to OHG. phonology
dicitur de mensibus ". Sequuntur Calendarium, cornputationesque
is very superficial : no real translations are found, and only one ecclesiasticre et ailtronomicre.
rune has changed its place as a result of an adaptation (j > k). 5z': De Itineribus et mensuris agrorum.

Even such simple cases as beric and heil, which would not have 55 f : Argumentum beati Theophyli de epacta.

58f: Ratio sphene Pythagone philosophi quam Apuleius descripsit.

upset the order of the alphabet, were left untouched. 589: Ejusdem lunarium de legris. Lunarium Sancti Daniel de na6vitate
The underlying fuporc must have come from Anglian territory infantum.
rather than from Kent. The use of what is formally j for x 59f : De diebus Egyptiacis qui mali diOO sunt in anni circu1o. ~"
60': AIphabetum literarum Gl'lIlCarum cum notis numeralibus, inter.
connects our alphabet with the fuporc in St. Gall MS. 878 and qWlll habentur Runicre quzdam. [Litterae formatae]

with the last alphabet of Exeter MS. 3507, etc., as well as with 61 V : Miscellanea ad Calendarium pertfuentia.

Munich MS. 14436; with the latter and with Vatican MS. 63 f : Signa qure ostendit Delis Esdne prophetre.

66f : De numero annorum ab origine mundi usque ad adventum Christi.

Regin. 338 it probably shares the name of i for z. 68 f : De 7 sideribus errantibus, Zodiaco, januis czli, planetis, &c.

88f : Horologium.

9Z v: De positione 7 stellarum erranrium.

5. London, British Museum, Harley MS. 3QI7 (saec. IX).

97·: De aete, tonitru, fuhninibus, & ventis.

IOZf: De chronice vocabulo. Ex Isidoro.

J. M. Kemble first edited the runes in this codex (I). The IOOf : De 4 Anni temporibus. Ex Isidoro.

problems which they raise are similar to those met with in the II6 f : De compotu vel loquela digitorum.

IIQ': De saltu lunre.

Leyden manuscript we discussed on p. 192: here too the runes

seem to have been included because the compiler somehow
(I) C. W. JONES, Bedoe Pseudepigrapha, 12Z;
connected them with the Greek alphabet, and that Greek M. L. W. LAlSTNER. Thought and Letters, 189, note 3.
alphabet belonged to the regula Jormatarum. (z) Hence the date in the ca~ogne : ... scriptus, ut ex fine miscellaneorum
Red leather binding, gold tooling, with inscriptions on the baek : MISCEL­ ad calendarium pertinentium apparet, A.D. 9"0.
LANEA / DE COMPUTO, / ETC. / BRIT.MUS./ HARLEY MS. / 3017 . (3) Catalogue II, 7"5;
On the front end-paper and fly-leaf various older press-marks; on fol. If: 13 C. W. JONES, Bedae Opera de Temporilnu, Iz6 (note 3), 15z, 165, 330, 33",
die mensis Augusti / A.D. 1724. Parchment of varying quality, usually rather 35", 36 5. 369, 37", 375 f., 386•
Id., BeJoe Pseudepigrapha, izz, defines the codex as .. an excellent example
(I) J. M. K'.BM:BLli:, Anglo-Saxon Runes, p. 339 and Plate XVI, fig. 10. of ninth-century Fleury computi".

212 21 3
....
 119": De cursu &: eclipsi solis. names, but refers to other persons, etc. by N .:...- nomen etc., or
126": Ordo &: nomina ventorum. ill = ilk, etc., it must have been intended as a model, a regula
129': Ratio de torutruo, de arcu, de IIlStu oceani, de circulis teiTaJ &:
divisione sua. lormatarum explained by an example. On fol. 6Ir we find :
132": Schema .mundi. (a) a Greek alphabet with the names of the letters;
137': de cursu &: magnitudine solis.
143': de 4- Anni temporibua. (b) a Greek alphabet, rearranged in the order of the Latin
ISO': De fulminibus. alphabet (X c, H e, e t, {J = 0, 0 = tt).
ISO": Cur mare non crescit.
151': De Nilo. (c) without any transition there now follows a runic alphabet
151": De literarum Gnecarum potestste. without values nor names; the first two runes- are on one
152': Numeri per quos potest qui voluerit alterius cogitstiones de numero line with N Z of the preceding Greek. alphabet; the
quolibet quem animo conceperit explorare.
154r: De quantitate solis &: lunre.
remaining runes fill two lines (c n, 0 - 21). At the end
155": De lumine &: cursu lunre. of the alphabet there are three dots placed in a vertical
100": De nominibua astrorum. De Arcturo. line; a similar device fo'r marking off is found after the
I63r: De -Nive. De signis tempeststis vel serenitstis.
I64 v : De tenre motu.
first Greek alphabet (a).· But there is nothing to show
165': De indictionibus.
that the scribe was aware of the fact that (b) and (c) are
16S v : De opactill [sic] lunaribus.
two different alp.habets.
I68v : De temporibua, horis, &: momentis.
Ex Isidoro.
169': De die et nocte. De Hebdomada, &:
mense. (d) a Greek alphabet, with above each letter its numerical
170': De salatitio.
value. This is the alphabet which normally belongs to the
I70 v : De temporibus, &: annis.

regula lormatarum : in drawing up a letter of commen­

17 I': De bisexto.

172v : De arithmetica.
dation, the writer inserted in it various Greek numerals
178v : De inventoribus Geometriae. &: vocabulo ejus.
and at the end the sum. of these numerals, to prevent;'
I80v : Notre musicales.

forgeries.
181': De ponderibus.

181": Confiteor.
To judge from the form of some runes (g, m), the author
182": De generibus aromorum. Etymologica qwedam de puncto &: cannot have had any first-hand knowledge of the runes. He
minuto, ex Isidoro.
. 183': De inventione horologll &: quadrantis.
copied them rather carefully; though, but without· betraying
183": De diebus, &: quomodo apud Cbristianos nominantur. any understanding. The height· of the runes varies between
185': Nomina mensium spud Hebrreos. Ad sciendum cyclum solis. 4- and 9 mm.
186v : De signis (Zodiacis.)
188': De XII. generibu8 annorum.
189v : Ad embolismum inveniendum.
abc d ef 9 h ·k 1m
The text on the dies aegyptiaci ends on fol. 6or; it is followed
by a letter of commendation (litterae lormatae) addressed to one ~n..B "p WM~..r
0 q
N1 'V ~ KI

s ·t . u . x 'y . z
r
bishop Accus (or Acco) (1). Since it contains no·other proper

(I) Inc. Incomparabiliter et ineffabiliter amando fratri Acro episcopo ill satis
in ecclesiasticisnegotiis uigilanti N perpetut' felicitstis in Domino saluatore
1-ij ~~I\'1;f\A.}A%
FIG. 32
obtst salutem. Expl. IlllIuper uero adulterinis decolorationibua reiectis
epiatola sigillo nostro sigillats apparet habeIUI in se effigiem iDam. There Thirteen runes call for no comment, viz. those for a (= a),
was a bishop· Acca in the early eighth century (Hexham 7100731), but the b, I, h, i, I, n, 0 (= 0), p, r, s, t, u.
letter points rather to s later date.
21 5
214
'"
c for this rune we find the rather rare type with a straight to have worked directly from an OE. fuporc, which retained
lateral stroke. some archaic features (fl, open r). If we may really consider
d: although the rune is drawn somewhat awry, the original this alphabet as Fleury work ~ and I did not find any evidence
type with low intersecting strokeS can still be recognised. as to the contrary-we have here a good proof ,of the widespread
interest the runes enjoyed in the ninth century, and at the same
e: the form of this rline is clearly influenced by a Roman
time another indication of the part that abbey may have played
capital M.
. in transmitting and preserving runic material (cf. BemeMS. 207
g : the awkward form given here can only go back to a j-rune. and Leyden MS. Voss. lat. F. I2 8). As the alphabet is only a
The basis for assigning the value g to it must of course lie copy, it would be interesting to know where the original was
in OE. spelling: g was the current spelling device for /j/ written. Perhaps the litterae /onnatae might enable us to
(whatever the exact phonetic value of this sound may have settle that question; but as far as I know this brief text has never
been). been studied. Finally it is worth while noting the surroundings
k: the form we find for k is either :1:, or k turned upside d~wn. of the runic alphabet: this is the third time we find it together
On the other hand the x is rather like a not very successful with litterae /ormatae (cf. Salzburg MS. a IX 32, Leyden MS.
attempt at k. It rather looks as if the two runes had been Voss. lat. F. 128); the dies aegyptiaci, ratio spherae Pythagorae,
mixed up. A similar confusion is found in Cotton MS. etc. will also be met in the neighbourhood of runes in other
Domitian A 9 (scribe B); cf. also :it in Munich MS. 19410 manuscripts.
and q in Arsenal MS. 1169.
18: from the form of this rune we may perhaps infer that the 6. St. Gall MS. 270.
exemplar from which the scribe copied. was in a poor state. ~".
~.

q: somewhat similar q's are found in Munich MS. 14436 On this manuscript cf. p. 90 ff.
(' Arabic alphabet ') and in Oxford MS. St. John's Col­ The runic alphabet which is found immediately after the
lege 17. It is probably .derived from p, and may represent fuporc is not merely an alphabetization· of this fU}JOrc. With
an intermediate stage between that rune .and the q we the latter it disagrees on a couple of important points :
found in some fuporcs. The fully developed form of the
latter may never have reached the Continent. a a. b C ,d d e f 9 9 9 h k [
x: cf. under k.
y: the m-like subscript is p~obably the scribe's fanciful inter­ ~ ~ tl\ ~ ~M'~ ~1>lIlN I J' I

pretation of a subscript y in the original; cf. y in St. Gall

MS. 878 etc.
mnoo ppqrss ttuX7.
:it: will be derived from a Roman Z rather than from 3.

On the whole the alphabetization is quite straightforward.

f41~ r::rf~~ R~~1'HnXZ
FIG. 33
Apart from k and x (the confusion of which may go back to a
faulty fuporc, cf. Domitian A 9 as 'corrected' by scribe B) a: both a and z are given this value; in the fuporc the former
the procedure raises no problems. The difficulty of providing had the value aa.
a symbol for z has been got round with little effort. There d: the first d is It (cf. d, Mrn in the fuporc); the second is w,
are no 'traces of Continental influence; the alphabetizer seems which, notwithstanding the value flU in the fuporc, was
216 21 7
 confused with I» on account of the similarity of the forms .. but was afterwards dropped for one reason or other: it is
Perhaps the alphabetizer relied on the form because from the only rune of the fuporc missing in the alphabet ..
the point of view of the acrostic principle he found a
z: in this case, as with k, the scribe depended on the value
contradiction between the value and the name (uu : huun).
of the rune (z) rather than on the name (aer).
g: to the two g's of the fuporc (i. e. the special g, and j) the Among the runic alphabets there are only very few which
alphabetizer' has added a variant form of g; it differs from also assign a place to the • superfluous' runes of the fuporc
the first g in exactly the same way as the two s-runes (cf. Oxford MS. St. John's College 17). The scholar ~~­
differ, i. e. the middle parts of the broken lines are at a sponsible for the St. Gall alphabet obviously used the preceding
different angle. fuporc, but may have had another source for reference by the
k: in this case the alphabetizer relied on the value (k) rather side (a De inventione alphabet ?). The St. Gall alphabet did
than on the name (ih), and for obvious reasons: he had probably not belong to the isruna tradition for a long time: it
two other runes which could take the place of i (iis, inc). is not found in any other manuscript of the group; the runic
p, q : here we come to the most important difference. The alphabet in Munich MS. 14436, which is also derived from an
isruna fupore, shows a completely different structure.
first p of the alphabet has the value q (and the name yur)
in the fuporc; the second is identical with p' in the fuporc.
On the other hand the alphabet has an entirely new q. On 7. Exeter, Cathedral Library, MS. 3507 (saee. X).
thiS point, consequently, the alphabetizer corrected the
8. London, British Museum, Cotton MS. Vitellius A I2
fuporc. It is rather hard to believe that he did so led by
(saee. XP/XIP).
. his o;wn knowledge of· the runes : his way, of confusing
'",
I» and w shows that he cannot have been a very proficient 9. Phillipps MS. 37I5 (saec. XII ?). ~.

runologist. He may have known about the p-rune from

. some other runic alphabet; if .so, he probably borrowed The' first two-of these manuscripts are so closely related that
his q from the same source. His q is typical of the De the second has been considered a partial copy of the first. The
inventione alphabet. That the isruna tradition and the third shows .no such c~osehelationship with the first two,
De inventione tradition occasionally met is clearly demon­ except in itsrunie material. ~ Therefore the runes of these three
strated by. Vatican MS. Urbin. 290 and Munich MS. manuscripts will be discussed together.
14436, q. v,. Exeter MS. 3507 has eonnexions with Salisbury and Sher­
I: cf. under g. borne; it was probably writteQ in the same scriptorium as the
t: in agreement with the shift of values in the fuporc (d > t, Sherborne. pontifical (paris, Bibliotheque Nationale, MS. lat.
943) (i:), and perhaps a.witness of the reform movement of the
dtllg > tac) the alphabetizer lists two t's : t and d ..
(y) : in the alphabet the equivalent of Lat. y is missing. Two (I) N. R. KBR. Salisbury Cathedral Manuscripts and Patrick Young's Cala­
runes could have filled its place: either g, which plays logtUl. The Wiltshire Archleologicsl and Natural History Magazine. 53 (1949),
156, note 2. On the Paris manuscript see :~II
that·part e. g. in Munich MS. 14436 (also related with the R. BBoTANEK. Texte utul UntllTSJlChungen zur altenglisclum LiteratUT und
isruna group) and in Arsenal MS. n69; or else the q-rune KiTchengeschi.chte. Halle. 1913. 1-29, 33-149.
of the fuporc, on account of its name yur. It is not im­ M. F(}RSTER, Die altenglisclum Texte tier ParisllT Nationalbibliothek. E. St. 62
(1927/28), 113-131~(esp. Il6-1:l9).
possible that g originally took the place of y in this alphabet, Biblioth.e Nationale. Catar~genbaldes matlUscrits latins. I. 1949i 335 f.
218 Z19
...
\
 tenth century. Its runic alphabets were known to Wanley and Contents (I):
Hickes (I). fol. I ' Incipit pTologus hrabani peritissimi uiri / DILECTO FRATRI
Brown leather binding; on the back COD. ECCLESI. EXON. stamped MACHARIO / monamo hrabanus peccator in Christo / salutero.
in gold. On the front end-paper old pressmarks (N:O 2; N° 25), an ex-libris IV explicit prologus /Incipiunt kapitula libri sequentis (<:ont. 3r -'I')_
(CATHE: Ch Library. S': PETER'S EXON t Anno D :nl 1749) and a' Incipit ipse liber hrabani de compoto/ QVIA TE VENERANDE
bibliographical notes (a), which are continued on the first paper fly-leaf (3); / PRECEPTOR SEPE AVOIUI / de numeris disputantem, etc.
on the latter the present number '350'7 'and: DEC: ET: CAP: EXON. Hrabanus Maurua, De com:puto.
The verso side, and the second and third paper fly-leaves are blank. The 57' De mtatibus.
97 folioa are bound in 13 quires, marked ' a ' (7 V), ' b' (IS"), ' c' (23.), • d ' 58' A number of short computistical items: Versus de duodecim men­
(3 Iv ), 'e' (J9 V), 'f' (47'),' g' (55'),' h' (63'),' i' (7[V), 'k' (79'), 'I' (87'), sibus anni; Tetrasticon autenticon de singulis mensibus; Versus
'm' (95'); the lll$t quire is incomplete: IV (- I) [1-71 + II IV + 2 Cols. de singulis mensibus; Versus de singulis mensibus et signis;
The whole codex seems to have su1fered from dampness, especially the Versus de duodecim mensibus; Versus de cursu anni; De octo
first quire, the parchment of which hll$ become rotten. Repairs had to be tmmitibus circuli decennovalis; Versus de septem dierum vocabulis;
made and as a result the order of some folios WII$ disturbed (2' : .. misplaced / De septem miraculis manu fsetis; Duo sunt extremi vertices mundi,
sh[ould be] 3 d fo"; 3 r ; .. this is misplaced / It sh d be ad fO" in a nineteenth etc. ; De diebus egyptiacis; Ordo librorum cstholicorum in circulo
century hand). In fact the order should be I, 3, a, 5, 4, 6, 7. One (the anni legendorum.
first, blank?) folio seems to be missing. On the whole the parchment is white 64"' (no title) On pronunciation (ultimately from Martianus Capella).
or faint yeUowish, some pages slightly stained; the surface is rather rough. 6S r a Greek alphabet, with the names of the letters and their meanings,
Format ca. 315 x 195 mm, written area 228 X 148 (without double ruling in one column: A alia. agricola, etc.; a Hebrew alphabet, also with
left and right a28 X 132) mm; mostly two columns, 29 lines to the colwnn. the names of the letters and their meanings : K aleph, interpretatur
The text WII$ written by one scribe, probably in the second half of the tenth doctrina, etc.;
century. Glosses by several bands, into al. two OE. glosses on 98 v : luligines 65 v a Greek alphsbet, with the Greek numerals and the values in Roman
.i. merespin (i. e. mere.lfllin); delphini .i. si'l1Ul1let l seoltu. figures : A mia .1., B dia .11., till 'P niacurin DCCCC.; the Greek
numerals for' 1000' to ' 10000' (' 3000 ' omitted) : Chile - Myre.
immediately followed by IullS./ IHCOYC • IHC./ CPIC. TOye./
(1) H. WANLBY, Catalogus, 281 : co Hrabani Mauri de Computo exemplar Christw.xpa; a list of Greek letter names, including those of the nu- ;:-,
probe ante 700 annos exaratum, in quo habentur Alphabeta qumdam & voces merical symbols: epissemon, sincope, eneacoseS.
nonnulile Latinle Litteris Runicis scriptte. (Vide D. Georgii Hickesii 66~ Three runic alphabets (cf. infra).
Gramm. Island. p. 4). JJ 66 v De <:oncurrentibus; de sex letatibus nominis (= hominis).
G. HICKE'S, Thaaurus, Gramm4ticae Iila1uJicae Rudimenta, p. 4 : " TabeUa (s) 67r Incipit liber isidoris psalensis episcopi de natura rerum. DOMINO
tria continet, qUIe extant ad calcem codicis MS. Hrahani Mauri, de cOmputo ET FILIO SISEBVTO / esidorus salutem, etc.
ad dilectum fratrem Macharium, qui in Manu Saxonica exaratur. Hujus 6,.' explicit prefatio. Incipiunt capitulationes. On fo!. 68 r , 11. 1-3
codicis usum acceptum gratus refero reverendo doctissimoque Tiro natalibus, (same hand) ; Alii autenx prologum MUS initium tu fortem locu/
ingenio, & humanitate pneclaro, Ricardo Amresly dignissimo Decano, & reve­ lentis uaga ca:rmina gignis in hunc locum introdu/cunt. Alii nutem
rendis fratribus Canonicis Ecclesim Cathedr. &oniensis, & Leojrici·fundatoris isidori esse respuunt Sed gilde.
maximis omsmentis, qui pro suo in antiqulI$ literllll lIll'Ore, exarchivis ecclesiz Isidore of Sevilla. De natura rerum. The end is now missing, but
veterum menxbranorum quicquid ibi Sa:JtQnire scriptum esset, in usum nostrum, was there originally, as a note on fol. 76' (lower left) shows: " Isdem
ro{!ati faciles miserunt "; tabella II, no. 5. . circulus / in fine huius (libelli alia / figura scrip/tus est "; this rota
(a) Hraham Mauri de comjN.to ezemplar prciIH. ante 700 anno, ! exaratum, is no longer in the manuscript.
in quo habentur Alphaheta qutIIIdo:m I};f 'Voces rwnmdlm/ Latin2 Litteris Runicis
scriptre. vide .Gramm4tic2 Iiiammm/. pa,giTuzm quarta.m I};f Tabe&m qumta.m Fol. 66r shows the following arrangement:
sequentem in Tamo ct. Hickesii, to which another band added: Accedit Iridarw:
Hispalemis Episc: de Natura Rerum / cum SchematibllS tit Figuris, a C/aris,:
(a) II. 1-6, a runic alphabet, with above each rune its value;
Wtmleio / in Catalogo tuO Codd. Mss. Anglo-Sax: EccletitB/ Exoniensi.s omissus, at the end " Super sunt iste ", and four remaining
and a third : NB. 97 Folia et I94 PaginaJ/in hoc libro continentur. There can letters.
be no doubt that Wanley knew the manuscript, as the first note is almost iden­
tical with· the entry in his Catalogus.
(I) Cf. H. SCHllNKL, BibliothJlt;a Patrum Latinorum Brit(l1Jnica VI, 46. DO.
(]) BaJumus edidit Hunc Librum ex Vetere Mss./ Wl. MiscellamorumLib. sm / 3751. (Sitzungsberichte der philosophisch-historischen Classe der Kaiser­
(in pencil :) Paris I678.
lichen Akademie der Wissenschaften CXXXI,X. Wien, 1894).
220
. 221
(b) 11. 7-12 a similar runic alphabet, again followed by Usual Cottonian brown leather binding with gilt tooled arms; on the back
DIALOGUS /INSTITUT. ECCLESIAS. / TRACTATUS VARII / DE
.. Super sunt littere iste .IIII. " and four remaining RATIONE TEMPORUM/VERSUS ET / PC)EMATA./P<E NITENTIALE /
runes. ANTIQUUM. / MUS.BRIT. / BIBL.CO'ITON. I VlTELLIUS A.XII /
(c) n. 13-18 a third runic alphabet, with the values shown in PLUT. XXVI. C. The manuscript has been badly damaged by fire and by
moisture. Especially the upper and outer margins have suffered i the parch­
the same way.
ment was charred and partly chipped off, and what is left must have shrunk
(d) n. 19-24 an example in runes with the transcription considerably. Fortunately the repairs have been made carefully: each folio
PAX VOBISCVM ET SALVS PAX. has been set into II leaf of stiff paper, the corresponding part of which was cut
The runes are drawn with evident care, and measure I I mm. out, and fastened along the edges; only the transparent material used to fasten
the edges sometimes covers sman portions of the much faded text. The ink
seems to have been washed away in places, especially the rubrics. The page

* **
with the runes, however, is fairly legible. Three fly·leaves of the same strong
paper have been added in front and have been included in the new folio num­
bering; therefore it does not coincide ,with the old numbers (e. g. new fol.
Cotton MS. Vitellius A 12 consists of several originally 65 old fo!. 62.). Fo!. 3 : II strip of parchment (30 X 135 mm), with a later
independent manuscripts. The part which contains the runes script, has been pasted on the paper leaf; the manuscript actually begins with
seems to have some relation to Abbo of Fleury's visit to England fol. 4. As all folios are loose, only the old quire numberings can help us to
r r
reconstruct the composition of the codex: 12r ' I " 2.0 r I 2. " 28 ' 3 " 36 ' 4 '.
(986-988) (1); part of it was presumably written in Bury St. 44' ' 5 " 52.! ' 6 " 60' ' 7 " 68 r ' 8 " and, parallel, 12.' ' C " 20'" • D I . . . 68'
Edmunds or in the neighbourhood (2). but the portion which , J " 72.r ' K " 78r ' L " 87' , M " 93r ' N " etc. till 176< ' Z '; with 10<)' begins
interests us here is believed to have originated in the West another series, apparently older than the one with espitals ; 109" I', 1I6r I 2. "
126' ' 3', 132.' • 4.1 '; 133' , 4.2.', ending on 135' , 4.4 '. Consequently some
Country (3). Bishop Ussher used it for his edition of the
material may be lost in front, but that did not necessarily belong to the codex
Epistola Cummiani. It was described by T. SInith (4) and from the very beginning. MS. *A beerns to have comprised fols. 5-7 2 ; MS. *B
H. Wanley (5), but soon after sustaine-d considerable damage fols. 87-185, but was itself made up of different parts : 87 ff., 101 if., 10 9 ff., ~,
in the fire of the Cottonian Library. Wanley seeIns to have 13 6 ff. Between *A and *B an amount of material was inserted when the ~
two parts were brought together. In *B the parchment is as a rule thinner,
been the first to notice the relationship between this manuscript the ink darker than in *A.
-and Exeter MS. 350'], at least as far as the runes are concerned; The maximum measurements of' the irregularly shaped pages now vary
he copied them -from both manuscripts -forGo Hickes, who 'between 205 X 115 and 2.10 X 145 mm; written areaca. 195 X 100 mm;
used them in his ThesaU'l'US (6). *A has mostly one, *B two columna; 36 (fol. 4) to 69 lines (fol. 106) to the
page; fol. 123r has exceptionally 86 11. Judging from the handwriting, *B is
somewhat later than *A, though most of the codex may be dated in the late
(1) See e. g. A. VAN DB VIJVIllI, LeI aluwes inMites d'Ablxm de Fuury, 141.
I l th century (part of *B in the late uth).
(2) A. BoUl'l!MY, Latomus 1 (1937), 295. . .
(3) F. WORMALD, EnglUh KaLmekrrs be/ore A;D. IIOO. Vol. I; Texts. To the.survey of the contlmts as given in the Cataiogue of 1802 (I), I add
(Henry Brad8baw Society LXXII). - London, 1935, 85, ascribes the calendar some items and remarks :
on fols. 65"-71 to Exeter. Fo\. 4': Rabanus de compoto (later hand).
(4) T. SMITH, Cataiogus, 82 f. (9. Alphabetum Norwegicum sive Runicum). 4": Succinctus dialogus ecclesiastical institutionis a Dom. Egbherto,
(5) H. WANUn', Catalogus, 239: .. fol. 62. Alphabeta Runica tria, cum Archiep. Eborac. civitatis compositus. Editus a Cl. Jacobo Wara:o.
his verbis Runicis litteria PAXVOBISCUM ET 8ALVS PAX. Eadem
Dublini, 1664'
Alphabeta OC<:W'lUIlt in libro veteri Exonienais Ecclesiae,· ex quo ea descripsi
haud ita pridem, in usum D. Georgii Hickesii ".
(6) G. HZCKllS, 'rfwaunis, Grmnmaticae Islandicae Rudimenta, Tabena II : 5. quod respondet Lizti'llOf'Unt &! Anglo-Sax_ c, deaignatur per Cimbrorum
Alphabeto. 3. gum habetltur od calcem libri Hraban:i Mauri de Computo, irt antiquo veterum sigma [s] • ut in hac tabl:lla videre est ..... The first two alphabets
Cod. MI. E:roniemis Ecclesiz, and: Grammatica Anglo-Sev«mica, 148 : " ... se­ in this tabella do correllpOlld to nos. 1 and 3 in the Cotton manuscript, but
cundo notandum est i'I vocabulis Emperic & Iustice, c apud Angw-Nurmamws the third (i. e. that which is supposed to illustrate Hickes's point) is a Norse
in Gallo-Lati'llOf'Unt fine sonuisse ut s Rtmwnorum. Unde fOIllan in lbmorum alphabet taken from some other manuscript.
aJphabeto quod extat in bibliotheca Cottf»!. Vitellius, A. 12. tertium elementum (I) Catalogue, 379 f.

222 223
....
 8r : Sententia Abbonis, de differentia circuli et sphreI1lli, et de cursu 123r: Marbodi versus de laude castitatis, de dissuasione mundanre cupi­
septem planetarurn per zodiacurn circulurn. ditatis, &c.
10V: Rabani monachi de cnmputo liber. Sic ait rubrics; sed eat opus 124r : Versus de XII. Imperstoribus Romanorurn; de longitudineregni,
Gildre: INCIPIT I PROLOGVS RABANI PERITISSIMI VIRI et finibua eorum j item invectio in quendam abbatem monachalem
AD GILDAM MA!GISTRVM SVVM. SED OPVS EST GILDE dominam sibi surripere volentem, cum aliis.
(cf. Ex., I V -57r ). The incipil simply attributes the work to Gildas : 127r : Invectio Gualonia Britows in monachos, versibus rhythmicis.
Di.lecto fram Rabano mo1UUM Gildtu PeccatUT in Christo salutem. Marbodi versus de VII. primis diebus, de X plagis JEgypti, de
40': Versus de mensibus, signis zodiaci, cursu anni, octo trsmitibus cin:uli muliere mala et bona : alia que plura.
decennovenalis, septem dierum appellationibus. A series of short 133': Versus Hugonis Sotsvagina:, cantoris et archidiaconi Eccl. Sci. P~tri
poems on computi~tical and astronomical subjects (most items also Eboraci.
in Ex., 57"-64r ). 135': Versus Augustini Canonici.
4 2 ' : De septem miraculis manufictis; de duobus verticibus mundi; de 136<; Prenitentiale antiquum. ex vsriis canonibU8, item ex pamitentiali
diebus JEgyptiacis; ordo librorum Catholicorum in circulo anni Romano, Theodori. et Bedre co11ectum.
legendorum. Is,..v: Oratio dominies Nonnanno-Saxonice.
44': Oe vocibus literarurn., quomodo fonnantur (= Ex. 64'). 185": Narratio fabulosa de quodam episCOpOI qui celebraturus divinum
46r : Gildre, peritissimi viri, liber de compoto, de mundo, planetis, stellis, officium, ex vultu cognovit quinam digni essent, quinam indigni.
tonitruo, fulminibus, ventis, oceani reatu, Humme, terrre motu, monte ad communicandum.
JEtna, aliisque physiologicis : cum prrefatione ad Rabanum monachum
quam edidit Usserius in epist. Hibernicar. Sylloge, Dublini, The alphabetic material is distributed as follows :
ib. 2. p. 55.
INCIPIT LIBER I GILDE PERITISSIMI. DE NATVRA fol. 45 r : A collection of alphabets (= Ex. 65 r -v ) : a Greek
RERVM, i. e. Isidore, De natura rerum (= Ex., 67 r -97'). The
address too has been adapted to fit this attribution, the name Sisebuto
alphabet' with the names of the letters and their
having been erased; " Domino et filio (erasure) salutem ". Smith values, and a similar Hebrew alphabet have been
and the catalogue of 1802 seem to take this attribution serious. mixed together: Greek col. a, 11. 10-19, 30-36, col. b
Moreover they mix . up the computus of fol. 10" ff. with the
10-16; Hebrew col. a 20-29, col. b 17-34; 45 v another
...
">.
De natura rerum.
51 ; (blank) was taken fI'OIIl another manuscript, probably to fill the Greek alphabet with numerals and corresponding
place of a lost fol. Greek figures, additional numerals and graeca, and
65': Alphabeta Runica tria; cum his verbis, Runicis literis, "Pax vobis­ a list of the Greek letter-names.
cum et salus pax ".
6S v : Clliendarium vetus. r
65 : Three rumc alphabets and an example illustrating
72<: Versus de constellationibus, et ventis.
their usage (Ex. 6&'; similar arrangement):
73 v: Kalendarium aliud, cujus OInnes dies nominibus sanctorum signantur.
78 : blank. .
(a) 11. 1-4 : a - 0 I p - z, with above each rune its
79': Epistola Cummiani. directa Segieno 'abbati, de disputatione lunre.
Edidit Usserius in Epist. Hibem. sy11. p. 24. value; afterz : " Supersunt istCC" and four more
83': Epistola Bedre presbyteri apologetica, eo quod insimularetur a runes.
quibusdam de aetatibus secuJi non rerte gensisse.
87': Ube11us de computo; cum regulis ad inveniendum annum, indic­ (b) n. 5-8: same arrangement; after z: "super
tiones, epactas, retatem lunle, &C. sunt littercc istCC iiiior".
IOl r : Libellus alius de eadem argwnento.
109': Versus Serlonia Parisiacensis, ad Muriel sanctimonislem virginem (c) 11. 9-12 : a - p I q-z, for the rest the arrange­
Oeo dicaWn, de capta Bajocensium civitste; ejusdem invectio in
Gilbertum abbatem Cadomi; item versus ad Odonem Bajocensem ment is the same.
episcopum, cum aliis.
r
1I4 : Versusrbythmici Godefridi priom Eccl. S. Swithini Wintoniensis,
(d) n. 13-16: the example PAX VOBISCVM ET I
de moribus et vita instituenda.
SALVS PAX, with this transcription above the
1I7r : Vita S.' Marire iEgyptiacre, per Hildebertum; versibus.
runes.
122?: Ejusdem episcopi versus de XII. plagis lEgypti, &C.

225
224 ... 19
 The runes, consequently, do not occur together with the The arrangement of the. runic .material is the. same in the
other alphabets, as in the Exeter manuscript. Fol. 65 1 may two codices ... It consists of three alphabets .and .an illustrative
originally have been blank, a calendar beginning on 65 v ; when example. Above each rune its value is shown by. a Roman
the scribe had no space left 'on fo1. 45 v to add the runes, he capital. At the end of the first two alphabets follow four
inserted them on 65 r • The runes are on the average 5 nun suppletive runes, inscribed" supersunt iste". (first alphabet)
high; they are as a rule well drawn, though they show some and "supersunt littere iste. lIII." (second alphabet). The
measure of ' cursivation' (I); serifs have been freely added. . difference between the two manuscripts lies only in the 'style
of the runes : Vit has obviously less understanding for the runic
style than Ex, which shows an almost epigraphical severity of
*** line. Vit imitates the smallest peculiarities of Ex, e. g. in the
T of the third alphabet, and in the 0 of oobiscum.
The close relationship of the' two manuscripts, or at least
of the two sets of texts which they have in conunon, is evident
even from a brief inspection. Whether Vit(ellius A 12) is a
direct copy of Ex(eter MS. 3507), only a detailed investigation ***
of all they have in conunon could decide (2). In the part
which I· examined there are mistakes found in both manuscripts, Phillipps MS. 3715 was described and edited in 1846 by Sir
e. g. dentes moti (= morti); Vit has many mistakes not in Ex : Thomas Phillipps, the then owner (I). During the gradual
Ex C Molaribus ... exprimituT: Vit solaribus; Ex F Dentes: dispersion of the Phillipps library it seems to have been sold
Vit Mentes; Ex Beta: Vit Beata; Ex I iota.aurissive aurum: but I have been .
unable to discover its present location (2). ~
~.
Vit ... arum etc.,. but Ex only few not in Vit and they are often
(r) Letter from Sir ThOMAS PWLLIPI'S, Bart., F. R. S., F. S. A" addressed
such as could be corrected easily, e. g. Ex K cappa prudentitia, to Albert Way, Esq., Director, comrm.mil:ating a transcript of a MS. Treatise on
Vit ... prudentia. For our purpose it does not matter which the preparation of Pigm4llts, and on varilius processes of the Dt1Corative ArtS
of the two possible forms of relationship (Vit copied from Ex, practised during the Middle Agu; written in the twelfth cent.u.Ty, and entitled
Mapp18 CfmJicula. In: Archaeologia 32 (1847), 183-244.
or both descended from the same prototype) .will actually be In the catalogue of the Phillippslibrasy, however, the manuscript is dated
found to exist, as the runic material' in th,e two manuscripts is 'aleC. XIII' (Cat.aJogus Lihrorum Manrucriptorum in Bibliotheca D. ThomlB
practically identical. PhiUipps, Bart. A.D. 1837. Impressus Typis Medio-Montanis 1837 sseq.,
. In tracing the origin of the two manuscripts, the attribution p. 4l!). Cf. also (all based on the Letter) : M. BBRTHELOT, Adalard de Bath
et to; Mappae CIl.l'lJiqda (Clef de to; petntuTe). Journal des savanlll rl)OO, 6r-66
to .Gildas may be a clue. It is proposed only tentatively in Ex, ATchioiogie et science, r908, 172-177.
and only with reference to Isidore's De natura rerum (marginal H. DIBLS, Die Entdecktmg des Alkohols. Abhandlungen der KlSniglich
note on fo1. 68r ); in Vit not only this work, but also Hraban'us . Preussischen Akademie der Wissenschaften. Jg. 1913. Philosophisch­
historische CIasse Nr. 3, 6 fl.
Maurus's De computo is attributed to. that British scholar. L. ThORNDIKE, A History qf Magic and Experimental Science. New York
Therefore Jones is probably right in assuming British in­ 1922, I. 4fi8, 765 fl., II, 22 f.
fluence (3). C. H. RWUNS, Studiet in the History of Mediaeval Science (Harvard Histor­
ical Studies· XXVII). Cambridge (M4ss.), 1924. 30 f.
J. SVBNNUNG. CompositUmu LuclmSes. Studien:rum Inhalt, :<rUT TeJCtkritik
(I) They are at any rate less' decadent' than Hickes's facsimile might tmd Sproche. Uppsala Universitets Arsskrift 1941 :5.
lead to suppose. . (2) Officials of the' British Museum to whom I applied for further infor­
(2) N. R. KBR, Salisbury Cathedral Mamucripu, 156 note 2.. mation could not tell me the date on which the manuscript was sold, nor who
(3) C. W. JONl!S, Bedae Opera de tmnPoribus, Il9 f., note 4. is its present owner.
226 '227'
...
Consequently the discussion will have to be based exclusively building, glass. mensurstion of altitudes, warfare, etc. may be due to Adalard
of Bath, "the greatest name in English science before Robert Grosseteste
on Phillipps's account (1). Being over a century old, this
and Roger BllCon" (I).
account does of course not give all the information we could
The runes are found on two of the last folios of the codex,
expect in a modem publication; fortunately it includes fac­
amidst material not registered in the table of contents. Chapter
similes of the two runic alphabets found at the end of the codex.
cclxxxviij AzuT quomodo molatuT (p. 2 ..p) is followed by a
The manuscript was described as " a small duodedmo volume of sixty-seven mnemotechnic distich on weights, and a runic alphabet,. with
leaves of vellum, written in the twelfth century. It appears to be perfect,
except a leaf tom out between pp. 64 and6s of the modem paging, and a little above each rune its name, and below its value; to the lower
cropping in two leaves. It was purchased in Paris, in x8z4. from the Rev. left, the inscription super su[nt] and three more runes (2). The
Mons. Allard, Cure of the church of Saint Eustache, and had previously outer margin had been trimmed, resulting in the loss of pan;
belonged to Mons. J. Rabaut •.. The character of the writing appears to be
that used in England or -Flanders in the time of Henry II.; but from an English
of the name of the I-rune and of the last six letters of the in­
word being used in the work the presumption will be, of the two nations, scription supersu[nt isteJ (and also, as we shall see, of ther une
in favour of England •.. The passage which quotes the English word is in for :l and the fourth suppletive rune). Next comes a table
chapter cxc. where the shrub" caprifolium .. is translated " goat tree ". This with symbols for weights (Assis, Deunx, etc.); recipe" for
is a singular circumsm'nce. and seems to indicate, as I said before, that the
author or the transcriber was an Englishman, for had he been of any other making soap (without title: Dum partes cineris quercini, etc.)
nation he would most naturally have tranalated it by the language of his own and a table of Roman numerals from I to· L. On p. 243 we
country. Moreover, in the _very next chapter, he mentions the herb" grening­ find the remark" (A written leaf has been torn out here)"_
pert ", a corruption, 1 suspect, of " greni:ngwert H, the Saxon p being easily
mistaken for a Pi (•..) which I consider an additional mark of his being an What follows ~ust consequently have been found on fol. 65 :
English author" (z). Stagni -;. ix, c(u )pri -;- ar(genti) vj.. -:- simulfunde, per figuram
The manuscript must go hack to a much older original (saec. IX/X), written arragab, etc.; a runic alphabet, with above each rune its value;
in Kent (3). The title INCIPIT LlBELLUS DICTUS MAPPlE CLAVI­
a Greek alphabet of a late date, majuscules and minuscules, '",
'>.
CULA is followed by some introductory. material and a prologue (pp. x87-9) (4).
The list of the chapters (Im:ipiunt Capitula, pp. I90-z) does not entirely agree with the values and the names of the letters (alpha, beta vel
with the chapters in the text: Cap. I-Zag correspond to Text x-z6x (with uita ... zita, ita, (t)hita, etc.). Of this page, too, the margin
numerous additions), after which come 3Z more chapters (26z-Z93). The
had been cropped. Then comes chapter cclxxxviiij : Ad vitTum
table of contents consequently belongs to an older ' edition'. The expanded
version of the Mappae Clavicula, II treatise on chemistry, alchemy, painting, incidendum. Since. there is no gap in the text of the Mappae
Clavicula notwithstanding the loss of a leaf, we may suppose
that the folios with the runic alphabets etc. originally belonged
(x) His comment on the runes has only histolic3J. interest (0. c., 186): to another manuscript; their having had a different format
" One of me most valuable entries in me book is connected with philology,
namely, the alphabet of Runes, if they are Runes, but which I am inclined would explain why part of the text along the margins was lost_
to think are Oscan, or very early Greek, and derived originally from the Perse­
politan, or Babylonian character", etc.
(z) T. PHILLIPPS, Letter, 183. Since the mention of the missing leaf is * **
found on p. 243 of the teXt (which covers pp. IC;l3-z44), it was probably tom
out between lois. 64 and 6S. not between pages 64 and 6S. The first of the two alphabets in this manuscript (Ph) is no
(3) The form of the Old English words (223 : Accipe grana matuTa arbon doubt related with the first in Exeter MS. 3507 (Ex) and Cotton
caprilom, hie est Ang1ice gatetriu ...Accipe 1Jerbam que dicituT greningpert) points
to II Kentish original of the ninth or the tenth century, cf. H. DII!LS, Die Ent­
MS. Vitellius (Vit) A 12 :
de&kung des Alkohols, 7 note z. (d C. H.HAsKINll. StuJ.ies, zoo A German version which must be rather
(4) In PHILLIPPS'S edition the pages of the manuscript are not indicated; close to the prototype of the Phillipps manuscript is found in Schlettstadt
therefore I can 'give only those of his printed text, except for the pages with MS. IlS3 blo, cr.,H. DlI!LS. Die Enkhcktmg des AlkolwIs, 6 note 6.
the runes, where I have tried to reconstruct the order of the material. (2) Probably on fol. 64'.

228 229
....
 Ph!P i3 'h'.db4 ·r.rY·9*e. Ht~ ~
b rch U b: Ex Vit retain a rather pmrutlve type. of b, in which

the two loops are drawn well apart; Ph has a form adapted

s~l~ g h txt· M r )K NH[I K\f L~ to a capital B of its time. The name berck may be explained

as bere written by a High German scribe.

n
'A BCD E F.G d: the forms of d a,r;.d m are hardly distinguished~
Ex shows

Phi W·;:t~ P~d·~i.R·s~1. :{:" hn.~ .) some sort of an attempt in the first alphabet, where tile.

m
cross strokes of do not reach all the way down the ver­

tical shafts, but Vit does not seem to have noticed this

~~Itxl
M
tNO~ hP S?
QR
R '15 Tt VnXh rY
hlZ detail. The name derku indicates that there may be a

good many intermediaires between the original. alphabet

and the version in Ph; it differs as much from d::eg as

Super sunt fste from porn; yet it may perhaps rather be derived from the

Ph ~ ~ b­ latter in the form dhorn; cf. dkron in Paris MS. 5239.

FIG. 34­
~rt}Z Rrt e: To explain the form of the name eg, we probably have to

start from ek, cf. Berne MS. 207 egck and lege.

I: unless we derive leu from OE. leo (abstracted e. g. from

The second alphabet in Ex and Vit shows almost the same the gen.leos), it could be either ON. (cf. Leyden MS. Voss.
forms: lat. Q 83 :/eu) or OLG. (eventually OHG). The latter
ABC D E f G H l l < l MNOPQ. is more probable than the former, as there are no other .~~~
traces of ON. influence (tir is not conclusive) ..
~g
ht4MY t 1'1 I\f' ~ M}~~ ~
In Ex Vit the rune has only one lateral stroke, in Ph two.
Ph must then be derived from, another exemplar than the
R S T V X Y Z Super sunt littere fste .1111. common ancestor of Ex Vito

R1 tf1 h~ hf . ~ ~ t ~ FIG. 35
g: the obviously corrupt name may perhaps be read geuue

or geue, cf. geuo in Munich MS. 19410..

The rune itself appears in two different forms : with long

Therefore we are fully justified to discuss these five alphabets crossing strokes in Ex Vit Ph,' with short ones iIi the second

together. . alphabet of Ex Vito Although the name in Ph is that of g,

a: the rune chosen to represent a is a, as appears also from
the form is rather that of ~he· variant j (gear liar).
the name OK in Ph. ,This is the same form as that found k: Pllrt. of t~e name. may have been illegible in one of the an­
ce~tors of Ph (ke[gilj? be[gll? ke[iI}?) Damage. to the
in Munich MS. 19410, arid it wy have the same origin.
exeml?lar may also account for the absence of.a name for
(.) The following footnote applies to lo.... in I. I : "The ~argin of the
MS. is cut" (Letter, 241). PIDLLlPPS gives the runes and their values in
the .next r::une, i ..
facsimile, the names in type.
k; formally this rune is the x-rune (eolkx), but the name cer
(..) In Ph the rune for % is lost, together with itS name; the fourth extra
rune (Ie in Ex Vit) is also lost, and of the text on the suppletive runes only (cf. ker'in Munich MSS. 14436 and 19410, Vienna MS. 751)
tnlpe, su[•••J is left. is decidedly derived from that of another rune. There
:.t~o
. 23 1
 are two possibilities : either OE. gar was translated into have written it above the variant for u, etc.; or else this x
OHG., giving giir or kiir; or else OR ger (= WS. gear) goes back to It written upside down (cf. the 'double'
was gennanized to her. The latter explanation is the forms in St. Gall MS. 878 and Munich MS. 19410) or to k,
more probable: the j-rune had been given the place of g, the left lateral stroke of Which had become partly invisible;
hence its name remained unused; in a Gennan garb it fonnally this would account for the lateral stroke (in all
could serve as a name for k. Since g does not occur three manuscripts) not ending at the stem, but transecting
among the suppletive rune!', it is rather likely that it was it. The second solution is the more probable, especially
not found in the underlying fuporc. if we assume that the rune was an It turned upside down.
I: the name in Ph may have been lagu or [ago.. This latter assumption is rendered rather obvious by the
form of the k in the alphabet. In this way :x: is restored
m: cf. d. to its original value. We have here a good illustration of
n: net is probably an adaptation to OHG. phonology of OE. the way in which k and x could be confused. A rune k
ned. having been turned upside down, hesitation arose as to
the actual form of x. A middle course was chosen : for x
p: the name perd preserves the OR (Anglian) form. The
two forms were given, the regular x and the same form
adaptation of the dentals was not performed in· a uniform
upside down. Later on this led to a shift of the values
way.
as indicated above. The name of the rune is unique. and
q: here again rune and name do not belong together. The may have been invented ad hoc : it looks as if the resem­
rune is of course the UI-rune, and was chosen for q on blance to the c-rune had led the alphabetizer to derive
account of its similarity to Roman Q. The name cui a new name from cen : xen. ".
~.

is probably to be connected with the form CUT found in a

number of manuscripts. y: the x-rune has been combined with a name uir, which is
either an attempt to render OE. yr, or an archaic OE.
r: rat, too, is adapted from OE. rad. spelling for the latter (I). In Ex and Vit the y-rune is
s, t, u,: all three runes are accompanied by the regular OE. found under z; in Ph it has been cut away.
form of their names.
z: in Ex and Vit the rune which takes the place of z is a
x: one might be at a loss to identify the rune chosen for x, variant form of the regular OE. y (i. e. u with subscript i).
if it did not appear that from here on the runes or the The first of the suppletive letters il; actually the z of the
names and values ~ave been shifted one place: the x-rune alphabet. It looks rather: like a Roman capital Z, at least
takes the place of y, the y-rune that of z (in Ex Vit, but in Ex and Viti whether it was really derived from that
missing in Ph), and z is found among the suppletive letters. letter, or whether it is to be connected with 3, only the
The symbol found for x looks rather like a c-rune (espe­ name of the rune could have made clear; but that name·
cially in Ex Vit) , but I believe its origin lies elsewhere: has been lost together with the y-rune in Ph. The 'second
it is either a variant of the u (Ph would make this rather alphabet in Ex and Vit has a cursive z, which tends to
probable, but its forms are not very trustworthy), which prove that Roman Z, no~ 3, is behind the last letter of the
would at the same time explain why the names and values alphabet. .
were shifted one place: instead of writing x (and th~ name
of that letter), above the rune acting as x, one scribe may (1) E. SlI!VERS-K. BRUNNER, Altenglilche GTammatik, § 94 A.

Z3 Z Z33
>,
 The three remaining runes (the last of which is also missing (4) It is not possible to decide at what date this alphabet(at
in Ph) are g, ., and tE. The alphabetizer apparently had no least in the Ph version) was reimported into England. The
use for them. The fuporc which he used' may be recon­ shift of the values in all tlU:ee manuscripts and the rather
structed as follows (of the g-rune only the name was retained) : severe distortions of the names in Ph prove that their

(W)
for: f
r II t> ~ ~ h X- ~~I'~t.,~
u - 0 h 9'
reg - n i z1. p
common ancestor was written by a scribe with hardly any
knowledge of the runes; otherwise he would certainly
have corrected this material according to his own con­
ception of a runic alphabet.

for: T
rth ~ tt &b MM~I ~- M
s
~ ~ ~ ~ 'P
d q a.
em, y k ~
(5) As to the relationship of the ~ee versions, one may safely
state that Vit is' either derived from Ex, or both are de­
scended from a common ancestor. Either Ex or the
'FIG. 36 common ancestor gave up the rune-names, which Ph has
kept; Ph also has the correct form of f. ,The second
From the discussion of the details some general information alphabet of Ex and Vit may originally have existed in Ph
may be derived : as well j in that case,it was lost when a leaf was tom out
(I) The fu}:>orc used by the alphabetizer probably originated between fols. 64 and 65 of the latter manuscript. It
in Anglian territory (berch, eg, cer, net, perd) and may have shows such triRing differences with the first alphabet
reached the Continent at an eady date (sigil, UiT). (short cross strokes in g and a cursive type of z) that one
(2) The alphabetizer may have worked on the Continent; is surprised to see the scribes of Ex and Vit take the trouble
several peculiarities of his alphabet can only be explained to copy it. ;.
as being due to theinHuence of OHG. phonology. There
may be some remote relationship between his work and the
***
alphabets in Munich MSS. 19410 and 1#36. This High The third alphabet in Ex Vit and the second in Ph are again
German influence appears only ;from the rune-names in very closely related :
Ph, and it is of course a priori not impossible that the ver­
sions in Ex Vit never left England. But on the other
a.b 'c d e f g' h f kim
~i~} ~ ~ hp'Mf Lf,t,}"Kr M

hand 'the three versio~s have in common such striking

peculiarities,' that one can har~ly assume that Ph came
from the Continent without implicating the same for Ex
and Vito The neighbourhood of a work by Hrabanus
Maueus agrees ~en with this view. .
Pht' B,~ ~. M' r i,~lXr!X1
. '

(3) He proceeded rather freely, and not always with much n 0, p q r 5 t u x y. z

understanding fqr the typical features of the runic system;

thus he separated the names from the runes as. he ,thQught ~~}i- .~ ~ ·th t 1 t /\ >f{ t.y b

Ph{)\~ <hR L-r·1'A~~ ~.

fit and changed the values, sometimes on purely external

features (e: g. (2 takes the place of q).
(*) Cf. p. '188, note (*). , . FIG. 37 ' .
234 .... 235
· This alphabet differs from the first on the following points : Voss. lat. F. 12 8. Formally it is identical with:x. Per­
d g h k 0 q x y z. The last three runes come apparently in haps we may agairi. suppose that at the origin there were
the right places, as z is probably a cursive z (Ph)' or derived two forms of X t one of which was then interpreted as y.
from such a letter. The u-rune shows a slightly different, z: is derived from a Roman Z.
perhaps more archaic type. But differences with the first
On the whole the evidence for Continental (High German)

alphabet are found on practically all the critical points (d, g, k,

influence is smaller in this alphabet than in the first. The

etc.), therefore we can say a priori that this alphabet is the

absence of the names probably helps to convey this impression;

outcome of a different alphabetization. The curious addition

but the agreement of the three manuscripts is also an irnportant

to the right of the stem of r in Ex is also found in Vit : it is

factor, and in this case Ph may force the decision. One may

accidental, and just one more proof of the close relationship

of course raise the question whether these runes were not

of these two manuscripts.

revised after having returned to their homeland. For the

d: the case of d is a simple one : as we found in other manu­ first alphabet such a thought may safely be rejected, but for the

scripts, the distance from porn to dorn was small indeed, second, with its somewhat archaic features, the possibility may

and so was that from ., to d. be left open. Some connexion with the isruna fu pore must be

g: the type of g is that of the isruna group and of a couple admitted, though too little seems to be known of the Continental

of alphabets derived from that group. See also below. background of these manuscripts to allow of a definite con..

h: this h reminds one of the possible variant in Domitian A 9, clusion.

which may in fact be a ?I-rune, as in the isruna group. We

probably have to assume that the name of this rune at
one time was hine, cf. that form in Brussels MS. 93II-93I9
10. Vatican Library, MS. &ginensis lat. 338 (saec. X ex.). .,
".
and in Nemnivus's alphabet.
The runes of this codex were discovered by G. H. Peetz, who
k: the place of k is taken by the g-rune. This probably .communicated them to W. Grimm (I). They were noticed
requires an intermediate stage where the name of the rune again by C. Greith (2) and edited also by an anonymous author
had become car (cf. caar in Munich MSS. 1+4-36 and in 18.55 (H. Massmann?) (3). The codex consists of two pans,
19410). Consequently, it is rather likely that this alphabet written at about a century's interval but apparently in the
too was created on the Continent. same region. They may have been bound together before they
0: Continental influence may also account for the value of came into the possession of Alexandre Petau, the French
the m-rune (OE. mpel, epel = OHG. otlil, odal). humanist. Afterwards the manuscript belonged to the library
q: for q the alphabetizer probably chose the k-rune (cf. of Queen Christina of Sweden, who bequeathed her collection
Vatican MS. Regin. 338). to the Vatican Library.
Plain modem red leather binding with the arms of Pope Benedict XV and
x: in this case Ph seems to present a more archaic type of x of Cardinal A. Gasquet. The following data apply only to the latter part
than Ex Vit; at any rate it agrees closely with the x in
Munich MSS. 1+4-36 and 19410, as well as with that in (1) W. GRIMM, Zur Litteratur tier Rwum, 21 f. Kl. 8chrijtm III, 108.
the fuporc of St. Gall MS. 878. (2) C. GlUllTH. Spicilegium Vaticanum. Beitriigtl:ntr niihem Ke:nntnifs der
Vatikani$Chm Bibliothek filr deutsCM Poaie des Mittelalters. Fmuenfeld. 1 8 39.
y: 'this type of y is not well established for England, but is 45 f.
found in the De inventWne alphabet and in. Leyden MS. (3) Anzeiger fUr Kunde der deutschen Voneit 1855. 4. 77 ft.

23 6 237
....
 of the codex (I) : Pan:hment of poor quality wi~ many repairs, ca. ISo X
150 mm; written area 140-145 X 80-85 mm; 20 lines to the page.
(2) Another ,Hebrew alphabet with the inscription (fo1. 91v):
Composition of the codeX: A ,fols. 1-63 8 qUatemions, in the last 'Iudaicas uero quibru / & iam nunc utuntuT iu~i / isdem
of which one folio ia missing; B fols. 64-126 = 8 quatemioll8 numbered uocabulis eadem I uirtute forma immu/tata ZV'supra
on the last page; the first two folios are only " miserandae laciniae "; of fol.
loS" only a smaIl strip is left (z). . .me/morauimus & sunt istCf. / form~ qUQdmodo utuntuT.
Part B was written by an ,Engliah scribe, but probably on the Continent The. letters.A Aleph to m mem in one column on fo1. 9 1 ",
(Northern France) and in Carolingean minuscules, about the end of the tenth continued on fol.92r from N nun to T tau, also in one
century. About a C1mtury later a reader, also Engliah, added a number of column~ same arrangement as for the preceding alphabet
small items: an OE. charm against fever (fol. 91r), a note on bloodletting
(fol. IlIv), s prayer or antiphon for travellers (fol. II4r); he hitnself, or a (from left to right: values, letters, names).
.contemporary, added a number of short notes (3).
Contents : (3) A Greek alphabet with'the inscription (fo1. 92r) :
A :' Leges Ribuaria et Salics (saec. tx ex.).
Haec sunt caracteres / grecas iuxta numerum / XXllll.
v
B : fol. 64 J\. fragmentary metrical calendar of English origin. the so-caned senioru1!' sic eorum / nomina in ordine I posit€C form€C
manyrologium Bedtu.
ist~.
fn v Amalarius (?), Eclogae de offiCio misSQe.
Then follows the alphabet in two columns (values, letters,
88v
De decem praeceptis vel de decem plagis.
90' Horologii descriptio (on reading the sun-,dial).
names) from A alfa to z Z z&a and h H h&a to 0 w o.
9 1v Sevenalph:abets (cf. infra)

93 v Excerpts from a sacramentary. '

(+) A Chaldaean and Assyrian alphabet with the' inscription
96 v
Pontifical benedictions. (foI.9 2V) :
10" Liturgical items (nuptial benediction. etc.)
Haec'sunt caracteres que caldei I & asyrii utuntuT &
lOSv Breviarum Psalterii.

II5 v Monastic hyrnnary.

ante/quam XXlIn esse manifesta est I sicut & latini ita

in ordine I formf,l ita est.
The alphabetic material on fols. 91 v-93 r consists of the "'I:'
following items : The alphabet shows the same arrangement as the preceding
ones (values, letters, names); it is written in two columns:
(1) A Hebrew alphabet with the inscription (fo1. 9 1 v) : aalim - m morm, and n nithoin y sophy.
Haec sunt litteras hebreorum I iuxta. numerum .XXII.
(5) An Egyptian alphabet with the inscription (fo1. 92") :
librorom ueteri testamentorum & litteras I ipsas matrem
Haec sunt caracteras egyptorum quas / utuntuT ita
esse omp,ium littermum I quCf. per moysen data sunt /
sunt in ordine positf,l si-;:ut / & latinf,l. form€C autem
XXII. eorum forma ita est.
ist€C sunt.
The letters are written in tw9 vertical columns, with to the The letters A atohin ­ u ur follow on the saf!1e page with
right of the lett~rs their names, and to the left their values; the usual arrangement; the remaining four x xyron tau
in the first column a alep -:-ch caf, in the second llamech on fo!' 93r.On fo1. 92 v a later hand repeated the' Chal­
- th Thau.
daeo-Assyrian' k to the right of the original character.
(r) For full details see A. WlI.MART; Cridi.cel II, Z58 ff. In the lower left ffiargin an untrained' hand wrote the Latin
(2) There are two different foliationa : A 1-61, B 6z-1:23. in ink in letters bcdefghik, and perhaps' also the two scribbles to the
the right top comer of each recto page; this places the runic alphabet on
fo1. 90 ' (sic GlUMM; GIWTH is not quite clear; the author of the paper in
left, of the text on the 'Egyptian' alphabet.
' '

the Anzeiger also has the alphabets begin on. fol. 88v , the runes on 90'); and
(6) A runic alphabet inscribed RUNAS, without further
A = 1-63. B .= 64-1:26, stamped in the 'right bottom comer of each recto
page and used. by WII.MART, whom I have followed here. introductory text, in two columns a aac p pear and
(J) W.'STOIOlS, The Anglos_ hale and Glosus, 144. q yymoth ­ z gaaT.
23 8 239
'"

(7) After the .3'-rune one line blank.; then comes one more Assyrian and Egyptian alphabets which may have led to the
alphabet inscribed NORMA, still with the same arrange­ inclusion of these two). The introductory texts even allow us .
ment. The characters for a - k form one column, to decide to which of the two De inventione versions the Vatican
those for 1- .3' a second column. Below the .3' an untrained compilation is most .closely related : the fact that Hebrew is
hand made four -more or less successful attempts to write , the mother of all languages' (in the first paragraph we have
non continebit; it also repeated part of the name of the .3' probably to read linguarum for litterarum) and the expression
just above the original; finally the same or another hand per Moysen data sunt show that the Vatican manuscript comes
repeated the name of the o-rune between that rune and closest to our type B (cf. p. 349 f.); this may be due to the use
the name as written by the first hand, and the p-rune to of the same sources rather than to direct relationship. At any
the right of the original. rate this finding agrees well with the localization of the
Vatican manuscript in France.
The first Hebrew alphabet is closely related to that found It is rather tempting to read Normannicum or NormannOTUm
in the De inventione treatise. The second seems originally to for NORMA, the name of the last alphabet. But the origin
have implied a better knowledge of the Hebrew characters, but of this alphabet is quite obscure, and all one can say is that it
has evidendy suffered much in the course of being copied. has no connexion with the runes, even if a couple of names
Not only the forms of the characters in the two alphabets remind one of rune-names: aschot : alse; berit: bere (beriC?);
show important differences; the names too differ: (I) alep .3'at menu: mann; nut: ne(a)d (in Continental alphabets also nod,
teh lamee samet Thau: (2) aleph zai t&h lam&h samec tau. not); Sutiltu: sugil. H. Harder (I) has made an interesting
The Greek alphabet too shows traces of having been copied if not fruitful attempt to prove that this alphabet was in fact
over and over: the name of E is he; the value of (J: Tb; the based upon a runic alphabet, the names being distorted rune- ~.
name of I: ioth, of A : lo.uda, of 1J': spi. The next two al­ names. By eliminating or shifting a number of letters he
phabets are of special importance for determining the back­ obtains a more or less runic list of names, whilst the eliminated·
ground of the collection. They connect it with a group of letters, read in the right order, give the following warning to
manuscripts mentioned occasionally before and apparendy the reader : Tuto te peto I tu totum muta I ne flde te jappet. But
proceeding from a collection made up in the eighth century or to obtain this more or less appropriate and comprehensible
even earlier. In an appendix to this chapter a brief survey of warning, Harder has to 'reconstruct ' such rune-names as flu,
these spurious alphabets will be given. hail, lo.ku, quon, not found in the other versions of the • Hra­
- The text of the introductory notes is due to a very poor banic' alphabet from which he starts (2). There can be no
Latinist, who stumbled even over the simplest sentences : he doubt th~t this alphabet may be due to what Harder calls
repeatedly used the accusative for the nominative, wrote (I) H. HARDER, Zur Frage der hrabanischen Alphabete, 188 f.
testamentOTUm for testamenti,joT'1nll. ita est (for .,. sunt), etc. The (z) HARDER obviously wanted to prove too much, a9 the following extracts
paragraphs on the two Hebrew alphabets hardly make sense. show: " Als Urheber des Schetze8 hat Tuto seinen Namen nicht nur in dec
ersten Zelle versteckt, aschot cecut, sondem auch in der dritten· nut oto<:.
It looks as if the compiler had tried to condense a more extensive Auszerdem findet aich die erste Silbe seines Namens noch in terut, sutiUu
version, and in doing so had eliminated a few indispensible (why not twice here : sutiltu t), tucal, zepput... Die Form foppet setzt eine
words as well. This longer version was based upon the same scherzhaft gebildete Iateinisch-deutsche Mischvokabel foppare voraus ...
Hmweisen m6chte ich nocb darauf, dllSZ der lateinische Spruch in jeder Zeile
sources as the De inventione text, as a comparison with the texts Stablmg enthilt •.• Obwohl wir bine regelrechten germanischen Stabreim­
quoted in the Appendix to Chapter IV shows (in one of those verse vor una hahen, wird docb der Spruch vermutlich in einer Zeit entstan­
texts the onginal compiler also found mention of the Chaldaeo- den sein, als der Brauch des Stabreirns noch geiibt wurde " (0. c., p. 189).

:iIIO 241
240
...
" Der Spieltrieb eines Monches"; the inventor may occasion­ runes, so that it looks identical with the scribe's minuscule h.
ally have sought and found inspiration in a list of rune-names; The name een is the normal OE. form.
but Harder's reconstruction has no serious basis, for a glance d: this rune is clearly distinguished from the m, the inter­
at the other alphabets shows that the' Norman' names may secting strokes in' the former being placed lower. The
very well contain mistakes, and one or two would suffice to name probably retains the OE. spelling (= dtl!g).
upset the carefully conCealed warning (I).
e: the name eel. is rather surprising: OE. eh should be
The runic alphabet is in many ways superior to the surround':
expected to have short e. But long i ffil!.Y have developed
ing alphabets; it cannot be far removed from the fuporc
in case forms where intervocalic h was dropped, especially
prototype. On the other. hand it is not very likely that it
. in Anglian territory (I). Alternately one might think of
should be an addition by the scribe himself: we shall see that
an error for eel. or for eoh(2).
a couple of mistakes at any rate prove it to be a second hand
copy. I: the rune is inclined to the left. The name leI. is Anglian.
The inscription RUNAS is probably to be 'understood as a' The difference in spelling with the preceding eeh (WS. eoh :
Latin accusative plural, which reminds us of the plurals en­ leoh) may be due to various causes: some degree of incon~
countered in the introductory texts (2). The alphabet itself sistency is to be expected in a list of names which ~d been
shows the following forms and names ~ copied several times and was hardly understood by the
scribe of our version.

~aa..a.c ~we
b cen.
-he i><td M;rr x~t:~k t. rhk IT
dug eeh feh ge05 hegil liS ca.lc 1a.9°
g: the name' clearly reads geos. ~s is probably a mistake
for geol. which in its turn may be a Northern form for
geal (3). A form without ending is known especially in ;,
[><1m~n ~o ~p c.L 9 Rr ~s tt f)ut Ay~% the accusative of some Northern texts. Holthausen
reconstructs a form geolu (4), bU:t then a mistake for gell)
moun need oos pear yymotn ra.M1 sign tu ur ilfh fr gUl" might. rather have resulted ingeol > geos.
FIG. 38
h: in hegil e points to a non-West-Saxon origin, whilst final
a: the name aae keeps the OE. spelling, with double aa ~il (for ~el or -1) is an archaic feature (5).
indicating the quantity (aat occurs six times in the Epinal. i: in this name too double spelling indicates the length of
Erfurt and Corpus glossaries). the vowel.
b: the OE. form berc without breaking points to an old proto­ k: bo~h the rune and the njillle calc are pure OE.
type, or to an Anglian source (3)·
I: the. way in which the I-rune was drawn shows that the
c: 'the rune is drawn in a more diffident way than the other

(1) There can be- no doubt that these' Nonnan' names were not created (1) E. SIEVERS-K. BRUNNBR, Alterrgliuhe Grammatik, § 129. :II and note S
by the scribe of the Vatican manuscript: ptli and xri. are no very likely inven­ (p. 112 f.).
tions; the acrostic principle has been abandoned in six cases : g "'" hilod, II = (2) This is .the solution proposed by F. HOLTHAUlIBN. Alterrgliuhe Runen­
'IUZI7Ii!n. Perhaps one might also think of the merging of e (value) + ell (name),
~erut. l = _ . u = Wel. ;1;1 - symol, y xri..
(2) OE. riin wall probably a fem. D-stem. which could be easily Latinized cf. the ne:rt:name fen.
to a Lat. a-stem. liB *nm- fonna fem. nouns in all Gmc:. languages, .there (3) E. SIEVBRS-K. BRUNN1!R, AlterrglUthe G,-ammatik. § 3S !,.. 1.
can be no question of interpreting f'Ilnal as an OE. rnase. plur. (4) F. HOLTHAUSBN. Alte7lg1isthe RuMtmamen.
6) E. SIEVERS - K. BRUNNBR, Atte7lgliuhe Orammatik. § 120 A. I. (5) E. SIlMIllS-K. BRUNNBR, Atte7lglUthe G,-ammatik, § 15:11.

242 243
...

t.

 scribe had no great understanding for this type of writing. s: the unrounding of y to i before palatal g occurred at an
Cf. J, t, u. early date (1), hence sigil < sygil; for -il cf. h.
m: the name moun is in all probability mistaken formonn. t: the name of this rune was certainly intended to be tu, not
This moon is the more common variant for mann in the tii. Tii, however, was in all probability the original form
Anglian area, and also in Wessex and Kent till the ninth (for the spelling cf. TUg: Mars, Martis Epinal-Erfurt gl.
century (I). 663, Corpus gl. 1293) (2).
n: here too double spelling indicates the length of the stem u: this and the preceding rune reveal the scribe's ignorance
vowel; ned is the usual form corresponding to WS. nead, of runic writing, or at any rate his lack of training; ur is
nied. the normal OE. name.
0: the name cos, repeated by a later scribe, is probably a x: The vocalism of ilih is comparable with that of ilix in
mistake for oos (cf. Trier MS. R III 13 cos, Vatican MS. Brussels MS. 93II-9319 and ilcs in Vienna MS. 795 (these
Urbin. 290 eos); same spelling device as in aac, iis, etc. two being genitives *ilih-s, *ilh-5); I cannot see any reason
p: it is difficult to account for the form pear, as the rune-name for reconstructing a name ilch, as Holthausen seems to
itself has a somewhat uncertain status. An original propose.
*peor[d] seems· to be the most likely starting point. Cf. y: as in ngil, the y may have been unrounded to i. But since
perhaps geol (2). this change is against the acrostic principle, ir will rather
q: neither the form nor the name of this rune have parallels be a spelling variant for yr. The rune itself shows a
in other runic alphabets, except that the same form is curious malformation.
found in Leyden MS. Voss. lat. F. 12 8. von Grienberger z: here I must refer to other cases where g or its name appear ~,
explained the name yymoth as a mistake for *queorth (with in the place of z. In the runic alphabet of Munich MS.
suprascript r), which enabled him to connect it with the 19410 and in the so-called' Arabic' alphabet of Munich
OE. rune-name C'fJJe0r4 (3). There is, however, a far more MS. 14436 we find a rune-name caar (in the former it is
likely explanation. In the ' Chaldaean' and 'Assyrian' written ca \a' r). In the former of these two alphabets this
alphabet we find a character for q that is identical with this caar is inserted between the names of the a-rune (ag) and
so-called rune; its name is quimot. In the corresponding the b-rune (beTic); no corresponding rune is written below
alphabet of Cotton MS. Domitian A 9 the name is qymith, it, but at the en<1 of the alphabet (1. e. taking the place of z)
in Munich MS, 1#36: quimit, in Avranches MS. 107 we find a k (or a reversed x) without name. In the' Arabic'
quimithi. There can be no doubt that the pseudo-runic q alphabet the a-rune is 'followed by a g with value a and
is in fact no other than the pseudo-Oriental character; . name catlr; in this alphabet there is DO .a'-rune at all. To
yymoth is a mistake for qymoth. these anomalies we may possibly add the name .a'ar in
r: the name has the usual double spelling of the stem vowel Vienna MS. 751; perhaps also the use of i fQr k in Exeter
(cf. Corpus gl. 129 alfleus: streamTatld). MS. 3507, etc. The assumption that at one time the
gar·rune was chosen to take the place of z seems to fit all
(I) E. SnMllIs-K. BRllNNlIR, Altenglilche Grammiztik, § 79.
(z) The mixing up of ea end eo may be an additional indication of I. Nor­
thern origin: E. Sl1M!R8-K. BRUNNIllI, Altenglische GFant:ItI4tik, § 3S A I. (I) E. SII!VER!I-K. BRllNNlIR, Altenglilclle GrammaUk, § 31 A. z.
(3) T. VON GlW!NBJ!RGI!R, Die cmgelsiil:hsischm nmmreihen, 34; d. (zl Hence it is not necessary to reconstruct a fonn tir, as F. HOLTHAUSEN,
F. HOLTHAUlIBN, A.lUnglilche R _ . Altenglische Runemumum, proposes.
2# 245
...
 these cases equally well, but it is not so easy to find the the' other alphabets which use gaT for z connected with
reason for this choice. We may first think. of a sort of the solution offered by the Vatican alphabet? In other
mechanical accident : the insular g of the rune-name gar words, do these three lists go back to one attempt at
may have been interpreted as a form of z, resulting in the alphabetization? One glance at the two other runic
name zar and the transfer of the rune to the last place in , alphabets and at the Vienna name-list shows that they
the alphabet. Or else the rune g, not having found its represent independent alphabetizations. So we must
way into the alphabet, was used for z (to which no rune conclude that several scribes came to the same result,
corresponded anyway); the fact that g occupied one of perhaps (not to say probably) by different ways (I).
the last places in the fullOrc may also have played some
part. The other solution was as follows: OE. gaT (or
gaar, indicating the length of the vowel, cf. nahogaaT * **
'terebellus' Corpus gl. 2002, nflbugaar • id.' Leyden There can be no doubt that the scribe of the Vatican manu­
gl. 196, gaaTieec ' al[l]ium' Corpus gl. II3) may have been script, even if he was an Englishman, knew very little about the
written CoaT (I) by a scribe who spoke an Upper German runes. His characters show clearly how uncertain he felt in
dialect (cf. cast = gast). But in OHG. spelling c also handling them; the names 'geos, moun, cos, pear prove that his
rendered the affricate (usually written .Ii), mostly before knowledge of the language was not scholarly, as could already
e or i, but also elsewhere (2); and thus CoaT could be be inferred from his Latin text. He seems rather to have
interpreted as the name of the z-rune. This explanation copied mechanically, and probably not without adding a couple
would account for caar in the two Munich alphabets as of mistakes of his own~
well as for zaT in the Vienna list of names, but precisely On the other hand it is clear that the underlying fuporc must
the Vatican alphabet o,ffers a difficulty at this point: the have been far correcter and quite archaic; it may easily 'have ;.
value of the g-rune is z, but the name retains the original been from one to two centuries older than the Vatican manu­
OE. form gaaT. With the first solution this difficulty can script. No' other list of rune-names uses consistently double'
be avoided.
spellings to indicate the lengtliof the ,:owels (aac, iis, need, *005,
Here we have to choose again between various possi­ Taad, *tii, gaar); only older OE. manuscripts use this device
bilities. Either the scribe of the Vatican alphabet restored with some consistency (2). As a matter of fact no other list
a zaar or CoaT of his exemplar to gaaT; but this is not very of rune-names, neither fuporc nor alphabet, presents such
plausible, for then he would certainly have corrected geo! archaic features. This seems to have escaped Harder's notice
and pear too; and the other names show no such traces when he tried to normalize the rune-barnes. There are some
of Continental Germanic influence as would justify a ' indications that the fuporc came from Anglian territory.
hypothetical form caaT or zaar in the exemplar. Or else Apart from z = gar and q = yymoth, the alphabetizer seems
the gar-rune took the place of z because the' alphabetizer to have met no difficulties. The solution he chose for q pr(,)ves
found no other solution. . Then the next question is : are that the alphabetization was performed after the other alphabets
had been collected; for there he found a substitute for the
(I) The use of g for k would then represent an intennediate stage.
(z) W. BRAUNB, Altlwchdeutsche Grammatik, § 159. note a (p. 141): .. nur
(1) The fact that the Munich manuscripts do not connect the name ca(a)r
lilelten, in uncom:ct geschriebenen glossen u.dgl., begegnet c = !If (auazer
yor I. I) "; one would prefer a statistical survey of the c-apelling to this censure with the rune i indicates what devious paths the alphabetizers may have
of OHG. scribal habits. chosen. '
(a) E. SII!VBIIS-K. BRUNNl!.R, Altenglische Grammatik, § 8.
246
247
....
 IDlssmg q-rune. It is quite possible that he worked on the Fol. 39 v contains liturgical texts with musical accents; the
Continent, as the collection of alphabets enjoyed some popu­ first five lines are clearly visible, but the rest of the page has
larity there. From the poor state of the texts and the rune­ much faded. The scribe who wrote the upper part may
names we may infer that the Vatican collection is several also be responsible for the two alphabets in the margins (1).
removes from the original. The fuporc used by the alpha­ The runes extend from the very top of the page to I. IS, i. e.
betizer must at least have had thirty runes (fig. 39). a height of ca. 110 mm. Their size varies from 2 to 5 mm.
The letters of the Roman alphabet do not exactly face the
rll-t!~hX-~ll--r:'t' corresponding runes : A faces b, Z: s. Yet there. can be no
doubt that the two alphabet~ belong together; the discrepancy
for: f u - 0 reg - h n· r - - p x seems simply due to a miscalculation of the copyist. He may

~1'BMM~-H-~-A-th)8(
have inserted the runes after his interest in strange alphabets
had been aroused by the text on 39r •
for: s t b e m I - d - a - y - k Z The runes are written in the outer margin, very close to the
FIG. 39 edge of the page. Frequent handling has soiled the margin,
and as a result the runes are somewhat obscured. Their
II. Paris, Bibliotheque de ['Arsenal, MS. II69 (saec. X/XI). reading, however, has hardly been impaired:
,
a. b c d e f 9 h I k I m

~B I NMr*.t~I~~

Thus far this manuscript has been studied mainly for its
binding, its miniatures and its liturgical contents. The runic
alphabet, though mentioned in the catalogue of the Arsenal ~>
~.

library, seems never to have been edited. It is of special ,n 0 p q r stu x y z

N~~~'t'R'11'()*~
importance because it presents features which will return in
most of the late TUnica manuscripta (Mandeville, Wyss).
The binding consists of an ivory plaque (saec. III) sawed in half. The . FIG. 40
fine parchment is well preserved; 57 folios are arranged in 7 quires: 4 IV
(1-3 2 ) + II (33-36) + 2 V (37-56) + 1 fol. Format: 167 X 60 mm (written As soon as we assign a value to each rune, we are struck by
area on the average 130 X 45 mm); 20 to 22 lines to the page. Written by an anomaly: beginning with h all values have been shifted one
several hands; between 996 and 1024 as appears from acclamations for Robert, place in the direction of a. The explanation is obvious: there
King of France (996-1031) and for Walter, Bishop of Autun (977- 1024);
possibly in Autun. • are two g-runes. In the manuscript the two alphabets are so
Contents (I) : far apart that the anomaly becomes apparent only upon closer

I
fo1. I' Incipiunt tropi cum laudibus: the tropes for the liturgical cycle, inspection. It is quite possible that the scribe was not aware
from Christmas to Michaelmas. of it; he was probably copying from an older exemplar; if he
39' (inserted in this cycle) a brief extract on music; Latin written had any knowledge of the runes, it must have been slight
with Greek and pseudo-Greek characters.
v
39 in the left-hand margin, a runic alphabet; in the right-hand margin
indeed. As a result of his ~versight, there is no rune for z.
a Roman capital alphabet (cf. infra), both written in one vertical row. It was probably left uncopied after the number of twenty-three
49' a fragmentary prosary.
(1) Mr. j. 'BoUSSARD, Conservateur-adjoint of the Biblioth~que de l'Arsenal,
(I) Catalogue gbll!ral ••• BibIiothAque de I'Arsenal II, 320. Cf. kindly informs me that the runes are probably written in the same ink as the
L. GAUTIER, Histoire tk la pohie Iiturgique (paris 1886), 126 f. upper five lines of the page.
248 249
...
runes, corresponding to the twenty-three letters of the Latin ways: either this y-variant was used for x because the
alphabet, had been reached. Of course we cannot decide· regular x, having been mixed up with k. had received the
whether this omission is due to the scribe of the Autun manu­ value q; or else it is simply a doublet for y (cf. next).
script, or whether he found it in his exemplar. In the following y: the normal OE. V takes the place of y. We may then have
discussion of details the values are tacitly restored to their doublets, just as in the case of g. But here we cannot
respective runes. know for sure whether they were originally intenfied as
a: the rune stands in a rather worn corner of the· page, but variants of the same rune; in the case of g the inclusion
there can be no doubt that a was intended. of a second rune caused the values of all the remaining
c: I can find no trace of the side-stroke on the photograph 1 runes to move one place forward. The two ~-runes may
originally have been variant forms for y; but then there
at my disposal; it was probably omitted in the process of
is no symbol for x, or it must have been omitted, cf. the
copying.
case of.z. At this point it is difficult to decide how the
d: the distinction between this rune and m has been extra­ alphabetizer actually proceeded.
ordinarily well preserved : d has the intersecting strokes
at about middle height, with m they start from the top of
There can be no doubt that this alphabet is derived from an
the vertical shafts.
expanded English fuporc, even if only one of the new runes is
g: the first g-rune is in fact j, the second j; the regular g inciuded (0, and also q = k f). . At any rate it does not continue
does not appear here, but cf. k. a local tradition of which the inscriptions of Charnay and
k: at first sight the runic form which fills the place of k looks· Arguel might be remnants (I). The use of ag-rune for k may
identical with n •. but closer inspection shows that the imply that the alphabetizer spoke High German. In: view of ~.
slanting cross-stroke has longer serifs in k than in n. This the connexions of Autun with cultural centres in South Germany,
leads us to the assumption that the k-rune is in fact a g of the presence· of this runic alphabet in a manuscript of Autun
the type found in the isruna fuporc. in which the outer would not be surprising. As to the underlying fuporc, it is
ends of the broken line have been reduced to serifs. The puzzling in one point : no other fuporc has come down to us
choice of a g for k may be an indication of High German with two forms of the !I-rune (in Domitian A 9 the type with
influence. Usually it is the j-rune that takes the place two cross-strokes is a variant of h). . The runes for k and x may
of k. the OE. name ger being interpreted as OHG. ker. prove that our alphabet is related with the isruna group.
As there are no other traces of High German phonology.
this indication has only a relative value. 12. Munich, Bayerische Staatsbibliothek, Lat. MS. I4436
q: the rune which takes the place of q is either x or, more (saec. XI in.).
probably: a k turned upside down. The runes x and k
were occasionally mixed up, cf. Exeter MS. 3507, etc. This codex is of special interest because it embodies at least
x : in other alphabets this rune is found with the value y three different runic traditions, two of which we have met
(e. g. in Munich MS. 14436). According to the isruna before. Moreover we kno~ something about its origin, and
fuporc it is a variant form of the ~-rune; on account of the
(1) H. ARNTz-H. ZEISS, Runendenk:m6ler, 173 ff. (Charnay);
name inc it may have been chosen to take the place of y. J. A. BIZET, L'inscription nmiqtUI d'Argrud. :Etudes germaniques 3 (1948),
Here it fills the place of x, which may be explained in two I-Ia.

250 25 1
this still increases its value : few are the instances where we 67-82, 62·66 Boethius, CommentoTum in ysagflgas Porphyrii liber I (end
missing).
have more than internal evidence for the prehistory of the 108-113 Julius Severianus, Praecepta artis rhetoricae; Julius Victor,
manuscript runes. De memoria.
Thus far the manuscript has been studied mainly as a source 113" Bede, De temporum ratione, c. I (Loquels digitorum).
Il4 Gerbert, De flUmerorum divisione.
for the history of Mediaeval science (I). Part of it was collected, 11S" Bede, De temponmr ratione, c. IV (De ratione unciarum); Com­
and part written, by one Hartwic, monk of the Abbey of St. mentary on Gerbert's R8gu1a abaci.
Emmeram in Regensburg (2). This Hartwic studied for some 117 Herigerus, Regularum pars I (fragment).
I I 8 Rhetorical fragment.
time in Chartres under the famous Fulbert (1007-1029),
Hartwic, Vita et laudes S. Emmerammi.
perhaps also in Rheims. On returning to Regensburg he took
with him several volumes with texts he had collected during The seven alphabets on fo1. 1r are written side by side in
his stay in France. The runic material seems to have been vertical columns. Above each there is an inscription. The
copied by Hartwic himself. It was edited as early as 1750, but first two of these inscriptions were partly removed when the
seems never to have been properly connected with its back­ upper margin was cropped. Moreover the right top is rather
ground (3). badly spotted by moisture; part of the last alphabet can only
The codu consists of II9 folios, the original order of which wss: 1-9, be read with difficulty. Mter the alphabets were written, the
83-1l9, 10-62 (into which 33-34 have been inserted), 67-82, 63-66. right and lower margins were filled with diagrams which form
Format (after cropping) ca. 243 X 192 mm (written area 195 X 140 mm); a more or less logical concatenation (e. g. Vox = significatiua :
33 lines to tbe page. Partly written by Hartwic (int. al. fol. I). partly by
relsted hands. non sig[nificatiua] ; Significatiua = naturaliter: secundum
Contents: (,4) plac[itum], etc.) (plate VI).
fol. If Seven alphabets (details infra).
IV-9, 83-108 R.hetmica ad Herenmum in VllilrfO$.
(a) The first alphabet is inscribed Hebraice. It is a Hebrew ~.
10-32, 35-58 Macrobius, In S01fUJium Scipionis 1m II. alphabet of the type current in Western Europe in that
,58-61 Excerpts from Pliny, Naturalis historia II, 8; 21 f.; 15-18. period. The compiler seemS to have had two different
6I v Catalogue of books.
versions before him, for he adds variant forms for a number
32 Epistolae Senecae ad Paulum apostolum et ad Senecam Pauli.
34 Fragment on the quantity of syllables. of -letters (beht, gimel, he, uau, zai, thet, mem, chaph [chi,
h). This he explains by a note below the alphabet:
(I) H. P. LATrIN, The Eleventh Century MS Munich I4436: Its ContributWn Sunt quedam / litter« in isto / alfabeto bifor/mes. The
to the Histm-y of Coordinates, of Logic, of German Studies in France. Isis 38 doublets probably indicate where the two versions at his
(1947/8), 205-225. disposal did not agree. In this and in all the following
(2) B. BISCHOFF, Literari$ches utld kiinstlerirches Leben in St. Emmeram
(Regeniburg) wiihrend des friihen utld hohen MittelaIters. Studien und Mit. alphabets, the values are written to the left of the letters
teilungen zur Gescbichte des Benediktiner-Ordens und seiner Zweige 51 and the names to the right.
(1933), 102-142.
(3) J. G. ECCAlIDI V. c. de Origine Germanonrm eoromque fJetvstissimis coloniir, (b) The alphabet called • Syriac ' (MS. siriace) consists entirely

migTalionib'Ol ac .,elms gum liIni d'Oo... edidit, figuras aeri incisas adiecit et
praefatus est C. L. SCHlm>lvs. Goettingae MDCCL. FaCllimile on Tab.
of runes. It ends in y, but has twice i and k (cf. infra).

XIV, facing p. 188. The manuscript had been used by Johann Tunnaier
(AventinWl) as early as 1532, Bee P. l..EHMANN, Mitteilungen QUI HmulJehrijten (c) The third alphabet too is purely runic, although the in­

VI (Sit:zun8sberlchte der Bayeriscben Akademie der Wissenschaften. Philo­ scription styles it as .. Arabic' (Ara[bum ?], not Arabice as

sophiscb·historiscbe K1asse 1939, Heft 4) and H. P. Lattin, 0.£., 2og. Eccardus and all authors after him read). It ends in x,

(4) Ca~ II, 2, I7Z;

H. P. LATrIN, The Eleventh Century MS Munidt. I4436, 206.
but h~ two a's, b's and n's (cf. infra).

25 2 Z53
..,.

 (d) (e) (f) The' Egyptian' (Egiptiace). 'Gothic' (gothorum) for m; n. nod for n: p, perd for p; r, Tat for r; s, sigo for s;
and • Chaldaean ' (Chaldaice) alphabets belong to a group U, UUTfor u; x, elwc for x.
of fanciful alphabets· briefly treated in the Appendix.
a: the alphabetizer must have had before him a fuporc which
Below the Egyptian and Gothic alphabets there is a note :
did not mix up :e and a. as a retains its original name ac
pares nominibus sed / figuris differentes.
(in all the isruna fuporcs these two runes, and their names
(g) The last alphabet, purporting to be used by the Alans and values, are confused).
(Alanice), was never completed. The letters go as far as t, e: here he could choose between e and ea (eh, ear). He may
the values end with r and only six names are given (6 _ g). have preferred the latter because formally e looked too
These names are identical with those of the Chaldaean much like M. The f~)Tm ear shows that his exemplar
letters, but amongst the characters several are doubtlessly was more correct on this point too: all isTUna fuporcs
runic (cf. infra).
have aer, and give the value z to this rune.
The first runic alphabet looks· as follows : i :. unlike the related fuporcs. the alphabet spells the name
with single i. In uur. however, the double u has been
ac' berg cen dam ear feu gebohagj)J is calc ker ~i retained. The second variant for i is puzzling : the name
KBhl>"r[YiM·1
a. b c. d e,f g·h r
~cb~
calc rather points to a variant for k. The form of the
rune.we found in only one fuporc, that of St. Gall MS. 878,
r k k where we interpreted it as a regular k doubled. Noisruna
lago man nod odd perd qur rat 5i90 ta.c uur elux inc manuscript has this type, so Hartwic may have found it

t]><1 t ~%. hPl l1t

t m n 0 p 9 r 5 t
n~ fY
ux
in one of his other sources. The value i must be due to
an oversight, unless the rune had received the value y at ~.
some previou.s stage; cf. iT in Vatican MS. Regin. 338,
and the use of the same rune in Exeter MS. 3507, etc.
FlO. 41
k: Hartwic's fitst k goes back to a j-rune. In the isruna
A comparison with a typical imina fuporc shows immediately fuporcs the name was still ger, but then it must have
that we have here an alphabetization of such a fuporc. The undergone the same treatment as depicted for Munich
form of g, the use of p for q. tbe special type of g are unam­ MS. 19410 : ger > ker, hence the value k.
biguous proofs of this relationship. A few letters underwent The second k is the J-rune, but this transition has nothing
alterations in the course of tbeir transmission : I has a spurious surprising. In theisruna manuscripts the name was ih
horizontal stroke added at the lower end of the shaft; In the but the value k. Hartwic (or the scribe of his exemplar)
first k (i. e. the j-rune) the lower end of the shaft is missing; . simply regularized the name according to the acrostic
the lateral strokes in the second k and in 1 are curved; o. too. principle : ih > ki.
has a spurious addition to the right; r looks almost like an 0: here, too, the alphabetizer could choose, viz.· betwe~n
English y-rune. o and C2 (oos, odil). Perhaps the double 00 in the former
For most letters the alphabetizer's t,ask Was an easy one: a. (or a Inistaken eos, as'in Vatican MS. Urbin. 290, or· cos
ac fora; h, berg for b; c, cen for c; ~, darn for d; f. ieh for I; as in Trier MS. R. III 13?) led him rather to take the
g, gebo for g; h. hagal for h; i, is for i; 1, lago for I; m, man unambiguous rune with name odil.
254 255
....
 The so-called • Arabic' alphabet is still more composite :
p, q : Hartwic's prototype must have agreed with the other
imma fu}>orcs in exchanging the values (and names) of a.sc caarHrc.a.b~r'hcaondoroeorfeugeuoheillos kerr
p and q. His gur is the only name in the whole tradition
that conforms to the acrostic principle (St. Gall, Brussels
yur, Trier cur, Urbin. ru ?).
k.* n 'B k +~LI'* XY
a a b b e d ef 9 h f
;r.

k
r: for the value Hartwic first wrote $, then corrected it to r. Irn, men naut net os pern 9uor trr sol tau ur· elx
The rune itself looks rather like the y-rune. Does this
mean that at one time that rune really occurred in the
isruna fuporc? Cf. p. 124.
~
I
11"" 'f. J:
m 'n n 0
t'p ~q ¥r Iis,t utl :t
X
FIo.4Z
$: a truncated sigi meant nothing to Hartwic or to the scribe
of the prototype, and was • corrected' to sigo OHG. a: the rune and its name znay either come from an isru.na
id. • victory '}. fuporc or from a De infJentione alphabet.
t: here is another deviation from the isruna model. The (g) : the second a-rune is the gar-rune, with the same transfer
latter had two t's: one was the regular t-rune (ti), the as in Munich MS. 19410, q.fJ.
other an original d-rune, the name of which had been h: the first b looks rather like a Greek 1T, but this can hardly
germanized- (d;;eg > tag). In Hartwic's alphabet the be its origin. Perhaps it goes back to a Swedish-Nor­
two have been combined: the name tac (i. e. the Upper wegian b : t. The name birca is a Low Gerznan form.
German form corresponding to tag) + the original t. We find a strikingly similar brica (i. e. *birca) as a rune- ~,
This combination was not necessarily invented by Hartwic : name in the Abecedariu.m Nordmannieum. The name of ~.
some contamination may have occurred in the exemplar, the second b (i. e. the original rune b), berih, looks like
cf. tan in the Trier fuporc. an adaptation of beric, the form we met in Munich MS.
x: at this point Hartwic seems to have combined two different 19410 .
traditions. In the isruna fuporc the original x had been c: the rune has the English form, but the name caon is the
dropped in favour of a Roman X, which was anyway rendering, by a Gerznan scribe, of ON. kaun, the name of
identical with the regular g. From that fuporc Hartwic the k-rune. The spelling ao is a typical Bavarian way of
only retains the name, elux. The form of the rune is rendering Gmc. *au. before a dental (1). As this spelling
typical of the De infJentione alphabet. As we shall see, disappears in the ninth century, one znay perhaps conclude
there are other cases where Hartwic borrows from this that Hartwic's prototype reached Southern Germany not
source. Cf. Vatican MS. Urbin. 290, where the two later than shortly after Boo. The Abecedarium has chaon,
traditions also meet. with Upper Gerznan initial.
y: the choice of lJ to fill the place of y is not unique: cf. d: formally the rune goes back to lJ; the name must be a'

Arsenal MS. 1169, where two variants of this rune occur, mistake for dorn.

one for x and one for y.

(1) W. BRAUNE, Althoclul8utsche Grammauk, § 4S;
J. ScHATZ, Altbairische Gro:nrnratik, § Ill.

*** 21
257

256
e: the forme of the rune is rather puzzling : the vertical strok
k: the k has a form which is typical for the De inventione

is drawn so f!lindy, that it looks as if the writer had had

alphabet. The name belongs to another rune: heir is

some doubts. If the rune may be interpreted as j I see

probably a variant of ker, the name of k in the • SyriQC

no other plausible explanation-we can only connect it

alphabet'. But then the diphthong ei remains unex­

with ~ar, ror in the English fuporcs of Cotton MSS.

plained: Gme. *ai before r becomes OHG. e. I believe

Domitian A 9 and Galba A 2. Both in OE. (I) and in

we can explain keir only by aSsuming that it was influenced

OHG. (2), ro and eo are often almost variant spellings,

by geiTT. the ON. equivalent of OE. gar, OHG. ger • spear';

therefore the spelling eor need not cause. any difficulty.

Cf. under i. I: the rune shows a spurious addition; the name looks utterly

f: the rune seems to have been tampered with; yet we may corrupt. In connexion with his theory on the name of

be certain that it is not an I-rune of any sort. It rather this rune (supposed to be *laukaz, not *laguz), W. Krause

looks like the I in the so-called Egyptian alphabet. The would accept this lin as a true ON. form: it would have

name is identical with that in the Abecedarium. become the name of 1 through some sort of an erroneous

abstraction from magic formulae such as /ina laukaR on

g: we found the name geuo in Munich MS. 19410, q.v. The the Fleksand scraper and lini gceddr ok laukum studdr in

rune itself has the additional horizontal stroke typical of one of the VQlsi stanzas (I). This solution looks so far­

the De inventione g. fetched that I prefer von Grienberger's (lin mistake for

h: the name is identical with the form we reconstructed from lac, cf. e. g. Paris MS. 5239), although it is .not entirely

Munich MS. 19410: heih <*heil. This proves that satisfactory (2).

Munich MS. 19410 cannot have been Hartwic's source;

m: the name men is unique; why this apparent plural (OE.
he rather used ,a version derived from the same prototype. ~,
men(n), ON. menn) should have taken the place of man(n)
The form of the rune reminds one of j, but the vertical
I fail to see, unless it may be explained as a mistake for
stroke does not go down· all the way. As a similar form
mon(n).
is found for e, we cannot decide whether the vertical
stroke is a spurious addition (in which case we would n: the first rune shows the Norse ( ~) type, with the cross-stroke

have the h of the De inventione alphabet) or whether it slanting to the left; the name is a transparent adaptation

should have the full height (then we should rather think of ON. naujJ(r). The second n is that usually found in OE.

of the Danish h, which is formally identical with the fuporcs and the name· net is a German's interpretation

English j). of OE. (non-WS.) ned. The occurrence of these two runes

both the form and the name of this rune are enigmatic. side by side is very important for our understanding of

Perhaps the normal English IJ is the prototype from which the structure of this alphabet.

this i derives. The name ios, however, goes back to ior

rather than to ing (3). this case, but there is another difficulty: in the shorter Scandinavian fUPlirk

the e-rune was dropped, consequently iOr was no longer a rune-name. Then

(I) E. SIBVBRS-K. BRUNNIi:It, Altenglische Grammatik, §§ 35. 38. we must assume one of two possibilities: either the OE. rune-name eoh Wl!S

(2) W. BRAUNE, AltJwcluJeuw:he Grammatik, §§ 15 c. 47. 48. translated into ON., or else the ON. rune-name was still known as such by

(3) W. KRAuS!!, Untenudumgen tru: den Runennamen I, 63, proposes to connect the time the alphabets were comPiled. Either assumption unnecessarily

this ios with ON. iOr (Gmc. *ehwa-, OE. eoh = the name of the e-rune). complicates the explanation.

Fundamentally. I have no objection against assuming ON. influence also in (I) W. KRAUSB, Untm'suclnmgen ,lIN den Runennamen I, 63.
(2) T. VON GRIllNBERGER, Dill·ange&ikluischen 1'UlIenreihen, 22.
25 8
259·
..
0: the last letter of the name looks almost like f; but even if values are only indicated as far as T. It looks indeed as if the
Hartwic really wrote of, we have of course to interpret compiler intended to make a seventh alphabet out of material
the name as an original os. not included in the first six alphabets, but then gave up the
p: the final n of pern may go back to t; then we have a simple attempt. Some of the Alanic characters are decidedly non­
adaptation of OE. peril (perd) to OHG. phonology. runic: a, h, c, e,f. q and s rather belong to some or other fanciful
alphabet (I). The remaining letters show a more or less
q: the rune may either be compared with that in Harley MS.
obvious resemblance to runic characters: d = d; g j; h ':"':'·a
3017 (p. 216), or else it may go back to a Norse k-rune.
degenerated h P); k = j (the square form as opposed to the
The name quor reminds us of qtu:lT in Oxford MS. St.
rounded form which takes the place of g); I = a somewhat
John's College 17.
distorted 1; m m; n = either the English or the Norse type
T. s, t raise some involved problems. For T we find a typical of n-rune; 0 = re; pap-rune closed by an additional vertical
De infJentione r but the name belongs to the t-rune and stroke (?); T = an T-rune of the De inventione type (cf. Vienna
may be connected with tiT in Munich MS. 19410. For­ MS. 10IO etc.); t = a t-rune with lengthened lateral strokes.
mally s is almost identical with the p-rune (i. e. s + a As to i, it may perhaps be interpreted as derived from the IJ;
spurious curved stroke ?); the name is that of the ON. variant with tW9 transecting strokes.
s-rune, sol, and must have come from the same source as
caon. etc. As to the third rune of this group, one might
feel inclined to identify it with the younger Swedish­ * **
Norwegian t-rune, but that is turned the other way
round (I). Moreover the name does not support this We can now make the balance of Hartwic's runologicaI
'i.,
explanation: tau seems rather to be borrowed from a harvest. ".
Hebrew (or Greek) alphabet (2). A similar t occurs in (a) He had at his disposal a fuporc of the isruna type, or an
the cryptic alphabet of Vienna MS. 751, where its runic alphabet derived from such a fuporc (on this fuporc cf.
character is more than doubtful. In the' Alanic ' alphabet p. 122 ff.). In one or two points (ear; also the x-rune?) this
of Munich MS. 1#36, the s has almost the same form. fuporc retained primitive features abandoned by the versions
On the whole, the chances that this t is a rune are slight. of the isTuna tract that have come down to us (aer, .1'1'; x-rune =
g). It may perhaps also have had a y, but then with a name
x: the rune shows the same type as that in Munich MS.
so corrupt as to lead the alphabetizer to use it for T ( cr. Vatican
19410; the name elx is the Anglian form (*elhs) corres­
MS. Urbin. 290: TV ?). Another factor may have played a
ponding to WS. Kent. eolhs.
part. As a rule Hartwic (or his exemplar) seems to avoid
*­
runic. symbols which resemble Roman letters too closely : thus
** ea takes the place of e (e looking like M); instead of the x-rune
of the isruna fuporcs Roman X) we find a type borrowed
.The ' Alanic' alphabet is again composite. The six letter­
from a De in'llentione alphabet. Similarly he may have dropped
names (6 - g) are identical with the names of the corresponding
the T-rune (almost = R) in favour of another rune which had
Chaldaean letters. The alphabet itself ends with t. but the
R_.
(1) O. VON .FiuBBEN, 14. 146.
(a) The name tau may perhaps remind WI of tan in the Trier fuJlon:; but
a naine beginning with T 'or containing T. High German
(1) A is rather like the Chaldaean a in Vatican MS. Regin. 338; b may be
this comparison does not enable WI to explain the character itself. oonnected with the corresponding letter in Vatican MS. 166, etc.
260 261
~,
 influence is evidenced by ker ( : ger) and rat ( : rad); also Sici
ON. documents of this period written on the Continent: the
> sigo. The evolution o( ih, k to Ri is hardly surprising. In Abecedarium Nordmannicum. In that same brief text we find
the name of th~ i-rune the double vowel has been simplified
brica, obviously a mistake for birca; then feu, naut, sol, and
(iis > is), but in UUT it was retained. Deviations from the
chaon with Upper German affricate as opposed to caon the
isruna type are on the whole exceptional: a k-rune of a type original kaun. It can hardly be a coincidence that Hartwic's
known only from St. Gall MS. 878; the combination of the
calc-rune occurs only in St. Gall MS. 878, precisely the manu­
t-rune with the name tac (which originally belongs to the
script which also contains the Abecedarium. But here too
d-rune) is typical of the De inventWne tradition. Five runes
Hartwic probably used an older source than St. Gall MS. 878
and names were superfluous: re, asCi 0, oos; w, huun; e, eh;
or at least one which retained birca and caon (I).
d, ti. Of course these various alphabets may hav.e formed one
(b) There can be no doubt that Hartwic also had access to a collection when Hartwic came across them; or he may have cop­
De inventione alphabet, probably even in two different versions. ied them from different manuscripts and at diffetent moments.
From it he borrowed the :Ie of the • Syriac' alphabet, and the We cannot know where the collection as we have it was com­
g, k, and r of the • Arabic' alphabet, besides the • Alanic' r. pleted; nor can we decide what Hartwic's sources looked like.
Since the latter does not represent the same De inventione From the set-up of the page we cannot infer for sure whether
tradition as the ' Arabic' r, we may conclude that the com­ Hartwic had any notion of the real character of the alphabets
pilation is based on two different versions. The alphabet(s) he was copying. Yet it is quite likely that by ca. A.D. 1000
used by Hartwic mayor may not have contained the names of the runic character of the 'Arabic' and • Syriac' alphabets
. the runes: not one of the rune-names in the Arabic alphabet had been completely forgotten. Hartwic's collection is an
points necessarily to a De inventione prototype (although asc, important element for reconstructing the further history of the ,::.
pern, ur, and rat and tac in the preceding alphabet may of manuscript runes. No one less than Sir John Mandeville (or
course go back to such a list), not even the names of those whoever may have been the real author of Mandeville's Travels) .
letters which are typical of the De inventione alphabet (elux: used a collection of alphabets rather like Hartwic's, cf. the
Di *eloh; geuo: Di gibu; heir: Di kale, Rilc, etc.. ;) Appendix to this chapter:
(c) He must :!.lso have had before himan alphabet closely
related to that in Munich MS. 19410 : a caar, berm, geuo, heil, 13. British Museum, Cotton M.S. Domitian A 9 (Saec. XI-).
net and elx are unquestionable proofs of sucl?- a relationship.
But Munich MS. 19410 itself cannot have been· Hartwic's On this manuscript ct. P.. 3 ff.; on the runic alphabet p. 8,
exemplar: he has correct forms where the former shows obvious with full details about the arrangement.
. mistakes. In view of g, which is missing in Munich MS. 19410, The second line of scratched runes forms an alphabet
we must suppose that Munich MS. i4436 goes back to a different which, though incomplete (it ends with p), must be discussed
alphabetization of the same fuporc; here. . As I pointed out before, there is no way of deCiding
(d) ANorse fup~rk, or at least alist of names must have been exactly when these runes were added. They must be older
the fourth runic ingredient : caon, feu, riaut and sol are undoubt­ (I) Hartwic may have found such a source in France. Chartres MS. ZI4
edly Norse rune-names; heir seems to show influence of ON. contained a series of alphabetS, the last of which was called Zitterae DanaOTlmlj
but since thi:sii letters were also called 1'WUle, there can be no doubt that the
geiTT. It is.rather striking that the name birca, which is rather archetype had Dtl1IOTImI. instead .of DanaOTlml. Cf; the Catologue general...
Low German than ON., should also .occur in one of the rare Depanementr. T. XI : ChartTe~, log.

26z 2'3
...

than the sixteenth century"when a reader covered them with The two alphabets have thirty-seven runes in common,
his notes; the terminus post quem is of course the date of the which means that a number of variants have been included.
fuporc (saec. XI). The absence of Latin equivalents (the evidence of those in G
The runes are scratched with a sharp implement that cut has little value, as they were added at a later date) may in some
the parchment. They were apparently made without great cases give rise to doubts: it will occasionally be difficult to
care : b is hardly visible; one cannot make out whether k was decide whether a given rune stands for one letter or another (I) :
meant to have one or two cross-strokes.
The alphabet presents no surprising features : the choice of a. bcde .fg hi k Imno
a for a, that of 0 for 0 was obvious. For k the alphabetizer
chose the k-rune as drawn by scribe A. From this we may hf BhHM/~~+* ~*~*~ ~ INl~

conclude that he preferred A's evidence to that provided by B, pq r tux y z 5 &wth ae [e m s I

or else that he scratched his runes before B • corrected ' the
fuporc. Either explanation may be defended, the material hY ~~ ~~ i r11 t-ilZlilrt ~ fX]N~ FIG. 43
being too scanty to decide which of the two is the more plausible.
The answer to this question, however, is of minor importance : The runes for b, c, d, f, I, n, 0, p, t, u and y require no comment.
from the alphabet we cannot learn more than we can infer
from the fuporc on the same page. a: the first character is a; the second is the a-rune found
in the two Norse fUPllrks of the same manuscripts. The
rune a!; has been relegated to the suppletive group at the
14. Oxford, St. John's College, MS. I7 (saec. XI ex./XII in.). end of the alphabet. ~,
~.

15. Cotton MS. Ga/ha A z (saec. Xltx:II?). e: of the three variants for e (four in G) the first is the regular
e; the second is 3 (name eoh I), with a spurious lateral
stroke also found with that rune in the fuporc of G. The
On p. 26 ff. we examined the fuporc material in these two third e is either x or a variant form of ea, either of which
manuscripts; here we have to discuss their runic alphabets. could take the place of e on the strength of the acrostic
Galba A 2 (G) gives the values of the runes; the Oxford principle (eolhs, .ear). The former of the two solutions
alphabet (0) has only the runes. Since the values in G do not is formally the most probable, and is also supported by a
seem to have been copied from the manuscript, they will have regular ea taking the place of x. It looks rather as if the
been added by Hickes or W anley. G also has three variants alphabetizer mixed. up' these two runes, probably owing
not in 0 : one for e, one for m and one for s. The last of these to the similarity of their forms. The variant proper to G
is a trifling variant of the normal s, from which it is only distin­ . seems to have no runological value, cf. supra.
guished' by having the middle stroke horizontal. The variant
g: neither of the two g's is the original g: the former is j,
which G inserts after the regular m is in fact a d such as occurs
the latter g.
for d in the same alphabet. As to the additional e, it rather
looks like the mysterious e at the end of the second fuporc. h: besides the usual h, ~ere isa variant which may have
These variants can hardly have been found in the common been a type of h-rune borrowed from one of the Norse
ancestor. They will rather be additions restricted to G, and
(1) The variants only found in G have ~ added at the end; the valUe.!!
probably due to the scribe of that manuscript. are those found in ~ICKE!I'S facsimile of G.

264 2 65
...
 fup~rks, with a horizontal stroke added. There is another 35-37 these are three runes for which the alphabetizer found
possibili.ty: this might be the first i in the alphabet, in rio use: w, ., and reo Cf. the procedure in Exeter MS.
which case we could identify it as the j-rune, also with an 3507, Cotton MS. Vitellius A 12 and Phillipps MS. 3715,
additional horizontal stroke. Since there is 8, regular j where 1], ., and re are declared suppletive.
amongst the variants of i, I prefer the first solution.
the first certain i-rune is I] (ing) ; the second is j, the third * **
(according to the distribution of values in Hickes's fac­
simile) is formally identical with reo The equation There can be no doubt that this alphabet is derived from the
iE-rune = i occurs only here. We have the choice between fuporc with rune-names found in G: the 3-rune has the same
two explanations: either the iE-rune was mistaken for a accidental addition, q is identical, etc. One or two variants
v!lfiant of the ing-rune (cf. Brussels MS. 9311-9319, were borrowed from the Norse fup~ks in G and 0 : a and h (?);
p. 71) (I); or else what looks like re is in fact intended to one is fanciful (r); on the other hand the variant iE-rune was
be a variant for k. This' is not so unlikely as it appears left out.
at first sight : in a few cases re is used for q (Exeter MS. This complicates the problem of the origin of the collection
3507, etc.; Harley MS. 1772) and this may have led to the in O. G cannot represent a revised and corrected edition, as
inclusion of that rune among the k's. . On the whole, the alphabet in 0 is based on the correct fuporc in G. On the
however, the former solution is the more probable. other hand 0 cannot have been derived from G (or from one
k: the use of k for k was obvious. of its ancestors or descendants), because then one does not see
why the good fuporc-with-names found in G would have been
m: on the additional m in G, cf. supra. ousted for such a poor corrupt version as that in O. We must .>
q: as usually the symbol serving for q is derived from p. then· suppose that the compiler of 0 had before him a relative
r: the regular r is followed by a fanciful letter, derived either of G, which either did not contain the good fuporc-with-names;
from the preceding rune or from Roman R (cf. Nemnivus's or which he began to use only after he had copied the corrupt
d, p. 158). fuporc-with-names from another source.
On the whole the alphabetizer proceeded logically and
s: on the additional sinG cf. supra. The two runes found
skillfully. In some way or other he mixed up the runes x
in both manuscripts are the .original s and the st-rune.
. ap.d ea; perhaps also re and 1]. Otherwise his alphabet offers
x: cf. under e. The use of ea for x is also -.found in Vienna no difficulties.
MS. 1761 and in Leyden MS. Voss. lat. 128.
z: no rune, but a fanciful Roman Z: 16. Vienna, Nationalbibliothek, MS. I76I (saec. XI/XII).
33 : a ligature of e + t, cf. 34.
As the manuscript contains a full version of the treatise
34: the abbreviation mark 7 = Lat. et or OE. and, and. This
De inventwne litterarum, it will be examined at length in the
abbreviation seems often to have been added to the Latin
next chapter (p. 299 ff.) .. Mter the runic alphabet which we
alphabet as a sort of 24th letter (2).
normally find in the treatise, a second one has been inserted.
(I) In Nemnivus's alphabet, too, a symbol which is fonnally identical with It is the latter which we have to discuss. here. It occurs on
the I)-rune takes the place of O.
fol. . 100:V and is introduced by" Item". T.he compiler ob­
(2) Cf. the examples in A. C, PAUl'S, The Name of tM Letter J.
267
266
...

 viously intended it as a variant to the preceding alphabet. It Munich MS. 1+436 and in Harley MS. 3017 (also in
is rather badly crowded in; the values are written to the left of Oxford MS. St. John's College 17 i).
the runes, but two are at the end of a line whilst the rune to
which they belong begins the next line. The characters differ n: cf. g.
greatly in size, a and I being twice the size of nand :J, and y 0: the right hand part of the lower lateral stroke has been
three times. They show little runic style; the n-rune e. g. is much lengthened, a detail also found elsewhere (e. g.
identical with the g-rune in the preceding line and with the Exeter MS. 3507).
letter x in the next line. Yet some runes have well retained
q: formally d and q are very similar; the only difference is
their original forms: h, I, 0, p, t, x. The <;haracters for b, c, e, U
that in d the triangle is rather directed upwards. whilst
show some amount of cursivation. The whole looks as follows :

a k. 6Bck, d~ eM. f Fgx h~ d r

k It mtXl
in q it goes somewhat downwards. The origin of this q
is obscure. At the best it could be connected with the
(' Chaldaean ') q in Leyden MS. Voss. lat. F 12 a and in
Vatican MS. Regin. 338.
nXo~ p~ 9l rl\ 5'1 tt vA/fyt zv T: this rune is formally identical with a rounded u-rune. Yet
FIG. 44 such forms of r turn up occasionally, e. g. in the Dahms­
dorf and Britsum inscriptions (I).
a: the rune with this value is identical with the 0 of the
De inventions alphabet, which is a slightly modified o. In 11.: the archaic ' upside down V 'type of u.
the present case, however, this explanation cannot be x: the use of ea for this letter is also found elsewhere : cf.
adopted, as there is a regular 0 with the value 0 in the same a
Leyden MS. Voss. lat. F 12 and Oxford MS. St. John's~'
alphabet. Perhaps we have rather to start from ae or a, College 17.
and to assume that the 0 of the preceding alphabet in­
fluenced the original character. y: on this point, too, the Vienna alphabet agrees with that
of the Leyden manuscript. The choice of the rune x to
d: a similar type of d is found in the runic alphabet of Karlsruhe
611 this place may again be based on considerations of
MS. Aug. 176, but the Vienna character is closer to a
form': x was most like a Roman Y.
minuscule d. I explain the Karlsruhe type as derived
from a regular d, but this solution does not necessarily z: I see no way of explaining this character, unless it is an
apply here. Did the compiler (of the present manuscript imperfect 3 (formally the latter reminded of Roman Z).
or of its exemplar) turn around a .., to make it look more
like d? The alphabet is on the whole rather problematic : the characters
for a, d, k, q and z are obscure. It has some features in common
I: in the I-rune the upper lateral stroke has been curved back with Leyden MS. Voss. lat. F 12 8, but all in all too few for
to the top of the shaft. establishing some degree of relationship between the two. The
g: as mentioned before, the character for g was no doubt. Vienna alphabet may then safely be regarded as an independent
interpreted as a formal x, and 'so was the rune n. alphabetization..
k: this unique type of k may perhaps be connected with the
(1) w. KRAuim, RutuminschrifUn, 442 f.;
De inventi!me c,. or with q in the • Arabic' alphabet of H. ARNTz-H. ZEIss, RJnumdenkmIiJtT, Iff, J54 ff.; pl. I, VIII.
26'8 269
...
 The alphabet consists of the following runes :
17. Kassel, Landesbibliothek, MS. Theol. F., 65.
(saec. VIfVIll).

r~~~Mr~N~fln~~f~HHh
As a matter of fact the runic inscriptions found in' this manu­
script should be classified somewhere between manuscript and
epigraphical runes: they are scratched on the binding. For­
mally they come closer to the inscriptions, but since they FIG. 4S
belong to the sort of runic writing that was practised in scrip­
toria, I have classified them with the TUnica manuscripta. The characters for IX y z are missing. A number of runes
Kassel MS. Theol. F. 65 was written in the latter half caU for no comment, viz. those for a (i. e. a) b c e i I m T u.
ofthe sixth century. The teXt in half-uncials was corrected by
early Continental and insular cursive hands. The latter is d: formally this character is rather like Roman D, but there
generally supposed to be St. Boniface's own hand, or that of is also a chance for its descending from ." cf. d in Harley
one of his companions. This circumstance might lend special MS. 1772.
value to the runes, but we shall see that they can hardly directly
go back to the famous missionary or his followers. f: has only one lateral stroke, as that in Exeter MS. 350 7 •
and Cotton MS. Vitellius A 12. It owes its origin probably
The runes on the Kassel binding were first noticed by
J. Caesar in 1864; P. Lehmann edited them almost thirty years to a faulty exemplar.
ago, but runologists seem hardly to have noticed them (1). It g: the occurrence of the rare isruna type of g, also found in
is not necessary to give a full description of the codex, since Exeter MS. 3507 and Cotton MS. Vitellius A 12, would
the runes' are found on the binding. The latter belongs to an be of great importance if the alphabet could be dated more ~.
interesting group of Fulda bindings examined by Lehmann (2). accurately. As things are, it throws little light upon the,
Many bear inscriptions of some sort, which often escaped origin of the iSTUna tradition, except that it proves that
notice. Scratching with a dry point was at one time widely. this variant may be less rare than appears at first sight.
practised, but owing to the difficulty of reading, such inscrip­
tions have often been neglected (3). They are of course not h: seems to have only one transecting !ltroke. Whether this
easily dated either. In the present case the binding (saec. , is simply an omission, or whether the example of Roman H
VIll, Fulda) offers a terminus post quem, but the other limit can played some part, I cannot decide.
hardly be established with certainty. Internal' evidence seems
k: thi~ character is almost like a' Roman K; it also occurs

to point to the late eighth or the ninth century. The ~sel

amongst the runes of Sf. Gall MS. 878.

binding bears two inscriptions: on the front cover onelthat

seems to refer to the contents' of the codex, on the back a runic n: is not runic, but rather a half· uncial n, the first shaft of

alphabet. The formerlwill be examined on p. 414._1f1 which has been lengthened. I found a similar n on fo1. 49 v

(i) P. LBHMANN, FuJdaer Studien (I). (Sitzungsberichte der Bayeri.schen of Basle MS. F. III. ISC, where it is followed by a rune a

Akademie der Wissenachaften. Philosophisch-philologische, und historische

-(in the word Natiuitate). As the Basle manuscript also

Kluse. Jahrgang 1925, 3. Abhandlung). Miinchen, 1925. IS f. Cf. p. 4J4.

The manuscript WlIS lost in 1945. originated in Fulda, we may perhaps suppose that some

(2) Cf. p. 42J, footnote I. scribes' there mistook the half-uncial n for a rune. The

(3) Cf. e. g. B. BISCHOFF, Vb. EinritZll1'lgen i" Ho:ndschri/ten des .fri1hen character is found in ornamental inscriptions, e. g. in

Mittelalters. Zentralblatt fUr Bibliothekswesen 54 (I937), 173-177.

270
27 1
",
 British Museum MS. Royal 7. C. XII and in Dublin MS.
good reason for not attaching too great importance to it in the
Trinity College 60 (1).
chapter on the isruna tract.
0: the choice of (2 for 0 may point to a Continental alpha­ A few words about the inscription on the front cover of the
betizer. same manuscript, to which we shall have to return in Chap­
p: this character is either the rune w or a Roman P. It is ter V. It consists of three attempts to write the same word, the
not unlikely that the former was interpreted as p under second of which remained unfinished: i05eW3 I i ios i I
the influence of the latter. ioseW3. Lehmann transcribed the first and the last line by
10SEPI, .and this seems to be the only possible explanation.
q: either a minuscule q or, less probably, a Greek ICO'7MTIJ..
The inscription agrees on one point with the alphabet: the
s: it is almost certain that this is not a rune, but an insular rune w stands for p. On three other points it differs, viz. in the
minuscule s. I doubt that there is any connexion between choice of 0 for 0 (the alphabet has re), in the use of 3 for i,
this s and that on the Thames scramasax and St. Cuthbert's and of s for S. It looks as if this inscription is based on still
::i coffin (unless the rune in the latter inscription were derived another alphabet.
~~> from the same minuscule s).
t here we have at last a sure indication that the alphabet CONCLUSION.
is due to a Continental scholar : the rune d has received
the value t, just as its OE. name dreg, deg became OHG. After the discussion of the general problems of· alphabe­
tag, taco tization and of the special aspect of each single item, our con­
Considering everything, this alphabet- can hardly go back to clusions may be brief. .
Boniface or his immediate surroundings. An Englishman of First of all, the proportion between English and Continent~
the early eighth century would have known which characters alphabets is striking. Compared with that of the fuporcs, it is
+
'. entirely reversed. For one English alphabet there are four or
could take the place of n and p; he would hardly have used
d for t, or re for O. If some relation to a hypothetical fuporc five on the Continent. The explanation is simply this: fuporcs
brought over by Boniface is to be assumed, one must at least can hardly have meant anything to Continental scholars, who
attribute the alphabet to the second generation of the school had lost all touch with runic tradition. They could only
created by the English missionary. This rather. diminishes conceive an alphabet in the order of the Latin letters (hardly
its value; and the latter still decreases if we think of the possi­ even of the Greek or Hebrew letters). When a fuporc was
bility-which can hardly be ruled out-that it has nothing to presented to them their natural. reaction would be to turn it
do with the founders of Fulda. Mter all, the alphabet is not into an alphabet.
, much more runic' than that in Leyden MS. Voss. lat. F 12 8, This being the case, we need hardly consider all alphabets
and rather less than that in Harley MS. 3017. It is then safer as manifestations of one movement. The success of the

not to connect it with Boniface and his companions. We may . alphabetization, the degree of adaptation to Continental Ger­

regret this conclusion, the more so because this alphabet also manic phonology differ so much that there can be no question

contains the iSTuna type of g. But there is no indication of of " urredactionen" which were gradually perfected.
this alphabet being older than the ninth century, and this is a As to the English alphabets, they must all belong to an age
that had lost the sense of the runic system. They either go
(I) E. A. LoWE, Codices Latini Antiquiores II, %17, 277. On the Basle
maDU8Cript ef. p. 421. back to fuporcs we know, or were imported from the Continent.

Z7 Z
2% 273
 APPENDIX 7. Bamberg MS. Msc. pa~r. 130/2 (saec. XII): Chaldaean
(= runes).
spurious Alphabets and the Last Phase of the Ru'liica 8. Cotton MS. Titus D 18 (saec. XV in.) : Chaldaeo-Assyrian,
Manuscripta (I).' . Gothic, Persian.
Manuscripts without rpnes also contain such alphabets. A few
In a number of manuscripts discussed thus far or mentioned
are listed here, but there are no doubt many more :
in the coming chapters, we find alphabets which do not seem
to be related to those of the three 'sacred languages', i. e. 9. Munich MS. lat. 14725 (saec. IX in.) : Chaldaean, Egyptian.
Hebrew, Greek and Latin. I· have called these alphabets 10. Vatican MS. lat. 266 (saec. IX): no names (Egyptian,
, spurious' for want of a better term; perhaps further research Gothic).
may discover the sour~ from which they are derived. The II. Avranches MS. 107 (saec. XII): Chaldaeo-Assyrian,
most popular of those alphabets was Aetrucus Ister's, which no Egyptian, Saracen.
doubt deserves closer study. It was incorporated into the
treatise De invtmtione litterarutn and is to be found in many 12. Munich MS. lat. 14684 (saec. XIV, alphabets XV):
of the codices examined in the next chapter. Nemnivus's Chaldaean.
alphabet was another; as we saw in Chapter III, its origin is 13. London, British Museum, Addit. MS. 4783 (late saec.
not so obscure as one might imagine. But here we are concerned XlV) : Egyptian, Mrican. (+ Norse runes)
especially with the alphabets circulated as 'Chaldaean' , 14. 'Dublin, Royal Irish Academy, Book of Ballymote (late
, Egyptian', 'Mrican', 'Gothic', etc. I came across the saec. XlV) : Egyptian, Mrican. (+ Norse runes)
following instances whilst collecting the material for this
15. Berne MS. 762 (saec. XVI) : Egyptian, Mrican, Samaritan ~.
work (z) :
(contains a n':Imber of runes), Persian, Chaldaean"
I. Cotton MS. Domitian A 9 (saec. VIII leaf) : Chaldaean, Cathayan or "Pentexorie".
Egyptian.
It is worth noting that these alphabets enjoyed considerable
2. Berne MS. z07 (saec. VIIIflX): no names (Mrican, popularity in the North of France: nos. 2, 4, 5, 9, II originated
Egyptian). there.
3. Vienna MS. 751 (saec. IX): no name (Egyptian or Gothic). The last manuscript probably draws part of its material from
4. Vatican MS. Regin. lat.' 338 (saec.· X/XI ): Chaldaeo­ Mandeville's Travels (1).. In the latter work we find a whole
Assyrian, Egyptian. series of alphabets: one group of manuscripts has six (Hebrew,
Greek, Egyptian, Saracen, Persian and Chaldaean), another
5. Munich MS. lat. 14436 (saec. XI in.) : Egyptiari, Gothic,
three more (the above plus Tartar-Russ, Cathayan and Pen­
Chaldaean (runes are called Arabic, Syriac, Alanic).
texoire). There can be no doubt that the author used a coHee­
6.' Oxford MS. St. John's College 17 (saec. Xl ex.fXII in.) :
(I) Cf. M. LETTs, Sir John Mandeville. The Man and his Book. London
no names (Egyptian, Gothic). (1949), 15I fr., Ilnd
(x) Cf. my Uit de Geschiedenis fJan de .Rtmen, 5Z if. G. DB POERCK'S review in: Belgisch Tijdschrift woe Philologie
(z) Professor B: BISCHOFF informed me that he had got together an 'important en Gesch.iedenis, 30 (195z), 881-883.
collection of such. material, but had not yet found an opportunity to study it R. Dl!ROLBZ, Uit de GeschiedeniS 'IIan de Runen, 53 fr.
in detail. In the sUrvey I hs.ve felt free to dispense with the quotation marks : Mr. LETTs very kindly sent me proof sheets. with the IIlphabets from his
DO J:eIIder will take those high-sounding names on their fllce value. forthcoming edition of the Travel!.

274 ~Z5
.".....
 tion of alphabets made up of material such as is found in the or planned by an author who had a feeling for esoteric bits of
above manuscripts (except the last). The origin of most of learning.
these alphabets is still a matter of discussion, but the one that In the Avranches manuscript Aethicus Ister's alphabet ts
is of special interest to us offers no such difficulty. Mande­ listed under the heading SARRACENORVMILITTERAE;
ville's Saracen is a combination of runic characters with the in that of Munich we found runes presented as Arabic, Syriac
names of Aethicus Ister's letters. In many manuscripts, and Alanic letters; in the Bamberg codex they are termed
especially in the oldest (among the Brussels copies which I Chaldaean. The author of Mandeville' $ Travels, who displays
examined, MS. W+20-I0425 [fol. 49l ] is the most convincing), an extraordinary liking for fantastic combinations, if not for
the runic origin of the characters cannot be doubted (cf. the truthful reporting, may have combined material similar to that
names with those of Aethicus's letters, p. 352) : in the above manuscripts in order to provide a Saracen alphabet
missing in his collection. It is of course not impossible that
atmoy bechath cati delfror effoli fOrthi choiri galop

a R r NMI!. *tl> '" J-.X

hJc;him ioith ka.ili
he found such an alphabet ready-made in his source material;.
t: there is even a chance that a copy of such an alphabet has come
down to us.
abc d ~ f c.h 9 h i k
Munich MS. 14436 (or a codex with a related collection of
lothrm malath noblleth orc.hy corisel incola.ch renchi 51la.d totimus alphabets) has -been supposed to have supplied the runic in­
1'1>1 )(t~ TRl T
gredient of the Saracen alphabet, but a comparison of the latter
with the Munich runes shows that this is impossible. We
1m no P 9 r ~ t
have, however, a much closer relative in Arsenal MS. 1169.
an.l.och yrchom aLotiti sotizim th€ch It has exactly the same selection of runes as Mandeville:s

tlv I ~y 'Zt ItJNR Saracen: .'

abcdef j jhig lmnopxrstu:lJ:lJ
X z'1 ?? 7? Ars.: abc d e f g h i kim n 0 p q r stu x Y z
FIG. 46 Mand.: abc d e f ch g h i kim n 0 p q r stu x y, etc.
Three questions arise immediately in this connexion: In the Arsenal alphabets the value have been shifted one place
(I) How did the runes come to be called ' Saracen' characters?
,I forward beginning with h; with Mandeville i has a value ch
(implying that the character was interpreted as a Norse h? or
(2) By what way were they connected with the names of i!'
I rather ch = It! I ?). But apart from this detail the agreement
Aethicus Ister's letters?
is so complete, that there must be a connexion between the
(3) Where did the runic alphabet come from? two. Mandeville's alphabet may be said to represent a more
The answer to the first two questions becomes obvious if we genuine tradition than that in the Arsenal manuscript, which
briefly examine alphabet collections like those in Avranches is more than three centuries older. At this point Hickes gives
MS. 107, Munich MS. 14436 and Bamberg MS. Msc. patr. us a valuable clue. In his Thesaurus (I) there is a facsimile
130/2. In all three the' alphabets are arranged in vertical of a runic alphabet precisely of the same type; the names of
columns. Such an arrangement may have led to the letter­ the runes are again Aethicus Ister's letter-names, but they are
names of one alphabet being connected with the letters of written with Greek characters. It is quite unlikely that Hickes
another. Similarly the name of one alphabet may have been
shifted to another. Either move may have been accidental, l (J) G. HxCKllS, Thesaurus, G1'am'IMticae Islandicae Rudimenta, Tab. III.

~.

276 277
or Wanley should be held responsible for this combination;
the ductus of the Greek letters, the shapes of the runes point
to careful copying. If so, Hickes's facsimile probably represents
the missing link between the material of Munich MS. 1443 6,
Bamberg MS. Msc. patr. 130/2, Avranches MS. 107 and CHAPTER IV
Arsenal MS. ll~ on the one hand, and Mandeville on the
other.
In the new garb and with the new name which the author RUNIC ALPHABETS (II):
of the Travels had given them, the runes left the scriptorium DE INVENTIONE LITTERARUM.
and found their way into the printing office. The Saracen
alphabet turns up in some of the earliest collections of alphabets,
One of the eadiest examples of runica manuscripta that came
viz. those of Theseus Ambrosius and Urbanus Wyss (I). More
to the notice of scholars was a runic alphabet included in a
examples could perhaps be unearthed, but there we are no
short treatise on the history of the alphabet. Since Goldast
longer in the realm of runica manuscripta. It is rather improb­
edited it, this treatise is known under the title De inventione
able that these scholars knew anything about the real character
linguarum ab Hebraea usque ad Theodiscam, et notis antiquis.
of this alphabet. Its real identity was discovered by Hickes
It is usually attributed to Hrabanus Maurus. Hence the runes
(and Wanley, who had planned a study on the alphabet?):
in this treatise are usually called • Hrabanic '; at times this
he printed the facsimile referred to above on a plate containing

I
term is applied to all Continental runic alphabets.
only runic material. That meant the end of the eventful
, The fullest discussion is that by G. Baesecke (I). As we
career of a runic alphabet through seven centuries.
saw in the Introduction, he considers this runic material as due .>
to the activity of two scholars of saec. VIII/IX: Hrabanus
Maurus and AJcuin. At least eight of the fifteen runic al::­
phabets discussed by von Grienberger are connected with the

t De inventione tradition; to these Baesecke adds Heidelberg MS.

Salem 9.39. Of the paragraph on the runes there are two
versions, a longer text in Vienna MSS. 1609 and 1761, in the
Heidelberg MS. and in Goldast's edition (= my type A), arid
the other, shorter text in Vatican MS. Urbin. 290, Paris MS.
5239 and Cotton MS. Titus D 18 (type B). Berne MS. 207
J. and Munich MS. 19410 cannot be assigned to one or the other
group. Baesecke holds that in the longer text the references
to the Marcomanni. to the lingua theodisca and to divinatWnes
are interpolations; in the shorter the remark about the runstafas
would also be an addition (but cr. infra). The additional
remark about the Goths in Vatican MS. Urbin. 290 would be
(1) T. AMBRosIus, I11tToductio, fol. 204v. A Norse runic alphabet (" alpha­
betum Gotthorum" - Swedish) is reproduced on fol. 206'.
U. WYlIS, Lihelltu, fol. [NiiijV]. ( (1) G. BABSECKJ!, AbeceJanum.

278 279

I ".
related to the Gothic material in Vienna MS. 795, which in Baesecke distinguishes at least two authors, who treated their
turn is connected with Alcuin. Dn account of Tunstafas the DE. models in different. ways. The older one retained a
shorter version would be closer to the DE. original than the number of DE. forms; his successor must have been a German :
longer one. The N ordmanni of the two texts are probably Hrabanus Maurus. At this point it is difficult to follow
to be identified with the Danes, in whose country runic practice Baesecke's argumentation, because he includes alphabets which·
had been revived about the beginning of the ninth century. have no connexion with the De inventione tradition. The
The interpolation divinationes would have been borrowed from original fuporc might have reached Germany at the timeCif the
the Germania: in ch. x Tacitus describes the use of notae Anglo-Saxon mission, but Baesecke rather believes it was
among the. Germanic tribes and the compilator would have brought to the Continent by Alcuin. Hrabanus would have
identified these notae with runes (an explanation still advocated learned it in Tours when he was studying under Alcuin (801­
by some runologists nowadays) (I). 804). The other chronological limit would be given by
Hrabanus's De institutione ciericorum (819), where Baesecke
discovers an allusion to the De inventione text.
(x) G. BAmlI!CKI!, Abecedarium, 83 f. : " Die gemeinsame Aussage der lin­
geren und kiirzeren Texte, dasz die Runen zur Aufzeichnung von Gedichten
In the Introduction I have pointed out some weak spots of
und Beschworungen dienten, ist in den lilngeren von den Nordmannen auf die this argumentation : it is based on unproved assumptions; it
(Germanen) ausgedehnt, die jetzt noch Heiden sind, und es wird noch die throws together unrelated material; it exaggerates the im­
Anwendung der Runen auf divifll1~s zugefiigt. Aber wir haben in Deutsch­
land keine 80Iche Aufzeichnungen in Runen. Auch Hmban kannte nichts
portance of Hrabanus's runic studies (if they ever took place).
dergleichen: sonst hlitte er nicht diese angelsilchsischen statt der deutschen Cf. infra p. 374 ff.
gebracht. Er denkt bei dieser Nachricht also wohl an nordische Heiden. W: Krause, the latest author to discuss the Hrabanic rune­
Dort liesze sich vieles, auch Zeitgenossisches, als Beleg anflihren. Aber
alphabet, follows much the same procedure. He too considers.
divina~s fehlen auch dart. Dasz Nachrichten iiber die Runen der Nord­
mannen damals nach Deutschland kommen konnten, zeigt ja das dinische the treatise as a manual to be used. in schools, and probably'··
Abecedarium Nordmannicum. Die Dinen waren in der Tat noch Heiden, rightly so; but we can hardly accept his further implications :
namentlich hatten sie seit etwa 800 eine jugendkriftige Runenkunst : die a1te "Der ganze Traktat diente offenbar Lehrzwecken und mag
des Futharks der 24 .Zeichen war dort zweihundert Jahre zuvor ausgestorben,
die neue des Futharks der x6 wanderte nun von Norden her ein und somit eher der Fuldaer Zeit Hrabans als seinen letzten Mainzer
eroberte das Land. Von ihr und ihrer heidnischen Anwendung kam auch Jahren angehoren. Da femer die angelsachsischen Muster­
Kunde zu Hmban nach Fulda: er hatte mindestens durch den Dichter des formen der Hrabanischen Runen ersichtlich auf den EinHuss
Helland Beziehung zu Niedersachsen, das seit dem ersten Abte Sturmi Fuldas
Missionsgebiet war, Jnd Fuldaer Besitz reichte im 9. ]h. nordOstlich bis
Alhwines weisen, der 804 starb, so scheint mir auch dieser
Magdeburg und nlirdlich fiber Braunschweig hinaus. Auszerdem wuszte Umstand fUr eine verhiiltnismassig friihe Entstehungszeit des
man im Kloster, dasz die Sachsen an die heidnischen Nordmannen (also Traktats zu sprechen " (I): This brief quotation contains the
Dinen) und Obodriten grenzten und dasz ein Teil von ihnen " beinah .. ins
essentials of Krause's argumentation. He too distinguishes
Heidentum zuriickgefallen war: eben darum waren vom Papste hilfreiche
Reliquien erbeten; und Rudolf von Fulda, Hrabans Schiller, erzihlt, wie die two versions, and he also considers the shorter version (e. g.
des ID. Alexander (im Jahr 85 x) herangefiihrt wurden. So paszt auch Hrabans
Satz fiber den heidnischen Runenzauber, und wenn das Weissagen mit Runen
fUr den Norden unglaublich bleibt, so fiihrt gerade dies auf Fulda : dort lag
[ zu divinationes (nicht zu carmina und incantaticmes) etwa passende significare
fUr das sCTibere des kiirzeren Textes: er denkt an das Losen mit .. bezeich­
die Germania des Tacitus mit ihrem Bericht iiber Los-difJinatia; es ist das neten" Stibchen.
ganze Mittelalter hindurch auszer in Fulda keine Benutzung der Germania Nach dem zuialligen und gewisz sehr vorliiufigen Material zu urteilen,
bezeugt, und in Fulda sclirieb damals der Monch Rudolf ganze StUcke fUr seine wiire also ein iUterer und kiirzen;r Text von Hraban interpoliert, wie das seine
Tmnslatio Sti. A1exandri daraus abo Hraban setzte also die taciteischen flatu
des losenden Priesters richtig mit den Runen in Beziehung und iibertrug das
• Art ist, u:nd der illtere wiire, wegen nuutafas 6 [must be 8 = Cotton MS.
Titus D 18 in VON GRIENBERGER's list], noch angelsilchsischer gewesen ".
Losen auf "diejenigen, die jetzt noch. Heiden sind ". . Daher auch das nur (x) W. KRAUSB, Die Hrabanische Ru1letlTeihe, x75.
r
280 281
".::"
 that of the Paris manuscript) as the older; it would be due to
As I indicated in the Introduction, I do not intend to give a
an Anglo-Saxon, proof for .which Krause finds in the contents
of the Pat:is codex : the treatise on the alphabets is preceded critical edition of the De inventione litterarum text. Under the
by several of Bede's works. The implication is that AIcuin title De mventione linguarum, etc. (cf. infra) it was last edited
in 1606. The text in the complete works of Hrabanus Maurus,
served as an intermediary, since he was Bede's pupil (according
edited by Colvenerius in 16z6, is simply taken over from
to Krause). The author is aware of a difficulty: in the older,
more 'Eng:fish' version the rune-names show precisely the Goldast's edition, and Colvenerius's text was in its turn re­
printed by Migne for his Patr%gia Latina. Since Golclast
same amount of German influence as those in the version
supposedly recast and adapted by the German Hrabanus seems to have used only one manuscript, a new edition is
Maurus. He supposes that the Paris manuscript did .not badly needed indeed (I). Yet a full discussion of the Hebrew,
Greek, etc. material has too little to do with runica manuscripta
preserve. the original text, but is in fact a mixture, " '" eine
to have any right of bci.ng included here. I have, however,
durch spatere Abschreiber entstandene Mischung der Arbeit
given the whole text of the two or three versions that may be
eines angelsachsischen Autors und eines deutschen Bear­
beiters" (I). distinguished. In the first place this may help us better to
evaluate the position of the runes; for the same reason I have
Krause also connects St. Gall MS. 270 (one of our isruna
added a number of parallel texts. Some reference to the non­
group) with Hrabanus Maurus : the example corui in the iSTilna
runic material cannot be avoided because in a number of cases
tract would be an allusion to his name (OHG. hrahan, rahan
only those paragraphs and alphabets enable us to establish the
' raven '). This manuscript would in fact represent an older
relationship between the different versions and manuscripts.
stage; its runic alphabet Would be " ein runisches Probe-ABC
Of the fourteen 'Hrabanic alphabets' mentioned by von
mit Angabe verschiedener Nebenformen". This alphabet
Grienberger - Baesecke - Krause (excluding for the time ~
would come closest to Hrabanus's Urjassung, which he improved
being two printed versions) only six or seven actually belong to •
and made 'more German' at subsequent attempts: "Dies the De inventione tradition (z) :
Probe-Alphabet mag er dann in verschiedener Weise ausge­
staltet und geglattet und eins der so gewonnenen Alphabete in I. Vienna, Nationalbibliothek, MS. 1609 (v. Gr. no. 3).
seinem Tractat .. De inventione linguarum" aufgenommen z. Vienna, Nationalbibliothek, MS. 1761 (v. Gr. no. I).
haben. Die verschiedenen, in der geschilderten Weise auf 3. Heidelberg, Universitatsbibliothek, MS. Salem 9.39.
Hraban selbst zUrUckgehenden Runenreihen haben aber weiter­ 4. Munich, MS. A. Weinmiiller (first edited by Plassmann­
gelebt und .sind von interessierten Schreibern wieder und
Krause).
wieder nachgebildet, dabei mehr oder weniger bewusst verandert
worden" (2). 5. Paris, Bibliotheque ·Nationale, MS. lat. 5239 (v. Gr.
no. 5).
I do not think it is necessary to offer a full appreciation of
these reconstructions at this point; a number of remarks have 6. London, British Museum, Cotton MS. Titus D 18
been given in connexion with Baesecke's work and in the (v. Gr. n? 8).
Introduction.' The rest may be easily gleaned from the con­
(1) The extracts from various D:J.anuscripts given by some authors, e. g.
clusions of the present chapter.
BAllSI!CKl! and l{RAUSB, only contljin the text on the runes, and even so they
pan hardly.be considered as a critical edition of that one paragraph.
(I) W. K:!tAUSll, Die Hrabanisclui Rwumreilu!, 186. (2) The manuscripts are sometimes indicated by their numbers in VON
(2) W. K:!tAUSIl, Die Hrabanisclui. Runenreihe, 187. GRIENBl!RGER'S paper (Die angelsiichsischen nmenreihen, 23 f.) j in the above

z8z survey these numbers are added between brackets.

.f, 28 3

i
¥,
 Only one more manuscript amongst those mentioned by Krause to count it as a separate item. The points on which it differs
is to be connected with the De inventione tradition : from those manuscripts may be matters of copying and editorial
7. Vatican Library, MS. Urbin. lat. 290 (v. Gr. no. 6). policy. It is only to leave open the possibility of Goldast

On the other hand, the following six manuscripts not included having used a related manuscript that I have listed his edition

by von Grienberger-Baesecke-Krause also contain more or less separately. No such treatment was needed in the case of

complete versions of the treatise; in two of them the runes are Lazius, as there the chances of his having copied from a source

missing: that has not. come down to us are extremely slender. There

8. Vatican Library, MS. Regin. lat. 294 (its version differs is of course no need for referring to Colvenerius or Migne, as

considerably from those in the first six manuscripts). their texts have no independent value.

The other manuscripts mentioned by Krause and his prede­

9. St. Gall, Stiftsbibliothek, MS. 876 (no runes).

cessors, viz. Munich MSS. 14436 and 19410, Vatican MS.

10. Vienna, Nationalbibliothek, MS. 1010.

Regin. 338, Berne MS. 207, Vienna MS. 751 and Phillipps

II. Strasbourg, Bibliotheque Nationale et Universitaire, MS. 371 5 show such important divergences from the De in­

MS. 326. ventione type of runic alphabet (not to mention the other alphabets

12. Niimberg, Germarusches Nationalmuseum, MS. 1966. nor the introductory texts) that I have felt safe to examine them

13. Florence, Biblioteca Medicea Laurenziana, MS. S. in Chapter III with other independent alpnabets.

Marco 604 (no runes). There is quite a chance that the above list is not exhaustive;
Isolated runic alphabets of the De in'iJentione type are found in the recent discovery of no. 4 is a warning against any illusions
three manuscripts : in this respect. Yet the chances that versions older than those
listed here should still await discovery does not seem to be very
~,
14. Karlsruhe, Badische Landesbibliothek, MS. Aug. 176. "'.
great: the oldest manuscript that has come down to us cannot
IS. Karlsruhe, Badische Landesbibliothek, MS. Aug. 254.
be much later than the date of composition of the treatise.
16. Bamberg, Staatliche Bibliothek, MS. Patr. misc. 130/2. As we shall see, two basic types are reflected by our eighteen
To complete this survey I include versions : type A is represented by seven manuscripts (nos. I,
17. W. Lazius, De gentium aliquot migrationibus, etc. 2, 3, 4, 9, 10, 12) and by Lazius's and Goldast's editions (nos.
Basle [1557] (v. Gr. no. 4). 17 and 18), type B by fQur manuscripts (nos. 5, 6, I I, 13)· Of
18. M. Goldast, Alamannicarum rerum scnptores aliquot the five remaining manuscripts one offers a quite different
veteres. Frankfurt 1606. (v. Gr. no. 2) (1). text, but probably belongs to the A group (no. 8); four have
These last two versions, however, can hardly be placed on one only the runic alphabet, and. so it is practically impossible to
level with nos. 1-16. Lazius's text-he gives only the text on decide what type they represent (nos. 7, 14, IS, 16).
th(' runes and a runic alphabet the end of which is missing-is _Occasionally objections have been' raised against the title of
in all probability borrowed from no. I. Goldast's version this little treatise : it does not treat of the origin of languages,
closely resembles nos. 1, 2 and 9, and therefore I have hesitated but of the origin of alphabets. Our only authority .for lin­
guarum is Goldast, who probably invented the title; at any rate
() " Macromannorum sive Normannorum liteme Runicae" are also found it is not found in any of tl;te manuscript'J examined for this
in Copenhagen MS. Ny kgl. Sml. 1867,4", written in 1760 by OLAFUR BRYN­ study. Therefore I have felt free to alter linguarum to litterarum.
JUJ..FssON. This manuscript is a compilation made up into aI. of printed texts
The framework of the treatise consists of five alphabets :
(e. g. RI!sEN~t18'8 edition of the SnOTT4 Edda [1665]); its runic alphabet
reminds of those edited by LAZIVS and GoLDAST. Hebrew, Greek. Latin, Aethicus Ister's and runes, each with
28 4 28 5
~,
 a brief introductory text on the origin of these alphabets. This in Donati Artem mmorem, the chapter De littera in Berne
interest in • inventors' was quite current in the Middle Ages. MS. 207, the ars grammatica of the Irishman Clemens, etc.
Their names could be gleaned from Isidore's chapter De de­ The extracts given in the Appendix to this chapter present only
scriptiune temporum (Etymologiae V, xxxix) and from other the more obvious parallels to the De inventione text. References
passages in that work (II A, ii De inventoribus rhetoricae artis, to Cadmus, Carmentis, etc. could be found in many other
III B, iDe inventoribus geometriae et vocabulo eius, etc.); regular Mediaeval grammarians (Servius, Sergius, Pompeius, Vic­
lists of nomina inventorum must have enjoyed some popularity torinus). As was pointed out before, much of this lore must
(e. g. Vienna MS. 1761, fol. 108 v ; Munich MS. 19413, fot. have been so current that one can hardly use the term • sources';
127r, etc.) (1). De origine litterarum may then be characterized there can be no doubt that many scholars knew such matters
as a sort of a hybrid between such a list of inventors and the by heart and could use them freely without reference to a
treatment of the alphabet in grammars. written exemplar.
The study of ~is material offers some difficulties. It lent The paragraph devoted to Aethicus Ister's alphabet is drawn
itself easily to recasting, to reducing or swelling. For the from the first and last lines of that author's Cosmographia (I).
first three paragraphs at least-those on the Hebrew, Greek This reference to the Cosmographia provides us with a terminus
and Latin alphabets.-anylibrary or school of some importance post quem. According to K. Hillkowitz (2) that work must
must have offered parallel or additional material. So it is have been written after 768, since it uses the Continuationes
quite natural that the versions should differ more on the point Fredegarii, .and before 821, in which year it is mentioned in
of Hebrew e. g. than on that of Aethicus Ister's letters. Some the library catalogue of Reichenau. The oldest manuscripts
desequilibrium in my treatment may result from the fact that themselves seem to go back to ca. 800 (3), and so the period
I have included stray runic alphabets (if they showed the same during which the Cosmographia was compiled may be narrowed ~
' ..
type as those actually connected with a De inventione text), down to the. last quarter of the 8th century. It is supposed

but no stray Hebrew or Greek alphabets (not to mention Latin). to have originated in the kingdom of the Franks. The' phil­

Such a distinction was, I think, unavoidable. It is true that osopher and cosmographer Aethicus, born in Istria of noble

for Hebrew the Middle Ages had only a few sources (mainly parents', is a fiction, and so is Jerome's participation in the

Jerome); but the study of the transmission of elementary work. But whether this allows us to consider the whole work

Greek (a smattering of which was offered even by Latingram­ . as a huge joke, as Hiilkowitz proposes (4), I cannot decide. At

mars) would be far more complicated (2).

(I) D'AVIrZAC, Mlnwire sur Ethicus et sur le, O1I.'IJ'I"QgU comwg,aphiquu intituUs
Consequently, when the sources of De inventione litterarum de CB nom. Academie des inscriptions et belles-lettres. Memoires presentes
are examined, a distinction must be made between the first par divers savants. Premim-e serie, II (1852), a30-551.
three paragraphs (Hebrew, Greek and Latin) on the one hand, K. A. F. PBRTZ, De Comwgraphia' Ethici libri vu. Berlin, 1853.
H. WUITKB, Die Komwgraphie des lltrier Aithikos Un lateinischen AtuJrUflB
and the fourth (Aethicus's alphabet) and fifth (runes) on .des Hieronymus. Zweite vermehrte Ausgabe. Leipzig, 1853. It is unfortunate
the other. The text of the first three was mainly borrowed that the edition planned by HILLKOWITZ (cf. the next footnote) bas not yet
from Isidore of Sevilla's Etynwlogiae I, iii : De litteris commu­ made the Comwgraphia accessible in a more up-to-date form.
(a) K. HILLKOWlTZ, Zu, Komographie des Aethicus. (Dills. Bonn) K61n,
nibus; iv : De litteris latinis. Some gnu::n.p1atical commentaries
1934·
also contain striking parallels, e. g. the Commentum Einsidlense (3) K. HILLB:OWITZ does not Seep1 to have considered the OCCUl:Tet1ces of
Aetbicus's alphabet as a possible clue for dating the Cosmographia.
(1) Extracts from such a list of inventoR in Ghent MS. 92 (the so-called (4) K. Hn:.!.KOWITZ, zui Kosmographie, 73 : .. Aber sicherlich haben wir es
Liber jloridw).will be found in the Appendix.. bei Aetbicus mit einem Spottvogel zu tun, der eine Parodie auf Erdbeschrei­
(a) B. BISCHOFF, Das griechische Elemmt, 3:11 if. bungen verfassen wollte ". What could be interpreted lIB a joke in certain
r
286

I,
287
...
any rate Aethicus's paternity of the curious alphabet found' not at the beginning of the text, but between the fourth
at the end of the Cosmographia was not doubted by the compiler and the fifth abbreviation beginning with C. Mommsen
of the De infJenh'one, and the letters themselves were placed on explained the attribution to Julius Caesar as. due to the
one level with runic, if not with Hebrew, Greek and Latin. misreading of notae i. c. iuris civilis) as notae lulii
The riddle of Aethicus's alphabet has not yet been solved. Caesaris. According to H. Hagen, it would rather be due
The fifth paragraph, and the last to which De inventione to a Inisapplied reminiscence of Suetonius's famous
litterarum applies, is that on the runes. Research devoted to it passages on the system of cryptography used by Caesar
has been surveyed in the general introduction and at the and Augustus (I); and since the following paragraph of
beginning of this chapter. Further discussion will follow the De inventione actua~y treats of a cryptic system inspired
after the material has been examined. by Caesar's (or rather Augustus's) secret writing, Hagen's
In most manuscripts the text ends here j the list of Greek explanation should not be rejected a priori. A combination
numerals which a few append to it was obviously meant of the two explanations probably approximates the truth.
to complete the paragraph on the Greek alphabet. Butthree (2) The notae sancti Boni/atii, of which there are two sorts :
or four versions, viz. St. Gall MS. 876, Vienna MS. 1761 , one in which the vowels are indicated by one to five dots
Heidelberg MS. Salem 9.39 and Goldast's edition, add notes . her
(elt a = . , e = :, ,. = :. ,0 : :.' u :.: ,or a = : I
on several sorts of cryptic writing. e =: ,i = . , 0 = : : , u = :.:), or else by the consonants
(x) The notae Caesaris. These notae are also found in several which come immediately after them in the alphabet
of the laterculi notarum (alphabetical lists of notae) edited (a = b, e = I, etc.). The text informs us that" the
by T. Mommsen (x) : the notae Vaticanae (Mommsen, archbishop and. martyr Saint Boniface had shown these
p. 301 if.), the notae Papianae et Einsidlenses (Mommsen, to our ancestors when he came from the Anglo-Saxons; ~
p. 3 1 5 if.) and Paulus Diaconus's notae litterarum more yet we tend to believe that they were not invented by him, '
fJetusto (Mommsen, p. 331 if.). The selection of abbre­ but were used in this way by the Ancients". In his
viations found at the end ofthe De inventione does not form discussion of these notae L. Levison points out that there
a close group, but there is one for each of the twenty-three may be some fundamental truth in this statement (2). At
letters of the alphabet: AVG Augustus, BM = Bonae any rate they reInind one of the notae Augusti mentioned
memoriae, till ZEN Zenovius [i. e. Zenobius], followed by Suetonius. One example is given of each variety of
by LR Locus religiosus and CM = Comes; abbreviations notae.
with initial C and L are Inissing amongst the first twenty­ (3) A list of monograms, with a brief introductory text. A
one. This type of writing has of course little to do with monogram is said to be one character made out of a con­
Caesar, but the notae Papianae et Einsidlenses provide us glommeration of letters. The examples go from simple
with a welcome parallel: there too we find the inscription
(1) SUETONIUS, De 'Oita Caaarum I : Divull Iulius, e. S6 : .. Extant et [epis­
INCIPIVNT NOTAE IVLII CAESARIS; apparently
tulae1 ad Ciceronem, item ad familiares domesticia de rebus, in quibus, Ii
cb:cumstances must not alwaya have been so. The interpretation of the curious qua oeeu1tius perlerenda erant, per notal seripsit, id eat sic atructo litterarum
ltuiniw of the grllMmlllian Virgilius Maro offers sim.ilar difficulties. There ordine, ut nullum verbum efficl posset: quae ai quia investigare et persequi
is II. striking resemblance between Aethieus Ister, the Scythian sage, and Fenius volet, quartarn elementorum ~ id est D pro A et perinde reliqullS
Farsaidh, who .. was a sage in the principal languages even before he· came commutet ~': 11: Divus Augustus, e. 88: "Quotiena per now acribit,

from the North out of Seytbill. .. (G. CALDER, Arwaicept, 13; d. also the intro­ B pro A, C pro B Be deinc:eps eadem ratione sequentis litteras ponit; pro X
duction to this work). autem duplex A".
(l) T. Mol\IMSEN, Notarum lamculi, in: Grammo.tid Latini, Vol. IV, 26S ff. (a) W. LIMSON, England and the Continent, 290 ff.

z88 23 289

l,

~~.
 constructions 8.uchas Dominus, Sanda, Maria, to Salomon 782:"786 (?); 786 Abbot of Reichenau, .806 of St. Denis, later
Rex Pacificus. Bishop of Basle) and his immediate successors. It is mentioned
In St. Gall MS. 878 we had a good illustration of what may in the oldest catalogue of the St. Gall library :
have been the prehistory'of the De inventione tract (cf. p. 76). Partes asporii; item partes donati grammatici; item ars
There we found precisely those extracts from Isidore's Etymo­ . honorati gra~atici j, diomedis de metro; item bedre
logiae upon which the text of the first three paragraphs is presbiteri de metrica arte; item partes donati minores
.based; .these extracts are followed by a {uporc and by the maioresque; item tractatus pompegii in donatum et alia
Abecedarium Nordmannicum; There is as yet no introductory multa. Hec omnia in volumine I. (1).
text to the runes; neither Aethicus Ister's alphabet nor the Modem binding: back in white pressed Renaissance leather, wooden boards

notae and monograms have been added. But if B. Bischoff is not covered, clasp missing. Rather stiff, parchment, white and yellowish:

right in supposing' that Walahfrid' Strabo is responsible for many holes and old repairs. The codex is made up of z63 folios; they have

the usual St. Gall p!!-gination (pp. 1-sz6: the last page was at one time pasted

this compilatiori, it can hardly have direct connexions with onto the binding) and are arranged in 3S gatherings (no quire marks) :

the De inventione tradiiion : the tract is supposed to have been +

1 By-leaf [I-Z] + III (+ z) [3-18] III (+ 1) [I9-3Z] + S IV [33-1IZ]
completed one or two. decades before Strabo may have picked + III (+ z) [II3-1z8] + IV [IZ9-144] + II [145-148]+3 IV [149­
"up his alphabetic lore, and even so Strabo's runes are much 196] + 1(+ 3) [197-z06] + 4 IV [zO'/-z70] + IV(- I)[z71-Z84: cf.
ifffra] + IV [z85-3OO] + HI (+4) [301-316] + IV [317-33Z] + II
closer to the OE. and ON. models than the runic alphabet in (+ x) [333-340] + 7 IV [34I-45Z] + III (+ z) [453-468] + Z IV
the tract we are examining. On the other hand Walahfrid's [469-500] + IV (+ I) [501-518] + II [SI9-5z6].
collection shows how current most of the material in the tract Bruckner distinguishes at least seven hands : A =< pp. 3-33(aIso< 33-128,
must have been. It may also be a· warning not to forget that 144-145, (481); B = 12.9-144,146-147, zoo-z03«closely related to A); C =
. 149-199. zoS-z06; D = 156: E z08-Z77 (closely related to C): F = z85-338
such tracts could be composed in various places with not too (identical with C 1); G 341-5z6 (probab!ymore than one hand). He does "
different results, or that additional bits ·of information could not mention the handwriting of pp. z78-z81, preCisely the part containing the '
easily be 'added to them. De in'IJlI1ItWne text. There is little doubt that this part is approximately con­
temporary with· the ,rest of the manuscript, even if the handwriting cannot
In the following survey, the .manuscripts have been classified definitely be identified with anv of the vsrieties described by Bruckner. The
roughly' according to their relationship. Further details on manuscript was probably not 'plamied as a whole from the very beginning,
.p. 345 1£. but as~mbled gradually as the material could be collected: the large nUmber
of single leaves points in the lIIIme direction. It will be safe to assume that
compiling this codex took quite some time, perhaps two or three decades: so
THE MANUSCRIPTS.
the manuscript may be dated in the last quarter of saec. VIII, including (espe­
cially fox: the De i_tiline) the first years of saec. IX. . .
Format: z3z/z3S X 145/150 mm: written area 18z/19O X IIO/IZO mn:i.;
1. St. GaU, Stiftsbibliothek, MS. 876 (saec. VIII/IX). z7-z9 lines to the page, one column (on the special arrangement of the De
in'lJ/!ntillne text, cr. ifffra).
Contents (z) :
According to A. Bruckner (1) this manuscript was written
p. r probationes pennae, drawings : zmen with swords ( 1), inscrll;led Melib...
in St. Gall at the time of Waldo (first mentioned 770; Abbot TITIrusj Christ on ,Cross, inscr. Rex iudeoru[mJ.
3 IN NOMINE DEI SANCTISSIMI INCIPIUNT PARTES ORA­
(I) A. BRUCKNER, SeripWria II, I, 80. G. 8cmm.R:sR, VtIT:leic1mUI,. 303, also
dates it .. s. VIII/lX ~'. On this manuscript cr. also:
. TIONIS ASPORI = Donatus minor in dialogue form, here ascribed to
Julius Aaper.

H. HATl'lIMBR, Denkmohk 1,423 and pl. II.

30 Ilicipit ordocognoscendi nomen (after Donatus).

G. ScHERRBR., VtIT:leichnin, 303 fr.

<
E. STElNMB'YEll-E. SIlM!RS, Alt1wchdeubche GlOlleti IV, 454.
J. M. CI..AlUt, The Abbey of St Gall, 101, n+
290
I (I) P. LBHMANN et. aI., Bihliothe/ukatakJge I, no. 16, 81, 11. Z9-3Z.
(2) Cf. G. SCH:I!RRl!Rt VtlTllteichniu, 303-30S.

29 1

I ".

 3:1 HBl1AEKB8/JJ)-(H+'IIBA8 = Explicit, Domine + fiat.

the available space to the five alphabets : below the first line
33 INCIPIT ARS DONATI GRAMMATICI (commentary on Donatus).

86 De adverbio, etc.
of the text on the runes (p. 278) there was no space left for the
90 Exce1pta from Isidore, Etynwlogi.ae I, 6-13.
runes themselves. Therefore, on p. '1.79 he did not proceed
98 EXPOSITIO ARTIUM DONATI (from Setgius's commentary).
with the text of that last paragraph, but probably reserved it
104 DE BARBARISMO; 105 DE SOLOECISMO; 107 DE CETERIS
UICIIS: loS DE METAPLASMIS: 109 DE SCHEMATIBUSj III
for a later page, where he would be able to give the text and
DE TRHOPIS (from Donatus). the runes together. So he carried on the text of the first four
lIS Victorinus, De fouJ/ibw md'roTImI (no title). paragraphs till p. 280, having even to crowd it in on this last
124INCIPIUNT MAlORES PARTES DONATI GRAMMATICI (in­
complete at the end).
page. On p. 278 the text on the runes reads (the parchment
1:19 INCIPIT ARS HONORAT[I] GRAMMATICI / DE FINALIBUS being very greasy in places, much of the text has become
SYLLABIS (Semus Honoratus, De :finalibus syllabis ad Aquilimnn). illegible) :
13'1 DE SCANSIONE HEROYCI UERSUS ET SPECIE EORUM. LLitteras quippe q]uas utu[ntur marcomanni quos nos
1# INCIPIUNT NOMINUM EXTREMAE SYLLABAE DIOMEDIS
GRAMATICI.
nor]d[mannosJ uocamus infra /
146 INCIPIT MALLI THEODORI DE METRIS.
The text of the other four paragraphs shows the following
163 INCIPIT ARB DIOMEDIS DE METRO; ao'] blank.
arrangement (I) :
a08 INCIPIT LlBER BED~ DE ARTE METRICA.

-<>-as'1 INCIPIT DE SCHEMATIBUS SICUT ALII UOLUNT (Bede, De p.a'18 p.a'19 p. a80
, schew4ilnu et tropis sacTl'J8 scripturl'J8). Primo ......• csptiui / tatem ••••.... subtus/ ut inuenire .•.•... verius.
\ a'18 De inventione litterarum (no title) j a82, a83 drawings, a84 blank. thau .....••• samech I nun ••....••••••• zai / uau .......•..••.. aleph.
\ a8s IN NOMINE DOMINI NOSTRI IHBSU CHRISTI INCIPIUNT /
Litteras ...• ueniens I [iJn grecism . numeros / faciendos .....••.. possint.
OCTO PARTES ORATIONIS DONATI GRAMMA17CI URBIS
alfa •.•••.•..•• th&a I iota ........•.•••• ro / simma .•.••. nota numeri.
ROM~ (Donatus minor).
Littetas ... dicebatur/inuenisse ...... littere / .III.adiecerunt .•.. reliqua.
30a INCIPIT. DECLINATIO. PRIMA. DONATI I GRAMMATICI a ............... k/el ................ u/ix .................... Z. i.
URBIS ROMAE (Donatus major); 339, 340 blank. Litteras .. inuerumus I quas .•... perdu [xit] / quia ........... fallimur. '
398 INCIPIT TRACTATUS POMPEI IN I QUO 'DO'NATI ARTEM [alamonJ ..••• h&mu I iofitu ...•..• [ozechi] I [cho]rizech ..... zotichin.
MIRIFICE I COMMENTATUS EST.
463INCIPIT 'DE' BARBARISMO; 4'11 DE SOLOECISMO: 478 DE On p. '1.81, however, we do not find the whole text on the runes,
CETERIS UITIIS; 484 DE META/pLASMO; 490 DE SCHEMA­ nor even the part missing after the first instalment on p. 278,
TIBUS; 498 DE TROPIS. .
510 TRACTATUS POMPEGII DE MAIORIBUS PARTIBUS ORA­ ---. but only the end :
TIONIS (beginning missing). /nesque ac diuinationes significare procurant, qui adhuc
In this manuscript we find the De inventione text in its natural pagani ritus [inuolu]untur.
surroundings: a collection of grammatical writings (some of The space reserved for the runes remained blank; it is followed
which,e. g. Pompeius's, also refer to the origins of the Latin by the text on the monograms, (beginning also missing) :
alphabet). It may have been inserted to fill the gap between Jlitterarum unum caracterem pictores facere solitis (i. e.
two parts of the manuscript (note that it is followed by three soliti sunt) quod monagramma dicitur quorum signi­
originally blank pages, and that the text on p. 285 ff. partly ficat[ ...] per pauca adnotata monstrantur,
repeats earlier parts of the codex). by five monograms (Simon, Iudas, Iacobus, Bartholomew,
For the De inventione text the scribe chose an arrangement Mathias) and by eleven notae Caesaris (oppidum ... locus reli­
found nowhere else : he began the text of all five paragraphs giosus) ..
on p. 278, leaving sufficient space between each pair of text It is obvious that one folio is missing after p. 280. This
lines to insert the alphabets, and went on in the same way on folio must have contained : the middle part of the text on the
pp. 279 and 280. He had, however, made a mistake in allotting (1) Cf. the complete text on p. 349 1£.

292 293
....

 runes, and the runes themselves (? cf. infra); the tat on the reagent, but apparently without noteworthy results. In the
notae Caesaris, and the notae for Augustus - Nero Caesar; the white stain left by the reagent It number of runes may be made
first part of the text on the monograms, and the monograms out with more or less certainty (I). Only seven out of a total
for Dominus Jacobus (I). There is no way to decide whether of nineteen runes (including pseudo-runes?) are not doubtful.
it also contained the text on the notae saneti Bani/atii. As to The question whether this scribble has anything to do with
why this one folio was removed, and by whom, we can only the absence of the runes in the De infJentione must be left
guess.' The text of the De infJentione given by this manuscript open. Two or three runes in the scribble may be De inventione
is closely related to that of Goldast's edition, and consequently types.
it is not impossible that Coldast himself is to be held responsible Although this version is apparently the oldest that has come
for this depredation-as he was for many others at the expense down to us, there can be little doubt that it is se~ra1 removes
of the St. Gall library (cf. p. 303). from the original text :
The set-up of the text was probably due to the compiler's amoy seinuente (= a moyse inuente); post illorum capti­
desire' to have the corresponding letters of the various alphabets vitatem et reuersionis eorum; a1iquantas ... qui ad numeros
placed approximately one below the other. This would have faciendos habiles habentur; quia nonnulla uerba necesse
made the treatise suited for comparative purposes, and would habuerunt sicut in grecis habetur loqui ut Christus est
: have brought out the idea expressed in another version: ymnus (corr. from ymnis); gosmographi (cf.- Gadmus);
L) " Litterae enim grecae et latinae ab hebraeis uidentur exortae ". in istis adhuc in a1iquibus allis fallimur (for: si in istis
, But the arrangement would probably have required mQre adhuc litteris et in a1iquibus a1iis fallimur uos emendate ?);
careful planning than oUI scribe was capaJ>le of. pictores facere solitis (= soliti sunt); monagramma; per
A question now arises : did St. Gall MS. 876 ever contain pauca adnota(= adnotata) monstrantur.
the runic alphabet? Of course we cannot be sure whether it Yet, notwithstanding the absence of the runes and the rather
did or did not oCCUI on the missing folio. But since the other decayed state of the text, the St. Gall version is important : by
alphabets extend over three pages (pp. 278-280), we should at its age and by being one of the rare localized versions it gives
least expect to find the final runes of the alphabet below the precious indications concerning the transmission of the treatise.
text on p. 281 (I). There is an indication that the absence
of the runic alphabet may be due to the compiler's doubts 2. Vienna, Nationalbihliothek, MS. I609 (saec. X). -@
about its authenticity or correctness; the text on the Hebrew
( alphabet ends with a remark not found in other versions :
L-)" sed require caracteres carum uerius". The scribe (or his
predecessor) may have entertained similar doubts on the point
of the runic alphabet. .
In the right top comer of p. 281 there are some faint scribbles,
which were subjected at one time or other to the action of a
I This manuscript seems to have been. one of the first ever
examined for its runic material: Lazius probably edited part
of it in his De gentium aliquot migrationibus (ISS7). He prob­
ably acquired it at Freising during one of the journeys he
undertook to visit libraries (ISSI) (2). As early as 1576 the
manuscript belonged to the Imperial Library in Vienna. Its
origin is unknown. H. Menhardt has shown that both palaeo­
h) At firat Bight I believed to have found a rune Q and one more, very faint graphical and linguistic criteria point to Freising, but ultimately
rune in the blank space below the text. But closer inspection showed that
these were only the· remains of the monograms for Simon and Judas. which (I) cr. Chapter V, p. 4U
were el'lllled and, written again a few 1ines lower, apparently to leave free the (a) H, MI!NHAJmT, Die Oberliefmmg au ahd. I38. Psalms. Z. f. d. A. 77
space reserved for the runes. (1940), 76-84 (with facsimile).

294 29S

Ii
".
 a good deal of the contents seems to be derived from St. Gall, 53' Poem on the five senses; eight epigrams.

55' Definitions of crepusculum, vesper, etc.

into al. the De imJentione text (1).

55" Excerptum de epistolis Hieronymi et aliorum "irorum.

Parchment binding of 1755; imperial eagle pressed in gold on front IIIld _ 64' NotM:er, fragment of II letter to Chancellor Liutward.
back cover; inscriptions: E.A.B.C.V. (= Ex Auguatissima Bibliotheca 64' Pseudo-Methodius.
Caesarea Vindobonensi) and 17.G.L.B.V.S.B. 55 (= Gerardua Liber Bare 69' OHG. tra.n.alation of the 13Sth PsaIm.
Van Swieten BibHothecarius 1755). Parchment of poor quaHtv, OKen tom
(e. g. tol. a). It was formerly referred to as 'no. saS' IUld 'Theal. 73a '. The De imJentione litterarum text shows the following
The 70 folios are arranged as follows : arrangement :
V (- a) {1-8} + 5 IV {9-4.8} + IV (- 1) {49-5"'; 53 occurs twice : fo1. 2 r, 11. 1-2 : closing lines of the excerpt on pronunciation;
53a, 53b} + III (+ I) {55-61} + II {6a-651 + II (+ I) {6&-70}.
3-24 : the Greek alphabet, with above each letter its
Format ca. 185 X laa mm, written area ca. 160 X 100 mm; one column,
a4 lines to the page. To judge from the handwriting, the manuscript may name and the corresponding Greek numeral, to the
be dated in the early tenth century. According to Menhardt, this agrees with left its Latin equivalent, and below the Roman figure
the linguistic data provided by the OHG. tra.n.alation of the 13Sth Psalm. indicating its numerical value. The last three lines
ryThe De itltlmtione hand does not appear again elsewhere in the codex; yet it
is probably contemporary. The text on the alphabets seems to have belonged
give the Greek numerals from 2000 ' to '10000'
1

to the manuscript from the very beginning. (dischile mire mia).

Contents (a) fo1. 2v, 11. 1-8 : the text on the Greek alphabet; the initial of
fol. I" On pronuncistion (= MartiIlllUS Capella, De nuptiU III, § a61). [LJitteras was not filled in by the rubricator; 9-24 :
:z.r De inventione litterarum (incomplete).
4r The Muses, with their respective domains (from Isidore, Etyma­ the text on the· Hebrew alphabet, followed by that
logiae VIII, II, S7); the divisions of music (UI., III, 19, 1.5). alphabet from thau to he; above each letter its name,
",r Sequentiae or tropes, with neumes. to the left its value. On the last line the names of
9r Notker Balbulus, two letters.
1St Salomo III, Abbot of St. Gall and Bishop of Constance, Liber the remaining four letters, deleth to aleph. <.,

fontllllmu:m (3). fo1. 3r , 11. 1-2 : the remaining letters of the Hebrew alphabet
(1) H. MBNHAIIDT, Die Oberliefenmg, 80 f. ~~ ~~~>~i{~ 3- 1 1 : the text on the Latin alphabet,
Palaeographie musicale. Les principawc ntmIVScritt de chant gregorUn, am­ at the beginning of which the rubricator skipped the
I!'osien,
mosarabe, gallican, pub/iis mfac-similh phototypiquu par Ie. Bhledictins
initial of [LJatinas; 12-15: the alphabet itself, with
fh Sol.enn&. III (Solesmes, 1892), PI. l09 A (" neum<es-accents allemllllds
sangailiens "). above each letter its name (a be ce, etc.); 16-24 : the
On this manuscript cf. also text on Aethicus lster's alphabet.
E. DOMMLBR, Dos Fcmnelbuch des Bischofs Saloma III. von Konstamt. Leip­
zig, 1857.' fo1. 3 , 11. I-I I : Aethicus's alphabet, with the names above
v
K. ZEtlMl!R, Formulae M",owingici et Karolini Aevi (Monuments Germ. the letters and the ~ues to the left; 12 blank; 13-19 :
hist., Leg. sect. V). Hannover, 1886, 390-4lll7.
L, GAUTIKR, Iluttnr6' fh la poIsie litur~. Paris, r886, 13:1, no. 3'.
the text on the runes, followed by the alphabet from
M. MANITIUS, Geschichte I, 596. a to p on 20-25. with above each rune its name and
J. M. CLARK, The Abbey 0/ St Gall, 304.
to the left its value.
and on the runes :

H. F. MAssMANN, R.tmm, a56 ff. The rest of the tract is missing. This, I believe, is the strongest
W. GiuMM, Ueber deutsche Rrmm, 80 f. argument for identifying Vienna MS. 1609 as Lazius's source.
G. STBPHI!NS, M_ts I, 107; III, I3.
T. VON G.IUl!NBI!RGI!R, Die angelriichsischm nmmreih.en, a3. text, there aeems to be some confusion: according to E. STBINMEYER - E.
(a) Tabulae I, a61 f. SIIM!RS, Althoc1uleut4che Glossm IV, 569, the excerpt on pronunciation is
(]) This liber, fonm.dt.mJm is also found e. g. in l'v'unich MS. lat. 19413. followed by a " deutung der hebriiscben. buchatsben ", apparently = VJeD.na
But when MI!NHARDT says that 1his manuscript also contains a De inwntione MS. 176I, fal. I 05"-1 06' (cf. p. 30a).
296 297
",
 His runic alphabet, too, ends with p, Perc, for which he gives be a copy of the 'original. Its mistakes (e. g.Litteras ... quas

this explanation: utuntur Marcomanni ... scriptaS habentur; a quibus origiQ.em ...

C~terz literCi, in eodem antiquo codice desiderantur, & ob tradunt) will also be found in other versions. The whole is

nimiam uetustatem legi non possunt (I). written with evident care (although the scribe 'did not take the

To be sure, some readings in his edition differ from the Vienna trouble e. g. to restore the right order of the paragraphs);

version, in the alphabet itself : , therefore we can only regret that the end of the treatise is lost.
Ac (V. asc), Byrith (V. birith), Chilch (V. gilch), Fortunately we have fairly good substitutes in Vienna MS.J761
and especially in the text; but the former may simply have and in Goldast's edition.

~I -I(
arisen in the act of copying, and the latter mainly consist in
corrections to the faulty text. A more' serious objection may 3. Vienna, NationalbiblWthek, MS. I76I (saec. XII).
be found in Lazius's statement that the manuscript which he
used contained an annalistic history of the Franks and a genealogy The Latin and OHG. glosses of this manuscript have thus ~I ( , ( . , C
of Charlemagne: far been its main point of interest (I). Yet its runic alphabet ,/.J.?

Ac ne dubium ullum sit, Norrnannos eosdem illos Marco­ was known as early as W. Grimm's time (2). In fact, as far

manorum extitisse posteritatem, in antiquissimis Annalibus, as the De inventione text is concerned, it comes quite close to

Francorum historiam ac Caroli magni genealogiam conti­ Goldast's version; but I believe there are sufficient reasons

nentibus, membrana longe omnium antiquissima scriptis, for supposing that Goldast used a different exemplar (cf.,

h"c nominatim uerba excerpsimus (2). p. 370 f.). '

Yet I do not think we can doubt that Lazius used Vienna MS. For our purpose this version is of special interest. Here the

1609 for his edition. ,Marginal notes from his hand appear De inventione text has been incorporated into a mass of collectanea ",

on some +2 pages; from this same manuscript he printed part obviously intended for use in schools. Such collectanea have ~

of the 138th Psalm (in OHG.) in his De gentium aliquot migra­ unfortunately received little attention thus far, and when they

tionibw (p. 81). The chances that two manuscripts of De were examined, it was mainly in order to trace the sources of

inventione litterarum should be mutilated in such a way, and the constituent parts, rather than to study them as autonomous

that in both the lacuna should begin after the p-rune, are slight entities (3).

indeed. The deviations appearing in the text and in' the The manuscript once, belonged to Sebastian Tengnagel,

alphabet are easily accounted for by Lazius's editorial technique; prefect of the Imperial Library iq Vienna from 1608 to 1636

his text of the OHG. Psalm is also quite different from the (fol. Ir : Ex libris Sebastiani Tengnagel, LV.D. et Caes. Bibli­

manuscript original (3). His explanation of the lacuna in his oth.). Formerly it was numbered • 6+' and • Theol. 863 '.

edition of the treatise on the alphabets shows clearly that he Its origin is unknown. From the binding Steinmeyer inferred

did not have the manuscript before him when he WJ."ote his
comment. (1) E. STBlNMlmlR - E. SIlM!RS, Altlu:H:hdeutsche G/.oum IV, 643-646;
Although fairly old, the text of Vienna MS. 1609 can hardly v, 108 if;
(2) W. GRIMM, Ue1ur 4eutsche Rrmen, 80 f.
G. STBPBBN8, M _ t s I, I II; III, 13.
(1) W. Lwus, De gentium aliquot nrigrationibru, 645. T. VON GJU£rmI!:RGBR, Die angelsiIc1uisclum nmt:m't!iJum, 23.
(2) W. LAzIUS. De gentium aliquOt nrigratiottilnu, 644. (3) The e:l:cetpts are tnostly taken from Isidore, Eucberius or perhaps from

(J) H. MmmAImT. Die tlhrliefmmg, 79. From MENH.umT·S account we some other popular collection of mt'onnation. The~ 110 often without

may perhaps infer that Lszius mixe4 up Vienna MS. 1609 and abother Freising any indication of origin, that we may safely lSSIlUllle that they had become

codex (one containing Orro OP FtwSING'S Historia. ?). standing ingredients of an elementary cycle of Mediaeval studies. rI

Z9 8 299
",
 te~; IZI r De animalibusj IZI" De vestimentis; IZZ" Ornaments
that the codex came from Mondsee (I). Dr. F. Unterkircher,
muliebra (from 1I8r with OHG. glosses).
however, would rather place it in Western Germany. Through u3 r E:x:cerpts from Isidore, Etymologiae.
its collection of glosses it is closely related to Steinmeyer's C u6r {8]m et calJei unum III:IIt, etc. (glosses Jerome's ProhJgw galeatw). ,-'
on the one hand, to St. Gall MS. 295 on the other (2). With Iz6" Biblical glosses (continued) to Kings I, etc.
Z14' Hrabanus Maurus, De clericonmt institutitnu! I, 14-23·
to the latter it shares part of the didactic material referred to ZI7' Glosses to the Psalms; 241" Glosses to Hebrews, Apocalypse.
supra; more of it is found in St. Gall MS. 899, Munich MS.
lat. 19413, Vienna MS. ~732 and probably many more. The De infJentione material shows the following .order :
Fifteenth century leather binding over wooden boards with two clasps fol. 97v, II., 1-17: the Greek. alphabet from A to H, with
(damaged). Heavy parchment, well preserved. The composition of the above each letter the corresponding Greek numeral
codex is rather involved; according to Steinmeyer it is made up of two originally
independent manuscripts: A = fols. 1-63, B = fols. 97-183, to which further
and the name of the letter, to the left the Latin equi­
material was added when the two parts were brought together. Arrangement: valent and below the Roman figure indicating its
3 IV {I-Z4] + IV (+ I) {Z5-33] + 3 IV {34-57] + III {S8-63] + V numerical value. The last line contains the numerals
{64-73] + IV (+ I) {74-8z] + V {83-9z] + II {93-96] + Z IV {97-IIZ] enneca (9) - seranta (4-0), corresponding to the first
+ V {U3-Izz] + IV {u3-I30] + V {I3I-140] + IV {I+I-I48] + Z V
{I49-I68] + 3 IV {I69-I9I'; two fols.marked I8z] + IV(+ Z){19z-ZOI]
five letters on the next page.
+ 4 IV {ZOIb11-z3Z] + ? {Z33-Z39 : mostly single folios] + I (+ z) fo1. 98f, n. 1-8: the rest of the Greek alphabet (same arrange­
(Z40-Z43].
~~~.
ment): the scribe omitted the name of the M. At
At least one quire is missing after fol. 8 and one after fol. IZZ. Fonnat III x
90 nun; written area on the average 70 X S6 nun; one column; 11-17 lines the end of the alphabet there is a reference mark
to the page. The codex is written by severai hands; from the set-up it appears ,repeated to the left of 1. 15, where the Greek numerals
that the De invent:ume text belonged to the manuscript from the very beginning. for • 2000' - '10000' (dischile - mire mia) are
COntents (J) : given; 9-14: the text on the Greek alphabet, whic,h
fo1. I Y Biblical glosses covering Genesis-Ruth (including OHG. glosses).
is complet~d on fol. 98v, 11. 1-4·
_ _ _)97' On pronunciation (Martianus Capella, De -pUis III, § zI6).
97" De inventione Iitterarum (with accessory material; no title) and fo1. 98v, 11. S-II : the text on the Hebrew alphabet; the last
non-alphabetic itemll, IOZ' Hadrian, .. Animula nudula ibis ad word is incomplete: litte[rarum]; 12-15: the letters
loca pallidula, etc. "; explanation of at'temon; IOZv-I03': on the
nymphs (Isidore, Etymologiae VIII, II, 87); 103" blank. thau - he of the Hebrew alphabet, with for each its
108r De talentis; 108'1 Inventom (" Uulturnus. Im'entof ferrari~ artis,
etc. "); names of the winds in Lat. and OHG.; 101)' names of the
months in Lat. and OHG.; 101)" Amo" Db oculis oritur, etc.; on the
! name and value. In the last line only the names of
the remaining fou~ letters were written; the letters
four cardinal virtues; no" De medicina; lIZ", De X nominibus
themselves and their values are found on fo1. 9«1',
Dei (Isidore, Etym. VII, I); De diis gentium (Ibid., VIII, II, I.Z); 11. 1-2.
II3 t Excerpt from Augustine; 1I3" Nomina musarum; 114' divi­
sions of music (Isidore. Etym. III, 19, I); definition of epictll'tlt1UJ; f fo1. 99r, 11. 3-14: the text on the Latin alphabet, followed by
114" table with the degrees of consanguinity; IISr consanguinity that alphabet from a to x; above each letter its name:
terms, with OHG. glosses; n6" Xmodochiu:m. est lDcus _abilis,
etc.; 117' explanations of isagoga, natura, jJnSOfla, periernumi(!,
catagori(!; 117" the order of Creation; II8 r De omatu eccIesiae;
118" De edi5ciis; de vaais; II9" De fettanlentis; IZO" De cUlturis
! a, be, ce, etc.
fo1. 99 V ,1. 1 : x and % of the Latin alphabet; 1-10: the text on
Aethicus Ister's alphabet, with, on 11. II-14, the letters
(I) E. STBINMBYBR - E. SIEVERS. AltluJchdeu.tschtt GIOlUl7I IV, 643.
alamon - thiotimos with their values and names.
(z) E. STBINMBYBR - E. SIEVERS, Althoc1ukrttsche Glouen V, 108 ff. fo1. 1001', 11. 1-2: the rest of Aethicus's alphabet; 3",II : the
(3) Fun details in E. STEINMlMlR - E. SmvBRS, Althochdeutsche Glossen.
IV, 643 ff.; cr. also Tabulae 1,887 f. text on the runic alphabet; 12-14: the alphabet

300 301
~":,
 itself, a - q, with above each rune its· name, to the The Deinventione text in this handy little codex is not only
left its value. very full, but hal! apparently receiv.ed some additi.ons : a second
runic alphabet, not found in any of the.other versions; several
foLIoo", ll. 1-2: the rest of the runic' alphabet. Mter %:
cryptic alphabets, one of which is connected with the Greek
Item, and another runic alphabet (3-6); 7-14: mono­
numerical system, two perhaps ultimately with the isruna
grams, continued on fol. IOIr-v.
tradition; interpretations of the Hebrew letter:"names. Even
fo1. 104r , II. I-II: text on the notae Caesaris; the notae them­ some non-alphabetic material has found its way into it, en­
selves are not given, although the text refers to them dangeri.ng the continuity of the treatise; the notae Caesaris
(ut supra in pau.cis os,ensum est: yet fol. I03v i1 blankl); were omitted. It is not el\SY to decide whether the scribe of
11-14: the formula.for the dotted no.tae saneti Roni/atii . Vienna MS. 1761 may be considered.identical with.the compiler
(a :, e :" . ,0 =::, u = :.:) and an example responsible for those additions, or whether he found them in
(transcription: INCIPIT VERSUS BQNIFACII his exemplar.· If I prefer the latter explanation, it.is because
ARCHIEPISCOPI GLORIOSIQUE MAR/l04v/ it may account for the somewhat disorderly appearance of the
TIRIS). whole. The scribe seems to have pmceeded rather carefully,
fol.x04v, ll. I-II: the text on these notae, followed by the but .with little understanding for his text; this will appear
formula for the subl\titution variety (R for A, F for from the mistakes it has .in common with Vienna MS. 1609, etc.
E, K for I, Pfor 0 and X for U), and by an example .This also applies to the runes: the scribe took pains to copy/ f J
(transcription.: KARUS. CHR/STUS FORTIS irrelevant or spurious details as well· as essential· features. I
TIRO INSTAP [for I NSTAR] SAFFIRO /loSr/
ARCHI .TENENS . SCRIPTOR REGNI UT .4. Melchior Goldast, Alamannicat:Um rerum scriptores aliquot <
DEC US AURI). veteres (1606) (I). .
fol. 10Sr, II. 3-7 : a key for cryptic writing based on the Greek

;Melchior Goldast, or, as he termed himSelf, Goldast von

figure!!, It .shows exactly the same simplifications

Haiminsfeld, was one of the wandering scholars typical for his

and misunder&tandings as that in Vatican MS. Regin.

period (2). He was born in Espen near Bischofszell in 1576

lat. 421 (cf. the Appendix to Chapter II); 8-IS: a

or 1578 and studied in Ingolstad and Altdorf. For some time

key to a system of cryptic writing copied from the

he lived in St. Gall and Geneva, received a doctor's degree in

isruna : the Latin alphabet is divided into three groups

Heidelberg' and collected a remarkable library, apparently

of six and one of five letters (see also the. Appendix

not always by the most honest means. He was actually brought

to Chapter II).
to trial by the Council of St. Gall for having damaged manuscripts
foboSv, n.I-S: a fictitious cryptic alphabet; 6-11 : a key to and for having removed folios and even entire manuscripts
cryptic writing, in which the Roman figures I -XXIII and books. Goldast denied the charge, but was found guilty
are substituted for the letters. a - %; 12-13 : a. con­ and condeInned to a fine. During the Thirty Years' War
cordance of the Greek letters A - K with the Roman (I) H. ARNTz's entry in his Biblwg"aphi.eder ~, 69, no. 1067
figures I - XI (error for XX ?). These two devices
I are also found in the Vatican manuscript. L. 14.
•
J
seems to mix up three things: GoLD~8 sruWicamm r~ Icriptorel (1605),
the first edition of the Alammmicamm rerum Icripiores, and the 1730 edition of
I[ begins an explanation of the names of the Hebrew
\ letters. (ALEPH InteTpretatur doctrina, etc.).
i
the latter ~ork, where the runes actually occur in the volume and on the page
indicated by ARmz (vol. II, pt. I, p. 67). .
(a) D1etWnnaire historique et biographiqu8 die la Suisse III, 485­

302 30 3

,
....01'

.(
Goldast, who had not returned the produce of his activity to notae Caesaris precedes the text on that device. The
St. Gall, found a safe refuge for it in Bremen. After his death paragraph on the notae saneti Bonifatii, continued on
in l635 it was acquired by the authorities of that town, and as
a result a good deal of material relating to St. Gall is still kept p. 93, is preceded by the dotted variety and followed by the
there. a b variety of that script. The whole IS completed
The second volume of bis Alamannicarum rerum scriptores by the text on the monograms and a series of eighteen.
contains, besides the Lex Alamannorum and a collection of examples.
charters, a few short texts attributed to Hrabanus Maurus and Apart from the substitution of v for consonantic u, and of at
Walahfrid Strabo, and the Keronian glossary. Goldast is our only for (, there are very few differences between the first edition
authority for the title under which the treatise goes, and for of l606 and that of l730, and none in the text relating to the
the attribution to Hrabanus Maurus. The inscription in his runes, nor in the runic alphabet (1). The engravings cut to
edition is: Hraham Mauri Abbatu Fuldensis, De inventrone facsimile the runes and the letters of the other alphabets seem
li1fgUaTUm ah hebraJa usque ad theodiscam, et noti; antiquis. to have been made with great ~are. The only peculiatity is
Since Goldast's original does not seem to have come down to presented by the 21St and 23rd runes. The o.ther versions all
us, we cannot be absolutely sure whether he did actually find have the OE. ea-rune for:t, and a form derived from the x-rune
some or other title in his exemplar, or whether he simply f9r Xj in Goldast's edition th~ two forms have been interchanged.
invented one. But the latter is no 'doubt the more obvious On the relation of Goldast's text to the manuscripts· known to
explanation: no title occurs in any of the versions that have uS t cf. p. 345 ff.
come down to us.
The De inventrone text is found on pp. 91-93 of Goldast's
work. It is arranged as follows : 5. Heidelberg, Universitiitsbibliothek, MS. Salem 9·39
(saec. XIP).
P·9 l the text on the Hebrew alphabet, followed by that
alphabet j each letter is. preceded by its value, the
names are written above the letters. . The order of the The runes of the Heidelberg manuscript were first edited
by K. Bartsch eighty years ago (2). The De inventione text
.letters is thet - alephlphe - iothl tltau - sade. Then
itself, however, seelDS not yet to have been examined; in fact
comes the text on the Greek alphabet, followed by the
the whole manuscript has received very little attention thus far.
letters, with above each its name, to the left its Latin
It seelDS to belong to the oldest stock of the library of the Cister­
equivalent, and below the Roman figure indicating
cian monastery at Salmansweiler (' Kloster Salem ') founded
its numerical value. The text on the Latin alphabet,
which begins on this page, is continued on in Il34. Some of the manuscripts of this oldest group came
from France, others were written in Germany (3). The origin
p. 9 2 ; then comes the alphabet (capitals) with above each of MS. 9.39 is unknown; out we shall see that its Deinventione
letter its name (a be ce, etc.). The paragraph on
Aethicus Ister's alphabet comes next, followed by
{I} In the 1606 copy of the KoninkJijke Bibliotheek in Brussels the , of
those' Istrian ' letters' (name above each letter, value NordmatItIOI fell out; 1 could not ascertain whether. this also happened in other
to the left). The text on the runes, with the runic copies of the same edition.
alphabet (same arrangement of names and values) (a) K. BARTSCH, HandscJrrift mit H,abamu' Alphobete. Germanill 17 (1 87 2),
407 f .
. closes the series of alphabets. A set of twenty-three (3) P. LEHMANN et al., BibliotlUllukataloge I, aS4 ft.

.30 4 ::t.+ 30 5
 text is based. on type A, which probably means that it originated 26-27 : blank; 28-32: the text on the notae lulii
in Germany. Caesaris; 33 blank; 34-35 : the punctuated. notae
Pressed pigskin binding over wooden boards, two clasps. Parchment of sancti Bonifatii; 36 blank; 37-41 : text on the notae
varying thickness, well pmJerved;· 133 folios, mostly bound in quatemions. sancti Banifatii; 42 blank: 43-46 : the substitution
Format 345 X 235 mm, written area 265 X 80 mm (x 2). 2 columna with
45-47 lines each. The manuscript was written in the latter half of allee. XII.
variety of these notae; 47-52 : an alphabet with the
An insular prototype (or one showing insular i.n1iuence) may be postulated : numerical values of the letters: a I to z =
twice we find an abbreviation H, which can be explained either as h- autem, XXIII.
or as .. =mim; 7 for et is fairly common.
fo1. Ir Twelve monograms in an ornamental frame (Matheus
Contents :
fot. :IV Capitula; Isidore. Episklla ad B,aulWnem.
- Paulus), followed by a list of notae Caesaris
I2V Isidore, EtymologiaJ!. (Augustus - Marcus Caesar) [fol. I v : a map of the
132v Greek-Latin gloaasry (I). world, in a circle].
I 33'-v , Ir, 2 r De:invan~ne Iitterarwn (no. title).
fo1. 2' Six monograms in an ornamental framework (Do­
~, .. The De inventione text was added by a different but con­
f', minus - Iacobus, with Hebrew, Greek and Latin
temporary hand, which, since fol. I 33 did not offer enough
) equivalents, e. g. hebraice hel, grece 8EOC, latine
, space, inserted the remaining monograms on fols. Ir and 2r.
<"
deus); notae Caesaris (Nero Caesar - Comes); one
Owing to rubbing, the text has suffered so as to become almost
more monogram under the inscription Pacificus fecit
illegible in a couple of places. It shows the following arrange­
(but the monogram contains M, :N, Land 0); a
ment (d. plate VIII a) :
rota with the inscription spera pitagori; a Latin alpha­
fol. I 33 fb , 11. 16-35 : the Greek alphabet, each letter accom­ bet with numerical values (A = III. B = XXV-h
panied by its Latin equivalent, name, numerical etc., cf. p. 335).
value in Roman figures, and the Greek numerals
mia - chile; these last are continued to mire mia on This arrangement indicates that the De inventiane text was
36-37; 39-44: the text on the Greek alphabet; added after the manuscript had been completed, and in such
45-46, and fol. 133 V &, 11. 1-10: the text on the places as the set-up of the codex still afforded. This led to
Hebrew alphabet, followed by the .alphabet itself some confusion: the text .on the monograms was omitted. It
with the names and values of the letters (II-19); is not impossible, however, that the lacuna occurred in the
20 blank. exemplar of the Heidelberg manuscript: the same text is
missing in Vienna MS. 1761. THe text of the Heidelberg
fol. 133 V &, 11. 21"28: the text on the Latin alphabet; 29-30 version is very full, at least in the ;initial paratraphs. It shows
blank (or erased e); 31-37 : the Latin alphabet, with a shifting of elements which we hive met 1,efote : instead of
the names of the letters; 38 blank; 39-44 : the text opening with the text on the Hebre\iV alphab.t and that alphabet
on Aethicus's alphabet; 45-52, and itself, it has the Greek alphabet first; the explanatory text
fol. 133 v b, n. 1-6: Aethicus's alphabet, with the names of the follows. This finding has some importance for the history of
letters; 7 blank; 8-12 : the text on the runes; 13-25 : the De inventione treatise. Equally important are the traces
the runic alphabet, with above each rune its name; of an insular prototype (or one exhibiting insular peculiarities)
in the Heidelberg text. The alphabets show as a rule very
(I) Related to the ~ta Einridlenria and the /unmeneumata Vaticana;
quite close to the fragnumtum BTWl:elieme (CGL III, 393 ft.), but offering a little wear; the runes e. g.appear in remarkably genuine forms,
better text. especially if one considers the late date of the manuscript. It
306 30 7

i'
 was probably copied from a much older exemplar, which can text on the Greek alphabet, and that alphabet itself,
'only increase· its value. with above the letters their numerical value and
their name, to the left their value. After the symbol
for' 1000' : mutatori{ Htterl';. ! :Ab. Fe. hi. Po. xu, i.
6. Number-g, Germanisches Nationalmuseum, MS. I966 e. the key. to the substitution variety of the notae
(RI. 69IZ) (saec. XI!'). sancti Bonifatii (the first pair should be Ba).
Finally, the text on the Latin alphabet, which ends on
This manuscript formerly belonged to the library of Freiherr fol. I2.2ra, and is followed by a majuscule and by a minuscule
von und zu Aufsess, the founder of the Museum. Its further alphabet; above the letters of the latter the Roman
history' is unknown : former marks of ownership seem to have figures I - XXIII were inserted, this giving the
been erased or cut away. As dosely 'related versions were key to a variety of cryptography.
written in Southern Germany or Austria, the same may probably The text on Aethlcus Ister's alphabet, and that
be assumed for the Niirnberg manuscript. I
alphabet itself; the text on the runes, followed by
Modem binding (19th century). Pl"Ircbment of medium thickness, well the runic alphabet.
preserved, slightly tumingyellow. The IZ4 folios originally fonned fourteen
quatemions lind. one sextemion; two folios are missing, one after fol. 91 and. The exemplar used by·the scribe seems to have been a rather
one after fo1. nI, 80 that actually IZZ folios remaiJi. Fonnat zSs X 195 mm, poor one; especially in the last two alphabets there are obvious
written area zao X 140 mm; one column fols. I-IZO', two IZO""-UZv ; 34 lines
to the page. Two scnbes seem to have wri~ this codex: A = fols. I-IZOr , signs of 'wear'. In Aethicus's alphllbet the letters c;and d
B IZOv-IZZv, not' later than saec. XII med.; they an:: practically contem­ have become one symbol; the next three were shifted one place
poraries. to fill the gap, and the order was restored only by dropping the
Contents : name of the g; at the end of this same alphabet, too, the scribe .::
fol. l r fragment on vices P). had quite some trouble to. get the letters under the right names.
IV list of Latiit words. Then (in uncials).: Incipit in moralis /
The runic alphabet shows a spurious vertical stroke (practically
Job Beati Gregorii I Pap~ Urbis Rome I Pars secuqda / Liber
sextus = Gregory the Great's Moralia: in Job, Books VI - X. a second i-rune) between m and n; b, k, n, p, r, s, t have decadent
UOVI\..b Conftictus virtutum et vitiorum (poem, 54 11., inc.: Uos forms, and this first impression is strenghthened by the rest
qui sub Christo roundo certstis. in isto I Discite uirtutum of the treatise.
confticnu et u,iciQrum).
Izovb Liber sententiolarum.
Yet the scribe (or a contemporary corrector) took some
UI Ya De in~tione litte~ (~o title).
trouble to give a correct text : in the paragraph on the Latin
uz rb Greek numeraJ8· with corresponding Roman figures: I mia,
alphabet greci [litterasJ was corrected to grecas, in that on the
1ld1a, IIUriD H.D.CCCe niaeusin. One line blank. Notes
on v~OWl ~rds.i soltmmitlU, scoria, nnapis, 1tu:TiJegium.
runes teotiscam to theotiscam.
The tflirty roads' of Ronxe; the seven hills (end. missing), Many of the errors and peculiarities enumerated above are
!liZ' was ~y blank; later were added a German poem (s I/Z also found in the other manuscripts; cf. 'the general discussion
lines) and two lines of Latin.
on p. 345 if.
The De mYJentione text is arranged as follows':
fo1. 121 va: the text on the Hebrew alphabet, followed by the 7. Vienna, Nationalhibl!othek, MS. IOIO (saec. XII).
first seven letters' (the order being reversed thau
ain); above each letter its n~me, to the left its value. This manuscript has received very little attention thus far,
fo1. 12Ivli: the remaining Hebrew letters (samech ---"-.aleph); the at any rate from runologists.

3 08 30 9
".
,r
 It may have originated in St. Florian (Upper Austria); it century. There is even a better reason for supposing that it
belonged to the library of this Augustine Abbey in the 15th was part of the codex at a very early date : in the closely related
century, as appears from entries on fot. I r and 78 v • As early Munich fragment it is also found in the neighbourhoo d of the
as 1576 it belonged to the Imperial Library in Vienna. From Physiologus Chrysostomi. This can hardly be a coincidence,
Tengnagel's hand it received no. 99, from Blotius's M 3869. so the two texts must have been found together in the common
The binding, white calf over beavy wooden beards, is of the 14th century ancestor.
(repaired 1911); two clasps are missing. Strong parchriient, well preserved.
The 166 folios show the following arrangement: The text of the De inventwne litterarum is crowded in rather
10 IV [1-80J + ? [81-89J + I [go-9IJ + 3 IV (9'l.-1I5J + II [II6-119J badly:
+ ? [1'l.o-l'l.4J + I [1'l.5-r:l.6J + 5 IV [u7-166J. fol. 9or, II. 1-4: the text on the Hebrew alphabet, followed by
*1 - fly-leaf from a 13th century hymnsryj *'l. = a fly-leaf from a I'l.th century
Hturgicalmanuscript. Format'l.70 X 190/195 mm, written area 'l.00 X I30mmj
that alphabet (5-7), with the names and values of the
one column, '1.6 to 32 lines to the page. letters (order: thau - aleph); 7-12: the text on the
Contents (I) : Greek alphabet; 12-15 this alphabet, with the equi­
fol. If Caecilius Cyprianus, Duodedm almsiva.
valents, the names of the letters and their numerical
lOY Augustine, Dicta th discip/ina ekristimwrum. values in Roman figures; 16-20 : the text on the Latin
18· Martin of Braccara, I..ilHIllw th qIUlttuoT viTtu.tibw. alphabet, followed by the alphabet, majuscules and
23' Ansbert of St. Vincent, S_o de cupiditate. minuscules (21-22); 23-28: the text on Aethicus
35· Augustine, Soliloquia.
62r Baudemundus, Vita S. Amandi (excerpt). Ister's alphabet.
63 Y Pturio S. Columbae. fot. 90v, II. 1-3 : Aethicus's alphabet, letters, values and names;
. 65 r Phytiologw, ascribed to St. John Chrysostomus.
74r Pseudo-Jerome,- Epistola ad Pmdam fit Ewtochiwn de asmmptio1Ul 4-7 : the text on the runes, and the runic alphabet;
S. Mafiae. runes and names (8-II); 12-16: the text on the mono- ~'.
gor De inventione Iitterarum (no title). grams, with one example (DOMINUS). Ll. 17 ff. :
90v De vocibus variwn animantiwn.
92r Ambrose, De paradiso (from De myltmis).
Voces variae aniIl13ll,tium.
I 17v Ambrose, Disputatio th servl'Ulda ani_ puritate. The letters of the various alphabets (including the runes)
II9r De Salomone (excerpts from Augustine, Gregory, etc,).
120' Ambrose, De poenitentia. have been partly filled with red, or ornated with red dots. This
uS r Ambrose, De mysteriis (cf. fol. 92'). evident care hardly compensates the poor state of the text in a
133' Ambrose, De poenitentia (continued from fol. I'l.4Y). number of places. Aethicus'salphabet can hardly be deciphered
163r Ambrose, De sat:f'amenm (fragm.).
165' A parallel between the Ten Commandments and tbe Ten Plagues
as names, .letters and values are badly entangled. But a com­
of Egypt. parison with closely related, manuscripts will show that most of
As the De inventione text occurs on an independent double the real blunders in the text· come from the prototype. More­
folio, it is not possible to decide for sure whether it belonged over a number of mistakes were corrected by the scribe : fot.
90f, I. 10 litterarum has a from u; I. II adiuncti from adiuncte.
to the manuscript from the very beginning. At any rate it
was written by a contemporary hand, and must have belonged to The runes received rather more space, and the names are
the bulk of the codex when it was bound in the fourteenth carefully written above them. Their forms show little under­
standing on the part of ~he copyist; yet they were probably
(I) Tabulae I, 175; imitated fairly carefully.
H. J. H.i!.RMANN, Die tUJutschen romanischen Ranihchrijten. (Leipzig, 1926),
.85-188, with further bibliography.

310 3 11
..
"~
 8. Munich, Ms. A. Weinmiiller (saec. XII in.). 9. Vatican Library, MS. Regin.lat. 294 (saec. XI/XII) (I).

This fragment was identified some ten years ago and edited The collection of alphabets in this manuscript stands· in a
(with facsimiles) by I. O. Plassmann and W. Krause. I was class by itself: it is obviously derived from De inventione
unable to collect infor~tion on its origin and history. The litterarum, but the text has been much shortened; on the other
antiquary A. Weinmiiller lent it to Plassmann, who gave a brief hand new items have been included. In fact this version is an
account of it (I). It consists of 10 folios measuring 28 5 X attempt to adapt the old treatise to the needs of a period which
205 mm. Contents: had a better knowledge of Hebrew and especially of Greek
fol. 1 : De inventione litterarum (no title). than was current at the end of the·eighth century.
fol. 2 : Physiologus, aScribed. to Chrysostomus. The manuscript once belonged to the Cistercian Monastery
of Langheim near Bamberg, founded in 1133 by monks 'of the
According to Plass mann, the manuscript was written in Austria Abbey of Ebersberg in Bavaria (fol. Ir " liber sancte marie i(n)
ca. A. D. 1100. The De inventione text is arranged as follows: lanchheim "saec~ XIII). In the sixteenth century it was in
fol. Ir, ll. 1-3 : the text on the Hebrew alphabet, followed by the possession of the humanist N. Petau (fol. 3 : Nuro 6o.N.Pet.
that alphabet (letters, values and names: 4-5); 6-9: 1656, and below: volumen LX. Non Petauianum); afterwards
"
the text on the. Greek alphabet, and that alphabet it belonged to the library of Queen Christina of Sweden (where
(10-12, same arrangement); 13-16: the text on the it was no. 1896 or 1898) and with that collection it was incor­
Latin alphabet, followed by a minuscule alphabet porated into the Vatican Library.
(17- 1 8; only one capital, A, is given); 19-23 : the text Red leather binding (last quarter of the eighteenth century), stamped with
on Aethicus Ister's letters, and those letters them­ the anns of Pope Pius VI and Cardinal F. X. de Zelads. Finn, somewhat
greasy parchment; the last quire has much suffered from dampness. Com- ,"
selves with their values and names (24-27); 27-3 0 : position: 15 quires, nwnbered on the first page (saec. XVI) :
the te:x:t on the runes. +
I (1-2) +
7 IV (3-58; fol. 21& is a single folio inserted after fo1. 21) II
fol. 11. i-6 : the runic alphabet, with above each rune its
1 v, (59-62) + 6 IV (63-UO).
At the end one quatemion seems to be missing (cf. infra). Format approx­
name. It is followed immediately by the text on the
imately 260 x 180 mm, written surface 210 X 133 tnm; one column, 53 lines
monograms (7-II) and one monogram (Dominus). to the page. Written by one scribe (except fol. 21&, which is by another but
The text: shows a few corrections : hebraictJ. has been added in contemporary hand), probably in Bavaria, in the late eleventh or early twelfth
the first line (mark between littertl. and lingu't,); in the paragraph century. The same scribe made a nwnber of marginal additions (int. al. on
lol. 29v the OHG. names of the months: Wmtannarwth, etc.); other marginalia
on the runes Jteotiscam has been corrected to itheotiscam. . A in.saec. XIII and XV hands. In the seventeenth century a reader tnmscribed
couple of letters were repeated in the margins, into al. the i-rune some rubrics which had much faded (fol. IV).
at the top of fol. 1 v, just above tb,e corresponding rune of the Contents (2) :
alphabet. fol. IV A collection of alphabets (cf. injra)

The whole is written with care; some effort has been made to 3' Isidore, Epistoloe ad Brauliommt, et fJice versa (Letters IX, X, Xl,
XII, Xlm.
give an artistic touch to the initial of ea~h paragraph, and the 4-r Isidore, EtymologiDe.The text breaks off in Book XIX, so at
letters themselves (e. g. the runes) were obviously Copied without least one quire must be missing.
Joss of details, nor with malformations of any importance. The inclusion of the alphabets may be due to the scribe's
r
(1) A. WXLMART, Codicl!5 II, US ff.
6) J. O. PiAssMANN, Die Hrabanilche R1mI!tIreihe. (2) A. WILMART, CtHliCI!5 II, 125 ff.
3 12
313
wish to have a blank first page. Since the text of the Epistolae scribunt. dyptClngos scilicet. ut ai dyptongus per e.
began on a recto side, the simplest way to obtain a blank page breuem ut kai quod sonat ke. & oi pro i. & oy pro
at the beginning was to insert one leaf (two folios). This then u./ & u uocalem sono i pToducte exprimunt.
left 3 pages (Iv, 2r-v) which could be used for various notations.
(5) CASUS GRECORUM: ONOMATIKHC nomi­
Our scribe filled them with a number of alphabets which would
natil,lus, etc.
come handy at the l'tage where the Etymologiae were used in
school. Of course there can be no question to call the runes (6) ARTICULI IPSORVM: 0 hic, TOY huius, .etc.
in this manuscript" Isidori-Runen ", no more than in the case Each of three lines gives the forms for one gender,
of Brussels MS. 9311-9319 (cf. p. 66). as indicated at the end of each line: Masculini/ ."
The renovated De inventione version shows the following Femin[ini] / ... Neutrales.
arrangement : (7) LITERE quibus utuntuT marcomanni (I) id est
fol. I v (I) a Hebrew alphabet, inscribed: Litere hebraice northmanni a quibus originem tTahunt qui theotiscam
.XXVI. quibus nunc utuntuT iudei. The alphabet locuntur. This much faded rubric was repeated
is of the square Hebrew type and has in fact 27 in the lower margin by a seventeenth century hand
letters, with above each its name, in the order thau (var.: Nortmanni; loquuntur). The runes are
- aleph. A number of variant forms are included arranged in ·two lines, a - n, 0 - z, with above
(zadic andzadi, phe and pe, nun and nun, mem and each its name, and to the right (except a : to the
mem, Tchaph and kaph). left) its Latin equivalent.
(2) a Greek alphabet with the inscription LITE~ (8) L[I]TER[E] ethici philosophi quas Hieron[imu]s
GRECORVM. The letters are in~erted in the usque ad nos perduxit, repeated in the right margin
middle of their names: alApha, beBta, etc. The by the same seventeenth century hand (var.:
names show that this alphabet belongs to the De Aethici ... Hieronymus). The alphabet covers two
inventione tradition, e. g. epismon lauta simma; the lines : alaman - nabaleth / ozechi - zeta.
symbol ICo7T7Ta and the numerals at the end have (9) Notes on the declension of the names of the letters
however, been omitted (but cf. fol. 2r). and of numerals : Omnia literarum nomina tam apud
(3) a cursive Greek alphabet, inscribed LITE~ hebreos quam apud grecos & latinos indeclinabilia
VVLGARES GRECORUM.Above the letters sunt, etc.
their Latin equivalents have been indicated, in three
fol. 2 r , in 8 vertical columns,:
cases the names: ita thet psi. This alphabet is
obviously of a later date than the preceding one, (I) the Hebrew alphabet usually found in the De
although the· two were copied at the same time. inventione tract; here it is inscribed: Hebraice
liter~ / XXII primo inuente / a moise 7 ab ezra /
(4) the series of graeca is continued with a'list of the
renouate. The letters are ornated with colour
Greek diphthongs: nYPTONGI GRECORUM :
patches; to the left of each is its Latin equivalent,
At = e, Et = i, Ot = y, Oy . u, Ay = au, Ey = "eu
(repeated each with initial C: Cat = se, etc.),
to the right its name.
followed by a short comment : .Sciendum est quod (I) A. WILMART, Codices II, I2S, reads marcomannici, obviously misled
gTeci quasdam sillabas pTonunciant / aliter qulVn by .i. = id est which comes immediately after marcomanni.

3 14 31 5
".
 (2) the Greek alphabet, this time including the numerical itself better to ,an arti$tic treatment than that on fo1. IV; more­
symbols omitted on fo1. Ir; same ornamentation and, over, the scribe and the compiler need not have been one: person.
same arrangement as for the Hebrew alphabet; the Although both the text and most of the alphabets are obviously
inscription reads, : LiteTe;; grece / numervm / demon­ derived from De' in'fJentione litterarum, close inspection detects
stran/tes. a. number of mistakes not found in other versions. Some of
(3) the Greek numerals from Mia to ,Mire. Mia- these, especially in Aethicus's alphabet, point to a prototype
CHile correspond to the letter.s A "Y' of the using open a : chata (for *chatu), effosta (*ejjostu), perhaps also
alphabet. delfa (for deljvi ?). This also explains the rune-name laga (for
lap). Perhaps this prototype had insular features, as this
(4) Numeri / cardinales : Vnus - Mille. would account for a mistaken chopiceph (elsewhere chori-) (1).
(5) Ordinales: Primus - Millesimus. But in view of the forms betraying a transmission through many
(6) Aduerbiales: Semel - Milies. stages (gagijod phiorin agathin caim), it will perhaps be safe not
to press this point (2).
(7) Disperti/ui : Singuli Milleni. The runes are neatly drawn, but some show rather decadent
(8) (no inscription) Singularis - Millenarius. forms. At first the scribe had skipped t~e symbol for q, but
The Latin material in (4) - (8) does not correspond to the afterwards he inserted it in the right place. '
Greek numerals in (3); in the former the numerals' I I ' - ' 19'
are included, whilst they are not given in the latter. ' In (3) 10. Vatican Library, MS. Urbinas lat. 290
Nia Deca Ecusi Trinta and Tesserenta (9, 10, 20, 30, 40) are (sae~. XI in.).
written on an erasure. '.
Fol. 2 v is also divided into columns, but only the first three This manuscript has been described in Chapt~r II, wher:e

are filled, and these o'nly partly: I also examined its background. Its isrun~ fuporc is preceded

by a runic alphabet, which is derived from a De in'fJentione

(I )Ponderales : Simplum - Sedecuplum.

collection, but the introductory text to the alphabet is unparal­

(2) Complicatiui: Simplex - Sedecuplex. leled: Literas sequentes / cumminio colore nota/te (I)

,(3) Aduerbia ex his : Simplici~er Tredecupliter. 'nordmanni in suis usitant (2) / carminibus & uocantur

The rest of the page is blank. ,apud eos / rune;;. Sunt autem nonnulli qui opi/nantur

quod quando gothi & / uuandali gentes de finibus Jnord­

The order of the De in'fJentione litterarum elements has been mannorum eg'redientes / per germaniam & italiam ad /

completely upset: [Greek] Runes - Aethicus - Hebrew mare uenientes'perque iUvd I transuecti in affrica co~istel

- Greek, whilst the Latin alphabet has simply been dropped. bant; crescente apud eos christi/ana religione christiani

As far as the runes and Aethicus's alphabet 'Yere concerned, ex parle / effecti (3); doctores eorum tam nouum / ,quam

the compiler had no improved versions which he could substitute uetus testamentum in suam /linguam hoc (4) in theotis­
for those in the De in'fJentione collection j but he had a wider and
better knowledge of Hebrew and especially of Greek. It is (1) Cf. also the abbreviation 7 fpr ttl in the inscription of the second Hebrew
somewhat striking that the scribe bestowed more care upon the alphabet (fol. :ar ).
(:a) Heidelberg MS. Salem 9.39. with typical insular features, gives a l)llI11.e

I
traditional De in'fJentione Hebrew and Greek than upon the new, garfod for the g in Aethicus's alphabet; this garfod reminds one immediately
more scientific material. But the arr~ement of fo1. 2 r lent of gagifod in the Vatiam manuscript.

3 16 317
",

 cam (5) uel I intheotonicam conuerterunt I cum istis information on the history of the Goths from various sources;
litteRiS.
that of the Vandals would be harder to trace; but I know of no
(I) notateJ for notatas. one source where the two are combined as in the Vatican text.
(z) the scribe first wrote ua (cf. following uocantur), The equation theatiscam vel theotonicam is of special im­
then erased a and put s instead. portance: we find the formula teutonica vel theodisca lingua
(3) some verbal form like sunt is apparently missing. (and similar ones) from about 880 on; the examples have their
(4) in this explanatory subclause we expect est;perhaps. origin in Mainz, Fulda and St. Gall (I). In the case of the
the exemplar had .i. (= id est),· which was skipped Vatican manuscript this may be an additional indication that
for following i (= in). its material came (directly or indirectly) from St. Gall (cf. the
isruna tradition).
(5) corrected from theoiatiscam (dots above and below
The runic alphabet has suffered less damage than the fulJOrc;
the. first i and the second a).
yet a number of items must be examined in detail, as quite a
Several elements have been mixed up in this text : few readings are doubtful [M = Massmann (2)].
(al The emigration of the Goths from Scandinavia and their a: of the name only as (with accent on the s) is clearly visible,
settlement in Italy. but there is little doubt that c followed.
(b) The migration of the Vandals through Germany and their d: the rune shows an almost pedect form (M has a form
settlement in North Africa. They made raids on Italy practically identical with the OE. m).
from their Mrican homes, but did not march through it
f ~ the reading of the rune-name is doubtful; at any rate fech
on their way south (which led through Spain).
is more probable than M's feh. ..J
(c) The conversion of the Goths, and the translation of the
Bible into their language. g: M's gubu is based on a misinterpretation of the lower loop
of g; gibu is far more likely.
(d) The compiler may have had some vague notion of the
Gothic alphabet, but he identified it with the runes. . k: the name is clearly kal; no letter seems to be missing after t .
. Gothic was not altogether unknown in Germany about the q: there can be no doubt about the reading chon.
time our manuscript was compiled : there were the gothica in T: the name is partly hidden by a blot; only rek, and one more
Vienna MS. 795. obviously baSed on a Gothic Bible, and the letter which cannot be identified for sure (t? i ?), remain
latter was perhaps available in more than one copy : the Code:t: visible.
ATgenteus makes its first appearance in Werden, the Codex s: the last two letters of t~e name are faint, but probably it.
Car?linus in Weissenburg (I). The compilermay have had his.
Theotiacae exierunt (from Frecbulf of Lisieu:x. see W. KRAUSE, Die
Hrabmrische Rutum.reihe, 178) .
(I) W. STIIl!ITBEltG, Die gotische Bibel I, XXIV f. See especially the extmct and
from Walabfrid Strabo, and d. also
Golpbilas Gothonun episcopua adinvenit Gothicas litteras et quamvis
Qui primus· litteras Guticas inuenit? Goulphyla, Gothonun epis­ esset Arianus utile tamen opus fecit, quia per illas litteras transtulit
copus. Fuit autem da principium mundi usque quod Langobanli divinaslitteras in Gothicam linguain. Fuit tempore Valentis impera­
inltalia.praeaiderunt V milia DCCLXX et II anni, tempore Justiniano. toris (Sigebert of Gembloux, De wu iJIwtrilnu, cf. M. MANITltlS,
impemtore (from Joca numaclwmm, see G. BAI!BI!CKE, VoCilbularilu, r Guchichu III, 348).
Ia. f.);
(1) L. WmsGI!RBER, Die geschichtliche Kroft. tier deuuchen Sprache. DUssel­
- Alii Yero aflirmant eos [Francos] de Scandza insula, quae vagiI),a dorf, 1950, 48.
gentium est, exordium habuisse, de qua Gothi et caeterae nationes (a) H. F. MAssMANN, ~,. as) ff.
3 18
I ". 31 9

r
I
 Aug(ustin)i Ieronimi Gregorij et cetera. On fol. I the owner mark: tiber
x: M read beluch, or, alternatively, beluth.; :Ill a matter of fact
Augie maioris. The manuscript consists of three parts
the first letter. could just as well be h, which would give us A = fols. 1-71; B = fols. 72-152; C = lols. 153-213.
a normal form heluch. Holder dates A saoo. IX. B and C 8&00. VIII/IX. The 61 folios of C fonn
8 gatherings marked 'qI', 'II ',' qUI', -, 'V', 'VI', 'VII ',-j the number
y: M was right in reading horsi:; the third letter looks like y of single leaves is extraordinarily great :
because it seems to form a unit with the accent on i in the II (+ 3) [153-159] + III (+ 2) [160-167] + III (+ 2) [168-175] +
following name, zia. HI (+ 2) [176-183] + III (+ 2) [184-191] + III (+ ~) [192-199] +
z: there can be no doubt about the reading of the rune­ III (+ 2) [~OO-207] + 1(+ 4) [208-'1.13].
Format 187 x 121 mm, written area ca. 120 X 90 mm; one column, except
name: zia. <Ill fo1. 212T (three), 213" (four), 213v (two); 19-20 lines to the page, but 28-30

As to the forms of the runes, their peculiarities are .rendered <Ill foI. 212Y and 33"35 on 213 t •

in fig. 50 as far as they can be made out in the manuscript. Contents (1) :
All in all the runic alphabet shows a fairly independent tra­ fo!' 1": Annales Augien8e8 brevissimi (posteriores).
dition : its d-rune preServes an astonishingly correct form; for h IV: Gre'gorius Magnus, E:cpositio EfJart{leliorum.

it has hagal as opposed to hagale, hagalc, etc. in most other 72": Isidorus, De e«;lem.uticis ojJiei.is.
153": Incipit collectarlo de diuems sententiis = a collection of homilies,
versions, for i is, for k kol, for n nod.(elsewhere not, once noth). ending with Methodius, lWvelationu.
Not all these forms point necessarily to a better exemplar: 2II r : A runic alphabet,'

hagal, is and nod occurred in the isruna fuporc, and may have 21 IV: Incipit orolegium dierum. <

212V : Incipit l\ltania kyri\ll~on.

ousted the forms in the alphabet. Two forms may point to a 213v: Hebrew.Greek-Latin glosses: He1 Ebreum, theus Grecum, dominus
prototype still using open a: heluch (the other versions have a Latinum. .
in the second syllable) and zia (elsewhere ziu). The runic alphabet is written on fo1. ZIIf immediately after ",
the final sentence of the preceding text (Beati qui parati / sunt ~
II. Karlsruhe, Badische Landesbibliothek, MS. Augiensis 254 in ilia hora quando hoc fiunt / Eritque sanctorum & regnabunt
(saec. VIII/IX-IX) (1). cum Christo in saecula saeculorum amen). Liihgin found a
resemblanCe between the handwriting of the rune-names and
In this manuscript the runes are obviously a later addition. that of th~ first entries in the Liber confraterni.tatum. The
They may well have been.written at Reichenau, as the manuscript alphabet looks somewhat like a hasty scribble: the sizes of the
appears to be mentioned .in an' early catalogue (before 838) of runes differ greatly, the rune for 3C being almost three times as
the library of that. Abbey (z). Judging from the handwriting high as that for a. Although some forms seem to be rather
only w.e may place the runic alphabet-for this is all the manu­ clumsy imitations, the runes. have on the whole a fairly genuine
script off'ers-at a fairly early date: probably the second half appearance.
of the ninth century, but certainly not later.. The runes of this The alphabet is written over two lines, a - n / 0 - z; the
manuscript were first edited by T. Langin (3). values are written to the left of the runes, the names above
The codex is bound i,n leather-c:overed wooden boards; two leather clasps them. Langin's rendering does not always agree with the
with m~ta1 mountings. It be.a~ the inscription: Exporic:io Gregory pape in
photograph at my disposal; in the case of e, however, which
40 r / ewangelia/ Ysidorus de officys. I CoUectarium de diuersis untsncys I
often looks practica1ly like il, I have accepted his interpretation,·
(1) A. HOLDER, Die Imchenaqer Handschri/tsn It 513 fl.; .as only autopsy of the manuscript can be conclusive:
H. BEI!SON,lddontudim, 50•.

(2) P. LEHMANN e! al., BibtWtluilukataloge I, 252.

(3) T. LANOIN, Altalemannische SprachqrulJe,tJ., 700. (1) A. HOLDER, Di£ RNhenlllleT Handtchri/tsn I, 573 fl.

320 2S 3 21
",

a: the rune is less like R than Langin's facsimile implies. treatise; there is no indication that it existed independently
c: Langin's reading cheri is right; *chen is clearly impossible before the treatise was written.
(although cheri of course goes back to such a form).
d: throm (with suprascript h) could perhaps also be read 12. Karlsruhe, Badische Landesbibliothek, MS. Augiensis I76
throm, but the former reading is the more probable. (saec. IXI- Xl).
h: the name looks rather like hagalc.
The runes in this manuscript were also first edited by
k: here too I would rather read gilc than gile. T. Langin (1); they seem to have received as little attention
0: Langin's facsimile has a more decadent form of 0 than the from runologists as those of the preceding codex.
manuscript itself; in the latter the right hand strokes do According to B. Bischoff (2) the manuscript was written in
not meet. Freising at the time of Bishop Hitto (812-835). The Reichenau
r: The scribe might seem to have written rcmt rather than library catalogue of 821/822 mentions "in epistolas Ioannis
rehit. evangelistae homiliae X in codice I" (3), but to connect this
y: Langin's reading hu"y can hardly be justified. The third entty with our manuscript causes some difficulty: palaeograph­
letter is precisely the same dotted y as that to the left of ical considerations are in favour of dating the codex in the
the rune. It is hard to decide what letters come after latter part of Hitto's rule (4). It is not known when or where
+
huyr- : either a ligature s t, or y followed by a vertical the runes were inserted: judging from the handwriting, they
stroke. A form huyry is the most likely in view of the may be dated saec. Xl. Whether at that time the manuscript
occurrence of such forms as huyri (Vienna MS. 1761, was already in Reichenau, we cannot know for sure (5). Fol. 3r :
Heidelberg MS. Salem 9.39) and huiry (Karlsruhe MS. liber Augie maioris. '"
'.
Leather-covered wooden boards with two leather clasps and the inscription:
Aug. 176).
Episu.1a Iohannis apposu.1i / TractatIU X super certi.s / verbis eiusdem epi.su.le /
The two Reichenau alphabets are good evidence for the sam::ti Augustini. The parchment has suffered much from moisture, and is
often crumpled. The 123 folios are boWld to form fifteen gatherings; fols.
knowledge of runes (of a rather heterogenous and artificial type, r-8 = one quaternion, with foIs. 4-5 originally 5-4; 9-Il2 thirteen quater­
to be sure) in that Abbey. They also show that this knowledge mons; 1I3-123 == one sexternion, the first folio of which is missing.
cannot have been very thorough, as even a fairly early version Format: 270 X 167 mm, written area 195/200 X ItS/I20 mmj I column.
22 lines to the page.
such as that in MS. 254 has already forms of very doubtful
Contents (6) :
qUality.
fol. 2' A riddle in runic script, a runic alphabet, and various scribbles
We can of course not know for sure whether the Reichenau and additions (c!. infra).
alphabets (and that in Vatican MS. Urbin.290, for that matter) 2" Incipit epistola beati Johanis apostoli (= John I).

are extracts from a complete De infJentione treatise, or whether 9r Augustine, In epi.stolam Johmmis ad Partlws t.ractatus decem.
123' Probationes pennae.
they go back directly to the alphabet tha~ was incorporated
into that treatise; but there are two considerations that allow (I) T. LANGIN, Altalemtl1ll'ti.sche SprachqueUen, 701.
(2) B. BISCHOFF, Schreibschulen I, Il3.
us to favour the former alternative: there are certain peculiar­ (3) P. LEHMANN et aI., Bibliothekskataloge I, 245, 15.
ities in the Reichenau alphabets which also turn up in regular (4) B. BISCHOFF, SchreibschuJen I, II3.
De inoentione alphabets, and which can hardly go back to the (5) A Reichenau library catalogue of S&eC. IX· aIao mentions " In epistolas
Johannis evangeliste II" (p. !.JmMANN et. al., Bibliothekskataloge I, 2 63. 40),
first version of the alphabe~; and, second, the De infJentWne but this entry is still less conclusive.
alphabet is probably a construction especially made for that (6) A. HOLDER, Die Reichenauer Handlchriften I, 412 f.

322
t.
32 3

,(

 In the left top comet: of fo1. 2 r the press-mark CLXXVI (in The runes now missing in the alphabet (6 c d e) can fortu­
pencil), in the middle of the upper margin an a-rune. A large nately be supplied from the text of the riddle, but several of
dark stain extends from the right border tiU about the middle the rune-names remain doubtful or are even completely illegible .
.of the page; aclear patch going from line 5 to 10 has practically Some of the runes have a peculiar form:
removed several runes on U. 6, 7 and 8 and part of the alphabet; a: the original C2 has been interpreted as consisting of two
along the left margin the parchment is much soiled, most of V's placed upside d.own and more or less carefully con­
the runes along it are more or less damaged. nected. In one case (the seq,nd a of causas in 1. 2) the
Ll. 1-8: a runic inscription, obviously based on the alphabet upper A happened to be drawn straight above the lower
written below it. It is a Latin riddle, which may be one, the result being rather like ceo
transcribed as follows: d: this rune is not visible in the alphabet, but in the riddle
NON TIBI SIT NOSTRAS INDIGNUM I it looks rather like a fJ turned to the left. A d derived
NOSCERE CAUSAS I SEX SUMUS IN LUCEM from fJ would be unique in the De inventione tradition.
GENITE I [SI]NE LUCE SORORES I SAL­ which has only the name thorn, but not the rune itself.
TAMUS CANIMUS LUDOS SINE METE (I) I Since the • loop' is a narrow and rather long triangle, this.
[M]ONE[MUS] I HOC [NOBIS] MORS POSSE sign may rather be explained as a d-rune from· which the
DEDIT QUOD I UITA [NEG]AUIT. left half was lost. Perhaps the rune had been damaged
The lines of the poem are marked off by sets of three in the exemplar (cf. Vienna MS. 1761, p. 268), and the
dots (11. 2, 4, 6, 8), ll. 4 and 8 also by a foliate ornament. copyist could make out only part of it.
The word divisions are indicated by dots, but these I: the lateral stroke is sometimes placed so low, that this rune
:~;
are n.ot used everywhere, or else have become invisible. almost coincides with the 'u-rune (cf. infra). ~~
On il. 9-12 foll.oWS the runic alphabet, with above each rune m: there is a rather important difference between the form of
its name (9, II : names; 10,12: runes), below each its value. the rune in the alphabet and that in the riddle; the former
After the last rune there is a foliate ornament similar to those agrees with the other versions of the alphabet, the latter
.on 11. 4 and 8, but apparently by another hand, which may also is much more like an English d. From this we may
have scribbled a few letters more to the right. A later hand perhaps infer that the runic material was written by two
wrote AB and the runes e-g on 1. 19i these are followed by some hands: A the' riddle, B = the alphabet. Actually the
scribbles now practically illegible; 1. 14 marlinus m. orm:mus ductus of the runes in the riddle is firmer than that in the
in me in a big clumsy hand; l. 17 letge (contemporary with the alphabet; the forms of the e- and g-runes also show minor
runes (). differences. But at the same time there cannot be the
In the lower margin a hand not much later than that .of the slightest doubt that the riddle and the alphabet are based
runes wrote a variant riddle (now much faded) : ,.on the same prototype : there is no other way to explain
Sex sumus qu~ ludimus, qu~ numquam lucem uidimus; the peculiar d-rune occurring in both.
Nunc mortui agimus, quod uiui non potuimus.
11 : , is shaped rather like X. but with the end of the lower left.
The solution of the two riddles seems to be I the six strings
stroke turned vertica~ly down .
.of a musical instrument' (2).
Q: the lower lateral stroke has grown into a regular x.
(I) Read: ME[N]TE.
(2) F. MoNE, Anzei.ger fiir Kunde der teutschen Vorzeit 7 (1838),. col. 39
u: the right hand part of this rune has been red~ced to a.
n. 39 .. Die sechs Saiten einer either". short stroke slanting down to the right; the whole looks.
32 4 32 5
"­
 rather like an English c-rune, but the lateral stroke sets off l04V: Libtr S. Michahelis in Monte Babenbergmsi). On fol. I' is the old
press-mark M. a; the codex also has the no. B - IV - a9.
at various heights. Binding: wooden boards covered with pigskin; there were three clasps,
The names of c and d have completely disappeared; those which are now missing. The codex consists of 165 folios and two fly-leaves,
arranged as fonows:
of m and T are partly illegible. That of m is partly obscured by I,a + 8 IV [3-66J + III [67-7aJ + 4 IV [73-104J + 7 IV [loS-160J
the dark stain, so that only the final n is clearly visible; that of T + II (+ I) [161-165].
may be read reht or rehit, but there is hardly enough space for i With fol. 105 a new section begins : the quires comprising fols. 3-66 are num­
bered • XXI ' - ' XXVIII'; in the part beginning with fol. 105 the quires
between h and t. In the margin to the left of this name there are marked • I • - ' VIII'. The whole. however, is written by one careful
is a letter which may be R. In some cases the scribe seems to hand.
have had some trouble to get the right name above the right Contents :
rune : the name of the i-'rune is written below the line, as the 01. I': EpistoJa B. Gregorii ad Secundininum epiacopum, etc.
scribe found no space on the line itself between hagale and 3': the last chapter of Richard's De interiore hominis statu (continuing
fol. IS8v of MS. 130/1).
gikh; at the end the %-rune stands below the name of the y-rune, 3": Capitula, 4"" the text of Richard's De sommo Nabuchodonosor.
whilst its own name stands above the foliate ornament. The 73': Richard of St. Victor, Tractatus mper quosdam PSal1lW1.
scril?e probably first drew the runes, after which he found out 98v : Mystical interpretations.
104v-IOS': Alphabets, etc. (cE. infra).
that he had calculated too little space for the names. A couple
IOS v : S. Augustinus; De haeruibus liber.
of corrections seem to be due to the scribe himself. Above laI V : S. Augustinus, De eccksiastici4 dogmatibw liber.
the 0 there is an erasure, and the name othil is written somewhat 126': Gandulphu8, Flores sententiarum.
higher. In the name huiry the i has been underdotted, and above 14a': Hildebertus Cenomanensis, Versw de Sacramento.
De pontificibus Romania.
it y or r was added, thus giving the reading huyry or hurry 144': Haymarus monachus (1), RJryth'fIIfU de expediiione Hierosolymittmn.
(the latter being less likely). 151': S. Mathiae apost. Acta; Inventio et Miracuili. .;
The alphabets have been written on two pages which remained
13. Bamberg, Staatliche Bibliothek, MS. Msc. patr. I30/z blank when the two parts of this codex (1-104 and 105-165)
(saec. XIII in.) (1). were bound together. They are written with the same care as
the rest of the manuscript, even though the scribe does not
The alphabetic material of this manuscript seems never to seem to have found the right arrangement from the very
have been examined. It is a much revised edition of the beginning:
original collection: the Latin alphabet and that of Aethicus fo1. 104va.: a Hebrew alphabet, inscribed "Hebraicum Alfa­
Ister have been dropped; at the end a Slav alphabet has been betum " and offering (in vertical columns from left
added; the runic alphabet has received the epithet 'Chal­ to right): die values of most letters, the letters
daean '. The introductory texts have practically disappeared; themselves, their names (aleph-taph), and inter­
yet we shall see there can be no doubt that this collection belongs pretations of the latter (aleph Doctrina, etc.).
to the De in'fJentione group.
104 vb: at the top the scribe first wrote "A1fabetum
The Bamberg manuscript is the second part of a two-volume conection
of the works of Richard of St. Victor, to which other texts have been added.
Grecum" which he then corrected to" GRECVM
At one time it belonged to Michelsberg Monastery near Bamberg (e. g. fol. Alfabetum "J "GREGUS Numervs H, "Latinus
Numerus". Below these inscriptions we find in
(1) Kataklg tier Ha:tUhchri.jten tier kOniglichen Bihliothek ""u Bamberg. Bear­
beitet von F. LBrrscm1H. I. Band, Erste Abteilung, III. Lieferung : Kirchen.­
vertical columns from left to right : the values of
vlter und Iltere Theologen (Bamberg, 1895 if.), 514 if. the Greek letters, these letters themselves, their

3 26 32 7
",
 names, the corresponding Greek numeral,. (Mia. he found them not to agree with the names of these runes
Dia, etc.), and their values expressed in Roman (helach, hu; that of u he could not separate from the name of
figures. At first the .. Numerus GRECUS" the t-rune) and therefore preferred not to write them. He
section had beell written across the whole width acted in the same way with the Hebrew alphabet, writing only
of the page below the alpbabets. It was then the values he felt sure were right.
erased and replaced by the higher Greek numerals Considering the late date of the manuscript, the runes are
and their Roman equivalents. A reference mark' remarkably well made. In the. rune-names there is of course
at the end of the last vertical column (+) is repeated an amount of misunderstanding (mo, kile, olkel, %in), but a form
at the head of this list of Greek numerals.' laga may point to an exemplar using open a. At any rate the
IOSra: a runic alphabet, with the inscription" Alfabetum evidence provided by the Bamberg manuscript is of greater value
chaldaicum .. : values, runes and names arranged than its date might imply.
as before. Finally, there is the name " Chaldaicum alphabetum .. under
. IOSrb: an "Alphabetum Ruthenicum ", i. e. a Cyrillic which the runes are found. This reminds us immediately of
alphabet giving also the Latin equivalents ~d the Munich MS. 14436 : there runes were found under the headings
names of the letters (1). , Syriac " 'Arabic' and • Alanic " and in their immediate
neighbourhood we also' met a • Chaldaean' alphabet. The
The runic alphabet has been corrected in two places : th~
Bamberg manuscript must then go back to a De inventione
name of the d-rune has been altered from thron to dhronj the
version to which a Chaldaean alphabet had been added.
first two letters of the name of the k-rune (kilc) are written over
an erasUre, but it is not clear what letters the scribe first wrote.
Perhaps he began to write the name of the next rune (laga), 14. Paris, Biblioth.eque Nationale, MS. lat. 5239 ,
.... '.
and after the first a he became aware of his mistake (2,). On the (saec. X med.)
other hand the values of the runes for u, x and y are uUssing,
and to the right of the u-rune the name also. The latter, 'fhe runes in this manuscript were discovered by Graff and
however, has been merged with the name of the t-rune, resulting communicated by him to W. Grimm., who first edited them (1).
in tachuT. The Bamberg alphabet must consequently have Since rio full description of this codex seems to have been printed
been copied (directly or indirectly) from an exemplar where I have included a number of details which may help to define
the material was arranged in horizontal lines instead of vertical its relationship with Strasbourg MS. 326 (2).
columns. The scribe obviously proceeded with great care.
(I) W. GRIMM. Zur Litterahn" tier RJuum, 22 ff. "" Kl. Scm/ten III. 108 f.
He must have realized the importance of the acrostic principle Cf. I. O. P1..AssMANN-W. KRAUSE, Die Hrabanische RUftenreiJre. esp. 173 f.
and therefore corrected thron to dltTon; similarly the name of and fig. 4; and also: .
the i-rune is is (in the other versions usually his). This may G. STRPHBNlI. M _ u I. I I I, no. 44.
T. VON GRIBNBBROBR, DUI angelsikluilchen nmenreihen, 24­
also explain why he gave no values to the runes for fl, x and y : (z) On this manuscript see into Dl. :
C. W. JON1lS, Bt4ot! Opera tk Temporihus. ISS, 166, 168. 171, 368.
(1) Ahnostidentica1 with that giVen by D. DIRINOBR, TM Alphabet, 476 f. C. W. JON1lS, Bedtu Pseudepigrapha,3I, 35. etc.; 1:&8.
1'be Bamberg alphabet seems not to have been noticed by Mias A. BJEKLUND, A. VAN DB VljVER, 1M _ _ mediUs d'Abbon tk Fleury, lSI and note 7.
Dtu StodIUJlmu Aiiecedarium (Spr8kvetenakapliga Sill.lakapets i Uppsala Id., Les plus andemru Traductions latinu medievales (X!-Xl~ siecle,) de.
ftlrhandlinga.r 1940-1942 =; Uppsala Universitets Araakrift 1942 : 9, IIS-I4B). Traith d'Astronomili et II'Artrologie. Osiris I, (1936), 672 ff.
(2) Or did his .exemplar have a name for k with a different initial? Cf. the H. P. LATI'lN, TM El.efJenth Century MS Munich I4436. Isis 38 (1947/48).
furms with initial g- in a number of manuscripts. 217, note 91.

328 32,9
".
 Modern red leather binding; on the back: DIONYSII / EXIGUI /CYCLUS 119V INCIPIT PROLOGUS PASCHALIS QUEM BEATUS FECIT
/ PASCHALIS / ETC., and LAT. SZ39. Old ownen' and press marks on HIERONI/MUS. 126f bla.nk.
fol. 2': Cod. Colb. IOao and Regiua 3823/4. Parchment of good quality, 126· Computistical tables; 129' Calendar; 135r Cornputistical tables.
almost white, well preserved except the first and the last folio, formerly pasted 138. DE NOMINIBUS VENTORUM; VERSUS DE VENTORUM
onto the binding. The codex consists of.239 folios and three Sy-Ieaves foIs. NOMINIBUS AC SPERIS SUBPOSITIS; rotae with winds,
*1 *3 (I paper, 2 parchment fly-leaves) bla.nk except *3" (in a modern hand) : lunar phases, tides; astronomical items; rotae with four elements,
.. Codex scriptus sleCUlo XI. exeunte aut ineunte XII ". The 32 quires are four humours, etc.; orologium viatorum.
numbered [I], II - XXXII «(I II, q III, (I UII, the rest in pIsin figures) on 145' DE TEMPORIBUS ISmORI (Etymologiae V, xxviii f.)
the last page, and by a later hand [a], b - z. - , et, ae, aj - gj on the fint lSI' ITEM ISYDORI. I. De astronomie nomine, etc. (Etym. III,
page of each quire. The composition is as follows : xxiv ff.)
I (+ I) {I-3} + 2 IV {4-19} + U {20-23} + 14 IV {24-I35} + IV 158f ITEM ISYDORI. I. De mundo, etc. (Et,.".. XIII, i fl.)
(+ I) {I36-I44} + 6 IV {I4S-I92} + III (+ I) {193-199} + IV {200­ 162v Dies aegyptiaci; instructions for letting blood (cf. fol. 2" 238v)
2O'J} + III (+ I) {208-2I4} + 2 IV {2I5-230} + II (+. I) {23I- Z3S} 163' Heredis nomen imposuit census heris (Isidore, Ee,.".. IX, v f.)
+ II {236-239}. 166v PETOSIRIS NECHEPSO. REGI SALUTEM.
FolS. If and 239v were formerly pasted onto the binding. 167' GRECAE LITTEru:; CUM NUMERIS SUlS; rota (I); 167v,
168 blank.
Format ca. 322 X 235 mm; written area ca. 226 X 163 mm; 30 11. to the
169' Bede, Chronicon nfJe tit sex h:u.im sacculi aetatibm (no title).
page. Written by sevemI bands, which change frequently and some of which
1:93v ADBREVIATIO CHRONICAE.
are decidedly more archaic (e. g. using open a; ra, rect, rt ligatures, etc.)
than othen, although the manuscript seems to have been compiled at once; 196< ARGUMENTUM AD ANNUM MUNDI INVENIENDUM;
occasional traces of insular influence. INCIPIT COMPUTACIO CIRCULI MAGNI PASCHE (532
yean' cycle).
l'he catalogue of 1744 dates this manuscript in the tenth century; analistic
notes (e. g. fol. 17·, ZIf) show that it must have been written shortly after 95 0 , 200' CURSUS LUNARIS DE DIEBUS SINGULIS PER XVIIII
and probably in the Abbey of Saint-Martial at Limoges (I) (d. Strasbourg ANNOS, with tables.
MS. 326). 204" CURSUS LUNAE UNICUlQUE PER / DECEM ET NOVEM
ANNOS,and other computistical and astronomical items and tables.
The list of the contents is more variegated -than the catalogue of 1744 sug­ 212r VERSUS DE SINGULIS MENSIBUS, VERSUS DE DUO-·
gests (2) : DECIM SIGNIS, etc. 214' bla.nk.
fol. I' Fragments on winds, climates, weather; Greek - Latin glosses; 215' The signs of the zodiac, with explanations.
dies aegyptiaci; instructions for bloodletting.
225 v Antus's prognostics, with comment.
3 f
(later hand) obits; gifts to Saint-Martial. :t.31v (!ster hands) Deeds to Saint-Martial; 232v, Z33' blank.
-4-' INCIPIT LIBER DIONYSII EXIGUI (De ratimul Jla$clwlis). 233' ARGUMENTUM QUOTA SIT LUNA PER DECEM ET
6< INCIPIT CYCLUS DECENNOVENALIS QUEM GRECI NOVEM ANNOS, table (= fol. 137", and other computistica1
ENNEA CAl DECA DERICA VOCANT. items.
20' Annals for 830-930, continued by later hands 1025-1520; 21' bla.nk. Z35' De inventione litterarnm (no title); 236v (later) [V}ETERIS
22' DE DIFFERENTIA COMETARUM; 22v, 23 bla.nk (partly QUESTIONIS NOVA SOLUTIO; 237' blank; 237v Deed (later).
lined for tables). 238v Dies aegyptiaci (= foIs.2v, I62V ) ; [H]eredis nomen, etc. 163'.
24' Bede, De natura rt:rUm.

The fact that some texts are repeated twice or three times
32' INCIPIT LIBER DE TEMPORIBUS (Bede); 39v bla.nk.

40' PRAEFATIO BEDAE PRESBITERI IN LIBRUM DE TEM­

shows that planning was not perfect in the scriptorium where
PORIBUS (Bede, De temp,,",", ratimul). this codex originated; so do the many blank pages (some of
96' RATIO CALCULANDI, PRIMO DE DmBUS MENSIUM, which were used at later dates).
and other computistica1 and astronomical items.
1I6v INCIPIT PRAEFATIO BEDAE PRESBlTERI (Bede, Eputola Fol. 23Sr if. show the following arrangement:
ad Wicthetla tit paschaIJ ulebratW1UI). 23S f , n. 1-9: the text on the Hebrew alphabet, followed by
that alphabet (10-12; letters, values and names); 13 blank;
(I) L'art roman d Samt-MartU:d tit Li1flogu. Us mamucrit, d pei.nturu, etc.
1 ...-19 the text on the Greek alphabet, followed by that
Catalogue de 1'!DtPosition 17 juin 17 septembre 1950. Limoges, 1950, 58.
(2) Catalogt4 codicum manwcriptorum Bibliotlwae RtgUB. Pan Tertia. (I) With Greek inscriptions YIIEPrEIA, ZUH MIKPA, ZUH MErAAH,
Tomus Quartua. Parisiis, MDCCXLIV. 59. etc. = Oxford MS. St. John's College 17, fol. 8'.
33 0 33 1
",

 13 IV [1-1041 + III (+ I) {IOS-UI} + 7 IV [uz- I67} + 3 {168- 170}

alphabet (20-26; letters, names; and numerical values in + 1[171, 17z} + IV [173-I80} + I {18I} + Z IV [I8z- 1971 + 4
Roman figures); 27-30 blank. {I98- zOI }.
The only old quire mark is ' XXII' on fol. 167, which is either a mistake for
23S", 11: I-S : the ten on the Latin alphabet, followed by that , XXI' or an indication .of a quire being lost; the text ends incomplete on
alphabet (6-9; letters and peculiar numerical values); fol. l67r;167" is blank. Either II quire is lost here too, or else thllt text was
10-14 the text on Aethicus Ister's alphabet, followed by never completed. Fol. 181 is a smaller and thinner leaf. There is a later
that alphabet (IS-20; letters and names); 2I blank~ (seventeenth century 1) quire numbering E I - E 34·
. Format: ca. 3z0 X zSS mm; written area ca. :uSfz3S X 170 nUi!; ruled
22-26 the text on the runes; 27-30 blank. for 3 1-33 linea till fol. 167; fols. 168-201 have 44 linea to the page. The
manuscript is written by two hands, the first covering fols. 1-167, the second,
1-3 : the runic alphabet (cf. infra); 4 ff. Roman figures.
236r, 11.
much smaller, fols. 168-z01. They may-be fairly contemporary. Somewhat
and Greek numerals (I Mia - MM Ecathostochyle). younger interlinear glosses. On the ground of the calendar on fol. 176',
M. Porcher, Keeper of Manuscripts in the Bibliothilque Nationale in Paris,
The treatise is carefully written in a clear though somewhat
bellevea the manuscript to have been \lIntten either in the diocese of Angouleme
unequal hand. It has close a but open g and long i; uncial d <lr in that of Limoges (I). This agrees very well with the close relationship
only in the letter-names delta, delfoi and in numerals. The between this codex and Paris MS. 5z39·
runi~ alphabet fills the first three lines offol. 236r in such a way, The manuscript is certainly made up of two originally independent manu­
1;Cripts. The text of the first ends on fol. 167' in the 'middle of Bede's Epis
that the runes for a - n are on 1. 2, and 0 - z on 1. 3; the names· tom ad Wictlwia; 167" is blank, and on 168' there begins a different text.
of the first thirteen runes are written above the corresponding Cf. also the differences in handwriting and lineation.
runes on the first line; those of the remaining runes had to be The contents of this manuscripts being analysed very fully in"the Catalogue
general, only a summary is given here, arranged so 88 to simplify comparison
inserted interlinearly as space permitted. For further details with the contents of the Paris manl,UlCriptj the corresponding folios of the latter,
see the discussion of the next manuscript. preceded by P, are indicated between brackets.
fo1. I r Bede, De natura ,..mDft (no title; P Z4')· ,;. ­
IS. Strasbourg, Bibliotheque Nationale et Uni'lJeTsitaire, MS. 326 10' INCIPIT LIBER DE TEMPORIBUS (Bede; P 32r).
16' PRAEFATIO BEDAE IN LIBRUM DE TEMPORIBUS (Bede,
(Latin 275) (saec. X). De tempo1Ilm ratione; P 40r).
83r Bede, Chnmicon rive de,e:!: huiru saeculi lUtatibru (no title; P 169').
The runes of this manuscript do not seem to have been edited I04V ADBREVIATIO CHRONICAE (p 193')·
I07T INCIPIT DE ORTU ET OBITU PATRUM (Isidore).
thus far. My attention was drawn to them by the description 100v De inventione litterarum (no title) (P 235')·
in the Catalogue geniTal (I). HOv Bede, De mithmeticis propositionibru liher.
It is not known how and when this manuscript came into. the lIZ Astronomical tables.
119v DE NOMINIBUS VENTORUM. with figures, and other astro­
possession of the Bibliotheque Nationale et Universitaire
nomical items (partly P 138", partly P 38 ').
(formerly: Universitaire et Regionale). A pencil mark at the 1Z7' DE PRAESAGIIS TEMPESTATUM. ­
end of the volume may indicate that it was acquired through 129' GENERA NUMERORUM IN SENSIBUS SECUNDUM
exchange (T 227; T = Tausch 1). AUGUSTINUM, etc.
130" De fusione metallorum.
Modem leather binding (nineteenth century). without inscription. Parch­ 131v AMBROSII MACROBII THEODOSII DE MENSURA ET
ment of good quality, which baa suffered somewhat from moisture at the MAGNITUDINE TERRAE, etc. (Macrobius, Commentariorum
beginning and the end of the volunie. The codex consists of ZOI folios, in Somnium Scipionis liber I, xx, 14 ff.).
arranged 88 follows:

(I) Ca~ gblbal ••• Vol. XLVII. Mamuaiu de la bilitio~ uni­

'f-,erritaire lit r~ de StTasbourg, 139-143. (1) M. J. ROTI', Librarian of the Bibliotheque Nationale et Universitaire,
C. W; JONilS, .8eJae Opera de Ttlmporibus, HZ, 161, 167. 171. kindly informed me of M. PORcmm's finding, and also provided me with a
C. W. JONJIS, Bedae Pseudepigrapha, 3i, SI, 93. 133. detailed description of the codex.

33 2 333
......
 133' FELiCIS CAPELLAE DE MENSURA LUNAE (De nuptits. but the handwriting is rather crowded and somewhat hesitant;
lib. VIII), etc.
I33 Y Astronomical and computistical items (partly P 96' ff.).
it lacks the firmness of the Paris hand. It has no long i; mostly
143" INCIPIT PROLOGUS PASCHALIS QUEM BEATUS FECIT close g; no uncial d except in the numerals. The runes are
HIERONIMUS (p 119"). distributed over two lines : a - sIt - .11'; their names are
144" Astronomical and computistical items (partly P lOll', 112·). written above them, their values to the right. The forms of
159' INCIPIT COMPUTATIO CIRCULI MAGNI (p 196').
164v ITEM DE DIVISIONIBUS TEMPORUM (related to Alcuin's the runes are practically identical with those in the Paris codex.
De tur.m et saltu lu.puu ac biuexto). The mistake erisenon for episerwn (Gk. ~1TlCT7Jp..ov) may point to
16,.· INCIPIT PRAEFATIO BEDAE PRESBITERI (Bede, Epistola an insular prototype, the p of which was read as r. In the list
ad Wictheda, P 116·), incomplete; .67 v blank.
168 Aratus, PrognostU:a, with commentary. (p 225·). of rune-names the form bira implies an exemplar with open a,
17." DE ANNIS DOMINI IN QUIBUS TANTA DUBITACIO so that the final ic of biric could be read as a.
EST INTER CALCULATORES. The puzzling numerical values assigned to the letters of the
173" Calendar, followed by astronomical tables, into al• • 82" CURSUS
LUNARIS DE DIEBUS SINGULIS PER XVIIII ANNOS
Latin alphabet are found only in the Paris and Strasbourg
(p 200"). manuscripts, not in the other De invenhone versions :
185' Poems on the. months, signs of the zodiac, etc. (p 212').
A D, B = CCC, C = C, D = DCC, E = CCL, F XL,
186' De differentiis dierum et noctium per circulum anni.

191' Isidore, Etymologiae V. :xxviii f. (no title; P 145').

G = CCCC, H = CC, I I, K = C, L quinquaginta,
194' ITEM !SYDOR!. I. De astronomie nomine (Etym. III, xxiv ff.;
M = II, N = CX,O =.LX, P = cccc, Q = CCCC,
P IsIr). R = LXXX, S LXX, T = CXL, V L, X = decem,
197" ITEM YSIDORI. I. De mundo (Etym. XIII. iff; P Is8r ).
199" Dies aegyptiaci (p 2v, 162", 238").
Y = CL, Z DC.
200' Heredis nomen imposuit census heris (Isidore, Etym. IX. v f.; The text explains this as an older, fuller use of the alphabet for
P 163', 238·). numerical purposes. Although there is some degree of cor- .;
The text of De inventions litterarum extends over two pages : ruption in the numerals (the value CCCC is given to three
fot. 109 v , 11. 1-9 : the text on the Hebrew alphabet, followed different letters), there can be no doubt that we have here a
by that alphabet (10-13; letters and names); 14-19 system of indications of distance derived from that of the ancient
the text on the Greek alphabet, followed by that agrimensores, as is found e. g. at the end of the Notae Papianae
alphabet (20-25; letters, names, and numerical et Einsidlenses (I). The numerals are those which apply to the
values in Roman figures); 26-30 the- text on the letters having a horizontal stroke over them : A = Passus D,
Latin alphabet, with that alphabet (letters and Passus CCC, etc.
peculjar numerical values) on I. 31 and in the lower
margin. 16. London, British Museum, Cotton MS. Titus D IS
fot. 1lOr, U. 1-5: the text on Aethicus Ister's alphabet, (saec. XV; saec. XIII).
followed by that alphabet (6-9; letters and names);
10-13 the text on the runes, followed by a runic This manuscript shows how the De inventions It.'tterarum
alphabet (14-17; runes, values and names); 18-31 could live on for centuries and provide the framework for further
Roman numbers and corresponding Greek numerals, alphabetical studies. Here it has grown into a regular liber alpha­
in five vertical columns; I Mia - MM Ecathos­ (I) T. MoMMSEN, Notarum LDtercu.li. (H. Keit, Grammatici Latini. IV.
tochyle (plate VII a). Lipsiae, MDCCCLXIV). 330. A similar list of numerical values, but going
only from I (- Z) to XXVI (~ B) is found in Heidelberg MS. Salem o. 39,
The alphabets show better planning than in the Paris version, fo1. 2', cf. p. 307.

33+ 335
--,

hetarius, in which material gathered from different sources has margins (esp •. the uppet) have been curtailed, and a strip of parcbment, approx­
imately S mm wide below and 12-17 at the top, has been pasted to the outer
been more or less skillfully fitted into the original frame (1). edge of each folio. Format of B : c:a. I'S3 -x ;;zo mm, written area ca. IIO X
This liber alphabetarius is bound up with a collection of Middle 90 mm. The number of lines varies between 21 and 31; two columna. The
English texts with which it had originally nothing to Qo. At quire maries show that the beginning of B is missing. A was written in the
fifteenth century, B- in the thirteenth (ca. 12:tOy.
one time it belonged to one AndreaS Davidsonus (fot. I v).
Wanley was the first to mention the runic alphabet (fol. 7r-v); Contents :
he informs us that the manuscript was formerly in the possession fol. *2' ElenchUs contentorum I in priore parte I -eodicis (later).
1'_ IZv Liber alphabetarius, etc. (details infra).
of Thomas Allen (z). 13 Index Contentorum (later).
Leather binding (nineteenth century), with gilt stamped Cotton anllll. 14' - lOS' A1fQ'en Rifde.
Consists of two manuscripts: A fols. Ion; B ;= Cols. 13-148, gathered as 105-' - lIZ-' SawIN Warde.
follows: ­ 112v - 127' Bali Meitlenhod.
VI (I-n,) + 1 {13} + V {14-Z3} + VI {24-3S} + V(+ 1){36-46} +VI 127' - 133' Wohtmge.
{47-S8} + V(+ l){s9-69} + V {70-79} + VI {8G-91} + VIlI{9N07} I33v - 147· Seint Kath#i7u.
+ III (+ I) {IoS-1l4} + IV {US-IU} + VI {IZ3-134} + VI (+1) The first part of the codex A) contains the following items
{135-147} + 1 {I1,8}.
(those taken from De imJentione litterarum are marked (x» :
In front there are 4 paper and Z parcbment fly-leaves (not counted). The
first ot the parchment fly-leaves was at one time pasted -onto a bindirig with fo1. I v Hebrew alphabet with introductory text (x).
two clasps. Fol. 13 was added when the two parts were brought together.
Fol. 148 is a larger single folio, loosened from a binding with two clasps and
fol. zr-v AlphabetumEbraicum hie incipit (square Hebrew).
folded in three. There are press-marks on fol. *Iv: Titus D 18; XII A, fo1. :z.V Notes on the Hebrew alphabet, from ]eroIil'e.
and below to the left: XXII.A in pencil; quires are marli:ed u follows: 14'
• C '; 24" D-', 36" E', 47" F', 59'" G', 70" H ',So" 1',92' f K ',-loST fo1. 3r- v Hec pagina instruit greco(rum al]phabe/tu~ grecum
• L', lIS' M', IZ3' • N ',' 13S' '0.'.
f cum numeris [ .. it]otis propriis. (Greek alphabet "
The present state of A shows that it must have been independent for some with the names and Latin equivalents of the letters,
length of time: the rather thin parchment is much worn, especially fols. I
and la. It has suffered some damage from moisture: the rubrics are often
their numerical values in ROniah figures', and the Greek
hardly legible. numerals).
Format of A: approx. ISO/ISS X 122 mm; the si2:e of the written area and
the number of lines wry, as the pages are not ruled; when fully used they
fol. 3 v Aliud alphabetum wecumsive aliam formacionem
contain 22-Z4 lines of writing, in one column except fol. 6' (table). - A must litterarum (Greek alphabet with equivalents).
originally have had a different size, higher and narrower; the upper and lower
fol. 4r Notes on diacritic signs: diagamon, digamma, figura
(I) T. SMITH, Catalogus, 129.
cesaris, et~. _(cf. Isidore, EtymologifU I, xxi).

W. GRIMM,
J. M. KBMBLB,
U.,.
Catalogue (1802), 565.

deut:sche Rtmen, 153 f.

On Anglo-Smcon Runes, 334, 338; PI. XV, I, Z.
Alia formacio litterarum grecarum (Greek letters in the

order ofth~ Latiii _alpha.bet).·

G. STEPHENS, Momnn.flllts I, 112, nos. 52-54. Greek-Latin glosses, with the transcription of the

T. VON GRIKNlIERGBR, Die ang4/sdchsisclum f'tlmRfI'eilum, 24. Greek words: ArY.QC, agyos, sanctus; .Q 8H.QC, 0

C. L. WRENN, Late Old English Rune-Namel, 30 f.

J. BLOMIlTI!lD, R_, 221. theos, deus... (4 v) ... MAKP.QC, macros, longus.

In KBMBLB's paper the codex is quoted as "Cotton MS. Tib. D.XVIII ", fol. 4 v'Sr Greek alphabet (values and names, and for a - K
which led STEPHENS to include it by the side of Cotton MS. Titus D. 18
(G. STBPHBNB, M _ t s I, 107, no. 24).
the nUtIlerical, villues in Arabic figures) with intro­
(z) H. WANLBY, Calalogus,247. In the upper margin of fol. Ix''' there is ductorytefCt. (x)' .
another owner's name (partly cut away) : Simond [...Jey. On Thomas Allen's
ownership ct. N. R. KElt, Thomas Allen', Mamucripts. The Bodleian Library
fo1. Sr Incipit NVMERVS grecorum ab vno vsque ad [m]ille­
Record II, no.
z7 (1948), n~. narium(Greek numerals: Mia, dia, etc. and Roman

33 6 26 337
......
 figures). Item aliud alphabetum grecum sive aliam anglica ' and' / est vocata & pomtur pro istis sex
formacionem litterarum (Greek minuscule alphabet, coniunctionibus : et, que, at, / atque, ac, ast. Hec
with the names of the letters). Iittera p anglica 'wen' est / vocata & ponitur pro
fol. 5 v Hec pagina docet latinos, chaldeos, sirOl!, gothos, / w. Hec littera 'fj anglica ' thet' / est nominata &
perFos eciam et scottos. Littere Latinorum secuntur ( ?) : pomtur pro quod (corr., first th). Iste tres P, B, a
Latin alphabet with introductory text. (x) / tittere ' thorn ' sunt nominate & ponuntur prq tho
Item aUud alphabetum latinum (minuscule alphabet). Below, notes in later hands.

Tabula alphabete quae / multum valet ad instruendos fo1. 9v

UERSVS SIBILLE DE ADVENTV DOMINI.
pueros (text in many places illegible, especially along fol. 9 v Hec de Ch:risti natiuitate, passione et resurrectione.
the left margin). Explanations of IHCDS. XPICTOC. 8ED. lOS.
foJ. 6 Syllabary, in two columns: ba be bi bo bu to /la /Ie /Ii
r CD8HP (transcribed IESUS CREISTOS TEU lOS
/10 tlU, and bla ble bli blo blu to rna sne mi mo mu. SOTEP).
Littere caldeorum et sirorum (alphabet of 25 letters, fol. lOr Poems on climate zones, etc.
no values nor names). Below: non parebant due fol. I I v Poems titled: DE ARTICO CIRCULO; DE SOL­
littere [ ...J. STICIALI CIRCVLOj DE EQVINOCCIALI
fo1. 6 v Littere Gothorum (alphabet of 20 letters); alphabetum CIRCVLO j DE HIEMALI CIRCULO; DE
persicum (21 letters). ANTARTICO CIRCULO; DE DVODECIM
RVNES : a Scandinavian runic alphabet, with at the SIGNIS &; RELIQVIS.
end the word .. amen" written in these runes. fo1. 12 YMNVS GRECORY ANTE CANONEM (+ II."
V

foJ. 7r Aethicus Ister's alphabet, with introductory text. (x) Greek written with Roman letters); INCIPIT'
De litteris Nortmannorum / quarum forme sequntuR. LETANIA GRECORVM (same writing, incomplete).
Runic alphabet (letters, values and names), with As the above description shows, the liber alphabetarius has more
introductory text (x), ending on than twice the volume of the De inventwne tract. Many of the
fol 7 v, and immediately followed by Item / aliter and Nem­ .additions, especially the Hebrew (one), Greek (four) and
mvus's alphabet (letters, values and names) j after /I : Latin (one) alphabets will be hard to trace (cf. p. 286).
Item de diptongis. The lower half of the page is blank. The Hebrew alphabet on fols. 2v-3 f shows a curious order,
fo1. 8f Incipit orano dominica grece et latine scripta (Greek probably reflecting the arrangement of the exemplar: hheth
text of the Lord's Prayer, with interlinear Latin gloss). :lain va'll he daleth gimel / beth (two forms) aleph nun (2) mem /
IIATEP HMON . 0 ENTIC . OPANIC ... : Pater mem lamech kuf(2) ioth theth / quo! tsade (2) phe (2) / ain tsamehh
noster qui es in celis ... . nun (3rd form); then various combinations: tha, tta, fa, etc.
fo1. v
8 Incipiunt grece prepositiones (6 11.). This jumble of letters may be a precious help in tracing the source
Alphabetum Anglicum: a minuscule alphabet (ex­ of this alphabet. But for the rest the additional Hebrew and
cept A, 1;) with the names of the letters: a be" ce de ". Greek material can only prove how the knowledge of these
x = ix, y wi, 7 = and, p = wen, P = thorn, languages had progressed .since De inventione litterarum was
B = thorn, a = "thorn and commentary: written.
"Iste Anglice littere & proprias figuras ac nomina Nor will the English alphabet on fo1. 8 v be of great help in
propria pos/sident vt supra notatur. Hec Iittera 7 tracing the sources of the additions. The knowledge of the

338 ,,..,,,
339
additional letters must have been widespread, even though
~ to the runic alphabet, it has a number of peculiarities not
becoming obsolete by the tim,e the manuscript was written. In
found in the other De inventione versions:
this connexion the absence of the letter yogh maybe noted (1).
The name thorn was one of the last to survive; wen for w may c:' .the name een corresponds exactly to the original OE. rune­
perhaps be interpreted as pointing ultimately to a Kentish : '. niirne; all other versions have a mutated initial (eh-).
origin (cf. p. 16). h ':' 'n~ other version has hegl, which is again a genuine OE.
The spurious alphabets-if we may call them so for the time form (as opposed to Continental Gmc. hagal elsewhere (1».
being-will. probably provide us with pre~ous indications as. q, x, )': the compiler seexns to have paid special attention
soon as they have been studied in detail. These alphabets are: to the acrostic principle; hence his names qhon, xelach, yri
(1) the Chaldaean and Syrian alphabet (fol. 6r) corresponding to chon, helach, hyri (and variants) in the
other manuscripts. Similarly he may have dropped
(2) the Gothic alphabet (fo1.. 6 v)
initial,h- in (h)is and (h)tI.T.
(3) the Persian alphabet (fol. 6 v)
s: the name sigil too may have been anglicized (the other
(4) Nemnivus's alphabet (fol. 7v). versions have tI., ai, etc. in the first syllable) (2).
The text at the top of fol. SV promises the following alphabets : In other respects we shall find the Cotton manuscript to provide
Latin, Chaldaeo-Syrian (i. e.Assyrian), Gothic, Persian and remarkably archaic material, but the features enumerated above
• Scottish '. This last name is not found elsewhere in the liber have no counterpart in any' of the De inventione manuscripts.
alphabetariui.· At first sight onemighi conclude that the al­ Even if we do not consider the cases where the compiler simply
phabet going under this name' was dr.opped or forgotten by the restored forxns jn agreement with the acrostic principle, een and
scribe, but there is another possibility: could not Nemnivus's hegl (and also runstafas in the text) seem hardly to fit into Conti- "
pseudo-Welsh alphabet be meant? It is found here without nental Gmc. surroundings as evidenced by ase, gibu., not, othil,
the introductory text (cf. p. IS7), a~d this may have led to the tac, ziti.. Therefore we must suppose that the compiler had an
confusion of Welsh aJld Scottish. Nemnivus~s alphabet can English runic alphabet (or a fuporc) at his disposal; it enabled
hardly have been added to the collection of alphabets except him to make a few corrections.. The presence of Nemnivus's
in England: it is quite unlikely that it ever reached the Con­ alphabet clearly indicates that he used old material available
tinent. A peculiarity of the alphabet as found: here is, that the in England only. There remains a difficulty: why should this
letters coming aJter !At form a group apart. with the heading Englishman have corrected hagal to hegl, and not the other un­
item de diptongis. The compiler who used De infJentione litte­ English rune-names? In a number of cases the acrostic prin­
TaTUm as a framework may have chosen to insert Nemniyus's ciple may have prevented him from doing so: he could not
alphabet after the runes in order to indicate that he was aware change ase to aJse (or esc, cf. hegl), dTom to thorn or porn, ot(h)il
of the real character of the Welshman's creation. The alphabet to epelor ethel, tac to daJg or deg, and ziti. to ti(r), without up­
in our manuscript is closely (elated to the oldest v.ersion in setting the whole system. We saw he demonstrated his respect
Oxford MS. Bodl. ·Auct. F. 4.32. Cf. p. IS7 f. for the acrostic principle by changing chon to qhon and helach
The other alphabets, except the Persian,' are also found else­ to xelach. He may, however, have used the same excuse for
where and have been briefly discussed in the Appendix to the
preceding chapter. (I) The fonn hegl Pllints to a prototype originating in Kent or Mercia. cf.
E;•.SIlM!lIS-K. BRUNNER, Altenglische Orammatill, § S:.I.
,(z) E. SllM!lIS-K. BRUNNER. Altmgbsche Orammatill. § 31 A. :.I: the Y Df
(1) Cf. A. PAUllS. The Name of the Letter J. on the ~rrence of this letter.
'sygil was unrounded to ,.
340 34 1
....<1'
introducing kegl: if he found in his exemplar the form agalc Welsh-English border and afterwards at Glastonbury (Oxford
(which occurs in the closely related Paris and Strasbourg manu­ MS. Bodl. Auct. F. 4.32), Thorney (Oxford MS. St. John's
scripts), he had to substitute for it a name with initial k. Thus College 17) and perhaps Canterbury (Cotton MS. Galba A 2).
he would be led quite naturally to replace it by his native kegl.
With een and sigil the case is not so clear; perhaps here too the
17. Florence, Biblioteca MediceaLaurenziana, MS. S. Mareo 604
prototype presented some monstrosity similar to aga/c. As to
(saec. XI in.).
runstafas, a form like runstabatk in the related versions would
of course have to be corrected. The form in the Cotton manu­
script may either be the regular English form (nom. acc. pI'. of Although this manuscript does not contain the runic alphabet,
runstref) or a Latin acc. pI. it may throw some light upon the history of the De infJentWne
A few words remain to be said about the composition of the text and consequently also upon the runes thexnselves.
liber alpkabetarius. It is made up of the following elements: Modem binding (seventeenth century n, wooden boards and leather back
with title and number stamped in gold. Parchment of good quality, well
DIL: A De infJentione litterarum text as known from French preserved; II7 folios = 1$ quires numbered in the lower right comer of the
manuscripts (here we have to anticipate the results first page of each quire (14th -- 15th century), several quires are missing;
of the discussion on p. 345 ff.). 1 = fols. 1-5 : ternio, the fourth folio of which is missing; II-XV =- fols.
6-II7: 14 quaternions, with lacunae after fols. 29 (U hic deest quinternio,
CA : A collection of alphabets such as we found circulating quo perierunt septem sequentium Pontificum vitae "),45 (" cetera desunt "). 53
in France (Chapter III, Appendix). C'mutts desunt "), 61 (U multa hic desunt") and 77.
N: Nemnivus's alphabet, known only from England. Format: 223 x 140 mm: ",{ritten area 170/185 X 100/105 mm; .a4-a6 lines
to the page.
AM: Additional Hebrew, Greek, Latin and philological
material. Contents (I) : ",
It is rather probable that DIL and CA had been brought fol. I" Epitome Joannis Diaconi de Episcopis Neapolitsnis.
6' PontificaIe Romanum falso adscriptum Hieronymo, a S. Petro,
together before the two were imported into England (cf. Mu­ usque ad Leonem II, cetera enim desunt. quemadmodum etiam
nich MS. 14436), although we cannot know how far they had abscissae aunt, nonnullae chartae unde desunt vitae VII. Ponti­
been integrated. N can only have been added in England. ficum, qui Hormisdam secuti aunt.
4:6' Catslogus virorum IUustrium ab Hieronymo editus, mutilus in
As to AM, that may have been inserted either in France or principio.
in England, and either at once or by ·steps. This may then 66v Epistola Quod vult Dei Diaconi ad Sanctum Augustinum; 68<
be graphically represented as follows: Responsio Augustini ad ipsum; fi9" Aliae duae Epistolae eorumdem.
70v I.J.ber S. Augustini de baeresibus.
DIL CA 90" Compendium IX priorum Capitum Cassiodori de Institutione
~. ,/~? Divinarum LitteI'llI'Wli.
/ /"
94' Concilium Romanum sub Damaso. Christi nomina.
. /"' ,./"" ....

x/ 94v Gellillii Decretum de recipiendis et ahiciendis libeis.

_____ N 98v Opusculum de vitii. et virtutibus absque nota Auctoris.
I
r:-------
Lib. Alph.
AM
?
lOS'
10&"
De inventione litterarum.
Alius Opusculum adscriptum Isidoro. quod probabilius est frag­
mentum alicuius libri Ethimologiarum.
Written by several Beneventan hands of the early eleventh
We have no indication as to the place where the liber alpkabetarius
was written. Even Nemnivus's alphabet is no clue. The other (I) As found in MS. S. Marco 945, pp. 5"53. of the same library, with

versions of' this pseudo-Welsh creation are localized on the additional notes kindly given by Dr. TlnulsA 1.001, Librarian.

342 343
".
 century, probably in Naples; the script betrays some insular
r

E -+I- literarum in gente nordomannorum fertur I primitus

affinities, e. g. -+I- enim(I).·· inuente quibus ob carminum eorum memoriam digeste sunt.
Fol. IOSr-v shows the following disposition of the text: . The initial E i!1 in red, as if it were the la.qt letter of
(I) In the upper half of 10Sf there is a figure, consisting of Aethicus's alphabet (the letters of which are rubrics, the names
concentric circles enclosing four rings with inscriptions. in black); the second H-like sign is the insular abbreviation
The central field bears the inscription: Spera pitagoris / phi­ for enim. We come close to the' version in the Cotton
losofi apulegio dedit. / de qua re consulere uolueris. / ut manuscript: Bee etenim, etc. So we have to interpret E either
puta de egris qua die ebdoma/dis incubuerit. et quota luna a!! & = et, or perhaps as E . Bae. The latter mistake would
eo die I fuerit. et nomen ~us per litteras infra scrip/tas not be impossible if the scribe (or the scribe of his exemplar)
singillatim numerando in unum col/lige. dein per XXX . wrote from dictation; but since the E is a rubric, probably
diuide et quotus numerus re/manserit in spera inspice in drawn after a given model, it will rather be a misreading for &.
ex~eri/ori uita. in interiori mentem i1'!/uenia (sic). ' After literarum some word like genus would be required by
(2) In what little space is left free to the left and the right of jertuT; but considering inuente, it is much more probable that
this figure, begins the text on the Hebrew alphabet, ending . a word figuTae, found in the related versions, was skipped, and
with the alphabet itself on II. 19-20 (letters, names and jertur mistaken for jerun.tuT. The form nordomannorum remains
values). The order of the letters has been disturbed: problematic.
from right to left first aleph - nun, then thau res and.
on the next line, sanech coph. THE TRADITION.

(3) Lt. 21-24, and fol. 10SV, I. J : the text on the Greek alphabet.
with the alphabet on 11. 2-3 (no names nor values, except The following sigle!! are used in the discussi on and in the o'
episimon, copi, ennacose ("900 '), enna (M= mille». .stemma on p. 349 :
(4) Ll. 4~8 : the text on the Latin alphabet; the alphabet itself a) hypothetical versions
is not given, although ,the next line is blank. A = the type of De invennone text circulating in Ger­
(S) Ll. 10- 1 9: Aethicus Ister's alphabet with introductory many.
text (letters and names,. no values). . AI, AlII, As = prototypes or groups of manuscripts derived
(6) Lt. 19-20 : a shortened text on the runic alphabet; the latter from A.
is omitted; 11. 21-24 are blank. B .'. the type of De inventione text known from France.'
BI the common ancestor of manuscripts P and S.
The text shows a number or scribal errors, some of which X ..:..:. . the' archetype.
have been corrected by a contemporary hand (Abranam to
. Abraham, sciptas to scriptas, but also diluuium to dilubium); a. b) manusc~pts (including one edition)

few remain, especially in the sphera (the text of which is entirely Ba 13. Bamberg MS. Msc. patr. 130/"2..

corrupt) and towards the end of the treatise (literaru[m] carac- ' F 17. Florence, MS. S. Marco 604.

teres, tradere curaui[t]; cf. the text on the runes). " G 4. M. Goldast's edition.

The passage on the runes offers some special difficulties : H = . 5. Heidelberg MS. Salem 9· 39.

N 6. Niimberg MS. 1966.

(1) E. A. Loaw. The Beneventan Script; A History 0/ the SCNtI!'Italian P. 14. Paris MS. 5239.

Minuscule. Oxford. i914. 73. 184. 194 r.. 266 note 6. 339. R = 9. Vatican MS. Regin. 294.

344
..... 345
 S 15· Strasbourg MS. 326.
even in the alphabets (e. g. Aethicus Ister's z has the name zeta
SG I. St. Gall MS. 876.
in A, but zothichin in B). .
T 16. Cotton MS. Titus D 18.
Within these two main groups we may again distinguish
U 10. Vatican MS. Urbin. 290.
several subtypes. Thus A consists of Al SG 1609 1761 G,
W 8. Munich MS. A. Weinmiiller.
and Ai! = N 1010 W, which differ on a number of points of
176 12. Karlsruhe MS. Aug. 176.
minor importance :
254 I I. Karlsruhe MS. Aug. 254.

1010 7. Vienna MS. 1010.

Al (SG) All (N)
1609 = 2. Vienna MS. 1609.

1761 = 3. Vienna MS. 1761.

Primo omnium litere hebraic" Primo omnium littere lingua­
lingu~ a Moyse inuente sunt rum a Moyse invente sunt et
Among these seventeen versions two main types may be distin­
et ab Esdra post illorum capti­ a~ Esdra post illorum capti­
guished:

uitatem et reuersionis eorum uitatem et reuersionem eorum

A = SG 1609 1761 G N 1010 W

renouate sunt, etc. renouate sunt, etc.

B = P S T F.

The difference between these two types appears e. g. from the In Al the two Vienna manuscripts are again very closely
text on the Greek alphabet (I) related (AlB)' In both the order of the first two paragraphs has
been upset, resulting in the following sequence : Greek alphabet
A (e. g. SG) B (e. g. P) -text on the Greek alphabet-text on the Hebrew alphabet­
Hebrew alphabet, etc.; both add omnia opera to the text on the
Litteras uero Grecas in primis Grecas litteras Cahtmus e Greek alphabet, and et ea cum summa cautela ego uobis ilecertaui
Cadmus Agenoris filius a Fe­ Foenic~ Agenoris filius primus
conscribere to that on the Latin alphabet. This last addition is ~.
nice ueniens in Greciam non­ inuenit decem et septem. His also found in Goldast's edition, but there the order of the para­
nullas attulit. Post quem ali­ PalaInides tres adiecit; post graphs is quite regular. This may of course be a normalization
.quantas alii adiecerunt qui que quem Symonides alias tres due to Goldast. The latter's editorial technique was rather
ad numeros faciendos habiles adnexuit, Pythagoras SaInius I, free, except as far as the alphabets were concerned (these seem
habentur; earumque literarum quae in summa XXIIII con­ to have been copied carefully); therefore it is difficult to decide
cum qui bus scribi potest sum­ sistunt; cum quibus uerba what his exemplar was like.
ma ad XXIIII peruenit. Cetere orationis componunt. Sunt­ The inclusion of H in group A may at first sight be questioned.
caracteres III adiuncte sunt ut que omnes ad numeros habiles Its paragraph on the Hebrew alphabet is a quite long text; it
ad Inillenarium numerum ve­ componendos cum adiectis offers some information also found in the manuscripts of group B.
nire possint. nonnullis caracteribus ut ad Similarly its text on the Greek alphabet states that Cadmus
Inillesimum numerum perue­ borrowed seventeen letters from the Phoenicians, a detail not
nire quaeant. given by the other versions of group A. But its first two para­
graphs show exactly the same disorder as 1609 and 1761, and
The two types differ in the text of all five paragraphs, and that can hardly be a coi~cidence. H also agrees with the
A group on a number of details which can only be explained
(J) These te:r.ts have not been emended; I hav~ only put in capitals and as due to close relationship. The passages in which H agrees
regularized the punctuation. with B do not necessarily point to a common heritage, but rather
346
347
""~
 to the use of the same or of similar sources (cf. the texts in the work, and therefore the stemma given here should ,only be
Appendix to this chapter). considered as a first rough outline (I). We may return to it
Five other verslons have also heen classified under A, but after having examined th!! runic material. .,
mostly on a narrower basis. R mentions the • renovation> of X
the Hebrew alphabet by: Ezra, a detail found in A but not in B;
similarly its paragraph on the runes, with the equation Marco­
manni = Northmanni and the reference to the origin of the B
Germanic peoples, connects it with A rather than with B. U has.
a completely different text.on the runes, but the mention of the
/\?
theotisca -Lingua again reminds of A. For the remaining three
manus~,rip~s we have to rely entirely on the runic material.
Here it is mainly the absence of some features typical of B
(e. g. the rune'-name ,hiraj the runic forms for a b f $ t) which
allow us to group them with A.
In group B the versions P and S are very close relatives, so
AAI
P S F T SC C 160~ 1761 H R 251 m; u
;f\

Sa: N 101Q W
much so that we should have to examine all texts they have in
THE TEXTS.
common (and not only the two De inventiotui versions) to ascertain
their degree of relationship. For the time being I have con­ In the Introduction I gave my reasons for not offering a
sidered them as being derived fro~ a C01ll.Illort anCestor,· R1 • critical edition of the whole treatise. The texts which follow
'Manuscripts F and Tagree on a number of points which ,illso here will mainly help to understand the position of the runes. ".
indicate a common origin B., but it is not possible to decide I have only included the most important variants, and felt free
whether Bl or B2 comes closer to B. As we saw already, T has to correct obviously corrupt readings without detailed justi­
been reVised by an Englishman (or at least with the help of fication. The alphabets are those of the A .group;
genuine English niaterial). F is a somewhat atrophied form, § I A: Primo omnium litterae Hebraicae linguae a MOYFe
from' which the runic alphabet has been omitted and the text inventae sunt et ab Ezra post illQrum captivitatem et
on the runes curtailed. Its main interest lies in its being written reversionem eorum renovatae sunt; quarum elementa
, in Naples (or at any rate in a region with Beneventan script) : litterarum subtus ut invenire quivimus adnotata hahe­
this is an indication of the widespread interest in this treatise. . mus, earumque summa XXII constat litterarum.
The approximate localization of A and B offers little diffi­
B: Primus omnium litteras ante diluvillm invenisse Enoch
culty : A was obviously circulating in Germany, especially in
dieitur, et post diluvium Cbamfilius Noe. -Deinde
the South-East, and B in France (mainly in the West ?). The
Abraham Syrorum et· Chaldaeorum literas invenit.
analysis :of the rune-names will show that the common archetype
Postea vero in monte Sinai Moyses ubi Lex I;Qnstituta
must also be assigned to Germany; this may bea factor in the
est invenit digito Domini scriptas. Litterae enim
discussion of the priority of A or B. In the following stemma,
Graecae et Latinae ab Hebraeis videntur exortae; apud
however, I have left that point undecided. I believe that a
illos erum prius dictum est .. aleph", deinde ex simili
'more detailed study of the whole treatise (and not just of the
'part concerned with runes) would be needed fully to justify a (I) Yet I am confident that it will be found right on most points'by future.
research. ­
stenuria.' Such a study hardly fits into the scope of the present
'348
349
...

 annuntiatione apud Graecos tractum est "alpha"; ad XXIIII pervenit; ceterae caracteres III adiunctae
indeque apud Latinos " a ". Translator enim de simili sunt ut ad millenarium numerum perveniri possit.
sono alterius linguae litteram condidit, ut nosse pos­ B: Graecas litteras Cadmus a Phoenice [veniens] Agenoris
simus linguam Hebraeam omnium linguarum et litte­ filius primus invenit :fCVII. His Palamedes III adiecit;
rarum esse matrem. Sed Hebraei viginti duo elementis postquem Simonides alias III annexuit, Pythagoras
Jitterarum secundum Veteris Testamenti libros utuntur. Samius unam; quae in summa XXIIII consistunt, cum
H: Ante diluvium primus litteras Enoch invenit, quibus quibus verba orationis componunt. Suntque omnes
textum suae prophetiae scripsit illis temporibus, testante ad numeros componendos habiles cum adiectis quibus­
Iuda apostolo in epistola sua: "Prophetavit autem de dam caracteribus ut ad millesimwn numerum pervenire
his septimus ab Adam Enoch dicens : Ecce Dominus, quaeant.
etc." Sed tamen illae litterae qua figura et quo numero
~Ifa. 61lh. gil.mma. delta. e brltvls czpisinon zeta. czta. theta
fuerint Scriptura minime memorat. Post diluvium

vero Cham filius Noe litteras invenisse fertur. Deinde

Abraham Syrorum et Chaldaeorum litteras invenit. aA bB 91 dfleE ) zZeHth-6­

I II III 1111 V VI VII VIII VIlli
Hebraeorum vero litteras digito Domini in monte Sinai

scriptas primus omnium Moyses invenit quibus Legem

populo conscriptam. ( ?). Sed Hebraei nunc illis non

iota kappalaub. moy noy xi 0 brevis pI coft ro
utuntur litteris. Esdras autem scriba legisque doctor

post captam Hierosolymam et restaurationem templi sub dx kK Il\mt\ nN X\ 00 pIT ~ rP

xx xxx XXXX l LX LXX LXXX LXXXX C. ",
Zorobabel alias renovavit et invenit litteras qui bus nunc '.
utuntur Hebraei; quarum elementa figurarum subtus
Simmil.. t~u oy phi chi psi 0 \onga ennacose chile.
ut invenire quivimus annotata habemus, earumque

summa XXII constat litterarum.

a.le:ph beth 9/mel d~leth he uau zaS 12th t~th ioth ca.p~
sC tT yY fct> c.hX pfoW t
ce ee.c ecc.e D De Dee DCCC DCCCC
crM

a.Z bJ 9) de; ;r u\A d<ef1 ~ ,I c.h4 FlO. 48

§ 3 A: Latinas namque litteras Carmentis nympha Evandri

lamech m4m nun samech .In 'phe sade cof res s;n . thau mater quae alio nomi?-e Nicostrata dicebatur invenit et
in Italiam eas prima attulit. Et postmodum Graecas
I L ll5n ~ s3 e~f~ raU c:P r\!\!(l)thT litteras III adiecerunt Latini propter necessitatem, quia
nonnuna verba necesse habuerunt sicut in Graecis
FlO. 47
habentur loquelis, ut est Christus (ips), ymnus ac zelus
§ 2 A: Litteras vero Graecas in primis Cadmus Agenoris filius et reliqua.
a Phoenice veniens in Graeciam nonnullas attulit [H: B: Latinas litteras Carmentis nympha Evandri mater quae
id est numero XVII]. Post quem aliquantas alii adie­ alio nomine Nicostrata dicebatur invenisse perhibetur
cerunt quaeque ad numeros faciendos habiles habentur. et in Italiam primum attulisse. Carmentis'autem dicta
Earumque litterarum cum quibus scribipotest summa quia carminibus futura canebat. Quaeque!ad numeros
35 0 35 1
''"'>
 apud antiquos .conficiendos habiles. aestimabantur; sed Ianuarias. MCS. Marcus Caesar. NC. Nero Caesar.
modernis temporibus paucieassumptis nostrum nume­ OOP. Oppidum. PRS. Praeses. (ST. Quintinus. R. Rex.
rum conficimus. SN. Senatus, TIB. Tiberius. VAL. Valerius. *
Xynodus. YMN. Ymnus. ZEN. Zenovius. LR. Lo­
ABCDEFGHIKLMNOPQRSTUX~Z cus religiosus. CM. Comes. .
Istiusmodi genus descriptionis notae Iulii Caesaris
. § 4: A: Litteras enim Aethici philosophi et cosmographi natione appeUatur, quod cum litteris quae antiquamanus
Scythica nobill prosapia invenimus quas yenerabilis appelatur perficitur; Cum quis Romanorum in aedi­
Hieronymus presbyter ad nos· usque cum suis dictis ficiis, parietibus, vel ill turribus aut in monumentis
explanando perduxit, quia magnifice illius scientiam saxeis ob memoriam sui suorumque aliquid litteris
atque industriam duxit; ideo et eius lltterail maluit commendare scalpendo curaverat, eas cum punctis et
promulgare. In istis adhuc litteris [sil fallimur et in titulis obligat, ne statim quis ignarus possit legere; ut
aliquibus· vitium agimus vos emendate. supra in paucis ostensimus.
B; Aethicus philosophus et cosmographus litterarum carac­ § 7 Ala+ G; traces in As_
teres quos adinvenit ita ut infra notatum est distinxit, A E IOU 'NC'P-T V:RS:.:S B::N'F:C"
quosque beatae memoriae Hieronymus presbyter cum :: :: :.: :RCH': p·SC::p· GL::R·::S·Q:·::
libra suo cosmographo Latinis tradere curavit; non llt M:RT'R'S
illos .imitemuiscripturis nostris, sed ut, sciremus Genus vero huius descriptionis tam quod supra cum
industriam indagationis· illius, qui illos adinvenit. punctis V et vocalibus, quam subtus cum aliis vocalibus,
a.!a.mon· boc.ah cathu delfoi eFfothu fomethu 9arfou hetmu iofithu quam solitum est informatum continetur, fertur quod'·
sanctus Bonifatius Archiepiscopus ac Martyr ab Angul­
a~ bO ~l iO. e-tt ~ f1.\ 9~ htd- I <I> saxis veniens hoc antecessoribus nostris demonstraret;
quod tamen non ab illo in primis coeptum est, sed ab
hJ~u ll/.\:hfu malat.hr nabalo.th ouchi; chorizech pithkin salathi antiquis istiusmodi usus crevisse comperimus.

pc€ 9~ s~
KBRXS. XPP. FPRTKS. TKRP. KNSTBR.
kY; 11\ m33 nA o{s)
A E IOU
B F K P X SBFFKRP. BRCHKTFNFNS. SCFPTRP
RFGNK. . XT. DFCXS. BXRK, etc.
intaleth theotimo$ uathot req/pror yrc.holm zothkhin/zeta

xX t·t· rt ~ y'V' z~
§ 8 G and All; monograms also in Hand 1761.
Litterae enim monogrammae scriptae nonnullis in locis
inveniuntur, ubi pictura cum museo (I) in pariete
FIG. 49
imaginis aut in velis, vel alicubialiter facta fuerit j ibi .
§ 5 Cf. infra. eorum nomina cum congerie litterarum unum carac­
terem pictores facere soliti sunt, quod monogramma
§ 6 Ala + G; partly in All' dicitur; quorum significatio subtus per pauca adnotata
AVC . Augustus. BM. bonae memoriae. DM. domus
monstratur.
mQrtui. EPM . epistolammisit. FIt. forum. GCS.
Gaius Caesar. IMP . Imperator. k. kalendas. IAN. (J) Cf. musivum in Du CANCE, G1oS$a:rium _dire et injimfB latimtatis.

35 2 27 353
,.#
 [Monograms for Dominus, Sancta, Mana, Sanctus. 3 et incantationum] om. F, which ends: digeste sunt;
loharmes, etc., ca. 18 in all]. "'" runsta~] -stahath P S, -swfss T.

* ** * **
THE TEXT ON THE RUNES. Research devoted to the texts and the runic alphabet has been

surveyed in the general introduction and at the beginning· of

This is the text on the runes in the two versions, with the this chapter. As far as the texts are concerned, we may accept

variants from each manuscript (I) : Baesecke's analysis (I) as a starting point. Version A tells us that

§ 5 A: I Litteras quippe quibus utuntur Marcomanni, quos. " The letters which the Marcomanni use--we call them

nos Nordmannos vocamus, infra scriptas habemus. Nordrnanni-we have written below; from them [1. e. the

(a quibus originem qui Theodiscam loquuntur lin­ Nordmanni] those people des.cend who speak Germanic;

guam trahunt); cum quibus carmina sua incanta­ with these [letters] they , signify , their songs, incantations

5 tionesque ac divinationes significare procurant, and divinations, [for] they are still given to pagan practices. "

qui adhuc pagano ritu involvuntur. Version B:

I Littems] Litteme H. [.]itteras 1010; quippe] om. N W 1010; "These forms of letters are said to have been invented

quibus] quas G 160«) 1761 ;-nos] om. N W 1010. among the people of the Northmanni; it is said that they

:a Nordmaonos] Nonnannos N W 1010; 8criptas] scrip~ H. scrip­ still use them to commit their songs and incantations to

tas N; habemus] habentur H 160«) 1761.

3 TheodiscanIJ T'h'eotiscam N. itheotiscanI W 1010; trahunt] memory. They gave the name TUnstabas to these letters,

"'" tradunt H W 1010 160«) 1761; incantationesque BC divinationes] I believe, because by writing them they used to bring to

-que ac divinationes om. W 1010. light secret things. " '.

s significare] om. N W 1010; In version A several layers can easily be distinguished. The

6 qui] quia 1761: pagano ritu] pagani ritus H 16(1) 1761. paga­
norum ritu N W 1010, pagams ritibus G;involvuntur]inuoUuun­ first probably contained only the definition of the runes as letters

tur 1010. used by the Marcomanni;~ the purposes for which this sort of

R; Litere quibus utuntur Marcomanni id est Northmanni II quibus

originem trahunt qui Theotiscam loquuntur. writing could be used were probably also indicated. The term

SG: Litteras quippe qUII$ utuntur [.•.] nos nordmann[.] Marcomanni was then defined as equivalent to Nordmanni j at

UOcanIUS infra [...]nesque etc. inuoluntur any rate, I believe this is the simplest way to explain how the

two names were brought together. Baesecke proposes to

§ 5 B: I Hae quoque literarum figurae in gente Northman­ consider the term Marcornanni as a later addition, but he is led

norum feruntur inventae; qui bus ob carminum to do so by his assumption that"B (which only knows the North­

eorum memoriam et incantationum uti adhuc dicun­ manni) is older than A. The Nordrnanni were then charac­

turj quibus et runstabas nomen imposuerunt, ob id, terized in turn as the people whose descendents speak the

S ut reor, quod his res absconditas vicissim scripti­ lingUl.l TheQdisca, i. e. Germanic. Lastly, the use of magic and

tando aperiebant. oracles with the Marcomanni was explained by their still living

I Hael Ha~ P S, Hee T. F(?); quoque] -II- (=enim) F, etenim T;, in paganism.

Nortbmannorum] Nort- T. Nordo- F.

As Baesecke rightly remarks, the Marcomanni of the De
:a fenmtur] fertur F; inventlle] primitus inventae T F.
im;enti~ne text can hardly have anything to do with the people
(I) Purely graphical variants (u;: Vi ae: 1B: , : 6'; qu: c; litera : littera) have
as B rule not been recorded. (1) G. BA:I!sBcK:R. Abecedarium Nordmmmicum.

354 355
-.
 ..

of the same name known from Roman history (saec. II-III). that the opposition of the -Saxons· to Charlemagne's attempts
Lazius connected them with Ditmarschen and Mercia; his to incorporate them into the Empire was connected with some
discussion of this relationship has now only historical interest. form or other of Danish support. As early as Willibrord's
Baesecke points out that there was a people called Marcomanni days missionaries had occasionally visited Denmark. Willibrord
in the twelfth century : with the Holzati and the Sturmarii it hiIflSelf "made a journey to Denmark, bringing back thirty
was subject to Count Adolf II of Nordalbingia. Baesecke Danish boys, whom he baptized, no doubt in the hope of estab­
supposes that from an original frontier guard (against the.Danes) lishing a kind of seminary for future work, thus anticipating
they had developed into a tribe, and compares them with the the English missionaries of the next centuries in Scandina­
Marcomanni of Antiquity and the ON. Markamenn. The via " (I). The late eighth century is exactly the period during
argument is in fact W. Grimm's (I). He argued that Scandi­ which runic writing is revived in Denmark (2); it may have
navians could not be meant here, because Hrabanus (i. e. the served as a means to "demonstrate Germanic character and
author of De-inventione litterarum) would not have equated their independence" (3) against ever stronger foreign influences.
language with the original Germanic language. But if the clause These new runes were not uriknown in Germany, to judge from
referring to the TModisca lingua is· to be connected with some the Abecedarium NOTdmannicum. We shall see that the runic
of the" numerous reports of historians, from the Goth Jor­ alphabet of the De £nflentione contains some ON. elements.
danes (550) to the Frank Frechulf (about 830), according to In the text significare is somewhat ambiguous: as far as
whom at least the Goths, or even all Germanic tribes, came carmina goes one might simply translate it by , to write',. but
from Scandinavia " (2), there is no reason for doubting that the in connexion with magic (£ncantationes. and especially Jiflina­
Marcomanni were Scandinavians. It will riot do to show that tiones) one would rather expect something like' to effect by
Nordmanrii or Nordliudi could also indicate the Nordalbingians, signs = runes'. In versionB, too, the reference to £ncantationes
and that these may have been pagans at the time when our is awkward; it looks indeed as if the magic use had been stressed.
treatise was written. I rather believe that the name Marco­ on an afterthought.
manni is connected with Denmark, or at least that the scholar The remark on the Theodisca lingua need not retain us very
who added the gloss quos nos Nordmantws vocamus believed in long. The text on the runes is quoted in a good many works
such a connexion. The runic alphabet which follows has discussing the word tModiscus (4). It is one of the instances
nothing to do with German runes : it is an artificial mixture
(x) W. LIMSON, England and the Continent, 64.
of OE. and ON. runes, whilst the rune-names have been more
(2) O. VON FruBs!!N, Runcmra, 1I4 if; 0. BBBNDUM NIBLSEN).
or less successfully forced into a German garb (cf. infra). (3) A. BJEKSTED, Mdlnmer, 328 (where this applies to the origin of the
Some information about Denmark must have reached Ger­ runes).
many by the end of the eighth century. It was certainly known (4) Among the numerous works diacussing the origin and early use of this
word, see:
W. KRoGMANN, Deutsch. Berlin, 1936.
(1) W. GRIMM, Ueber dBUtrc1re RIma, lSO·: "Skandinavier sind hier auf J; L. WBtsGliRBHR, Theudisk. ~ tleutsche Volkmame UM die we.illiche
keinen Fall gemeint; Hrabanus wiirde ihre Sprache nicht die wspriinglich 8pradigrt!Wle. (Marburger Universitiltsreden 5). MaIburg, 1940.
deutsche genannt haben. An die aus derfriiheten Periode bekannten Marko­ H. BRINlCMANN, Theodiscus. In: Alttleutschn Wort und Wortkunstwerk,
mannen, die erst im siidlichen Deutschland ihren Sitz hatten, dann oatlich (Fa. G. Baesecke), 20 if.
naeb Bohmen, MlIhren und Oestreich drangen, und endlich mit den Qu!\den T. FRINGS. Das Wort Deutsch. 114: Altdeutschu' Wort und Wartktmstwerk
dem Strom der Volkerwanderung naeb Gallien und Spanien folgten : an dieae (Fa. G. Baesecke), 46 tI.
lilngst verscbollene Markomannen ist nicht mehr zu denken. Dl'r Ausdruck E.: LERCH, ·Das Wort" Deutsch 't. 8ein Ursprung und stJine Geschichte bis
Nordmatmi deutet oifenbar auf em Volk im nordlichen DeutscWand ". (Juf Goetlu. (Daa Ahendland 7). Frankfurt, 1942.
(2) E. PROKOSCH, A Comparative Germmric Grammar, 26. Cf. infra. G. HimoUJ, Der Vo/.ksbegriJf in! 8prtuhschatz da Althochdeutschen UM

35 6 357
...
where there can be no doubt about the interpretation of this. the text I have corrected it to ~stabas, cf. OE. sta/as nom. acc.
term. As was indicated before, this remark may go back to pI. of stmf (I), but I admit that the distance between -s and -th
Jordanes's description of Scandinavia as an officina gentium, or is rather great. One could also try to connect -stabath with
to a related statement (I). ON. stafar, older stafaR, nom. pI. of stafr, but this hardly helps
Whether the reference to the divinationes need imply that the' to bridge the gap. The form runstafas in T is a poor clue, as
author knew Tacitus's Germania, I dare not affirm. Baesecke it is ambiguous: it may be due to the scholar who anglicized a
states that divinatory practices were unknown among the Danes number of rune-names (chen > een, hagal > hegl) in that
ca. A. D. 800. But we know hardly anything about carmina manuscript. Yet I believe that runstabas in B ultimately goes
having been written down with runes, and that Baesecke will back to information provided by an Englishman (of course not
not deny. If di!,;nationes was really borrowed from Tacitus, it the same as the one who adapted T). So Baesecke and Krause
may be an argument in favour of Fulda as the place where the' are right if they consider the text of B as more • English' than
treatise was compiled: Fulda is the only place where the that of A. This does not necessarily mean, however, that B is
Germania seems to have been known in the ninth century, closer to the archetype than A: its runes and rune-names
Note that version B (where the term divinationes does not occur) show certainly more traces of wear than those in A.
uses a circumlocution with approximately the same value: In the expression which describes the purpose of runic writing,
res abseonditas vicissim scriptitando aperi(re). viz. ob earminum eorum memoriam et incantationum, the reference
Version B is on the whole more straightforward, but it may to incantations looks like an addition. Runes may have been
also consist of two or even three layers. The clause explaining used to put poems on record; they may have been used in
the name *runstabas looks like a late addition. That explanation charms; but it is far less likely that they were used to record
can only be due to somebody who really understood the word : charms. ..Moreover the rest of the text does not state that ,.
he must have known that *rUn- meant ' secret, mystery' (res runic writing was a pagan custom at the time when this version
absconditas) and *stab- 'letter' ('lJicissim scriptitando). The was written, but it may have been so at an earlier period (uti
use of the compound *rUnstab- may perhaps indicate that his dicuntur: aperiebant). If runstabas goes back to a revision of
native language opposed that word to *bOkstab- I Roman letter, the text by an Englishman, it is not impossible that the same
bookhand (2) (although OE. boestmf is occasionally used for
J scholar also toned down the characterization of the runes as a
, rune '). The form -stabath in P and S is quite puzzling. In pagan form of writing: in his home country they had not fallen
under the anathema of the Church, as is shown by their use in
AltttUderdeutsclum (Junge Forscbung 8). Halle, 1941.
G. BAlISECKE, Dm NationaJbewuutsei'll der Dmtsclum tin Karolingerrei.t":Ms Christian monuments.
1It:lCh den lIIeitgeniJssUclum Benenmmgen iIrnr SpracM. In: Der VertrtllJ tlOIJ
V~dun 843. Leipzig, 1943. II6 if.
(1) " Ex hac igitur Scandza insula quasi officina gentium aut certe uelut
uagios nationum cum rege suo nomine Berig Gothi quondam memorantur THE RUNES AND THEIR NAMES
egressi " (cap. 4); d. also: .. Habet quoque is ipse immensus pelagus in parte
arctoll, id est septentrionali, amplam insulam nomine Scandzam, unde nobis The runes and the names found in fifteen versions (SG and

sermo, si dominus iuuauerit, est adsumendus, quia gens, euius originem F have only the text) are reproduced in the two tables following

fiagitas, ab huius insulae gremio, uelut examen apium erumpens, in terram

Europae aduenit" (cap. ·1) : IOlIDANIS de origine actibtuque Getarum ed. A. here:

HOLDER (Germaniscber Biicherscbatz S), Freiburg-Tiibingen. 1882, 6; 3·

(2) H. KUHN, Das Zeagnis der Sprache iiJnr Alter una Urspnm, der RUMII­ (1) Intervocalic -6- may either be an archaic DE. spelling, d. hlabard =

schrift. In: Beitrlge zur Runenkunde und nordischen Sprachwiasenschaft, hlaford, 'E. SllMIRS - K. BRUNNER, Allengtische GrtmmuJtik, § 191; or else it

54 fl. may be explained in the same way as that in gilm, d. irifra.

35 8
-. 359
 ~
.0­
0

I.
a
.::..
:t
('\

t
ZI ~
?
....0
-i -E 3 ::=' ~ oS' "fb ~ 1"­ 7'0
" ..•
~

""
'"..
~

-"
•
VI

.--4

p:;t)iJi­
, '4

1609 !!------­
I1761
_H_I~J~.i~.I_R_I~I~_!_~_L~ I-p_1~2-
G
asc asc ase ase lISe asc ase a.sc a.so a.sc , ase asc

asc ' asch asc

birith biriht birith birith birith
birith birith birith birie birie birieb berith bim I biN bim
eben eben eb.n eben eben eben ebenll ebon cheri chen ch! ch! cen
eben
thorn thorn thorn thron thron thron thron dhorn t'h'rom dhron dhron dhron dram
thorn
ech eho ech
fech
ech
fech
I ehe
fech
ebe
fehc
ebe
fehe
ech
fech
ech
fech
th
feeh fech
ech
fech
eth
fee
eth
fee
ech
fech
fech fehc
gibu gibv gibu gibu gibu
gibu gibu gibu gibu I gibul gibul
hagalc hagal hagal
gibul
hagal
gibu
hagalc
gibu
hagal
gibu
hagale hagate hagalc agate agate hegl
hagale hagale hagale
his his
gilch. gilch
I his
gile
his
gile
his
glie
his
glic
his
glie
his
kale
is
kol
his
gile
his
gileh
is
kile
his
kile
his
kile
is
kale
lagu lagu lagu lagu lagu lags !ago lagu !agu laga lac lac lagu
lagv ,iagv
man man man man man man man man .an man mam mam man
man man
not not not not not. noth nod not not not not not not
not not
othil othil othil othil othil othil otil othil othil othil othel otil oill otil
othil
pereb perc perc perch perch perch pert perc percJ perc pert perc perc perc
perc
chon chon qhon
chon chon chon chon I I
chon :chonl chan :chon chon chon chan
rehit rehit rehit rehit I
rehith rehith reiht reb ..
sugil
rehit
sugil
.•h.t rehit
suhll suhil
rehit
sugil
rehit
sugil
rehit
sigil
suhil sugil suhll siugil siugil siugil suhll
tae taeA tae tac tae­ tac tae tae
tac tae tae tac tact. tae.Jll
hur hur hur hur bur , bur hur hur hur -hur hur hur ur
hur
helae helae helach heluch he!aeh helach helach helae helae xelach
he1ahe halach :helaeb helae
huyri hurn hum hum huuri horsi huyry huiry hu hyri hyri yri
huyri. huyri
! ziu zia I zui
ziu zin ziu ziu ziu
i ziu ziu ' ziu i ziu ziu zui

w
I
I
Q'I
'"'
',.
We can now discuss the runic material in detail ( I ) in T (cf. infra). U stands apart in having the same name
a: There can be no doubt that tb,e alphabetizer chose C2 to for c and q: chon. In the first case it may be a mistake
fill this place. It is usually well preserved. In B, however, for chen. The sct'ibes of P and S seem not to have dared
it has been distorted beyond recognition, becoming prac­ expand the abbreviation in their exemplars (cf. d and m).
tically like a Roman P in versions P and S, and like R in T. d: Formally the rune is an English d. It is mostly not
The name mlc was adapted to German asc, Ioistaken for ase distinct in form from m (= m); in U, however, we find a
in 1010 W, and for aso in Ba; 1761 has a unique further very pure type of d, but that may have been borrowed
adaptation to asch, which agrees well with the late date from the fuporc in the same manuscript. In 176 there
of this manuscript (2). occurs a form which is either the right hand half of a d,
b: The rune itself is no problem; yet it appears with curious or a It turned to the left. The same form is found in the
distortions in group B. In T one may still easily recognize rune riddle which precedes the alphabet. The former
it, but in P and S it has become a Ioinuscule e. In N it explanation is probably right, as there are no traces in the
seems to be truncated at the top and below. The name other versions of a It having taken the place of d. The
must originally have been biric or birich. The form biric name is that of the English It : porn or thorn. The latter
is an archaic OE. form corresponding to beorc, bere, d. is the form we find in 1609 1761 G H; thron in .~ and R
biram Epinal gl. 792, biricean Bosworth-Toller, Suppl. 7 2, must be due to a misreading, which was then further
byric Napier Gil. 56, 364 f. This biric appears in U and distorted to become throm in 254. This form with final -m
254, but it is not possible to decide whether it was the form is probably due to the Ioisinterpretation of an abbreviated
of the archetype, or whether final h was dropped. The name of the exemplar (0 with a stroke above it); the same
former is rather probable, considering that it must also be explanation goes for drom in T. The forms with initial d ~.
at the origin of the name bira in group B (bira Ioisread may be explained in one of two ways : either thorn was
biric, perhaps in the neighbourhood of open a's). The made to conform to the acrostic principle; or else dh points
germanized form birich (176) was in tum Ioisread as birith to a region where the sound p became dh (at least in spell­
(1609 G H N 1010 W R) and this again as biriht (17 61 ) ing), i. e. some parts of the Franconian area (I). The
and berith (Ba). In OHG. the word for' birch' has a explanation need not be the same in all cases. Thus the
final a : pircha, pirieha, bircha. As a rule the adaptation acrostic principle is probably behind drom in T and dhron
seems to have been rather mechanical, so there is no need in Ba (two versions where it plays a part in other cases as
for assuming that~a final a was dropped, as von Grienberger well), but not necessarily behind dhorn in U, which comes
supposes. from a region where the spelling dh may have been used.
e; The rune is probably ON. k(on the name cf. infra q); e: The form of the e-rune is quite stable (its resemblance to
formally it offered little difficulty. Somewhat fanciful Roman M may have helped in preserving it); only Ba has
forms appear in 1761 R U 254. The name is obviously a decadent form. The name ech is a transparent adaptation
derived from OE. cen, the initial of which has become an of OE. eh, which may occasionally have been spelled so in
OHG. aspirate. The name may have been understood England _too (2). It hl}.s been misspelled in various ways :
as OHG. chen, ehien ' torch '. The OE. form was restored
(1) Cf. T. V9N GRl1!NBIIRGBR. Dk angehacluisclum I'UMnreihen, zs iI. {I) W. BRAUNE, Alt1wchdeutsche Gram71ltUik, § 167.
(z) W. BRAUNE, Alt1wchdeutsche Gram_tik, § 146. (z) E. SIIMIRS - K. BRUNNER, Altengli.sche Gram_tik, § U3 A. I.

362 363
-.
 eth in P and S (but not in T), ehe in N, and, deriving from -corresponding to OE. luegl, hegl; the latter form has been
the latter, ehe in 1010 W and eho in 1761 . restored in T. In most manuscripts the name hagal has
/ : Formally this rune too has been well preserved, except an additional letter, c or e. This anomaly is not found in
in P and S. In R it has received a spurious additional the versions which add a spurious I at the end of the name
stroke. The name must at first have been/ech, cf. OE./eh. of the preceding rune (gibul). If the names in the archetype
It became /ehe in three Versions (1761 1010 W) and lost were arranged in a vertical column, a mark after those for
its final h in two (~ S). g and h may have been interpreted as an additional letter
either of the name of g or of that of h. The form hagal
g: The rune which takes the place of g is enigmatic. It does in U does not necessarily contradict this hypothesis: it
not occur in any other runic alphabet and seems to be may be related to that in the fuporc which follows after
unparalleled in inscriptions, unless it. were a simplified. the alphabet in that manuscript. The initial h is missing
for.m of g. The problem is complicated by the fact that in P and S. There is another way of explaining the addition­
the next rune is formally identical with g, whilst that for k ·aI letter in most versions: we could perhaps start from
reminds one rather of j. We are led to conclude that the ON. hagall, the nameofh in ON. lists of rune-names (differ­
alphabetizer was not able to distinguish the various guttur­ ent from ON. hagl' hail " which is the equivalent of OE.
als which the OE. fuporc offered. If he really found a luegl, etc.). But this hagall is only found in late sources;
similar g in his material, it may either have been a g or in the earliest manuscript versions (Leyden MS. Voss. lat.
a Norse h. But the latter solution again complicates Q. 83, St. Gall MS. 878) the name is spelled with single I.
things, as a similar form (Norse h and English j are iden­
tical) takes the place of k. The former must consequently i: The rune itself caused no difficulty for obvious reasons.
be preferred, although I admit that it remains somewhat The name has a spurious initial h in most versions, but not .:­
doubtful (I). The name gibu is a transposition of OE. in Ba and T, which uphold the acrostic principle as far as
ge/u, geofu.. gieju into OHG. phonology (but· -b- was not possible; nor in U, where the example of the isruna fuporc
unknown in the oldest OE. texts). In three versions may have had some influence. This h will also be found
(N 1010 W) a spurious 1 has been appended to the name. ·in other rune-names originally beginning with a, vowel.
It may go back to a misunderstood detail in the archetype, Therefore one' can hardly explain this h as belonging to
cf. the next rune. the preceding rune, the value of which it was meant to
indicate; such an explanation could not apply to the other
h; As was mentioned. before, this rune is formally g. Unless. · cases where this h appears.
it. was accidentally mixed up with the preceding rune, we
must assume that the alphabetizer picked his runes rather k: This rune may be expl;Uned in several ways. It may be
haphazardly (2). In a number of versions the rune shows related to an English j or a Norse h (cf. especially the
a fanciful stylization, as if serifs had grown into short forms in 1010 and W), but also to x in St. Gall MS. 878.
vertical strokes (cf. k, x). ' The name is the OHG. form At first sight any of these may serve to explain our k, the
more so !'ince the name of the rune does not help to solve
(1) We might also try to connect the special g with the symbol for the prefix the problem. The fact thatea probably was the last rune of
ga-.k4- used "in a number of German manuscripts (d. p. 389); but then we
have again the difficulty that this symbol looks more like the k of our alphabet
the underlying fuporc (p.372) seems to rule out j; but there
than like g. . may well have been fuporcs where j took the place of j.
(:2) A poS8ible explanation for such mixing up is given on p. 377. If there it had the name ger or gear, this may have given
36 4 36 5
...
 rise to the value k in High German territory. For another of P and S had obviously little notion of the value of h.
possibility see p. 377. The simplest form of the rune m: The English m-rune has usually well retained its form.

occurs in 1010 and W; elsewhere we find more sophis­ In All it is followed by an obviously spurious vertical

ticated forms, mostly with the same sort of stylization as stroke, whilst in B (or at least in P and T) the right hand

in h, i. e. with additional vertical strokes. vertical shaft has been lengthened at the top (and in Ba

The name occurs nine times with initial g : three times downwards). The name should have caused little diffi­

each gileh and glic, twice gile and once gile. This can culty; yet P and S spell mam, which must be due to the

hardly be the original form: the acrostic principle requires misinterpretation of an abbreviated form mii.

an initial k (I). von Grienberger was probably right in n: In the cases where the form of the rune is clear, the oblique

connecting the name with MHG. kil(i)ch 'chalice '. It is stroke slants down to the left, i. e. a type of n known from

true that no OHG. forms are recorded with i in the first Old Norse inscriptions rather than from England. In a

syllable (z) (only e: kelih, etc.), but there seems to be no few cases (B; also 1609 1761) it has changed to an indifferent

other possibility. This kileh would then simply be a X, whilst elsewhere the oblique stroke has become more

translation of OE. cale 'chalice', i. e. the name of the orles scurved; the extreme forms evolved by this trend are

rune k. In two instances we find forms which seem to go . found in the P-like shapes of N W Ba.

back directly to the OE. model. In T kale is probably The name is the OHG. word not, the equivalent of OE.
a mistake for kale, which form may have been reintroduced nead, nied, ned and ON. naulr. The forms noth in Rand
by the scholar who also changed hagal to hegl. In R kale nod in U are dialectal variants (d. nod in the isruna fuporc
is not so easily accounted for, as there are no traces of old of U).
forms having been retained, nor of corrupt forms having
been corrected (cf. birith, thron, /aga, otil, reiht).
0: The rune is the English 0; it is well preserved in H R Ba. -.
Elsewhere the lower broken line has been more or less
I: As a rule the rune retains its original form, except in P fancifully elaborated. In U it has become much like a
and S, where it looks rather like a Greek A. This may be capital R, in the B group rather like an f-rune (in P and S
an accidental change, of which intermediate stages are it is practically identical with thef-rune of those versions).
found in Rand T. In 1010 and W it has received an addi­ The name is that of (E, in the German form othil (OHG.
tionalloop to the left of the shaft, an affectation also found odal, uodal is more common, but odhil is found in the
in the g's and h's of the same manuscripts. The name is OHG. Isidore). It has lost its h in R and in group B; in
OE. lagu, which appears as /ago in U (in connexion with Ba it appears as othel (cf. berith in the same manuscript).
/ago in the fuporc of that manuscript ?). R and Ba have
.p: Here it is again the English rune which gave both form

/aga, a mistake which may point to an exemplar with

and name. The form was obviously unfamiliar, as we see

open a. P and S have lac. I fail to see what von Grien­

from the malformations in All' Ba and group B. In All

berger means by calling this form' German '(the OHG.

the rune has received a spurious horizontal stroke, perhaps

word corresponding to OE. lagu is lahha): I would rather

originally meant to mark that the two broken side-strokes

connect it with fee in the same two manuscripts and explain

should not meet. For the name we have to start from OE.

it as a mistake. The scribe responsible for the exemplar

peorl, perl, perhaps written perth. The forms in our

(I) On the use of g for k in OHG. cr. W. BRAUM!, Altlwchdeutsclul Gram­

manuscripts go back to pert, cf. R Ba. Final t often be­

_tik, § 143 A, 4­ came e (1609 G H U Z54 PST), and this led to the extreme

(2) At least not in E. G. GRAFF's Spracluchau IV, 388.

366 367
...
'
 OHG. form perch (1761 N 1010 W). The latter could of
form remt occurs in eight versions: a mistake is found in
course descend directly from perth, but that is less probable.
R (remt); an additional h appears in 1010 and W (rehith).
q: The origin of this rune (if it is a rune) is obscure. It may In two manuscripts only fragments are legible (U 176).
be an ON. k turned to the left, or else a Greek K&1l'1I'1I.
s: The s-rune is the same in all versions, except in G : there
The latter is found in most Greek alphabets of the De
the upper vertical stroke is on the right hand instead of on
infJentUme tradition. with the name cophe, cole, cope, etc.
the left. This variant may be accidental and need not go
and the definition nota numeri. The other possibility,
back to Goldast's exempla:r. In a number of versions the
however, should not be excluded. The name chon may
rune has been interpreted in a fanciful way (1609 N, and
be connected with kaun, the ON. name of k, cf. coon in
especially group B). The name appears in four different
Munich MS. 1#36, con in Oxford MS.· St.John's Col­
but closely related forxns: ,ugil (G U 254 P S), dugil
lege I7 and Cotton MS. Galba A 2, and chon in Munich
(N 1010 W), suhil (1761 H R 176 Ba) and dgil (T). On
MS. 276. von Grienberger considers such a connexion
account of the unanimity of the other versions, this last .
impossible because there. are no other traces of ON. in­
form will again have to be considered as due to a later
fluence; but how are we to explain the c-rune and the name
corrector. Su.gil and IUigiI are probably attempts to, render
of r without reference to such an influence? In view of
OE. sygil, cf. the name of y. The form with intervocalic h
the intermediate forms of the name found in other manu­
may be connected with the gradual disappearance ofgin
scripts, no doubt is possible; so we may also consider the
that position (I).
Norse rune as the most likely model. The name appears
as chan in. R and Ba, and as qhon (i. e. with a correction t: The rune is a well preserved tin' 1761 H U 254 176 and T.
for the sake of the acrostic) in. T. In N 10IO W it seems to have been interpreted as a sort
T: The archetype'must have had a somewhat cursive type of r,
o{ A, whilst one can hardly reco~nize it in R P S. The
rather like those in 1761 G H 254 R Ba S. The stroke to name tac is the OHG. equivalent of the OE. rune-name
the lower right was then developed into a p.orizontal (or hg, deg. In Ba it has coalesced with that of the following
rune (tachur).
nearly horizontal) stroke: AD U P T. von Grienberger
identified the name with OHG. reita 'curru~, vereda', a u: We can easily recognize the form of the rune in most
translation of OE. rad. In the alphabet reita would have versions (though it may have been interpreted by the scribes
lost its suffix, and would have been spelled with the h as a sort of n). In 176 we find a shape rather like c (i. e.
which is sometimes inserted between t~o vowelS (I). The ON. k), ill P and S an utterly decadent form (identical
absence of final -a is not explained by comparing rehit with Hebrew mem in most De infJentione versions). The
with biric(h), as the latter is a direct adaptation of OE. name has a spurious initial h except in T, where it was
biric. Therefore I prefer to connect rehit withrei4r, the probably dropped on account of the acrostic principle. On
ON. name of r; cf. rechet in the Norse fupl!rk of Munich the, situation in Ba ,cf. the preceding rune.
MS. 276. von Grienberger;s explanation of the intervocalic :JC: In most alphabets the rune looks as if it were derived from
h may be accepted, but it is not necessary to refer rehit to an English :x by the addition of two short vertical strokes
reita, as some of the names had obviously no meaning ,for below the upper ends of the side-strokes. Such a modi­
the scribes who copied them (ech,/ech, his, pert, etc.). The fication may have been necessitated by the use of a form
(I) w. BRAUNS, Althodadeutsche Grammatik, § IS:I A. 3·
(I) W. BRAUNB, Althochdeutsche Grammatik, § 149 A. sa.
3 68
.....
:18 369
 which looked like a regular x for y, cf. infra. The new which we met before (isruna fuporc, Berne MS. 207). The
rune may occasionally have been interpreted as a capital M form of the rune m:ay be easily recognized in all versions.
with an additional vertical stroke in the middle, e. g. in T. von Grienberger argues that ea was chosen for z on account
G stands apart in having z of the manuscript versions for :Ie of the similarity of that rune with t, OE. t becoming z in
and vice versa. Since there is no manuscript evidence to the name (cf. infra). But in the isruna group and in Berne
support this arrangement, we must suppose it came about MS. 207, where the same rune is used for z, there is no
accidentally. The name is probably a translation of OE. trace of a shift t > z, so this can hardly have beerithe
eolk, elk, *eluh, into OHG. In the latter the word is usu­ reason for which ea was selected. Moreover von Grien..:
ally an n-stem : elaho, elko, helaho, ace, elahun, helahun (I). berger's explanation implies that the name of the rune t
The form in the De inventione alphabet rather belongs to had been attached to the rune ea before the Continental
the strong declension which appears in MHG. and perhaps alphabetizer set to work. The same confusion is found in
even in OHG. (2). The name of :Ie was probably at first Cotton MS. Domitian A 9, but there it is due to scribe B.
elach. It received an initial h as did other names beginning There is no proof that such a confusion was ever made
with a vowel (except in T, where the acrostic principle is on the Continent. I rather believe that the alphabetizer
upheld in xelach). A form without final h is fairly common chose ea for z because it was the last rune, and the OHG.
(A z P S). G has halach and U heluch; the latter may equivalent of OE. ti (also tiT un<:ier ON. influence; in the
perhaps point to an exemplar with open a. oldest glossaries Tiig , Mars, Martis ') or ON. Tyr for the
y: The rune is OE. x or ON. R (or y). The latter has usually name of z because it began with z. Therefore tts coinciding
side!'trokes which go downward; moreover it has the with the ea-rune is purely accidental. The name is in­
value y only at a late date (3). Therefore the other expla­ teresting because it is one of the few real translations (cf. me,
nation must be considered the more probable. The not). Although ziu is hardly known from other German
resemblance between x and Roman Y may well have played sources (1), the alphabetizer must have had some notion
a part in this choice. In the manuscripts the lateral strokes of its meaning. The mistakes found in four versions are
are usually curved; they have become a full circle in 1761. purely mechanical: zUi (R 254), zia (U), zin (Ba).
Only 254 and T have them straight. The name does CONCLUSION.
not help to solve the origin of the rune, as yr is the name
of ON. R as well as of OE. y. The vowel sound of the We can now make the balance of this investigation. The
name obviously caused some trouble to German scribes : bulk of the runic material is no doubt of English origin : the
we find huyri (1761 G H), huiry(176), huyry (254), hyri runes for a d e mop z and the names of abc d e f g kim 0 p
(P S); huuri in R is less appropriate, whilst hurn in All s t x are English or go back to English models. In a number
and horn in U are no doubt mistakes.. Ba has a fragmentary of other cases such an origin is at least possible : the runes for
Dame nu, and T again conforms to the acrostic principle bflrstuxy and thenamesofhinuyz. The Norse ingre­
with yri. dients are few, but their presence cannot be denied : the rune
z: For his last letter the alphabetizer chose ea, a solution for c and the names of q and r point to ON. models; the runes
for g i k n q and u and the. names of h i n u y z may have a
(x) E. G. GRAFF, Sprachsch4tz I, 235.
Northern origin. The proportion between. the OE. and the
cr. M. LExxR, MitteUwchdeutJchel
(2) E. G. GRAFF I. c. also quotes a form elah;
HandwiJrterbuth I, 538. • (x) E. G. GRAFF, Sprachsr.hntz V, 578, knows of no oth.er occurrence than
(J) L. JACOBSEN: - E. MOLTKB,. RurJeindskrifter (Text), 980. in Zieftac • Tuesday'.

370
-. 37 J
ON. material is not at all what we should expect after having precisely those used by the Nordmanni at that moment, but
read the introductory text: that text does not even mention the that was of minor importance: since all Germanic peoples were
existence of English runes. In fact the runic alphabet is a of Scandinavian origin, their native systems of writings might
highly artificial construction, which was certainly never more also be supposed so. We sho~d not lose sight of the relative
than an object of ~holarly discussion. Only about half of the importance of these runes: they Can hardly have had a greater
runes hav~ retained their original names. In four cases at least value than Aethicus Ister's fantastic letters. They could perhaps
OK runes have been combined with the names of other OK be used as an ornamental script; at the best they may have been
runes: d + )J, 0+ m, t + +
d(t), ea t(%). As far as the a help to missionaries going to work among the pagans of the
rune-names are to be trusted, we might conclude that the runic North. Nothirig in the ,treatise indicates that the author hoped
material came from the Anglian area (ech fech) and that it runic writing to be practised again in Germany. In teaching
belonged to the eighth «entury (sugil). But the adaptation to them he will rather have added a remark like that which the
OHG. phonology, which is far from uniform, makes it difficUlt manuscripts of group B apply to Aethicus's letters : .. ... non
to evaluate the evidence. The alphabetizer must have used a ut illis imitemur scripturis' nostris. sed ut sciremus industriam
fuporc with at least twenty-eight runes, since i2 and ea are indagationis illius, qui illos adinvenit".
found in his alphabet. We cannot decide for sure whether his It is also important to note that not one of the runic alphabets
fuporc extended any further: the answer to that question studied in Chapter III goes back to the same alphabetization,
depends on the interpretation of the symbols for g and k. Only except some of the material in Munich MS. 1#36. Therefore
if these were really to be identified with g and j, they would all reconstructions which draw arguments from those alphabets
prove that the fuporc had reached its full development; but the may be abandoned. The runic alphabet. of the De in'Oentrotli!
choice of ea for % rather indicates that ea was the last rune. The tradition may be 'more German' than any other (although >

case of g = g must remaiQ. doubtful, but in that of k there is this is perhaps rather a subjective impression); yet this. does not
a simple solution : we have only to assume that in the fuporc oblige us to consider ' less German' alphabets as representing
the place of j was occupied by j (cf. p. 365). Perhaps the form intermediate stages, nor even justify such a way of reasoning.
of g was invented ad hoc (for k see wso p. 377). By considering each alphabetization by itself, we obtain a more
We saw that the adaptation to OHG. phonology was performed varied picture, and one which is also closer to reality than the
somewhat haphazardly. One wonders what a German may reconstructions referred to before.
have understood by ech fech lagu pert chon remt sugil flir. There­ Before we come to the problems of authorship, place of origin,
fore I can hardly believe that this alphabet is the result of a etc., a few words must be said about the relationship between
long years' study and of several attempts to germanize the the different versions. Inthe discussion of the runes we found
English runes. It can hardly have been the cO.Illpiler's purpose no fact which requires the stemma on p. 349 to be changed.
to reintroduce into Germany a system of writing expressly A fuller examination of the treatise may perhaps lead to R and
stated to be connected with paganism. I would rather assume Ba being more closely related; these two versions may perhaps
that he had heard something about Norse (especially Danish) come closer to type B than the other manuscripts of group A.
runes and wanted to treat them briefly in his work. For com­ We met nothing which might support the priority of type B:
parative purposes he needed an alphabet of twenty-three letters, all points on which T prov~d superior are not in the much older
whilst the Norse fuJ;y.p-ks of his time had only sixteen runes. versions ofP and S. So it is far more likely that T was corrected
So he constructed an alphabet out of what material he had at with the help of English material, than that it should have re­
his disposal. He may have known that his runes were not tained those features from the original.
'j7Z 373
'.
 There can be no doubt that the treatise was first written in Alcwn's share is of course still less certain, and Bede may be
Germany: the forms of the rune-names are such as could only forgotten altogether.
be due to a German scholar. On the other hand the extensive There remains only the use of the Germania to be considered
use of English material points to a centre with Anglo-Saxon seriously (cf. p. 358). Germania X describes the practice to
influence; this is perhaps corroborated by the traces of insular which our texts would allude as follows :
.script found in some manuscripts. The notae sancti Bani/alii Auspicia sortesque ut qui maxime observant. Sortium

probably point in the same direction (I). Some information on consuetudo simplex. Virgam frugiferae arbori decisam in

Norse runes and on conditions in the North must also have surculos amputant eosque notis quibusdam discretos super

,been available. But at the same time it is clear that the author . candidam vestem temere ac fortuito spargunt. Mox, si

can have had only second-hand information on runic matters. . publice consultetur, sacerdos civitatis, sin privatim, ipse
Several arguments have been put forward in favour of Fulda pater faIniliae, precatus deos caelumque suspiciens ter
as the place of origin. Hrabanus Maurus, the supposed author singulos tuIlit, sublatos secundum impress am ante notam
of the treatise, was teaching there about the time when the interpretatur. Si prohibuerunt, nulla de eadem re in
treatise is believed to have been written; Fulda showed an in­ eundem diem consultatio; sin permissum, auspiciorum
terest in Germanic matters, several OGmc. works (int. al. the adhuc fides exigitur.
Heliant!) being credited to it; and, lastly, the paragraph on the I cannot find any close parallelism between this text and either
runes is believed to be partly based on a passage of Tacitus's version A or B. It is not even sure that the notae referred to
Germania, and that work was known in Fulda only. Impressive in the text were runes (I), but let us for a moment agree that
though these arguments may look, they do not stand a critical this is the right interpretation.' In the excerpt from Tacitus
exaInination. the words carmina, incantationes, divinationes do not occur.
First of all, Fulda's monopoly in Germanic studies has become I believe, however,' that there is another sort of texts from which "
questionable since Drogereit has stressed the importance of these terms were borrowed: penitentials, canonical works,
Werden (2). The rune-names cannot be considered conclusive: sermons, in a word, that immense literature which arose from
they point to High German territory, but the evidence they 'the Church's war against heathen practices. To find parallels
provide is too vague to help us choose a location. Especially we have only to tum to Boudriot's collection of excerpts relating
such artificial material as a set. of rune-names made ad hoc to Germanic paganism (2), e. g.
should be used with caution. As to Hrabanus Maurus's et si adhuc videtis aliquos ad fontes aut ad arbores vota
authorship, not one of the manuscripts we examined supports reddere, et sicut jam dictum est, sorli1egos etiam et divinos
this attribution. Goldast is our only authority, and we do not vel praecantatores inquirere, phylacteria etiam diabolica et
know what grounds he had for attributing the work to that characteres, aut herbas vel succos sibi aut suis appendere....
scholar.' Therefore that authorship is at the best a possibility; si. quis presbyter aut clericus auguria vel divinationes aut
it can hardly serve to reconstruct the history of the treatise. somnia sive sortes seu phylacteria, id est scripturas obser­
vaverint, scilicet se .canonum subjacere vindictis.
(I) W. LEvISON, England and the Continent, 290 fl.
(2) R. DROGERBIT, Saclull'lt. und Angelsaclull'lt.. Niederslicbsisches Jahrbuch
nam quicumque ad friguras non solum incantat, sed etiam
fUr. Landesgescbicbte 2I (1949), 1-62.

In. Werden und tier Heliand. Stud~.aur Kulturgeschichte der Abtei Werden
(r) A. BlEKllTED, Mtilnmer, so r.
.and :w.r Herkunft dill Heliand. Essen, 195 1.
(2) W. BOUDRIOT, Die altgemumische Religima in tier amtlichll'lt. kirchlichm

Cf. B. BISCHOFF'S review in A. f. d. A. 66 (1952), 1-12, and W. FOERSTE'8 Literatur des AbendlmuJes oom 5. his II. jahrhundert (Untersuchungen ZW'

in Niederdeutsches Jahrbuch 15 (1952), 142-147. allgemeinen Religionsgescbichte 2). Bonn, 1928, 62 fl. (esp. 62, 65, 66, 67).

374 375
-.

 scribit, qui angelorum vel salamonis aut caracteres suspendit k- and k-runeS of the De inventione runic alphabet. If the com­
aut lingua serpentis ad collum hominis suspendit; aut aliquid piler mistook the rune coming after the name chaon for the
parvum cum incantatione bibit, non christianus, sed paganus k-symbol, and that following hagal for the h-rune, he would
est. precisely have obtained the types we find in the alphabet:
quicumque super sanctum simbulum et orationem domi- . h = x, k = *' Yet I would nQt claim this collection of
nicam carmina aut incantationes paganorum dicit, in material as the source of our treatise. The strongest objection
animalibus mutis aut in hominibus incantat, et prodesse is, that the English rune-names are not found in the St. Gall
aliquid aut contra esse iudicat.... manuscript. Two sol~tions seem then to be possible:
Carmina vel incantationes, quas diximlis, haec sunt: ad (a) St. Gall MS. 878 is descended from the same source(s)
fascinum, ad spalmum, etc. as were used for the treatise on the alphabets; as far as the
To point out that no such practices are known from the Nord­ runes are concerned, it has preserved the material unadul­
manni mentioned in the text on the runes does not disprove my . terated.
explanation. The author of the treatise knew that those who
(b) The material provided by St. Gall MS. 878 was completed
used runes were still heathen (at least according to version A).
by a list of English rune-names (and also by extracts from
To describe their usage he simply relied on those pagan customs
Aethicus Ister's Cosmographia, and from canonical litera­
and beliefs which he knew from the official literature on the
ture, etc.).
subject. He probably had no first hand knowledge of the
Nordmanni and their religion, writing, etc.; otherwise he Which solution 'We have to adopt depends largely on the date
would certainly have been able to give us a "more Norse' of the treatise, or rather on the relation of the date of the runic
alphabet. There is consequently no reason for supposing that material in St. Gall MS. 878 to that of the earliest De infJentione ,
he relied on the Germania for· his information, and so this manuscript, St. Gall MS. 876. Strabo's death in 849 gives the "
argument in favour of Fulda mUstbe dropped (I). One might terminus ante quem for the former codex; he may have begun
still locate the treatise in Fulda: if it could be proved that this it ca. 820. The latter codex belonging to the beginning of the
abbey played a special part in missionary work among the same century, solution (b) is not altogether impossible. But
Norsemen; but it is not certain that such was its purpose (2). St. Gall MS. 876 contains a De inventione version which seeIns
Thus far the criticism of the views on the origin of the treatise to be removed several steps from the original. Therefore I feel
De invention" litterarum has been mainly negative. . There is, inclined to adopt the first solution, although I admit that there
however, a manuscript which might be considered as repre­ are no really conclusive arguments in favour of one or the other.
senting such a treatise in formation : St. Gall MS. 878. There The comparison of the tylro St. Gall manuscripts has its
we found precisely those extracts from Isidol'e which were incor­ importance for reconstructing the origin of the treatise, in that
porated into the text of the treatise; we also found.an English it makes a transmission Alcuin > Hrabanus > Strabo entirely
fuporc, and a Norse fUPllrk arranged into German doggerel. improbable: Strabo's runic material is mor~ genuine than that
By referring to this last item we might even explain the curious of Hrabanus (or that ascribed to the latter). If we had to base
our judgment on the two St. Gall manuscripts, we should con­
(r) The use of Aethicua Ister's Cosmograpliia offers no clue either, since clude that Hrabanus had his runes from Strabo. This shows
that work must have found its way into a good many libraries at aD. early date. again with how much caution such reconstructions should be
(a) E. E. STBNGIlL, Zur FriJhgeschichte d8r &ichsabtlli Fulda. Zflgleich iii"
Literaturbericht. Deutsches Archiv fUr Erforschung des Mittelaltera 9 (19sa),
attempted.
513-534 (esp. sao).' Since we no longer consider Hrabanus Maurus as the author,

3'76 377
",
the extreme dates proposed by Baesecke must be abandoned,
plum vitae humanae primus formavit; cuius virgula sub­
and the origin may be laid anywhere in the first third part, not
tedor pdmam aetatem significat, quippe et quae adhuc se
to say the first half of the ninth century. Perhaps an investi­
nec vitiis nec virtutibus dedit. Bivium autem, quod
gation of the whole treatise would afford a safer basis for
superest, ab adolescentia incipit : cuius dextra pars ardua
narrowing down this period; that I shall try in my critical
est, sed ad beatam vitam tendens : sinistra .facilior, sed ad
edition of the treatise.
labem interitumque deducens... Omnes autem litterae
apud Graecos et verba conponunt et numeros faciunt. N am
Alpha littera apud eos vocatur in numeris unum. Vbi
APPENDIX.
autem scribunt Beta, vocatur duo; ubi scribunt Gamma,
vocatur in numeris ipsorum tres; ubi scribunt Delta,
I. Isidore of Sevilla, Etymologiae : vocatur in numeds ipsorum quattuor; et sic omnes litterae
III. DE LITTERIS COMMVNIBVS. ... Litterae apud eos numeros habent. Latini autem numeros ad
Latinae et Graecae ab Hebraeis videntur exortae. Apud litteras non conputant, sed sola verba componunt, excepto
illos enim prius dictum est aleph, deinde ex simili enuntia­ I et X littera, quae et figura crucem significat et in numero
tione apud Graecos tractum est alpha, inde apud Latinos A. decem demonstrat.
Translator enim ex simili sono alterius linguae litteram IV. DE LITTERIS LATINIS. LatinaS litteras Carmentis
condidit, ut nosse possimus linguam Hebraicam omnium nympha prima Italis tradidit. Carmentis autem dicta, quia
carmi~bus futura canebat. Ceterum proprie vocata [est]
linguarum et litterarum esse matrem. Sed Hebraei viginti
. duo elementa litterarum secundum Veteris Testamenti Nicostrate ... X littera usque ad Augusti tempus non dum
libro8 utuntur; Graeci vero viginti quattuor. Latini enim apud Latinos erat, [et digne hoc tempore, quo Christi
inter utramque linguam progredientes viginti tria elementa nomen innotuit, quod per eam, quae crucis signum figurat,
habent. Hebraeorum litteras a Lege coepisse per Moysen : scriptitatur,] sed pro ea C et S scribebant, unde et duplex
Syrorum autem et Chaldaeorum per Abraham. Vnde et vocatur, quia pro C et S ponitur, unde et ex eisdem litteris
cum Hebraeis et numero et sono concordant, solis charac­ conpositum nomen habet. A Graecis [autem] .duas litteras
teribus discrepant. Aegyptiorum litteras Isis regina, Ina­ mutuavit Latinitas, Y et Z, propter nomina scilicet Graeca,
chis filia, de Graecia veniens in Aegyptum, repperit et et haec apud Romanos usque ad Augusti tempus non
Aegyptiis tradidit ... Graecarum litterarum usum primi scribebantur, sed pro Z duas S ponebant, ut' hilarissat ';
Phoenices invenerunt; unde et Lucanus (3,220) : pro .Y vero I scribebant. (Etymologiae I, iii, 4-7; 10-I1;
iv, I; 14-15).
Phoenices primi, famae si creditur, ausi
2. Some ,versions will be found to follow more closely the
mansuram rudibus vocem signare figuris.
Commentum Einsid/ense in Donati Artem maiorem :
Hinc est quod et Phoeniceo colore librorum capita scd­ Litterarum igitur diuersi repertores fuere. Primus namque
buntur, quia ab ipsis Iitterae initium habuerunt. Cadmus Enoch septimus ab Adam litteras repperisse dicitur, unde
Agenoris filius Graecas litteras a Phoenice in Graeciam in epistola Iudae legitur: sicut scri.psit septimus ab Adam.
decem et septem primus attulit; A. B. r . .d. E.Z. 1. K. A. Deinde Cham filius N~ praenoscens cataclysmum esse
M. N. O. n. P. C. T. q,. His Palamedes Troiano bello tres futurum fecit duas columnas, unam marmoream quae in
adiecit H. X. D. Post quem Simonides Melicus tres alias aqua seruaretur, et alteram latericiam, quae in igne duraret,
adiecit If'. B. 8. Y litteram Pythagoras Samius ad exem­ qui bus septem artes liberales tradidit ut post diluuium
378 379
-.

 stoliditas hominum pelleretur et acumen ingenii exerceretur. Cadmus (chatmus) Agenoris filius Gr(a)ecas litteras inuenit,
Postea Moyses alias repperit litteras in monte Sinai. digito non tamen orones, sieut nunc sunt XX et 1111, sed aliquae
dei seriptas, quae usque ad tempora Hesdrae durauerunt, postea additae sunt, de quibus dieendum est. Latinas
qui bus nune Samaritae utuntur. Deinde Hesdras inuenit litteras repperit Carmentis' (earmentes) nimpha Nicostrata,
alias leniores et faciliores ad scribendum, quibus Iudaei mater Euandri ... (I).
utuntur. Graecorum uero litteras Phoenices reppererunt, 4. The Ars grammatica of the Irishman Clemens also offers
unde in initiis librorum Phoenieeo colore idest minio close parallels :
scribuntur litterae. Latinorum quoque litteras Carmentis XXX. A Quis primus litteras ~te diluvium invenit?
nympha Nicostrata mater Euandri inuenit. Carmentis M Enoch videlicet. hie enim ante diluvium nonnulla scrip­
autem dicta, eo quod futura ear minibus canebat; nympha sisse fertur per easdem litteras, quas ipse invenit et quibus
dicitur quasi lympha, idest aqua, quia sicut aqua sapientia teXtum prophetiae suae illis temporibus ostendit testante'
difHuebat. Nicostrata uero idest uictoriosa. NlK"II enim Iuda apostolo in epistula sua: 'prophetavit autem', ait,
Graeee uictoria Latine, uel gladiata, eo quod ingenii , septimus ab Adam Enoch dicens: ' Ecce dominus veniet '
acumine uigebat. ... Sieut media est Latina lingua inter et cetera. A· Quid custodivit eal!' litteras in diluvio ne
Graecam et Hebraeam, sie medium tenet numerum in perirent ? M In duabus etiam columnis erant scriptae.
litteris. Graeci uero XXIIII, Latini XXIII, Hebraei Iubal enim filius Lamech et soeii eius, ne dilaberentur ab
XXII habent litteras. Inde XXII libros habent ex uetere hominibus ea, quae ab eo inventa videbantur, aut ne depe­
testamento (I). ment, antequam venirent ad notitiam, duas columnas
3· . Another strikingly resembling version is found in Berne testante Iosepho fecerunt et in ipsis ea quae invenerant
MS. 207 (the runes of which were discussed on p. 178 ff.), conscripserunt, sed tamen ilIae litterae quo numero vel qua
fo1. IUr , INCIPIT DE LITTERA : figura fuerunt scriptae scriptura minime memorat. post
Genera litterarum diuersa sunt, quia (qui a) diuersis locis diluvium vero Cham filius Noe litteras inveniss.e fertur.
et auetoribus sunt inuentae. Nam ante diluuium Enoch deinde Abraham Syrorum et Chaldaeorum litteras invenit.
inuenisse litteras credimus, quibus textum prophetiae suae Hebraeorum vero litteras digito domini in monte Sinai
illis temporibus ostendisse non du~ium est testante epistola, scriptas Moyses invenit quibus legem populo eonscripsit.
quae catholica nominatur. Prophetauit autem: Ait sep­ sed non iUis nune Hebraei Utuntur litteris. Hesdras enim
timus ab Adam Enoch dieens: .. Ecce dominus ueniet " scriba legisque doctor post captam Hierusalem et restaura­
et reliqua, quae ibi seribuntur; per quas etiam studia sua tionem templi sub Zorobabel alias invenit litteras, quibus
in duabus eolumnis scribere post diluuium homines nunc utuntur Hebraei. Graecarum vero litterarum usum
curauerunt; sed postea repperit litteras Cham filius N oe ; primi Phoenices invenerunt. deinde Cadmus Agenoris filius
deinde Abraham t.Syrorum et Chaldeorum litteras inuenit. Graecas litteras a Phoenicia in Graeciam XVII primus
Hebreorum litteras- Moyses inuenit in monte Sinai, quibus attulit. sed ab allis auctoribus aliae postea sunt additae,
et legem seribtam digito dei suseepit, sed non.illis litteds ut sint XX et quattuor. Hebraei enim XX duobus ele­
nune utuntur Hebrei, utuntur autem eis litteris, quas mentis litterarum secundum Veteris Testamenti libros
inuenit Esdras propheta sub Zorobabel, filio Salathihelis. utuntur, Graeci vero XXIIII. Latini autem inter utramque
Gr(a)eearum litterarum usum primi F(o)enices inuenerunt : linguam progredientes. XXIII elementa habent. Latinas

(I) H. HAGEN, Anecdota Helvetica, a:n, a25. (I) H. HAGEN, Allecdota Helvetica, XXIV.

380 381
~.
 autem litteras Carmentis nympha prima Italis tradidit.· nouit. Abraham patriarcha primus ~ptiis qu~ sunt de
Carmentis autem dicta, quod carminibus futura canebat. astrologia tradidit. Ninus rex Niniuitarum filius Beli
Cetelilm proprie vocata Nicostrata. sed intuendum est, primum idolum (with Bd written above it) in Babylonia
quod Pompeio teste undecim tantum modo litterae inventae fundauit. Ysis regina Inachi filia ~ptiorum litteras
primitus apud Romanos fuerunt; postea tamen crevisse repperit; h~ sunt litterft : ... ; & Mercurius lege!' eis tradidit.
usque ad decem et septem, deinde usque ad viginti et tres Foronftus rex Grecis leges & iudicia primus omnium regum
idem prompte refert. sed ex his XXIII, ut praediximus, . instituit. Cicrops Atteniensibus primus leges dedit & postea
una aspirationis nota est h, una duplex x, duae supervacuae Solen iudicia tradidit. Cathmus Agenoris filius Greds
k et q duae Graecae y et z (I). litteras a Phenic~ primus omnium attulit: ." Postea
Palamides de Troiano bello reuersus Greds tres litteras
5. The paragraph devoted to Aethicus Ister's alphabet is
drawn from the first and last lines of the Cosmographia : addidit: ... Moyses legis lator dux Hebreorum in regione
INCIPIT LIBER ETIDCI, TRANSLATIO E PHILO­ Pharairoth eis leges tradidit : ... Hebreorum littere. Apollo
& Scolapius & Ypocrates medicin~ artis auctores extiterunt.
SOPHICO EDITO[S] ORACVLO A IDERONIMO
PRESBYTERO DILATVS EX COSMOCRAFIA ID Salomon rex primus templum Domino in Iherusalem
EST MVNDI SCRIPTVRA. EDICTA AETHICI ftdificauit orbis anno IlIIcLXX. Ligurcius Lac~demonis
primus omnium iura studios~ confinxit.. Romulus & Numa
PIDLOSOPHI COSMOGRAFI. cap. 1. Philosophorum
scedulas sagaci indagatione investigans mihi laborem tan­ Pompilius atque Postumius Romanis leges diuersis tempo­
tundem obposui ac hic de iniciis tanto studio indagare et ribus tradiderunt. Appius Claudius Genucius Veturius
altiora magnatimque [h]ac cursim tam astrologiam fasti­ Julius ManiliusSulpicius Sextus atque Curatius leges &
giaque excellentia, quae nec·dum cerni quis non possit ... iuditia Romanorum proferendo iudices extiterunt.. Car­
mentis nympha Italis litteras Latinas prima tradidit, qu~
cap. Il3. SUOS CARACTERES LITTERARUM
QUOS ADINVENIT, ITA DISTINXIT (then follows sunt, etc.
the alphabet). EXPLICIT LIBER AETHici philosophi
CHOSMOGRAFI NATIONE SCHITICA NOBILE
PROSAPIA PARENTUM. AB EO ENIM ETIDCA
PHILOSOPHIA A RELIQUIS SAPIENTIBUS ORI­
GINEM TRAXIT (2).
6. This is part of a list of ' inventors' from Ghent MS. 92,
fol. 1 v:
Caim filius Adam primus ciuitatem primam quam Effrem
uocauit condidit.Tubal filius Caim musicam artem primus
inuenit & omnia eius instrumenta. Obal filius Lamech
opera ferri ~ris auri & argenti primus inuenit. J onitus
filius No~ primus astronomiam inuenit & sydera c~li cog­

(I) J. TOLlUEHN, Clementis ars grammatica. Philologus, Suppl. .vol. XX,

fllSC. III. Lipsiae, [928, 18 f.
(a) H. WUTrKE, IJie Kosmographie US Istrier Aithihos, I, 84.

382 38 3
~.
 CHAPTER V

NON-ALPHABETIC RUNICA MANUSCRIPTA

in the Introduction to the present study I distinguished five

types of manuscript runes. All the material examined in the
firSt. four. chapters belonged to the fifth category: fuporcs and
runic alphabets. This last chapter is devoted to the other four
varieties, viz.
(a) runes used as additional letters;
(b) runes substituted for their ~ames;
(c) runes used as reference marks, quire marks, etc.;
(d) ,scribal signatures and notes, titleti, short scribbles, etc.
in runes.
At first sight the reader will be· surprised to find such different
types of runic usage discussed in one chapfer, whereas four· were
needed to cover the one remaining type. There are, however;
various reasons for which it is impossible to treat these four
categories as exhaustively -as the fuporcs and the alphabetS.
Type (a), to begin with,- is found in most OE. manuscriptS
(and manuscripts containing OE. material, e. g. glosses). To
obtain a re~ult of any value, one woUld have to excerpt dozens,
not to say hundreds of manuscripts, and present the data thus
obtained in the form of detailed -statistics. From this enormous
labour. we could only expect to learn something on the use of
two runes, wand.,. Therefore this undertaking, valuable
though its results might be, could only be justified as part of
some vaster project including the examination of 80 many
manuscripts, -e. g. a catalogue of all codice~ containing OE.
material; or a detailed study of OE. spelling.
From what evidence is available, type (b) must be extremely:
rare; only two authors seem ever to have used it systematically,:
I1nd even then its scope is restricted to a few runes~ Wi~ (c),:
,~ ~
the main difficulty will again be to collect the material. A In the Introduction it was said that no OE. manuscripts
systematic search with so limited an object will hardly ever be written with runep have come down to us, nor do they seem
undertaken. The same applies to (d). Such usage appears ever to have existed. Wanley was aiready aware of this fact (I).
only on close examination, and even if discovered it will often Occasionally we find whole words or parts of words spelled
not be found mentioned in manuscript catalogues. To this with runes; as a rule this usage falls under the heading , cryptic
last category belongs a wide range of runic writing, going from writing', somewhat similar e. g. to the notae sancti Bonifatii
ornamental pages with carefully drawn symbols "to hurried, (cf. the Riddles of the Exeter Book, p. 417 f f . ) . "
hardly legible scribbles and meaningless ppeudo-runes. Here The typological distinctions set up in the Introduction and
too the occurrences "are so few that it is rarely" possible to repeated at the beginning of this chapter were intended in the
reconstruct the background of such runic writing. first place to circumscribe the notion TUnica manuscripta. They
Consequently this chapter should not be considered as a full may also serve as a framework for this chapter, although we
c.ollection of non-alphabetic runic material but only as a first must keep in mind that sharp limits cannot always be drawn.
rough map of this territory, to be completed as more information Nobody will, I think, doubt that the use of w and II in Anglo­
becomes available. The reader will" see that this map does not Saxon script is a phenomenon quite distinct from e. g. the use
present the same density of detail for all areas. The reason for of the rune w where we shoUld expect. its name, rrynn, written
this situation is obvioulI: I had to depend entirely on what in full. But when the m-rune is used systematically instead of
material had been recorded. The discovery of such items the word mann, and equivalent to. the abbreviation iii found"
depends largely on a very close scrutiny of all manuscripts, a elsewhere, we are coming closer to the ' additional letter ' class.
scrutiny which omits no details as trifling or irrdevant. Studies Similarly there are cases where one may hesitate between (a)
of this sort are rare. As long as manuscripts are studied mainly and (d), perhaps also between (c) and (d). Several types may
for the texts they contain, or for their artistic value, the material !>e found in one and the same manuscript. Thus the Exeter
which interests us here will be in danger of remaining Book has
unrecorded. Therefore I expect that new items will become (a) w and ):J as extra letters throughout;
known as the continuation of Lowe's Codices Latini Antiquiores (b) runes written for their names in the poems Christ,
is published (in the period covered by that first series, TUnica Juliana, Ruin and Riddle 91; :
manuscripta must have been very rare). Works like B. Bischoff's (d) some words written in runes, e. g. in Riddle 75;
study on the scriptoria of SE Germany are still more useful,
In one case it· is impossible to decide for sure which type is
because .there the material is examined in its geographical and
meant, namely in The Husband's Message «b) or (d»?
historical context. How much the present survey benefited
from Bischoff's work will become evident in section (d); that
(a) RUNES AS AD"DITIONAL LETTERS
section would no doubt have become much longer if a larger
area had been studied equally exhaustively.
In England the use'of the runes w and ):J in forms more or
Since the material offered here is only fragmentary-for (a)
less adapted· to .cursive writing is widespread. From most
I could give no more than a general appreciation-I have not
grammars one would get the impression that this usage is almost
tried to give full descriptions of the manuscripts. Such a re­
striction was unavoidable if a minimum of uniformity was to be (I) H. WANLEY, Catalogus, Praefatio, (II) : .. At Librorum ChartBnimve a
recepto Christianisrno per Anglo-Saz01U!s nunquam quicquam esse lWnice
~aintained. Tenor tweJlty years from now it may become scriptum, saltern jamjam scriptum superesse, tam certum est, ·quam quod
possible to give some of these matters a more adequate treatment. certissimwn ".

386 .. 38 7
 general; but the real 'situation is, more complicated (I). The thatp was occasionnally written instead of w, the scribe having
,rune w is first used as an additional letter in an uncial charter only a vague notion of the latter. There are in fact two cases
of 692. By its side UU, It and fJ continue to be used, the last where the form of the letter is rather, like p, but in both a diacritic
;two especially in Northumbria. The Epinal glossary, e. g., stroke over the letter shows that w was intended (I). Baesecke
has only ten instances. of w. The ,first dated text in which It also records three occurrences 0.£ It for w in his transliteration,
appears is a charter of 8u; but in the Corpus glossary (first but as far as I can judge ftom his facsimiles these characters·
half of the eighth century) it is already used more often than tho hardly differ from the regularw; at any rate the differenc~is
It had a dangerous rival in ii, which had been lierived from d too small to read ~ (so there'are 30 instances of w in all). In­
'by means of a distinctive stroke. In many texts, e. g. the: stead
. of ~ the Hildebrandslied has iI (as far as the sound in
'

Lindisfarne Gospels, the' Vespasian Psalter, .the earliest West question is distinguished fr~m the corresponding stop). Since
Saxon manuscripts of the Cura Pastoralis, iI is the rule; It appears the origin of the manuscript containing this fragment is laid in
. only rarely (esp. in the abbreviation., = P2t). The. situation Fulda, this use of the w-rune is easily accounted for i it is paral­
is in fact exactly' the opposite of what we should expect: the, leled by the use of ii, re, ~ andother marks of English influence.
frequency of the runes w and ~ increases in the cOUlire of time The so-called Basler Rezepte, where w appears as well, are
(at least during the DE. period), and on the whole it is lowest also found in a Fulda codex. A few more instances are recorded,
in the region where runic writing seemed to be most firmly but on the whole one may say that Continental scribes did not
established: Northumbria.· This impression is corroborated adopt runes as additional letters (2).
by the appearance of runes on Anglo-Saxon coins : no runes Only on one point did Continental scribes venture beyond
are found on Northumbrian coins before the reign of Eanred the eXample set by their Anglo-Saxon teachers. There are a
(807-841). Mercia seems to have played an important part in few cases where the OHG. prefix ga- or ka- is rendered by a ,
popularizing ruilic writing for new. purposes : there coins with charac;t:er identical with the English rune ;. In London MS.
runic inscriptions were minted before the end of the seventh Brit. Mus. Arund.' 393 it is found in some twelve instances (3);
century; there too runes appear in manuscripts at a very early the best known examples are ~hose in the Wessobrunn Prayer
date. But, as I mentioned before, no precise statistical data are (Munich MS. 22053). where it appears not only in ga-fregin
available by which the history of tl;U$ type of,runic usage could and ga-uuorahtos, but also in forgapi (4). At times the scribe
be traced in detaiL Yet ,such a history would be of interest
not only to runologists, but also to. students of Anglo-Saxon
sCript and to philologists:,' , . (x) In their Manuel de l'Allemand du. Moyen Age. Paris, 1947, 309, A.
On the Continent this' type of runic usage is very rare, and JOLIVET and F; Mossi! (who base their text on Steinmeyer's edition) mention
four cases; actually there are onJy two, and both are in the first hand.
a certain indication of English jnftuence. The fragmentary
(a) W. Km:.1:.BR, A,ngelsiicluis&he Palaeographie, IZ.
Hildebrandslid (2) has mostly w (27 instances), but also uu J. BLOMI"IBLD, RJUIeS, 185 and note (I).
(5 times, before vowels) and. Jl (5 times between a consonant (3) E. STEINMBYEll - E. SIlM!ltS• .AlthcH:luleutsche Glossen II, 149. 16. 18
and a vowel, twice between wand a vowel). Baesecke assumes (note). 48.50.53.55; ISO, 5.9 f. I7.aS.3 8 ff.; IV, 49 6•
(4) E. STI.!INMIIYl!R, Kleinert! althochdeu.tI&he Sprachtknkmlilllf', 16 ff. The
WessoTmmn Prayer occurs on fols.. 6S r-66v of the manuscript, of which a fac­
(x) W. KEi:J..BR, A,ngelsiichsiscM'P~gTaphie, Ia i., ~ ff. simile edition wu made by A. VON'EcKART und C. VON KRAus (Milnchen. I9ZZ).
J. R:utru, 184 :ft
BLOMI"IBLD, Cf. also B. BIBCHOl'l', Schreibschuler& I, 18 ff.,:and G. BAIlSECKB'S review in
(a) G. BAliSECK'E, Das Hildel1randlied. Eine geschichtJieM· Einleituni/ far A.f.d.A.'60, 1941, 16 ff.; two instances are found in the facsimile printed by
La:Um, mit LichtllilMrn der Hamlsihrift, alt- tmd netdwclukutscIrMJ TextBn. GERHARD EIB in his Altdeuts&he HamischriJt8n (MUnchen, 1951); his comment
I Halle, MCM'XLV. ' . on the rune (p. Z4) ~ nUsleading. .

388 "'0,
38 9
no longer understood this symbol, as in the gloss ·on fol. 63 l f (= feoh) by):> frofur fira gehwylcum.

of the same manuscript : Sceal deah manna gehwylc miclun hyt dmlan

poetica: *
kazungali,
where the prefix is written out after the j (I). The character
gif he wile for drihtne domes hlootan

u (= ur) by):> anmod and oferhyrned,

felafrecne door, feohte):> mid harnum,

is usually explained as a ligature of g and~i; it would have

mmre morstapa; ):>mt is modig· wuht.

indicated at first the OE. and OS. form of the prefix, gi-. OHG.
}J (= dorn) byp dearIe scearpj degna gehwyIcurn

scribes would then have adopted it for ga- or ka-, the form of
anfeng ys yfyl, ungemetun repe

the suffix in their language. Such a reconstruction is, however,

manna gehWylcun de him mid rested, etc. (I).

unnecessary. There is no proof of this suffix having been used

for gi-, and sO the explanation as a ligature is questionable. It is not always clear what the poet meant by his circumlocutions.
I would rather connect it directly with the rune j, which has Thus the stanza on 0 = os has been interpreted as referring to
the values g and k in runic alpJiabets. Woden (OE. *os < *ans- 'god '), but some scholars have
All in all this use of runes was very short-lived on the Con­ equated os with Latin os ' mouth' (2). But one should not
tinent: the Hildebrandslied is dated 810-820, the Wessobrunn forget that the name was SO essential a part of the rune, that it
codex ca. 814. The other instances ·also belong to the ninth (or its meaning) could not be changed arbitrarily.
century. Apart from Fulda, Augsburg may have witnessed The most famous examples of this usage are Cynewulf's .runic
this usage for a brief time (Wessobrunn codex); the London signatures. All we know. about Cynewulf is contained in the
manuscript also originated in the South of Germany (2). autobiographical passages of four poems. From internal
In England the use of w and }J continued for centuries. The evidence he ;s usually supposed to have written in the latter
character w was still fairly generally used in the twelfth century; half of the eighth century, and to have lived in Mercia or North- . .
}J went out of use only in the MoP-ern English period, till the umbria; but more recently arguments in favour of a date in
fourteenth century it was quite frequent. These two letters the ninth century have been put forward (3). In the autobio­
were also borrowed by Norse scribes (3). graphical passages Cynewulf indicates his name by representing
each letter in it by a rune, the name of which is at the same
(b) RUNES WRITTEN FOR mEIR NAMES time a word in the text of his poem (C = c cynn, etc.).
Two of these signatures are found in the Vercelli Book, i. e.
As a first instance of this usage we have of course to examine Vercelli MS. CXVII, written in England in the latter part of
the OE. Rune Poem, a brief account of which was given in
Chapter I (p. 19 f.). There I also indicated why an edition
of the whole poem could not be undertaken within the frame (I) E. V. K. DOBBIE, The AnglocSa=n: Minor Poems, 28. "Wealth (f)
is a joy for every man. Yet everybody should distribute it freely, if he wants
of this study. The following extract is sufficient to give an i' to obtain it by the Lord's judgment.
idea of the technique; the runes and their· values and names The bison (u) ia very fierce and has big horns; a,very dangerous animal,
have been examined in Chapter I. it fights with its horns, the famous inhabitant of the moors; it is a brave animal.
A tho n (11) is very sharp. Tou~hing it is painful to all men, extremely
hurtful 0 everybody who rests a~t it ".
(1) E. STBINMEYl!R - E. SIEVERS, AlthochdeutscM Glouen IV, 312, 7. (z) This explanation was first proposed by J. M. KEMBLE, On A1IIIlo-Saxon
(a) B. BISCHOFF, SchreiJm:lruIm, al (footnote). Runes,140; it was adopted by B.DICKINS and W. Km..LER, cf. E. V. K. DOBBIE,
(3) H. SPEHR, Der UTsjrrImg der islii.ntJisdum Ichrijt wrd ihre weiteTbildtmg his The Anglo·Saxrm Minor PoemJ, 154.
:ttUT mitee des I3. jahr1Hmderts. Halle (Saale), 19ac), 8 if., 3z if. (]). K. SIBAM, Cymwulf, 304 if.

390 39 1
-.
 the tenth century (I). This 'manuscript 'contains the poems The passage from Elene offers no such difficulties :
Fates oj the Apostles and Elene by Cynewulf, four other poems .A wres secg 04 <fret

(Andreas, Soul and Body I, Homiletic Fragment I, and Dream. cnyssed cearwelmum, c. drusende,

oj the Rood), a series of twenty-two Old English homilies and peah he in medohealle ma4mas pege,

a prose life of St. Guthlac. The runic signatures are found on replede gold. y. gnornode

fols. 5'V (Fates oj the Apostles) and 133t (Elene). The former .n. gefera, nearusorge dreah,

page has unfortunately suffered much from blotting: part of enge rune, prer him .e. fore

the text and most of the' runes have' become illegible. G. P. milpadas mret, modig prregde

Krapp reconstr.ucted the text as follows (doubtful or illegible wirum gewlenced. w. is geswiilrad,

runes are between brackets) : (2) gomen refter gearum, geOgOll is gecyrred,

Her ma:g findan forepances gleaw,

aid onmedla. u.· wres geara

se 4e hine lystell le04giddunga,

geogollhades gIrem. Nu synt geardagas

hwa pas fitte fegde. f. }>rer on ende standep,

refter fyrstmearce for4 gewitene,

eodas pres on eordan brucap. Ne moton hie awa retsomne,

lifwynne geliden, swa .1. toglided,

woruldwunigende; (w). sceal gedreosan,

flodas gefysde .f. reghw~m bid

.u: on ellIe, refter tohreosan,

Irene under lyfte; landes frretwe

Irene lices frretewa, efne swa .1. toglide4.

gewitap under ,wolcnum winde geliccost

ponne (c) ond (y). crreftes neosall •

(Elene 1256-1271) (I)
nihtes nearowe, on him (n) lige4,
. The runes of the signatures in the VerceUi Book are on the
cyninges }>eodom. N u 4u cunnon miht
whole well made. In that of Elene only 1 is somewhat conspic~
hwa on pam. wordum wres werum oncydig.
uous : its lateral stroke is almost horizontal. The' scribe took
(Fates of the Apostles 96-106) (3) evident care to copy the runes as faithfully as possible. Before
w and 1 there are traces of similar runes having been erased,
probably because they were judged unsatisfactory. In the
(I) M. FOJ!RSTBR, R Codice Vereellese, con Omelie e Paerie in Lingua A;,glo­
sassone. Roma, 1913 (facsimile edition). signature of Fates of the Apostles only tp.ree runes can be made
M. FOIISTIiR (= Id.), Der Vercelli-Codex CXVII nebst 'Abdruck einiger 'out now (fu 1); here again the stroke of 1 is horizontal. The
alttmglUcher Homilien tier HaruJ'cJuift. In: Studien sur tmglischm Philologie L text of the signatures is quite straightforward and the inter­
'"' Festschrift fiir L. Monbach ... Redigiert von F. HOLTHAUSBN und H. SPIIlSS.
Halle, 1913, zo-179.
G. p~ KRAPP, The VerceUi Book. (I) G. P. K.RAPP, The Vereelli Book, 100 f.
(a) G. P. KRAPP, The Vercelli Book" S3 f., I:I3 f. .. Always till then the msn had been beaten by surges of sorrow; a
.. (J) G. P. K.RAPp, Tlut Vercelli Book, 53 f • smouldering Torch (C) was be, though he received treasures in the mead-hall,
"Here can theIruUl, shrewdin perception, who delights insongs, discovetwho apple-shaped gold. He bemoaned the Evil (y), he, the comIllde of Sorrow (N);
wrought this measure. We6lth (F) comes at the end: eaxls enjoy it on earth; he suffered distress, cruel secret thoughts, though for him the Hone (E)
they may not always remain together, dwelling in the world; Our (U) Plemure measured the mile-paths, proudly ran, decked with adornments. Joy (W)
(W) on earth sbaI1 pass away; the fleeting adornments of the flesh shall after­ lessened, and pleasure, as the yean pass; youth has departed and the pride
wards perish, even as Water (L) glides away. Then -shall the Bold WariiOr (C) of old. 'The splendour of youth was once Oun (U). No~ in due time the
and the Wretched Orlll (Y) crave help in the anguish of the night; Comtraint old days have gone, the joys of life have fled, even as Water(L) glides away,
(N) lies upon them, the service. of the king. Now ~yest thou know who the moving floods. Wealth (F) is fieeting for all men under heaven; the
bas been made manifest to men in these words .. (R.. K. GoRDON, Anglo-Sa:wn adornments of tne earth vanish under the clouds like the wind ..... (R. K. GoR­
Poetry, 199). ' DON, Anglo-Sa:wn Poetry, a59)·

393
392
' ..
pretation of the runes .and their names offers little difficulty: .1. flodum bilocen, lifwynna drel,

It was of course not always easy to make acceptable poetry with .f. on foldan. poone frretwe sculon

such heteroclitic words as the rune-names were. In a number byman on brele ... (Christ 797-808) (I)

of cases Cynewulf is supposed to have substituted homonyms

for the actual rune-names (I) : thus in the two passages just
This is the paSsage from Juliana :

quoted the name of u, UT, would have to be understood as· the G~omor hweorfed

possessive • our' and not as ' aurochs', the actual meaning of .c. y. ond .n. Cyning bip repe,

the rune-name. Similarly the name of c, cen ' torch', would sigora syllend, poone synnum fab

have become Ctml! • bold, brave'. We shall see. that there are .e. w. ond u. acle bfdad

serious reasons for rejecting such interpretations. hwret him refter dredum deman wille

The poems Christ and Juliana are found in Exeter MS. 3501, lifes to leane .1. f. beofad,

the famous Exeter Book, also written in the latter half of the . seomad sorgcearig. (Juliana 703-709) (z).

tenth century; it is supposed to have originated in the West In this last signature Cynewulf is supposed to have followed a
Country (Crediton?) (z). Wanley knew it, and Hickes had
different technique, for the runes do not seem to make sense
facsimiles of most runic passages pririted in his Thesaurus if read by their names. . Many scholars hold that the runes
(runes are found in various other places of the manuscript). stand for words with the same initial as their names, but there
If we leave out of account the first seven folios, which originally
is no agreement among them as to which words may have been
belonged to MS. Ii.ii. I I of Cambridge University Library, the meant (3), and there is a serious objection against this sort of
,codex consists whoUy of an anthology of OE. poetry, the most interpretation: it may easily have missed the aim of these
extensive that has come down to us (some eight thousand lines). signatures. K. Sisam has given a good account of the impli.:
Cynewulf's runic signatures occur on fols. 19 v (Christ) and 76r cations of these signatures (4). Cynewulf wished to be reinem­
(Juliana). The former sho.ws the same technique as the two bered, in the prayers of those who read his works or (this wil.­
quoted before :
Poone .c. cwacad, gehyred cyning mredlan,

rodera ryhtend, sprecan repe word

(1) G. P. KRAPp-E. V. K. DOBBIE, The Exeter Book, as·
Then the Bold (C) shall tremble, he shall hear the King speak, the Sover­
. pam pe him rer in worulde wace hyrdon,

eign of the skies utter stern words to those who before hearkened heedlessly
pendan .y. ond n. ypast meahtan
to Him in the world, while Misery (Y) and Distru$ (N) could most easily find
frofre findan. Prer sceal forht monig
solace. There many a one accursed shall wait in fear in that place what dread
tonnents He will doom him according to his deeds. Gone is the Glodnus (W)
on· pam wongstede werig bidan
of earth's gauds. For long our PosseSsion (U), our portion of life's pleasures,
hwret him refter dredum deman Wille
our Fortune (F) on earth, was overflowed by Waterjloods (L). Then shall
wrapra wita. Dip se .w. screcen
gauds be burned in t;he blaze ..... (R. K. GoRDON, Anglo-Sazon Poetry, 163).
(a) G. P. KRAPp-E. V. K. DOBBIE, The &eter Book, 133·
eorpan frretwa. .u. wres longe

"Sadly will joumey Ch, Yr and Nld; the King of Heaven, the Giver
of Victory, will be stem when Eoh. Wyn and Ur, sin-stained and trembling,
, (1) Or at least to use them in other meanings as those recorded in the Rune await what will be adjudged them according to their deeds, as the earning of
Poem. life on earth. Logu, Feoh shall stand. and quake in misery .. (K. SISAM, Cyne­
(a) The Exeter Book oj Old English Poetry. With 111lr0ductory Chapterl wulJ,317).
by R. W. CIw.mm!s, M. FORSTER, and R. FLoWER. London, 1933 (facsimile 6) A survey of various solutions is given by G. P. KRAPP and E. V. K.
edition). DOBBIE in the notes to their edition of The Exeter Book (p. aS7 f.).
G. P. KRAPp-E. V. K. DOBBIE, The Exet" Book•• (4) K. SISAM, Cyrmoulf, 3zo f.

394 395
....
 usually have been the case) who listened to them being read meodoheall monig .m. dreama full,

or recited. We must suppose that in Cynewulf's time the runes opptet pfet onwende wyrd seo swipe (I).

and their names were still well known: an audience would This m must be read monn, first element of the compound
immediately grasp that the rune-names pronounced by the monndreama 'of human joys '. In Riddk 91 (fol. 129v) we
reader stood in fact for runes, and that these runes spelled the find a similar usage (11. 3-7)
poet's name. Without a listener's being able to solve the riddle, Oft ic begine ptet me ongean sticad,

and thus being invited to fulfil the poet's request, the signatures ponne ic hnitan seeal, hringum gyrded,

make no sense. Therefore that in Juliana will have to be read hearde wid heardum, hindan pyrel,

as conservativ~ly as possible, even at the expense of the meaning fora ascufan pzt mines' frean

of the verse. This applies of course also to the other signatures. mod .w. freopad middelnihtum (2).

The runes of the Exeter Book have on the whole a somewhat

more decadent appearance than those of the Vercelli Book. In Here w must be read wynnj it is part of the compound modwynn
a passage to be me~tioned later the scribe took g for a minus.. • that which rejoices the hem' = • treasure'. The answers
cule Xi in another u was so poorly made that all editors have to two or three Riddks are also supposed to have been marked
read 1. In the signature of Juliana c looks practically like U. with runes:
One more detail is worth mentioning: the I-rune has a hori­ fo1. 103r , at the end of Riddle 6 (3), there is an s-rune. As
zontal lateral stroke in Christ, and one. almost so in Juliana; the answer to this riddle is generally assumed to be
as we found the same in the Vercelli Book signatures, it may • sun " we may conclude that the rune stands for its
go back a long way. From the runes in Cynewulf's signa~ures name sigel == • sun '. But in other cases such
no important chronological data can be derived : he knew the
rune y, but that had probably existed for quite a while before
fol. IOZ v , at
.
an explanation, does not seem to work. On
the end of RiddkS (4). there is also a
he wrote.
There are several other instances of the same use of runes
;
runes. But ,the solution ,of that riddle is
in these two manuscripts. On fol. 99 v of the Vercelli Book • shield' j therefore we must accept that t~
(i. e. in Homily VIII) we find geJean m meahte, where m must time. s stands for OE. scield. • shield'. Again, on .
be read man. In Elene there are two more occurrences of the fol. 103r some scholars believe to find a runec at the close of
rune w, one on fol. uS" (1. 788 of the poem), one on 131v (1. Riddle 7 (S). The solution of that riddle seems to
1089). In each case we have to read wynn: fJJeToda w = fJJeToda
'wynn • joy of hosts 'j wuldres w wuidres arymn • joy of
(1) G. P. i<.ftApp~E. V. K DoBBIB,The Exet/!T Book, u8, 365.
heaven' (I). .. Bright were the castle-dwellings, lDIII1y the bath.houses, lofty the host of
In the Exeter Book the '\1Sl8e appears on fol. 124r in the poem pinn8c1ea, gre.t the tumult of men, IDIII1Y a mead-hall full of the joys of m'en:
Ruin, and on fo1. u9 v , in Riddle 9L This is the passage from till Fate the mighty overturned that " (R. K. GoRDON, Anglo·S_ Poetry, 92 ).
(:2) G. P. KRAPP • E. V. K. DoIlBlB, The Exeur Book, 241, 379· .
Ruin (11. 21-24) : .. Often I gape at what is fixed ~posite to me, when, girdled with rings,
I must need& thrust stoutly against the hard bolt)pie~ed from bel\ind I ~U8t
Beorht Wceron burgneced, bumsele monige,
shove forward that which guards. the joy of my lord's mind at midnight"
heah horngestreon, heresweg micel, . (R. K. GoRDON, Anglo-S_ Poe.try. 340)·
(3) G.P. KRAPp·E. V. K. DoBBIB, The ~ur Book, 184, 325·
(4) G.P. KRAPp-E. V. K. DoBBIB, The Exet/!T Book, 184, 3 25.
(1) G. P. KRAPp, The V/!Tcelli Book. 144. 147. (5) G. P. KRAPp-E. V. K. DOBIIIB, The &et/!T Book, 185, 32 5.

396 397
~.
 be 'swan' ; the rune would then starid for Lat. this finding to the above passage, we must read each rune by
cygnus. But if we accept this explanation, there is its name. .Only one commentator- has tried to give this solution
no reason for not interpreting the two s-runes just to the puzzle, and he can hardly be said to have convinced
mentioned as scUtum (Riddle s) and sol (Riddle 6) everybody (I); and yet the solution must be sought in this
respectively. Moreover one should not forget that direction, even if the meaning looks forced.
these runes are made very differently from those in Other runes from the Exeter Book will be discussed on p. 'P7..
the texts : they are hardly mOre than faint scribbles, It is worth noting here, however, that all runic bits poinnoa"
whilst the other runes are drawn firmly, with heavy fu1>orc of twenty-eight runes; we meet the four extra runes
,downstrokes. Similarly there is a scribble on a lie yea, but neither g, k nor ;,
fol. 10Sr which most editors have interpreted as a The rune CE is used several times for the word epel : three
rune b with a rune 1 above it (I), and several runes times in Be()fI)Ulj, once in Waldere and once in King Alfred's
are found in the margin of fo1. I2Sr (b u gr ... ?); . translation of Orosius. The Be()fl)U[f manuscript (z) (Cotton
but'in neither of these instances do we have runic MS. Vitellius A 1 S), consisting of two parts joined by a seven­
writing of the type discussed in this section. Still in teenth century binder, WaJ! written about the end of the tenth
the same manuscript, there is a doubtful instance on century. The runes occur on fols. 141V, l49 v and 167r; in
fo1. 123v in the poem The Husband's Messt;Zge (11. 49-53): the third the rune is part of the compound epelweard ' guardian
Gecyre ic a::tsomne .5. r. geador of the native land, king ':
.ea.w. ond .m. a1>e benemnan, pa hine on morgentid
1>a::t he 1>a wrere ond 1>a winetreowe on Hea1>o-Ra:mes holm up retbrerj
be him lifgendum lrestan wolde, aonon he gesohte swresne .CE. ".

1>e git on rerdagum oft gespra::conn (2). leof his leodum, lond Brondinga ... (11. 518-521)
The last of these runes has exactly the same form as that in the ... swylce oft bemearn a::rran mrelum

poem Ruin and in Riddle 19 and must consequently be read swiMerhJ?es sid snotor ceorl monig,

as m (most editors read d). M. Forster (3) made an interesting se l>e him bealwa to bote gelyfde,

observation on the arrangement of the runes in the Exeter pa::t 1>a::t deodnes beam ge1>eon scolde,

Book: whereve.r a rune stands for its name, there is a dot before fa::derre1>elum onfon, fole' gehealdan,

and after it; when a sequence of several runes is to be read in hord ond hleoburh, hrele1>a rice,

.CE. Scyldinga. (11. 907-913)

the same way, each pair of runes is also separated" by a dot.

On the other hand, when several runes are to be read as one pret, la,· mreg secgan se 1>e soa ond riht

word, i. e. only by their sound value, there is a dot before and fremea on folee, feor eal gemon,

after the whole group, but none between the runes. If we apply eald .CE. weard, 1>a::t des eorl wa::re

(I) G. P. KRAPp-E. V. K. DOBBlB, The E:xeter Book, 189, 330 f. (I) E. A. KOCK, InterpTetmi.mu and Emendatiom of Early English Te;)Its.
(a) G. P. KRAPp-E. V. K. DOBBlI!. The Exeter Book, zZ7. 363 f. VIII. Angfui 45 (19zI), IZZ : s r =;; rigelrad" the sun's road" .. hellven ";
.. I put together S.R.EA. W. and D. to assure thee with an oath that he ea w = eanoynn " earth's joy II " earth"; he translates" I place together
was there, and that he would perfonn, while he lived, the true ~aith of which Heaven, Earth, and Man, confio:ning by an oath that he would keep, throughout
you two often spoke in earlier days II (R. K. GoRDON, Anglo-Sa:ccm Poetry, 90). his life, the compact and the faith ", etc.
(3) G. P. KRAPp-E. V. K. DOBBIE, The Euter Book. xxiii. (a) F. Kt..um1!R, Beowulf, xcvff.

398 ... 399

 geboren beteral (11. 1700-17°3) (I) hrefde abrocen Tyrom pa burg, seo wres on rerdagum heora
Of the poem Wtddere, dealing with the legend of Walter of ieldrena (E, 7 ondredon pret he eac to him cuman wolde (I).
Aquitaine (better known from Ekkehart's Waltharius), only In the other manuscript of this text (Cotton MS. Tiberius B I,
two fragments have come down to us (2). They are preserved saec. XI) the word is written in full (eilel). There is one more
in the Royal library in Copenhagen under number Ny kgl. isolated example of w = wynn, viz. in Psalm 99 of the Junius
sanil. 167b, and were written about the year 1000. At the end Psalter, where w sumiap translates jubilate (i. e. 'lOynsumiap) (2).
of the :first fragment we read In MS. B of the Poetical Dialogues of Solomon and Saturn
Forsoc he &.:m swurde and, dam syncfatum (Cambridge MS. Corpus Christi College 41) the name Salomon
,beaga m.a:nigo; nu sceal beaga leas is twice written Salom (headings preceding 11. 39 and 63. on
hworfan from disse hilde, h1afurd' secan p. 197 of the manuscript) (3). The scribe was not content to
ealdne .(£.odde her rer swefan ... (11. 28-31) (3). write only the m-rune; above it he made the horizontal abbre­
viation mark which is sometimes written above the letter m
H. Sweet edited the text of King Alfred's translation of the when it stands for man, mono This prov~ that he was not
Historiarum adversum pagarms libri VII by Paulus Orosius; the fully aware of the ideographic value of .the rune.
introduction to this edition was never puhlished (4). It should In all these cases' the use of the rune is exceptional. In
have contained a full account of the !1lmost contemporary BeOf1JUI/ e. g. the word epel occurs eleven times by itself or as
Lauderdale (or Tollemache) manuscript, preserved at Helming­ part of a compound (not counting inflected forms such as eple)j
ham Hall, Suffolk. In Book IV, Chapter iv of the translation' in three cases only the corresponding rune was written. It is
(Sweet'!! ed. p. 168, l. II), the rune (£ is again used for the '\ hardly surprising that this use of the runes was not developed
word epel: systematically. Most rune-names were low-frequency words j
lEfter prem hierdon Cartainenses pret se m.a:re Alexander some were extremely rare (ur os cen eoh peorp). Therefore they
could hardly become a system of notaej only a few occurred
(1) F. KLAlmm, Beowulf, 20, 34. 63. so often that scribes could think of substituting runes for them
U Then in. the moming the sea bore him to the land of the Heathoremes. in a more or less systematic way. Such a systematic' use on
Thence, dear to his people, he sought his loved country, the lan~ of the Bran­
a small scaIe is found in the Durham Collector, also known as
dings... So also many a wise man whO trusted to him as a remedy for evils
lamented in fanner times the valiant cine's journey, that the prince's son _ , the Durham Rituai (+). Most of the Latin text was written in
destined to prosper. inherit his father's l'IIllk, rule over the people, the treasure
and the prince's fortress, the kingdom of heroes, the land of the Scyldings ••• (I) This paSslige translates" Post haec Carthaginienses, cum Tyrom urbem,
Lo I he who achieves truth and right among the people may say that this earl auctorem originis suae, ab Alexandra MlIgIlo captam lIVersamque didicissent,
was born exceUent (the old ruler of the realm recells aU things from the past) " tintentes tnmsitum ejus in Mricam futurum.....
CR. K. GolUlON, Angw-Saxon Poetry, 14, zz, 39). (2) H. LoGBMAN, 'I'hII NamtI of the Angw-Saxon Rune p. The Academy 39
(2) F. NORMAN, Walder". London.(1933). (1891), 284­
E. V. K. DOBBlB, The Angw-SaXotI Mi:tuw Poms, Xix fr. (J) R. J. MRNNER, 'I'hII Poetical Dialogues of Solo_ and Satu.m. New
(3) F. NORMAN, Walder", 39 (note: " A loop inside the rune. The scn1>e York-London, MDCCCCXLI, esp. 82, 84­
wrote part of an II first "). . E. V. K. DOBBlE, The Anglo-Sf£JIon M,noT POtrllU, I fr., 31 fr., 160 fr.
.' E'.;:V. K .•DOBBlE. ,T1J8 A7igw-Smton Minar P08nI$, 5, 139.•. (4) U. LlNDELOF-A. H. THOMPSON, Rituak EcclesimJ Dunelmeruis. 'I'hII
" he refused' the sword and the treasures, the many rings; now must he Durham Collector. A NeW and "ReTJised editU:m of the Latin Text With the
needs depart from this battle bare of rings; the lord must seek his old domain;· Interlinear Angw-Saxon Version. (Publications of the Surtees Society, Vol.
or here die before..... (R. K. GOlUlON, Angw-S=mr Poetry, 74); CXL). Durham-London, 1927·
(4) H. SWEET, King Alfred's Oroftus I. (Early English Text Society, Ori­ R. A. B. MYNORS. DuTham Cathedral MtmlUcripu to the End of the TweUth
ginal Series, 79) London, MDCCCLXXXIII. Century. Oxford, 1939. 25, no. 14·

400 401
-. 30
the South of England; a small portion is due to a Northern quires were numbered with runes: a 20Z v, b ZIOv, C (? the
,scribe. Between the lines a Northumbrian gloss was added, rune has been curtailed, only a vertical stroke remains) 218 v,
apparently in the second half or toward the close of the tenth lJ 226 v, e 238v, f 246v, g 254 V (I). Of these runes b shows
century. In this gloss the DE. words dreg and monn occur very a somewhat peculiar form : the shaft is lengthened at the top
frequently.- In most cases the scribe did not write them in and below (cf. that on the Britsum inscription) (z). None of
full, but simply drew a d or an m instead (d : 2, 2; 4, 14; 5, 10 the alphabets studied in Chapters III and IV shows the same
(twice), 14, etc., in all 42 times; m: 13, 13, 15; 21, 13, etc. choice of runes. The last alphabet in Exeter MS. 3507 and
10 times) (I). Here we find a still greater degree of integration Cotton MS. Vitellius A 12 has a rather similar alphabetization,
than in those cases where the rune (i. e. the rune-name for but a different type of g. The rune for d (i. e. lJ) is also found
which it stood) was part of a compound: in the Durham three times at the beginning of a sentence on fol. 128 v (a usage
Collectar we even come across des = dreges or de = drege. It rather belonging to our last category). In Vienna MS. I2Z4
is not impossible that this usage of runes for their names is (the Cuthbert Gospels), finally, the ligature et and the rune
found in other manuscripts as well, but instances are probably for a were used in numbering the quires (3). A thorough
very rare. search may perhaps lead to the discovery of more instances of
this usage.
(c) RUNES USED AS REFERENCE MARKS, ETC.
(d) RUNES IN SCRIBAL SIGNATURES, NOTES, ETC.,

Very little information seems to be available on this use of The material of this last category is not easily classified. As
the runes. Some of the material in which occasional' runes a first variety we may consider signatures of scribes. The type
might tum up has hardly been studied, e. g. the reference meant here is in no way comparable with Cynewulf's signatures. "
marks which serve to indicate the connexion between a marginal As far as they contain any text, the signatures in this section
gloss and its lemma (2). The insular glossator of Fulda MS. are invariably written in Latin. This proves that the scribes
Codex Bonifatianus I, who worked in the eighth century, some­ considered the runes only as an ornamental, perhaps also as a
times used runes for this purpose. The same usage is recorded cryptic alphabet. They probably used ready-made alphabets,
from Munich MS. lat. 14179 (saec. IXI), where tunes are used and had little or no contact with runic tradition. At any rate
by the side of other symbols (3). In Munich MS. lat. 629 1 (4), I found no reason for supposing that such inscriptions were
whi,ch we shall have to discuss more at length soon, the scribe made on the basis of fu pores. The earliest examples of Tunica
Madalfrid marked the last four quires of the codex with the manuscripta that have come down to us belong to this type.
runes for abc d (fols. 217v, 225 v, 233 v , 241 V). The first London, British Museum,Harley MS. I77z (4). This manu­
twenty-three quires of Munich MS. lat. 14561, written in the script belonged to the Royal Library in Paris till 1707, when it
South of Germany in the first part of the ninth century, were was stolen; it was probably written in the NE of France (5).
marked with the letters of the Latin alphabet ; the last seven
(I) B. BISCHOFF, Schreibschu1en I, 251.

(I) The figures indicate the pages and lines of the edition by U. LlNDBl.lll' (:/I) H. ARNTz - H. ZEISS, Runeru1enkmiiler, 159 f.

and A. H. THOMPSON (cf. preceding note). (3) B. BISCHOFF, Schreibschulen I, III note (I).
(2) Prof. O. HOMBURGIlR kindly called my ~tion to some Berne manu­ (4) Catalogue of ths Harleio.n. Mamucripts II, 2Il if.

scripts using a rich variety of such symbols; among th~ I saw, however, Catalogue of Ancient Manuscripts in ths British Museum II, 38.

I failed to recognize I'Wles. (5) G. MICHELI, L'enlumillure irlaniJajse, 88 f.

(3) B. BIScHOFF, Schreibschulen I, 235 and note (I). J. F. K.I!NNEv. Sources I, 659111118 :" written probably in Northern England I t
(4) B. BISCHOFF, SchreibschuUn I, IIO f.; d. p. 409. but the runes should be su:lijcient to disprove this assumption.

:40 Z 4°3
It is usually dated in the late ninth century, but as far as I can
judge it could almost be a century older. It contains the Pauline
and the Catholic Epistles (except III John and Jude) and the
Apocalypse (only as far as XIV, 16). The codex has been
~&h~MrIN
b d
Q. 'c
1- rt1V1~~p ~K7-n
f 9 h ,~ I' -
flo q t u
r
FIo. 52
m" n 0 r 8
(I)
damaged by fire. On fo1. 6 v, after six lines of introductory
text to Romans, there is a large ornamental inscription in runes, we obtain the following reading (2) :
alternately one line black and one red: EGO IUSUEUS HAC/SI INDIGNUS DIACON[US] I
[H]ANC LIBRUM QUEM I AD OPUS PECULIARE I
UOLO OFF[ERRE] S[AN]C[T]O.

M~ln?nMn7HRh

The alphabet used here is again different from any of those we

have met before.
. . The first three runes are quite naturally abc (in the last

711V1~I;KVln ~~I~~~H

the lateral stroke transects the shaft). No. 4 probably goes

back to It, the triangle of which has been given the full height
of the rune. The example of Roman D may have played a
part. The next two runes are again the normal types, but that

rVl~fi~Knm nMl11
for g is,the English j, cf. g in two alphabets of Exeter MS. 3507
and Cotton MS. Vitellius A 12; also in Phillipps MS. 3715,
Arsenal MS. 1169, etc. The symbol for h has a fancy stepped
cross stroke instead of two parallel strokes. For i we find the

~~~hn7hMhn I~RM usual type; k does not occur. The form of I points to a poor
model rather than to an artistic development. The next rune
is formally identical with the x of some De inventione alphabets.

n~~~f~~~'
If it has the same origin, it must have been chosen to take the
place of 111. on account of its rel'!emblance to Roman M (cf. d
and q). The symbol for n has nothing to do with the n-rune;
it seeIns rather to be a Roman N turned upside down (or left
FIG. 51 right). It may eventually go back to some form of 3, but this
is only a guess. The runes. for 0, p, r, s, t call for no comment;
It may be transcribed as follows : their peCUliarities belong only to the realm of style. The
choice of q, however, was again influenced by the Latin example.
ejoiusueushac

The rune (E was interpreted as a squarish type of Q, cf. q in

siiNlJijNusltiacoN
two alphabets of Exeter MS. 3507 and Cotton MR Vitellius
aNclibruMQueM

A 12, as well as in Phillipps MS. 3715. The points of contact

a It 0 pus pee u 1 i a re

uolooffs co
(I) The last three letters of the alphabet do not occur in the inscription.
By reconstructing the alphabet used in the im~cription in this (a) J. M. KEMBLE, On Anglo-S_ Runes, 366 and Pl. XIX, fig. aI. In
the first'line Kemble skipped the ninth character, reading IUSUESj the
way: Catalogue of Ancient Manuscripts in tlul British Mweum (II, 38) reads IUSEUS.

40 4 40 5
-.
,(

 with the group of the Exeter manuscript are rather striking, follows next, the latter part of which is in notae sancti Bonifatii
but'they cannot be interpreted as indications of a direct re­ (July-August 806). Below, the scribe writes again AGAM­
lationship. BERTUS SCRIPSIT and the anagram S UTREBMAGA.
M to the interpretation of this signature, the main difficulty One of the manuscripts of the Lex Salica also contains a signature
lies in the first line: Egoiusueushac/si indignus, etc. Mter ego with the same name Agambertus (but without runes). It was
we expect the name of the scribe or the donator. The next probably written in the North of France. But R. Buchner,
seven runes may perhaps be interpreted as luseuus Josephus; who compared this manuscript with Mangeart's facsimile, holds
but does such a form exist? (I) Another tempting analysis an identity of the two impossible (I). In Berne MS. I I 8
would be iussu eius, but this hardly fits into the context. The B. Bischoff discovered another signature of one Agambertus
author of the inscription was at any rate a poor Latinist, as (2). This is not the place to examine the possible relationship
appears from anc librum in 1. 3. between the three codices; yet I do not believe that Buchner's
Valenciennes, Bibliotheque Departementale, MS. 59 (2). judgment must be considered as final.
I came across the runes in this manuscript when looking through Agambertus's runes are well made: they show not the least
Mangeart's catalogue of the Valenciennes library. They are tendency towards cursivation. From his use of runic ligatures
not mentioned in the index but a glance at the facsimiles is (' Binderunen ') one might conclude that he was well acquainted
sufficient to recognize these firmly drawn characters. The with runic practice : in his name m b and e: r are written with
codex formerly belonged to the Abbey of St. Amand (Elnone); common. vertical strokes. But he uses the same device with
it contains Jerome's comment on Jeremiah and his Ad Paulinum his Greek characters (e. g. Hand P in fieri), and this somewhat
de studio Scripturarum. The scribe had a special liking for diminishes the value of this argument; after all, ligatures were
strange alphabets; he wrote part of the title with Greek letters; a rather common device, e. g. in monograms.
on fo1. 281' the words Deo gratias amen are spelled with Aethicus Munich, Bayerische Staatsbibliothek, MSS. lat. 6250, 6272,
Ister's characters (but r is a Greek P). His signature on fo1. 18 I v 629I. Although the runica in these three manuscripts do not
is a regular display of cryptic systems. In the left top corner belong to the same type strictly speakhig, they must be con­
is a monogram which must probably be read HLOTTILDIS sidered together because they use the same runic alphabet.
ABBATISSA (3). The next two words are in the scribe's Munich MS. 6250 (3), which contains Isidore's Etymologiae,
peculiar Greek: fPYHPY .QPJYNAOY8 = fieri ordinavit. was written at Freising in the second quarter of the ninth
Then he writes his own name in runes: agambe:rtus, after century. Five scribes worked at it; the last of them, Cundpato,
which he goes on in mixed Greek and Istrian, with an occasional signed his name with runes on fo1. 280r . His signature reminds
rune (and K for I in gratias) : fecit Deo gratias semper Domine us rather of that in the V~enciennes manuscript:
amen. The inscription is dated exactly by the text which EXPLICIT LIBER UIGESIMUS. /J.Q rPA8KAC AMFn
fPAX8.Q fPINH lAM / I.J.VIHCX.Q I.J.VIA AACCA/8VC
(1) One might feel tempted to compare it with ioseW5 on the cover of a HVX VHNI/ cundpato mE parte: scripsit in primis. b.
Fulda codex now in KlII!Sel; but in the latter inscription w ia used for p on
account of its fonnal resemblance to Roman P. i. e. Explicit liber uigesimus. Deo gratias amen.
(a) J. MANGEART, Catalogue descriptij et rairomul des mllnWcrits dela Bihlio­
theq,u de VaIemiennes. Paris-Valenciennes, 1860, MS. 52, p. 50 if. (I) R. BuCHNER, Kleine Untersuchungen 1IU denfrankischen Stamm8uec/lten I.
A. MOLINIl!R, CatoJogue gmbal des mllnWcnts des bihli.otheq,us publ~ Deutacbes Arcbiv fUr Erforscbung des Mittelalters 9 (l951), 59 ff. (esp. 68).
de France. Depa,.tements. T. XVII: ... Valenciennes. Paris, '1894, ::U5 f. (2) Prof. O. HOMBURGER (Berne) kindly brought this signature to my notice.
(J) I wish to thank here my friend Dr. M. GY8SI!LING, State Archivist in b) B. BISCHOFF, Schreihst:huJm I, 95 ff. Cf. also F. J. LAUTH, Das ge-rma­
Ghent, for having kindly assiated me in deciphering this monogram. nische lbmen-Fuda,.k, 9 f.

406 40 7
-.
 Facto fine iam quiesco quia lassatus huc Cundpato's signature) call for remarks. The character for g is
veni. Cundpato me parte scripsit in primis. no doubt derived from the English g, but the upper and lower
The first two lines were written by another hand. In the word ends of the two strokes have been curved inwards; the result
gratias we find exactly the same peculiarity as in Agambertus's is rather like a figure S. For h we find either a symbol identical
signature : K for I. In amen the third letter is rendered by its with d (the cross strokes intersect instead of running parallel),
substitute according to the notas saneti BOTiifatii; the last is an or else a Roman H in which the cross stroke has been lengthened
n-rune, but with the cross stroke slanting down to the left. In on both sides. This is also the form of H in the part written
Cundpato's Greek the use of V for fl, and occasionally that of H with Greek characters. For i we find the normal type (runic
for h are worth noting. The last line consists entirely of runes. or Roman I) but also a character shaped like K. We met this
Although they show some degree of adaptation to the style of same character in the Greek parts of the inscriptions in Valen-.
such subscriptions-most characters being lengthened and a ciennes MS. 59 and Munich MS. 6a50. It may go back to the
few made smaller so as to fit into the spaces between two notae sancti Bonifatii. Instead of I we find a more or less
characters-yet they retain their original forms to a remarkable cursive Greek A; instead of r Greek P is used in one instance.
degree. In one case a letter from another alphabet has crept In the Greek part of the inscription n is found in Iohannem.
in, viz. the uncial It of me; but Cundpato also knew the regular It is also used in the Greek glosses elsewhere in the manu­
runic e, d. parte. For b and p he uses very curious types. script, e. g.
Instead of lateral triangles, these have closedX's. In the case fo1. 58r in foro siue In ANrDPA (=w a'Y0p~)
of b this form may be an ornamental development of the rune b, Sot' MH8HnIICEDCHDC (=P.ETEM6xwUEws)
but then p should rather be connected with W'. For the latter 17ot' KATA An8IfPACIn (=l(ltTa. aJ!'Ttq,pauw).
there is, however, a far simpler explanation: it was inspired by In this last example f is also runic.
Roman P, Cundpato's 'runes' p. and b showing the same Munich MS. 6291 contains St. John Chrysostom's In Epistolam
relationship as Roman P and B. The symbol for b does not ad Hebrasos (I). It was written by six (or more?) scribes. The
occur in the text of Cundpato's subscription, but has been third of these, who was also responsible for the final quires of
written by itself to the upper right of the last word. the manuscript, gave his name in a runic signature (a) :
Munich MS. 6272 (1), containing Jerome's comment on
omnis labor finem habet premium eius non habet I
Matthew, was written at Freising about the same time as the
f£nem madalfrid scJ;"ipsit istam partem do I gratias
preceding manuscript, and partly at least by one of the scribes
Quod ego perfed opus meum.
known from the other codex. On the inside of the front cover
there is an inscription partly in runes, partly in Greek letters. The alphabet of this signature agrees with that used in the two
The runic part may be transcribed as follows (plate VII b) : preceding manuscripts, except on two or three points. For h
matheus in Kudea. maPcus Kn itaAia. lukas Kn acHaKa. we find a Roman H, the two shafts of which are connected by
. ioHannes. in asia euuangeAim predicauit. a broken line (angle turned upwards). It may have been derived
These runes are somewhat less skillfully made than Cundpato's, from the d-like h used in Munich MS. 627a. In the case of I
but the scribe obviously used the same alphabet. Most of the Madalfrid uses the regular· rune, not a Greek character. The
letters occurring here for the first time (i. e. those not in (I) B. BISCHOFF, SchreibschuUm !. IIO f.
(a) H. F. MASSMANN, Neue Rune1/, a8.
J. M. KEMBLE. On Anglo-Smwn Runes. 366, and PI. XVIII, fig. :10.
(1) B. BISCHOFF, Schreibsclnden I. 97 fr. Cf. also F. J. LAUTH, Dds IIW­ F. J. LAUTH, Das german:i:sche R:unen-Fudark, 9.
nmmsche Runen~Fudo.rk, 10. .

J. B. SILVESTRE, PaUQgraphie umvenelle IV, 87 ; PI. a31.

408
40 9
~.
K-like variant for i is not found here, but Madalfrid once writes St. Gall, Stiftsbibliothek, MS. IZ7 (I). According to Bruckner
an uncial e instead of i. The character for q, which does not this codex was written at St. Gall about A. D. 800 (2). It is
occur in the other inscriptions, is not of runic origin, but rather made up of two manuscripts, the former containing Jerome's
developed from the capitalis rustica Q typical of Freising manu­ Commentary on Matthew, the latter his Commentary on Mark.
scripts (I). Finally we should recall here that Madalfrid also At the end of the first part (p. 379) the scribe signed his name
marked the last four quires of the manuscript with the runes in runes (cf. plate VIII b):
for abc d. The alphabet used in these three manuscripts
presents the following characters :

R~NX)(D

~~hHMrXHI-~M~~rQR~t~ FIG. 54

H~)(~~ ~t There can be little doubt that this is the signature of the scribe
a b e d ef'S h rk I m MOP!:!" s tl.l
responsible for the first part of the codex: it is orderly written
FIG. 53 (20) below the explicit and in the same very dark ink. It has usually
been read Ratger, but all scholars who printed the name in this
The use of these runes seems to have been on the point of way prudently put a question mark after it. The interpretation
becoming a tradition in the Freising scriptorium. How did of the signature is indeed somewhat doubtful. Transcribed
they happen to reach this centre of learning in the South of according to the usual system, it gives an impossible reading :
Germany? Since Bischof,f's study of the Freising manuscripts ra!dggu. Only a comparison with Continental runic alphabets
the origin of that scriptorium is fairly well known (3). In its can help us to solve this puzzle. From the adaptations found
early period (i. e. the latter part of the eighth century) a scribe in such alphabets we may conclude that the first three runes
who calls himself Peregrinus, and who writes an Anglo-Saxon stand for rad or rat (cf. St. Gall MS. 270, Leyden MS. Voss.
(Northumbrian) hand, seems to have played no mean part. lat. I2 8). As to the fifth, which does not make sense with its
We can of course not know for sure whether the TUnica mentioned English value, it probably has the same value as in Munich MSS.
above really go back to Peregrinus, but the possibility exists. 14436 and 19410. In the latter we found the name ca(a)r
The alphabet we reconstructed shows in fact no traces of an (or: name car, value a); in the former g has the value aj its
adaptation to High German phonology. What peculiarities name is caar. The scribe of the St. Gall manuscript probably
we find are such as could easily spring up when direct contact knew an alphabet similar to those in the Munich codices, and
with runic tradition had been lost, without any conscious his g has the value a. The last rune is formally rather like u,
adaptation to new sounds and values.
(I) H. HA'l'1'I!MlIR, Denkmahle I, 4 10•
G. SCHBlUIER, Vertteiclmiss, 46.
(I) B. BISCHOFF, Schreibschulen I, 66. Perhaps the resemblance to the rune E. STElNMRYER-E. SIIWERS, Althoc}uJeutsche Glossen IV, 443·
V played some part in the adoption of this fonn of q.
(20) The last three letters of the alphabet do not occur in the inscriptions.
(
A. BRUCKNER, ScriptoNa II, 1, 64 and PI. XXVIII.
(7) K. LOFFLER, Dill Sankt Galler Schreibschule in der ersten Hiilfte des 9·
(3) B. BXSCHOFF, SchreibschuIell I, 60 fi.
Jahrhunderts, 31 , characterizes the handwriting as Rhaetian; at any rate the
Cf. G. BAEIBClCB's review: A. f. d. A. 60 (1941), 14 f.

410
I ~.
manuscript belonged to the St. Gall library at an early date.
4Il
but a similar form stands for r in the Britsum inscription (I) I fail to see what it can mean in this context, unless it stands
and in the second alphabet of Vienna MS. 1761; it may also for a (as a variant for the a of amen? Cf. Ratgar's signature).
have played some part in the genesis of the curious y-shaped r Then the scribe signed his text :
of the Syriac alphabet in Munich MS. 144-36. Thus we obtain Quisquis titulum legat hunc mihi & misereatur /
the reading Ratgar, which is satisfactory from every point of indigno precemque fundat rogitatue qui uocor alpunc./
view. This name was not unknown in St. Gall: a Ratgari(us) ERCHANBERTUS. ebs' ualeas uigeasue felix.
consecretari(us) i~ mentioned in charters of 820, a Ratgari(us)
Alpunc's runes are well made, especially those of amen. Those
camerari(us) in one of 828; Ratger(us) secretari(us) appears from
in the last line have been lengthened to match the tall, narrow
817 till 829 (2). There may be Some connexion between our
letters of Erchanbertus; from the point of view of runic style
Ratgar and the Ratgar or Ratger of the charters.
they are less satisfactory. They raise few runological problems:
One may ask whether the scribe used two different characters
for a (and r) on purpose, and if so, what was his purpose. Did
he wish to mark a difference (in quantity?) between the two a's?
Remembering the procedure followed by' the scribe of Beme ~ &M~ XIf' f4 't ~ ~:*.

MS. 207, who used as many different characters for the same abe f <J ,I m n a x ?
sound as he possibly could, I rather think Ratgar simply wanted
FIG. 55
to conceal his name the better by using two different characters
for a; this would also explain why there ar~ two different
types of r. The b-rune is used with the value p (eps episcopus). which
Ratgar's signature may be of the greatest importance if one may either be due to the influence of OHG. phonology (I), or
wants to trace the origin of the alphabet type found in the two else to some analogical influence of OHG. biscoj • bishop'.
Munich manuscripts. If Bruckner's localizatio~ holds good, The character for :Ie is identical with the variant e-rune of the
the St. Gall manuscript may indicate by what way those runes Franks casket inscription; it does not occur elsewhere, and has
reached the South of Germany. . . probably a different origin. It is most likely formed on ,the
Munich, Hauptstaatsarchiv, Hochstift Freising, Lit. 3a (3). model of Roman X. The i of uigeas begins straight but ends
On fo1. 397r of this first liber traditWnum of Freising ends the in a wavy line; I can see no explanation for this peculiarity,
record of a gift by Bishop Chunihoh, dated 850~ It is followed except a whim of the scribe.
by an investiture ending in the words finit jeliciter, and amen S~ill' in the same category I have to mention the signature of
written in runes (amen). The n-rune of amen is followed at the s'cribe Ercanfrit in a Wiirzburg manuscript (2).
some distance by a character which looks very much like g. What we found in Munich MS. 6272 (cf. p. 408) was in fact

(x) H. ARNTz-H•.Zmss, IbmnIdenkmiiler, x60;' d. alsa the Dahmsdorf r,

ibid., 1:11.
(I) W. BRAUNB, AltlwchMutsche Grammatik, § x3 6 ; d. also § 133 A. 3·
(2) H. WAR'I'MANN, Urkundenbuch tIer Abtei St. Gallen. Zurich, 1861-19°4. (2) Edited by F. D. GRATER: Inscriptio Runice Herbipofita, Runis Anglo­
I, 213. 23'7, 240. 26'7, 2'76. 280. 283, 293, 302, 303 (nos. 223. 246, 249, 285, sa:wnicis concepta. Idunna und Hermode I (1812). 55·
291l. 30 3. 306. 316, 32'7. 328). (Cf. H. AR."I'l'Z, BibliogTaphie, 6, no. 85; AImmT provided GRATER with a'
(J) T. BrI'TEllAUF. Die Traditionm des Hochsti/ts Freising. Miinchen,
" diplOmatische Abzeichnung" of the inscription).
1905-x909. (Quellen und Erllrterungen zur Bayerischen und Deutschen
W. GRII\otM. Ueber deu.lsche Runlm, x65 and PI. V.

I
Geschichte. Neue Folge, 4-5). I, xvii fr.• 601. J. M . .KuMBLE, On Anglo-Sa:con Runes, Plate XIX, fig. 22.
B. BISCHOFF. SchreiblchuJen I. II2 f.
,pz 4 13
~.

 not a scribe's signature, but an inscription with cryptic (hardly discussed it at length and tried to explain its peculiarities (I).
with ornamental) purposes. A similar usage is found in Karls­ On fol. 17 v there is a miniature of the Descent from the Cross.
ruhe MS. Aug. 176, where a Latin: riddle was spelled in runes Above the cross there are three inscriptions, one in Greek
(p. 3Z4). and in Berne MS. z07, where runes (and pseudo-runes) letters (preceded by G graece), one in Roman (L = latine),
were used for an ornamental title page (p. 18z). the third in runes (B barbarice). Hence the Hebrew (or
A sort of inscription which will easily escape notice is that Aramaic) is here represented by runes-obviously because the
of Kassel MS. Theol. F. 65, which contains Hegesippus in sixth artist's knowledge of Hebrew did not allow him to give the
century half-uncials, with corrections in Continental and insular text in that language, nor even in the characters of that language;
cursives (1). The latter are supposed to be from St. Boniface's for the so-called Greek inscription is only the Latin text Ihesus
own hand, or at least from somebody of his surroundings. This Nazarenus Rex Iudaeorum written with Greek characters.
lends special interest to the runic inscriptions scratched on the Therefore we may suppose that the runes spell the same v.ords.
. cover. That on the front cover must be transcribed as follows: As a matter of fact the word divisions appear exactly at the
points where we expect them; but the runes give a very different
iosew3' reading (the corresponding Latin to the right) :
i i 0 S i (?) f g d 0 Y (s) IHESU(S)
iosew3' j a Zan d j (us) NAZAREN(US)
n d x . g JJ c REX IUD
Lehmann proposed to read IOSEPI, and he is probably right
a dIn JJ AEORU
in doing so. The rune w could easily be interpreted as a
i M
Roman P. As to 3', its name ih laid near its use as a variant
for i, cf. Berne MS. Z07. We do not find these peculiarities At the end of the first line the character looking like Y is pro­
in any of the alphabets that have come down to us, but this vided with a curved abbreviation mark which must probably
is no reason for doubting Lehmann's reading: it is just one be read as s; at the end of the next line it stands for us. The
more indication of the extreme complexity of the history of runic character which takes the place of m in the last line looks rather
alphabets. like a capital L, but Selmer is probably right in explaining it
as a somewhat fanciful rune i. For u a capital Y is used in the
In one group with these items we have to classify a very first line; elsewhere it is rendered by the rune 1J.
puzzling inscription in Selmer suppol:'ed that this inscription resulted from the use
Munich MS. lat. I3067. This codex was formerly in the pos­ of a substitution key; there may be something in this idea, but
session of St. John's Monastery in Ratisbon. Its contents I have some doubts about the complicated system of substi­
consist mainly of liturgical items (psalter, breviary, etc.). It is tutions which Selmer had to set up to explain the inscription.
supposed to have been written at Hastiere, a dependency of the He reconstructed the Vorlage with the help of various fuporcs
Abbey of Waulsort near Liege; it is dated saec. XI/XII. The and alphabets (some Qf which are mutually exclusive); he
runic inscription was first mentioned by Lauth (z); C. Selmer supposed that this Vorlage had twenty-eight runes (including
variants for a d e f him). that it had been written in lines of
four characters each, and that in making the key one of those
(1) P. LEHMANN, Fuldaer Studim (I), 16, first edited this inscription: (x) c. SELMER, The 1W.nic lmcription of Codex Latinus Monacensi.s I3067.
cf. p. 270 if.

P. M. L. A. 53 (1938), 645- 655.

(2)· F. J. LAUTH, Dos germanische Rurum-Fudark. 10 if.
ID., Die Runenimchrift des CLM I3067. Wirkendes Wort 1 (1951), 140-X44.
414 4 15
~.
lines had been overlooked. Most of the values have in fact The runes found in London MS. British Museum Royal I2
been shifted, but it is hardly necessary to adopt the involved D XVII are more puzzling., They are part of a charm wi}
evolution proposed by Selmer: l.enctenadle 'against typhoid fever' (I). Various elements have
been incorporated into the charm: the four Evangelists are
Therunes:a--cd--gf-- i jl--no-gx-Z
prayed for intercession, Veronica is mentioned twice, and the
are used for: a - - d e - - h i - - m n 0 - - r s - u x - z. whole ends in a line written with Greek characters. Lines IS. tf.
This looks rather like an attempt to construct an alphabet by in Storms's edition are as follows (I give my own transcription
using every single rune of the fuporc, the equivalents being of the runes) :
added afterwards without reference to the actual values' of the Eft godcund gebed

runes: so z may have become the symbol for b, b that for c,

In nomine dei summi sit benedictum

etc. Of cour!'e only an inscription using all twenty-three

deerej;). N7. PTX derFw N7. PTX.

(or more) characters of the alphabet could give us certainty

about ·the alphabetizing method followed here, and anomalies
.Cockayne read this DEERElJ HAND lJIN DERElJ HAND
must be counted with. Thus the runes g and fmayhave been
lJIN II thine hand vexeth, thine hand vexeth "; Storms is prob­
interchanged accidentally, in which case we have to read Hiesus
ably right in considering this interpretation as rather doubtful.
-instead of Ihesus. The rune j probably stood for k, cf. the
Although the manuscript is not late (saec. X) the runes are so
De in'IJentione alphabet. Selmer is obviously right in supposing
poorly made that the scribe can have had no notion of what
that the scribe had no practical knowledge of the runes.
he copied (2).
Whether these runes reached Hastiere through Irish interme­
The only instances of a systematic use of runes we have still
diaries (Waulsort was a settlement of Scotti). as Selmer holds,
I dare not assert. At any rate I do not believe that they were to examine are those in the Exeter Book riddles (cf. p. 394 1£.),
responsible for bringing the Vorlage to the Continent: such and in the Poetical Dialogues of Solomon and Saturn. Riddle 19
alphabets w~re no doubt known on the Continent long before (fol. 105 t ) spells out the words which form the answer, but they
the dedication of Hastiere (Io35). Similar shiftings of values are written in runes and, to complicate matters, these ~nes
,must be read, backwards :
are found in Arsenal MS. 1169 and in the Mandeville group.
and the latter at least has some relations with the Liege region. Ic on sipe seah . s ro
. On the analogy V\'ith these alphabets, I would rather conclude h. hygewloncne, heafodbeorhtne,
that the shift of the values in the Munich inscription arose swiftne ofer srelwong swipe prregan.
accidentally. I"
Hrefde him on hrycge hildepryJ>e
Before proceeding to other inscriptions we must consider for I .n 0 m. nregledne rad
a moment the change of sutus which the runes underwent in .a g e w. Widlast ferede
the Munich inscription. It is true thafthey are still marked rynestrong on rade rofne. C 0
B = harbarice; but on the other hand their use .as pseudo­ f 0 a h. For wres py beorhtre.
Hebrew reminds us of the • Arabic' and • Syriac' in Munich
MS. 14436, and it anticipates the • Saracen' alphabet in
(1) G. STORI\fS, AngJo..S_ Magic. The Hague, 1948, 1.70 f. (no. 33).
Mandeville's Travels. In a way they are even more remote (1.) COCKAYNlI'S reading is 'accepted by J. lL G. GRATTAN and C. SINCBR
from the runic model: only two runes have kept their original in thclr work Anglo-S_ Magi{: mulMedicintJ, 34, but without reference to
values. STORMS'S opinion. They render lenctenadl by • tertian fever '.

4 16 ~.
31 41 7
 swy1cra sipfl.'et. Saga hWl.'et ic hatte (I), Lastly, in Riddle 7S (fo1. 1271'), which may be no more than

The answer: hors, mon, wega, haDfoc, i. e. probably • a man on the beginning of a longer riddle, the solution is given in full :

horseback with a hawk on his fist '. The third group of runes Icswiftne geseah on swape feran

is obscure. To solve the difficulty many scholars have supposed .d n u h. (1)

that rad in 1. 5 stood for the rune r, but that does not quite
give a satisfactory reading (wegar = flJiggar • lance' ?). The Thus far all scholars have read the third rune as 1. Actually

answer to Riddle 24 (fo1. 106 v ) may easily be found in the same the lateral stroke is a broken line, and the whole can only be

way, the only difficulty being that the scribe substituted a a poorly made u. This gives the reading hund, which most

minuscule x for the runeg : scholars anyway obtained through emendation.

... •x. mee nemnaa, To understand the part played by the runes in these riddles,

swylce . iE . ond .r. o. fullestea, we have only to compare them with Riddle 36 (11. 3-7) :

.h. ond .i. Nu ic' haten eom Hl.'efde feowere fet under wombe
swa pa siex stafas sweotule becnap. (ll. 7-10) (2). ond ehtuwe
Reading g for x we obtain the word higorm = higore • magpie' monn. h. w. M. wiif. m.x.l.kf wf. hors. qxxs.
or •jay'. ufon on hrycge;
The runes of Riddle 64 (fol. 1251) are not so easily inter­ hl.'efde tu fipru ond twelf eagan
preted. They are marked with dots as if their names had to ond siexheafdu. Saga hWl.'et hio Wl.'ere (2).
be read, and this is no doubt the only way of reading the, riddle The system used here to conceal the solution consists simply
aloud. But it is hardly possible to solve the riddle by m~ans of notae santti Bonifatii, but "It is evident that thCl scribe of "
of these names; therefore most scholars complete each group the Exeter Book, probably through inexperience with this form
so as to form words fitting into the context: wi = meg, be = of writing, has jumbled his text considerably. In .h.w.M.
beorn, etc. (3); but cf. p. 396. the wwas miswritten for P. and a s.eeond p has been omitted
Ic seah . w. ond .i. ofer wong faran, after M; in .m.x.l.kfw. the w is rrliswritten for r; the f before
beran .h. e.; bl.'em Wl.'eS on sippe hors has been written in the wrong pJace,and should follow
ruebbendes hyht . h. ond . a. hors. We have then the Latin words homo, mulier, and equus.
swylce prypa dl.'el, .J:».. ond.. e. corresponding to the A. S. monn, mij, and hors .. (3). The func­
Gefeah . f. ond.iE. £leah ofer .ea tion of the runes does not' differ from that of these notae.
s. ond .p. sylfes pl.'eS to1ces (4). The first of the Poetical Dialogues of Solomon and Saturn

treats mainly of the power of the Pater Noster (4). In one of

(I) G. P. KRAPP-E. V. K. DOBBlB, The Exeter Book, 189 f., 331 f.

" On a journey I saw a proud st'oh with a shilling head run very swiftly the two manuscripts, Cambridge MS. Corpus Christi College 422

over the plain. On its back it had a brave nom, a nailed road agew (?). (= MS. A), the letters of the prayer are represented by runes.

On a long journey, trsvelling fast on the road, he carried a strong cofoah. This codex was probably written in the late tenth century. It

The journey was very .fine, the course of these. Say what I am called ".
(2.) G. P. KRAPp·E. V. K. DOBBIB, The Exeter Book, 193, '334 f.Y'-·
"They call me I, also at and 1'. 0 helps, and h and i. Now I am named
(x) G. P. I<:JtApp-E. V. K. DOBBIE, The Exeter Book, 2.3.... 371.

as these six runes clearly signify ".

(2.) G. P. KRAPP-E. V. K. DOBBlB, The E:ceter Book, 198.

(3) G. P. KRAPp-E. V. K. DOBBlB, The Exeur Book, 367 f.

(3) G. P. KRAPp-E. V. K. DOBBIE, The &:eter Book, 341.
(4) G. P. KRAPp-E. V. K. DOBBlB, The Exeter Book, 2.30. (4) R. J. MBNNBB, The Poetical Dialogues oj Solomon and Satum, S ft., 80 ff.
418 ~~ 41 9
 Purely ornamental runes are also found in Basle MS. F. III
consists of two originally distinct volumes, the first of which
ISC, a manuscript written in Fulda (saec. VIII) (I) :
contains the dialogues. The runes are found on pp. 4 and 5.
Their function seems to be purely ornamental. They are h : 43 r h.a::c; 43 v Choros; 44r hymnos, hilariusj 45 r hic.
always followed by the corresponding Roman capital, and must a : 49 v Natiuitate; Sol aMEN.
be pronounced as the letters of the Latin alphabet, as appears <e: 43 r Da:Mino; 43 v <eRAtorium.
from this line, where f and m alliterate with vocalic initials t : 52r IN ILLO tempore.
(ef, em).: On fol. 5 I v the rune IJ stands for 0 : Mgses; on fo1. 34 v the
Bonne hiene .f.F. and .m.M. utan ymbdringaa (1. 127). word Noli is written with a st-rune for the first two letters.
The scribe must have interpreted this rune as a ligature N + <e,
These are the runes found in the poem: which we can easily understand if he took the !I-rune for <e
(N + IJ st). For another instance of <e and IJ being mixed
1. 89 : P; 93 : a; 94: t; 95 : e; 98 : r; III : s; lI8: q u;
up, cf. Brussels MS. 93II-9319, p. 71.
123 : 1 C; 127 : f m; 134: j (equated to G};135 : d.
Other brief Tu~a are found in a number of manuscripts :
In two cases we find only the Roman capital (106 : N, 13 8 : H); Cotton MS. Olbo B IO (cf. p. 19.): old w n x fog.
two letters have been skipped altogether (106 : 0, 123 : I).
Cambridge, Corpus Christi College, MS. 4I, p. 436: abc d;
R. J. Menner, the latest editor of the poem, argues that the
p. 448: xii. 7. xxx swi}»or (? s looks rather like an
appearance of the runes "represents the last vestige of an
insular g).
ancient pagan' Germanic tradition, according to which the runes
themselves possessed magic power. The native superstition Cambridge, Corpus Christi CoUege, MS. 326, at the top of the
of the pag.an English has been pressed into the service of Chris­ last leaf, in mixed nines and notae sancti Bani/atii: ? s t?
tianity, just as has the persistent Oriental. tradition of King 1 f w te m N ? s? r? n s? t. This has been read as -~m
Solomon's power over demons" (I). I rather believe that the uilfraf1!.tlo s(c)ripsit (2) ; I hardly believe this possible,
poet's notion of the runes was very vague-provided we may although at one time the writing may have been less faint
suppose that the runes were inserted by him. The fact that than it is now. On p. 105 there are more notae (ViVe
they must be read as Latin letters (and not by their names) VaLe FeLi[X] CuM CRiSTo aMeN).
.and that they are always followed by the corresponding letters,
is significant. Since they are 'no~ found in MS. B (Cambridge (1) P. LEHMANN, Fv1daer Studien (1), IS, refers to other Fulda manuscripts
with runes : H Nichtige Spielereien sind die alten Einritzungen auf den Fuldaer
MS. Corpus Christi College 41, cf. p. 401) and can actually Umschllgen nicht durchweg. Ich aah in Basel gelegentliche Zeichen, die ich
be. dispensed with, I am rather inclined to interpret them as fliT Runen hielt, hatte leider nicht me Zeit me mir zu kopieren ". I examined
ornamental, with at the most an archaic, pagan or cryptic, the Fulda manuscripts now in Basle University Library, but all I found were
the runes mentioned mFa, and the use of w in the OHG. recipes in Basle
flavour (2). MS. F. III.15a. As far as this last manuscript is concerned, I have some
doubts about LEHMANN's reading of the scratched letters on the cover. On
the front cover I read a1Irb = Aetbertus (rather than aykb = Aykbertus);
. (I) R. J. MBNNEa, The PoetiCal Dialolflles 0/ Sohmum and Saturn, 49. instead of TOY'tuomm: t<Jrl:uosum; and on the back cover SaJ;UT1fIU rather
(2) In his review of Ml!NNBR's edition (Medium lEvum 13 (1944), :8-36),
than atumunus (cf. LEHMANN, o. c., 14).
K. SlSAM also concludes: " Clearly,the runes in MS. A have nothing to do
(2) J. M. KBMBLll, On Anglo-Sax01l Ru7f/!$, 366, read l£diljfsd scripsit;
with the original poem. They were added beside the Roman capitals in one
the reading !pven supra is that proposed by M. R. James (A Descriptive Cata­
branch of the MS. tradition at a time when the heathen associations of the
IOlfllll 0/ Mamncripts in the Library 0/ Carpus Christi College• . Cambridge,
runes were forgotten or harmless, so that they had become mere literary
1909 11., II, 143 11.).
curiosities .. (p. 35).
421
420 ..
St. Gall MS. 876 (d. p. 290 ff.): e? m i (: i k efr n (13 b) be a small fragment of the total which existed at one time or
Imp? I? The values by which I transcribe are those other. As soon as we are able to reconstruct the alphabet used
given in De iwoentione alphabets. The inscription is prob­ in some of the inscriptions examined in section (d), we find
ably based on such an alphabet; at any rate k is a typical evidence of alphabetizations not met thus far.
De inventions type. The whole does !lot seem to make sense. We also find that the usages listed here are as a rule more
St. Gall MS. 9II, p. 290 : below a half effaced ih heb there are intimately connected with their environment than we could
some characters, the second of which looks very much like expect from what we experienced in the first four chapters.
the second r in Ratgar's signature; the next is m or d, We know a few names of men who wrote runes: Madalfrid,
the fourth perhaps s (or some form of q), the last n (or g ?). Alpunc, Cundpato, Ratgar : mostly scribes. Cynewulf belongs
A somewhat similar scribble with even more distorted to a completely different tradition. On the other hand the
forms has been added more to the right. Neither allows isolated items listed at the end of section (d) have hardly any
of a plausible interpretation. value. They offer too little material to teach us anything about
Munich MS. lat. 373I, fo1. Sv: m? t? r (without connexion the runes theInselves; and on account of their isolation they are
with the context ?); fo1. 7v : in the margin by the side of almost without value for the study of runic usage.
a line with a blank space (quia quanta (blank) gramori The third and last point I want to stress is this : only a fuller.
tufmultu cogitationem carnalium praemimur) there is this study.bf the material surveyed in Chapter V can lead to positive
inscription: maiorn. Perhaps the final n is a slip for i, results. Unless more precise data on the frequency of runes in
which would give a suitable maiori. England, more runic inscriptions in Continental manuscripts,
etc., are available, it will be impossible to give a really balanced
Munich MS. lat. 6239, fo1. 13Sf : U 0 f () n i (1).
account of runic writing at large. ;..

There are probably more, perhaps many m.ore such runic

inscriptions; their value, however, is very small, except in those
cases where they can be connected with other items (e. g.
St. Gall MS. 876 and the De inventione tradition), or where
something is k:nown about their background. Thus the runes
of Munich MS. 3731 point to a close contact with Anglo-Saxon
culture, ~hi«h is even more evident from the handwriting (2).
But mostly the ties between such scribbles and their environment
are so loose, that one cannot come to interesting conclusions.

CONCLUSION.

The first impression we gain from this chapter is no doubt

this : although the first four chapters have given us some idea
of the vastness of the material and of its variety, it can only
(I) B. BISCHOFF, Schreibschu.kn I. 72.
(2) B. BISCHOFF, Schreibschulen I, 10. considen this codex a Wiirzhurg
product.

422
42 3
""*
 GENERAL ESTIMATE AND OUTLOOK.

As the reader may have gathered from the general discussion

in the previous chapters, it was not, and could not be, my purpose
to come to a final statement on the TUnica manuscripta. On the
contrary, I eincerely hope that my study may call the attention
of scholars to this somewhat: neglected field, and may serve as a
,I
I
starting point· for further work. I could think of no greater
compensation for my work, than to witness new discoveries.
Yet a few general results may be briefly outlined here.
First, the reader will have noticed the difference in the
treatment of the material in Chapters II and IV as compared
With that in I and Ill. In the former an attempt could be made
to outline the background of the runic material and its history;
in the latter we got a collection of isolated bits, mostly independ::­
ent, and which could not be fitted together into a general
picture. It is tempting, indeed, to reduce this heteroditic mass
of material to a common denominator, to consider it as so many
manifestations of one great movement. But here lurks a danger.
Apart from Walahfrid Strabo, whose part in the transmission
of runica manusmpta I accepted with some reserve, the names
connected with it are not those of the great men of the age,
men we can follow on their journeys through Western Europe
and in their intellectual history. They were more obscure
people, scribes such as Madalfrid, Alpunc, Cundpato, Agam­
bertus, 'Iu~ueus'; one was perhaps a secretary of some im­
portance: Ratgar. All the great names that have been
mentioned before: Hrabanus Maurus, Alcuin, Bede-and why
not Boniface 1- cannot be proved to have taken an active part
in the history of manuscript runes. I shall be the last to main­
tain that no such connexioIl'S may be discovered sooner or later;
an inscription of a few runes may completely alter the picture
I obtained. I would really have enjoyed to come to more
positive results, to be able to amplify and complete those reached
42 5
~.
by other scholars before me; destroying attractive reconstruc­ originals of our fuporcs belong, runic usage (if we may use this

tions is poor sport. But I doubt whether such reconstructions expression for this period) seems to have been fairly uniform all

on too weak foundations are of any use. over England; at any rate there are no indications of local

We have at least learnt something about the level the runes developments. This impression may be due to the lack of

occupied in Mediaeval culture. Here we must make a dis­ material representing all areas; but also to the fact that, as far

tinction between England and the Continent. The material as runic lore was still cultivated, it was in the hands of a restricted

written in England has few ties with its surroundings. Loose class with many interregional contacts. We should of course

leaves or, worse, facsimiles, can hardly provide clues as to the not forget that the runes had acquired an archaic flavour, and

connenon of the runes with their environment. Perhaps this that they were on their way to become mere fossils. Only the

situation is not entirely accidental. In England the runes were few that had found their way in everyday script survived:

originally not in the first place an object of learning, something as a system they sank to the level of« Egyptian " • Gothic',

that belonged to the classroom and the scriptorium. People • Chaldaean' and «Istrian '.
knew them much in the same way as they knew the famous The situation on the Continent was very different. From the
deeds of their legendary heroes. Runes were part of the environment of the runes-flee the tables of contents of the
intellectual pattern which existed in each member of a class various manuscripts, and the place taken by the runes-it
or a community. They were learnt with the help of poems appears that they belonged to the scriptorium. There are some
such as the Rune Poem. There was something natural about indications that they also found their way into the earlier stages
these characters; their use in Christian inscriptions and as of classroom training-the stages represented in our manu­
additional letters is quite significant in this respect. They scripts by extracts from Cassiodore, Isidore and Bede. But it
belonged to the culture of the royal hall; a poet could resort. to is easy to exaggerate their importance. There can have been ;.

them to remind an audience of his name. Their being fairly no question of a systematic teaching of runes. They were
widely known may at least in part explain why they were so probably treated rather like Aethicus Ister's alphabet, as a
seldom written down for themselves. The people who read, curiosum or, at the best, as illustrative material. It can hardly
or who were read to, were almost sure to know them. Only be a coincidence that there are five manuscripts with a version
as time went on, and fewer people were fanuliar with these on runic cryptography, but no single instance of this crypto­
old characters, was there any sense in writing down fuporcs. graphy being used .for inscriptions (apart from those which
Alphabets came still later : they indicate that prospective readers illustrate the treatise). Systems of Latin cryptography may
could no longer be supposed to know anything about the origin­ have been inspired by it, but that is another matter. We must
al runic system. This statement is not contradicted by the not forget that the English runes were imported on the Continent
occurrence of a runic .alphabet in a tenth century manuscript, long after the native runic tradition had died out there, and
as it goes back to a Continental model. The use in Solomon that they belonged to an entirely different level. The new runes
and Saturn, mainly for ornamental purposes, also belongs to were transplanted into the scriptoria and classrooms, but their
a later period. survival must usually have been a question of luck. It is hardly
We have also noticed that, as far as the material goes, the credible that one should have tried to reintroduce them into
dialect of the rune-names is rarely uniform. We often found Germany in the same way as runic writing was reintroduced
a mixture of forms belonging apparently to different dialects, into Denmark during the eighth century. That would have
or to different ages, not unlike the language in which the bulk been to lend support to what traces of paganism might still
of OE. poetry has come down to us. At the time to which the survive. The idea that runes were taught systematically can

426 427
~.
 only arise when one looses sight of the relative unimportance deutschung ", of successive attempts to germanize the DE.
of the TUnica manuscripta. As survivals of old Germanic culture rune-narnes. Not one of the alphabets we examined can be
they are very precious to us, but can hardly have been so to said to have been successfully adapted to OHG. speech forms.
Carolingean scholars. Both in England and on the Continent The attempts to dress the English words in a German garb
. the runes may have played some part in pseudo-science (or in were usually abandoned long before the aim was achieved. It is
what appears as such to us) (I). They occur often in the neigh­ not improbable that a number of adaptations were accidental
bourhood of such matters as the ' sphere of Pythagoras', the rather than conscious. The great number of forms which do
'Egyptian days', i. e. in surroundings which may be charac­ not really seem to agree with OHG. orthography and phonology,
terized as magico-scientific. This magic has not much to do the often unequal degree of adaptation should always be kept
with that practised by the Germanic peoples before their con­ in mind when we talk about' germanization '. We must also
version. It is rather of classical and oriental inspiration, and remember that many OE. rune-names hardly made l'ense for
this may explain why runes became ' Chaldaean " ' Assyrian ' Continental scholars, even when they had grasped the system
and 'Egyptian' characters. of sound-changes by which a word of the one language could
The bookish character of the Continental TUnica manuscripta be translated into the other. After all, some of the English
is also borne out by the proportion between fuporcs and alpha­ names were hardly intelligible to Englishmen themselves.
bets. Only four fuporcs written on the Continent have come For the understanding of the alphabets th~re is still another
down to us, and one is incomplete; but there are over twenty point to be kept in mind. The material from which alphabets
different runic alphabets, i. e. alphabets which go probably back were made, was not uniform. Among the fuporcs that reached
to independent alphabetizations. Fuporcs must· have been the Continent we can distinguish two main types: one with
unwieldy to people who knew nothing of the system ruling this twenty-eight runes (Vienna MS. 795), and one with over thirty.
sequence of characters. Continental scholars could not use These two types may already have been distinct in England, at
them-were it only for brief inscriptions-as long as they were least if we may judge from the fuporc in Cotton MS. Domitian
not presented to them in the familiar order of the Latin alphabet. A 9. The difference between the two lay especially in the group
This explains the great number of independent alphabetizations: of gutturals that had been added in the longer fuporc. For
we got pretty far from von Grienberger's " zwei urredactionen ". J an alphabetizer who had to find characters corresponding to
It is useless to try to reduce the material of Chapters III~V to I Latin c k q, this addition was extremely important.
a series of consecutive attempts at creating a runic alphabet.
Nor need these attempts be traced to the great scholars of the
age. Any leisurely scribe who found a fuporc with the values
lI Most of the runic alphabets must have been very short-lived;
few seem to have been known in more than one place. The
few inscriptions examined in Chapter V duit were long enough
of the runes may have tried to make an alphabet out of it. This to yield a substantial part of an alphabet point to still different
is also proved by the unequal success of these attempts. Of the alphabetizations. Only two runic ilphabets seem to have
fuporcs only one-that in Brussels MS. 93II-9319-has been become fairly popular. One was incorporated into a treatise
consciously adapted to Continental Germanic phonology. This on the alphabet 'written in a region of Germany were Anglo­
finding too is significant. . Much has been made of the " Ein­ Saxon influence was or had been strong. It was given a place
between the fanciful alphabet of Aethicus Ister and various
(I) I fear that e. g. GRAlTAN and SINGER (AnglD-S=on Magic and MeM1IiI)
devices for cryptic and ornarnental writing. This can hardly·
have laid too m~ch stre~ on • pseudo • in this connexion : did contemporaries indicate that the author considered it especially important.
make a distinction between such matters and what we should call ' science' ? Moreover his alphabet is very peculiar. It is made up of
428 42 9
~.
 English material with some addition of Norse; the names of the written in a German speaking at:ea; this development reminds
characters have been adapted more or less successfully for Ger­ one in a rather striking way of Old Irish cryptography.
man readers> The whole has been arranged in a way that To end we may briefly examine to what further use the material
points evidently to the scholar's study; it is utterly remote from collected in this study may be put. In the Introduction I ex­
runic tradition. Yet the brief introduction to the alphabet is pressed the hope that it will be of some help for the study of
one of the very few texts written in the Continental Germanic the English inscriptions. It is true that the gap between
area that tells us something about runes. The author explicitly manuscript and epigraphical runes will be hard to bridge;· hut
states that these characters are foreign : he attributes them to there must be points where the two meet. In a fairly recent
the Normans. He does not even seem to have realized that paper the j-rune was explained as a ligature of g and i (1) ­
the material from which he constructed his alphabet (provided an explanation which we also met in Chapter V (p. 390). In the
he made it himself, but that looks rather likely) was almost discussion of Dickins's system for transcribing OE. runes
entirely Eng1ish. . I proposed j for c]l and j for *.The latter seems to be the
The other alphabet that Wm' destined to be copied for cen­ only form used in English inscriptions, whilst the former is that
turies, and even to survive the introduction of the art of printing, of· the manuscript fuporcs; these add j at the end as if it were
went under an assumed name. In Mediaeval collections of a later addition. As a matter of fact the names of the two runes
alphabets we find a number which are of obscure origin; they
seem to have been invented ad hoc. Isidore had written that
Abraham had invented the characters used by the Chaldaeans
are very similar: gear: ;'ar. This can hardly be a coincidence.
The latter name reminds one immediately of the ON. rune
which must have heen called *jar- till the sixth century, at
*,
and the Assyrians; so a Chaldaeo-Assyrian alphabet had to be which time it lost its initial palatal (2). It is not very likely
found. Perhaps some sort of pseudo-Hebrew or pseudo-Greek that there should he no connexion between a rune *
which was
took i~ place in alphabet collections. At an early date runes called iar in England and one that was called *ja,- in Scan­
too were circulated under such fancy names. One such alphabet, dinavia. But this rules out the possibility that j should be a
probably located in the North East of France and in which the ligature of i and g; the form was developed in Scandinavia
names of Aethicus Ister's letters had been substituted for rune­ from more primitive forms of j. The new form may have
names, assumed the name of ' Saracen' and was given a place supplanted the English j in some regions, after which the two
in Mandeville's Travels. The real identity of the characters t were accepted into an expanded fuporc. The va1ue of both
was to remain unknown until Hickes or Wanley (or 'both)
discovered it. j. was probably indicated by the most common English device for
rendering the palatal spirant: S. This being read as g by
Of all runica manuscripta the isruna group brought us closest
to genuine runic tradition. Yet here too there is an evident
" overlay of sophistication: ", and it looks as if this rather bookish
I Continental scribes, we are hardly surprised to find both runes
used for g and, in High German territory, for k. Read as a
letter of the Latin alphabet, this runic k could become a device
development of the runic system had in its tum influenced
runic practice in regions where one might have supposed there
would be a continuity, as in Iceland. The study of the isruna
I for writing the OHG. prefix ka-. I believe this is a much
simpler way of explaining its use in the Wessobrunn prayer than

group also showed how difficult it is to connect the manuscript

runes with the epigraphic tradition. YVe find similar develop­ (1) L. WHITBRllAD, The Thornhill Inscription. Notes and Queries 1948,
Is6 interprets j as either gi or igj the first name on Thomhill III should be
ments within a few decades on the Rok stone in Sweden, on the read 19i1suilJ.
Hackness stone in England, and in a short treatise obviously (2) O. VON F'RmsBN, R - , 45.

430 43 1
~.
to postulate a use. of a ligature gi for the OE. OS; prefix gi-,
of which no single instance has come down to us.
But further study is also possible within the field of the
TUnica manuscripta. The runic alphabets e. g. may be divided
into two groups according to their first rune : re or a. The BIBLIOGRAPHY
choice of the former implies that the OE. name ;esc had been
altered to asc; that of the latter may indicate that no. such (As a rule only the works of which the full titles are not .given"in

alteration had taken place, and that re consequendy could not the footnotes are listed here).

take the place of a. In the same way we could examine all the
points where alphabetizers had to choose between two or more AGRELL S., Ru11.01"1UU talmystik och dess antiko: flirebild. (Shifter utg.

characters, and from this we might learn a lot about their av Vet.-soc. i Lund 6). Lund 1927.

approach, about their understanding for a foreign but closely _, Zur Frage nach dem Urspmng der Runennamen. (Shifter utg. av

related language, about their being conscious of the peculiaritie!l Vet.-soc. i Lund 10). Lund 1928.

of their own language. We couId even try to find out how _, Die sfJatantike Alphabetmystik unddie Runenreihe. (K. human.

many different fuporcs mul¢ have reached the Continent, and vet.-samf. i Lund, arsberattelse 1931-1932, VI, I-56). Lund 193 2 •

_, Die Herkunft der Runenschrift. (K. hurnan.vet.-samf. i Lund,

what dialects their rune-names represented. But here I must

arsberiittelsel937-1938, IV). Lund 1938.
warn my readers not to expect too much: after the accidental
ALTHEIM F. TRAUTMANN E., Yom Urspmng der Runen. (Deutsches

changes and those made on purpose have been eliminated, we

Ahnenerbe. Reihe B: Fachwissenschaftliche Unt<:rsuchungen.

are left with an abstraction quite remote from the fuporcs

Arbeiten zur Germanenkunde). Frankfurt a.M. (1939)·

written do\VJl in England. One detail will make this clear. ALTHEIM F. - TRAUTMANN - NEHRING E., Kimhern und Runen. •.
As far as our material is concerned, the state of the vocalism Untersuchungen: ZUT Urspnmgsfrage der Runen. 2. erweitel'te
is our main criterion for finding out to what OE. dialect it . Auflage (Beiheftl .zu Germanien). Berlin 1943·
belongs. Diphthongs are very rare in rune-names written on AMBROSIUS T., Introductio in Chaldo.icam lingumn.. Syriacam. atque
the Continent. This can hardly mean that these Continental Armenicam f!:1 decem alias linguas. (pavia) MDXXXIX.
runic alphabets go back to fuporcs with very few diphthongs, ANDERSEN H.• Det yngre Runealfabets Oprindelse. Arkiv 62. (1948),
and if they did, there are always considerations of OE. linguistic 2°3- 22 7.

chronology that must refrain us from rash' conclusions. ANDERSON G. K., The Literature of the Anglo-Saxons. Princeton 1949·

Therefore the reconstructions and localizations which I ventured ARNTz H., Handbuch der Runenkunde. (Sammlung ku.rzeI Gramma­

tilten germanischer Dialdtte. B. Erganzungsreihe Nr. 3)' Halle/

should be understood as possibilities or at the best as probabil­

Saale, 1935. Zweite Aufiage, 'ibid. 19+1­
ities, never as certainties. This leaves open another road for
Das Ogom. Beitriige 59 (1935), 321 -41 3.

further research. But to obtain any results we should first -, Bibliographie der Runenkunde. Leipzig 1937·

make a much closer study of spelling at large than we find _, Wilhelm Grimm und die deutschen Runen. In: Beitriige ZltT

reflected even in the most detailed OE. (and ORG.) grammars. Ru.nenku:nde, 25-34.

I do hope that in this field, too, the present work may be an -, Die Runenschri/t, ihre Geschichte· Itnd ihre DenkmIiler. Halle!

incentive to those who are interested in these studies. Saale 1938.

-, Runen und Runennamen. Anglia 67/68 (1944), 17 2 - 2 5°'.

Eis- und Wasserrunen. Beitrage 67 (1945), 26.}-27J.

ARNTZ H. ZEISS H., Die einheimischen Runendenkmiiler des Festlandes

43 2 32 433
.~.
 (Gesamtausgabe der iilteren Runendenkmiiler I). Leipzig 1939. Viking Society for Northern Research 12 (1941-1942), 177-194,
AsKEBERG.F., Norden och kontinenten i gammal tid. Studier i fomgermansk 209-2 31. .
kultuThistoria, Uppsala 1944. BOELES P. C. J. A., Zu den friesischen Runendenkmiilern. Runen­
BlEKSTED A., Vore yngste Rllneindskrifter. D. St. 1939, 111-13 8. berichte I (1939-42), 116-122.
- , Islands Runeindskrifter. (Bibliotheca Arnamagna:ana II). Keben­ - , Friesland tot de elfde Eeuw. Zijn VOOT- en 'IlToege Geschiedenis.
havn 1942. Tweede drUk. 's Gravenhage 195i.
.- , Mdlruner og troldruner. Runemagiske stud;e,.. (Nationalmuseets BOER R. K., Oudnoorsch Handboek. Haarlem 1920.
Skrifter, Arkreologisk~Historisk Rrekke IV). K"benhavn 1952. - , Oergermaansch Hilntlboek. Haarlem 1924.
: cf. JACOBSEN L. - MOLTKE E. BOSWORTH J. - TOLLER. T. N., An Anglo-Saxon Dictionary. Sup­
BAESECKE G., Der althochdeutsche Abrogans und die Herkunft des plement by T. N. TOLLER. Oxford MDCCCLXXXII-{1921).
deutschen Schrifttums. Halle/S3ale 1930. BRATH E., Zur Deutung derRokn Inschrift. In: S. BUGGE, De,. Runen­
_, Der Vocabularius Sti. Galli: in der angelsiichsischen Mission. Halle! stein vOn Rok, 265-302.
Saale 1933. - , (Jstergotlands runinskrifter (Sveriges runinskrifter II). Stockholm
-, Die Herkunft der Runm. G. R. M. 22 (1934), 4 13-417. 19 11 - 19 15.
_; Vorgeschichte des deutschen Schrifttums. (Vor- und Friihgeschichte - , Ny Rokstenslitteratur. Arkiv 38 (1922), 294-304­
des deutSchen Schrifttums I). Halle/Saale 1940 . BRATH E. - WESSEN E., SOdermanlands runinskrifter. (Sveriges runm­
- , Das Abecedarium Nordmannicum.· Runenberichte 194 1, 76-9°. skrifter III). Stockholm 1924-1936.
-, Die Karlische Renaissqnce und das deutsche Schrifttum. Deutsche BRAUER H., Die BUcher 'Von S. Gallen und das althochdeutsche Schrifttum.
Vierteljahrsschrift fUr Literaturwissenschaft 23 (1949), 143- 216 . (Hennaea XVII). Halle/Saale 1926.
BASIlE E. J. : cf. HEuSINKVELD A. H.
BRAUNE W., Althochdeutsch und Angelsiichsisch. . Beitriige 43 (1918),
BAUGH A. C. et al., A Literary History of England. New York 1948 .
36i -445·
BEESON C. H., Isidorstudien. (Quellen 'und Forschungen zurlateinischen
-, A lthochdeutsche Grammatik. (Sammlung kurzer Grammatiken

Philologie des Mittelalters IV, 2k Munchen 19 13. germanischer Dialekte. A. Hauptreihe Nr. 5). Halle!Saale 1936.

Beitriige zur Runenkunde und nordischen Sprachwissenschaft. Fs. Gotische Grammatik. 12. Auflage bearb. v. K. HELM (Sammlung

G. NECKEL. Leipzig 1938.. kurzer Grammatiken germanischer Dialekte. A. Hauptreihe I).

. BETZ W., Deutsch' und. Lateinisch.· Die Lehnbildungen der althoch­ Halle/Saale 1947.

. deutscken Benediktinerregel. Bonn 1949· BRIX H., Studier i nordisk Runemagi. Runemesterkunsten, upplandske

BEYERLE K. et. al., Die Kultur der Abtei Reichenau. 2 vols. MOO­ Runestene, Rokstenen, nogle nordiske Rllnetekster. Kopenhagen

chen 1925: 1928 .

BIRDER T., .Geschichte der' Germanenforschung. 3· vols.. Leipzig BR0NDUM-NIELSEN J.: cf. VON FRIESEN O. ,.
19%1-192 5. BROWN G. B., The Arts in Early E.ngland. 6 vols. (in vol. 5 philo­

BISCHOFF B., Die sUdostdeutschen Schreibschulen und Bibliotheken in der logical chapters by A. B. WEBSTER). London 1915-1937.

Karolingerzeit. 1. Die. bayrischen Diozesen. (Sa~ung BRUCKNER A., Scriptoria Medii Aevi Helvetica. Denkmiiler schweize­

bibliothekswissenschaftlicher Arbeiten 49). Leipzig 1940. Tischer Schreibkunst des Mittelalters. Genf 1935 If.

-'--, Eine. Sam11ll'lhandschrift Walahfrid Strabos (Cod. SangaU. 878)· BRUNNER K. : cf. SIEVERS E.

Zentralblatt fur Bibliothekswesen. Beiheft 75 = A.us der Welt BUGGE S., Tolkningen af Runeindskriftenpa Rokstenen i (Jstergotland.

des Buclles. Festgabe ... G. LEYH. Leipzig 1950, 30 -48 . Antiqvarisk Tidsskrift 5 (1873-1878), 1-148.
_, Das griechische Element in der abendliindischen Bildung des Mittel­ Om Runeindskrifterne paa Rok-Stenen i (Jstergotland og paa

alters. Byzantinische Zeitschrift 44 (1951) 1/2 = Festschrift Fonnaas-Spand.en fra Rendalen i Norge. (Kgl. Akad. Handlingar

F. DOLGER, 27-55. 31, -Nr. 3). Stockholm 1888 (publ. 1893).

BLOMFIELD J., Runes and the Gothic Alphabet. Saga-Book of th~ -, Bidrog till den ~ldste Skaldedigtnings Histone. Christiania 1894.

435
434 -.
 Der Runenitein von Rok in OstergOtiand (SclmJeden) ... heraus­ Con H. 0., Catawgus codicum mss. qui in collegiis aulisque Oxoniensibus
gegeben '" durch M. OLSEN, unter Mitwirkung und mit Beitragen hodie adseroantur. 2 vola. Oxonii 1852.
von A.. OLRIK und E. BUTE. ·StOckholm 1910. DAUNT M., Old English Sound-Changes Reconsidered in Relation to
BUGGE S. - OLSEN M. et. al., Narges.Indikrifter med de ;eldre Runer. Scribal Tradition and Practice. Transactions of the Philological
4 vols. (Norges Indskrifter indtii Refonnationen. I. Afdeling). Society 1939, 108- 1 37.
Christiania 1891-1924. DEROLEZ R., Problemen in de Runenstudie. Handelingen van bet
BULBRING K. D., Altenelisches Elemento.rbuch. 1. Lautlehre. (Samm­ XVIIIe Vlaamse Filologencongres (1949), 193-197.
lung gennanischer Elementarbiicher. I. Reibe, 4). Heidelberg -, Uit de Geschiedenis van de Runen : 'Dan Agambertus tot Mandeuille.
1902. Handelingen van de Zuidnederlandse Maatschappii voor Taa]­
CABROL F. - LECLERCQ H., DictiomraiTe d'archeowgt.e chrit1.enne et de en Letterkunde en Geschiedenis 5 (195 1), 40-57·
Liturgie. Paris 1907 if. -, Ogam, "Egyptian", "African" and .. Gothic" Alphabets.
CAHEN M., Origine et diveloppement de l'icriture nmique. M.S. L: 23 Scriptoriu~ 5 (1951), 3-20.
(1923), 1-46. -, Richtingen in de Runenkunde. Belgisch Tijdschrift voor Phi1010gie
CAHEN M. - OLSEN M., L','nscription runique du coffret de Martam. en Geschiedenis 30 (1952), 5-49.
(Collection linguistique publ. par la Societe de linguistique de -, Dubthach's Cryptogram. L'Antiquite Classique 21 (1952), 359­
Paris 32). Paris 1930. 375·
CALDER G., Auraicept na n-eces. The Schokws' Primer. Edinburgh DICKINS B., Runic and Heroic Poems of the Old Teutonic Peoples.
19 1 7. Cambridge 1915.
Du CANGE C., Gwssarium media'! et infim;e Latinitatis. Niort 1883 ff. -, A System of Transliteration far Old English Runic Inscriptions.
Catalogue general des manuscrits desbibliotbeques publiques de FrtmCe. Leeds Studies in English I (1932), 15-19.
Dtfpartemmts. Paris 1885 ff. -, English Names and Old English Heathenism. Essays and Studies
A Catalogue of the Manuscripts in the Cottonian Library, Deposited in by Members of the English Association 19 (1934), 148-160.
the British Museum. London 1802. -, Runic Rings and Old English Charms. Archiv 167 (1935), 252.
A Catalogue of the Harleilm Manuscripts in the British Museum. 4 vols. -, The Sandwich Runic Inscription R;eh;ebul. In: Beitriige. ZUT
London 18o?-1812. Runenkunde, 83-85.
Catalogus codicum latitwrum bibliothecae regiae Monacensis. Monachii .: cf. FORBES M. D.
1892 ff. DICKINS B. - Ross A. S. C., The Alnmouth Cross. J. E. G. Ph. 39
CHADWICK H. M., (Revitlw of W. VIETOR, Die narthumbrischen Runen­ (1940), 169-178.
steine). Anzeiger fUr indogennanische Sprach- und Altertums­ DIRINGER D., The Alphabet. London etc. (1947).
kunde 9 (1898), 60-63. . DOBBIE E. V. K. : cf. KRAPP G. P.
CHROUST A., Monumenta palaeographica. DenkmIiler tler Schreibkunst DORNSEIFF F., Das Alphabet in .Mystik un.dMagie. 2. Auflage
des Mittelaltm. Erste Abteilung: .Schrifttafeln inlateinischer (l:TOIXEIA 7). Berlin 1925.
und dmtscher Sprache. Miincben 1902 ff. DROGERElT R., Sachsen und Angelsachsen. Niedersachsisches Jahrbuch
CLARK J. M., The Abbey of St GaU as a Centre of Literature f!:J Art. fUr Landesgeschichte 21 (1949), 1-62,
Cambridge MCMXXVI. EHRlSMANN G., Geschichte tler dmtschen Literatur bis zum Ausgang
CLEASBY R. - VIGFUSSON G., An Icelandic-English Dictionary. des Miltelaltm. 1. Teil: Die althochdeutsche Literatur. 2. Aus­
Oxford MDCCCLXXIV. gabe. MUnchen 1932.
COLLINGWOODW. G., Late and Magic Runes in Cumberland. Trans­ FALK H. : Cf. SHETELIG H.
actions Cumberland and Westmorland Antiquarian Society 19Q6, FEIST S., Zum Ursprung der germ. Runemchrift. A. Ph. S. 4,(1929),
30 7- 12 • 1-2 5.
COSUN P. J., AltfJJestsiichsische Grammatik.1. Hiilfte. Den Haag 1883. Vergleichendes Wiirterbuch der gotischen Sprache mit Einschluss des
43 6
-. • 437
 Krimgotischen und sonstiger ~erstTeuter Oberreste des Gotischen. 3. (Publications of the Wellcome Historical Medical. Museum.
Neubearbeitete und vennehrte Auflage. Leiden 1936-1939. New Series, NO.3). London, etc. 1952.
FOLZ K., Geschichte der Sal:iburger Bibliotheken. Wien 1877. GREIN C. W. M., Sprachschatz der angelsiichsischen Dichter. Unter

FORBES M. D. DICKINS B., The RuthweU and Bewcastle Crosses. Mitwirkung von F. HOLTHAUSEN neu hrsg. von J. J. KOHLER.

M. L. R. 10 (1915), 28-36. (Germanische Bibliothek. I. IV. Reihe 4). Heidelberg 1912 if.

-, The Inscriptions oj the Ruthwell and Bewcastle Crosse. and the VON GRIENBERGER T., Die germanischen runennamen. I. Die gothischen
Bridekirk Font. Burlington Magazine 25, No. 133 (19 14), 24-29. buchstohennamen. Beitrage 21 (1896), 185-224.

FORSTEMANN E., Altdeuuches Namenbuch. 1. Pmonennamen. Zweite -, BeitTiige ZUT Runenlehre. Arkiv 14 (1898), 101-136.

voUig umgearbeitete Aufiage. Bonn 1900. -, Die angeuiichsischen runenreiken und die sag. hrabanischen alphabete.

FRANCK J., Altfrlinkische Grammatik. (Grammatiken der althoch­ Arkiv 15 (1899), 1-40.

deutschen Dialekte II). Gottingen 1909. Neue BeitTiige sur Runenlehre. Z. f. d. Ph. 32 (1900), 289-304.

VON FRIESEN 0., Om runskriftens hiirkomst. Sprakvet. saIlsk. i -, Schriften Uher oos ags. Runenkiistchen. Z. f. d. Ph. 33 (1901),

Uppsala forhandl. 1904. 409-42 1 •

-, Upplands TUnstenar. Uppsala 1913. -, Zu den inschriJten des Clermonter runenkiistchens. Anglia 27 (1904),

-, Runenschrift. In: J. Hoops, Reallexikon IV (19 18- 19 19), 5-5 1. 436-449.

-, Riikstenen, TUnstenen vid RiJks kyrka, Lysing. hiirad, Ostergotland, Drei flJestgermanische RuneninschriJten. Z. f. d. Ph. 41 (1909),

laSt och tydd. Stockholm 1920. 4 19-437.

-, De smast Jranutiillda meningarna i fragan om TU1WT1UlS hiirkomst. -, The Thames Fitting. Z.f. d. Ph. 45 (19 1 3), 47-55.
Arkiv 47 (1931), 80- 133. GRIMM W. C., Ueber deutsche Runen. 2 vols. Gottingen 1821.
-, RiJkstenen. (Svenska fornminnesplatser, Vagledningar utg. genom -, Zur Litteratur der Runen. Nebst Mittheilung TUnischer Alphabetc
K. Vitt. Hist. och Ant. Akademien 23). Stockholm 1934· und gothischer Fragmente aus Handschriften. Wiener JahrbUcher ,
VON FRIESEN O. - OLSEN M. - BR0NDUM-NIELSEN J., RU1WT1Ul_ der Literatur 43 (1828), 1-42 = Kleinere Schriften 3, 85-131.
(Nordisk Kultur VI). Stockholm 1933. HAGEN H., Anecdota Helvetica quae ad grammaticam latinam spectant

GALLBE J. H., Altsaechsische Sprachdenkmaeler. 2 vols. Leiden ex bibliothecis Turicensi Einsidlensi Bernensi collecta ... (Grammatici

IB9+-1B95· . Latini Supplementum). Lipsia:: MDCCCLXX.

-, Vorstudien. zu einnn altniederdeutschen WOrterbuche. Leiden 1903·

liAMMARSTROM M., Om TUnskriftens hiiTkomst. Stud. i nord. filologi
Altsiichsische Grammatik. Zweite vollig umgearbeitete Auflage.
20 (1930), I-fry.
Halle-Leiden 1910.
HARDER H., Zur Frage der hrabanischen Alphabete. Archiv 163 (1932),
GILES J. A., Saneti Aldhelmi ex Abbate Malmesburiensi Episcopi Schire­
187- 204.
burnensis Opera qure extant omnia. (patres Ecc1esia:: AnglicaIllr!).
HASKINS C. H" Studies in the History oj Mediaeval Science. (Harvard
Oxonii MDCCCXLIV.
Historical Studies XXVII). Cambridge Mass. 1924.
GIRVAN R., Angelsaksisch Handboek. Haarlem 193 1.

HATTEMER H., Denkmahle des Mittelalters. St. Gallen 1844 ff.

GOLDASTUS M., Alamannicarum rerum smptores aliquot YJetusti. Fran­
cofurti MDCVI. HAUCK A., Kirchengeschichte Deuuchlands. Vols. I-II. Leipzig 1904 ff.
GORDON R. K., Anglo-Saxon Poetry. London-New York 1934· HAVEl" L., La lettre ui. M. S. L. 6 (1889), 79-82.
GOULD CH. N., Gematria. M. L. N. 45 (193 0 ), 465-468 . HEMPL G., The Old-English runes Jor a and o. M. L. N. II (1896),
GRAFF E. G., Althochdeuucher Sprachsch4tz oder WOrterbuch der 347-352 .
althochdeutschen Sprache. Vollstiindiger alphabetischer Index ... -, Hickes's Additions to the Runic Poem. Modem Philology I (1903­
von H. F. MASSMANN. Berlin 1834-1846. 190 4),135- 141.
GRATTAN J. H. G. - SINGER S. : Anglo-Saxon Magic and Medicine. -, The Runic Inscription on the Isle oj Wight. P. M. L. A. 18 (1902),
IlhutTated Specially from the Sntli-Pagan Text •Lacnunga '. 95-98.

'1-3 8 •. 439
HllNEL H., Studien zum altenglischen Computus. (Beitrage zur engli­ JENTE R., Die mythologischen Alisdrilcke im altenglischen Wortschatz.
schen Philologie 26). Leipzig 1934. (Anglistische Forschungen 56). Heidelberg 1921.
HENNING R., Die deutschen Runentl.enkrn.iikr. Strassburg 188g. JIRICZEK O. L., Der Lautwert des runischen R zur Wikingerzeit. E. St.
lIERMANN H.J., Die deutschen romanischen Handschriften(Beschreibendes 60 (1926), 217-237.
Verzeichnis der illuminierten Handschriften in Oaterreich I, 2). JONES C. W., Bedae Pseudepigrapha. Scientific Writings Falsely
Leipzig 1926: Attributed to Bede. Ithaca N. Y. 1939.
-, Die frii.hmittelalterlichen Handschriften des AbendJo.ndes. (Be­ -, Bedae Opera de temporibus (The Mediaeval Academy of America.
schreibendes Verzeichl1is der illuminierten Handschriften in Publication no. 41). Cambridge Mass. 1943.
Osterreich 8, I). Leipzig 1923. JONSSON F., Runerrze i den norsk-islandske Digtni7!g og Litteratur.
HEUSINKVELD A. H. BASHE E. J., A Bibliographical Guide to Old Aarbl!lger for nordisk Oldkyndighed og Historie 25 (1910), 283 ~3 08.
English. (Univ. of Iowa Humanistic Series, vol. IV, no. 5). KAPTI!:m J. M. N., Zwei Runeninschriften aus der Terp von Westeremden.
Iowa City 1931. . Beitrage 57 (1933), 160-226.
HICKES G., Linguarum Vett. Septentrimtalium Thesaurus Grammatico­ KEARY C. F., A Catalogue of English Coins in the British Museum.

Criticus et Archceologicus. I; II = Antiquce Literaturce Septen­ Anglo-Saxon series I, edited by R. S. POOLE. London 1887.

trionalis Liber Alter seu Humphredi Wanleii Librorum Vett. KELLER W., Angelsiichsische Palaeographie. Die Schrift der Angelsachsen

Septentrinnalium .. , Catalogus Historico-Criticus. Oxonire mit besonderer Rii.cksicht auf die Denkmtiler in der Volkssprache.
MDCCV. 2 vols. (Text, Tafeln). BeIlin 1906.
HOLDER A., Die Reichenauer Handschriften. 1. Die Pergamenthand­ -, Zur Chronologie der ae. Runen. Anglia 62 (1938), 24~32.
schriften. (Die Handschriften der Grossherzoglich Badischen KEMBLE J. M., On Anglo-Saxon Runes. Archaeologia 28 (1840),
Hof- und Landesbibliothek in Karlsruhe V). Leipzig-Berlin 1906. 327-37 2.
HOLTHAUSEN F., Altenglische Kleinigkeiten. z. Altenglische Runennamen. KENNEDY A. G., A Bibliography' of Writings on the English Language
Archiv 99( 1897), 425. from the Beginning of Printing to the End of I922. Cambridge­
-, Altenglisches etymologisches WOr~erbuch (Germanische Bibliothek New Haven 1927.
1. IV. Reihe, VII.). Heidelberg 1934. KENNEY J. F., The Sources of the Early History of Ireland. 1. Ecclesias~
-, Vergleichendes und etymologisches WlJrterbuch des Altwestnordischen, tical. New York 1929. '
Altnorwegisch-isliindischen, einschliesslich der Lehn- f!!:f FremdwlJrter KER, N. R., Medieval Libraries of Great Britain. A List of Surviving
sowie der Eigennamen. Gottingen 1948. Books. (Royal Historical Society Guides and Handbooks 3).
: cf. GREIN C. W. M. London 1941.
Hoops J. et al., Reallexikon der germanischen Altertumskunde. Strass­ KIRCHHOFF A., Das gothische Runenalphabet. 2. Durch ein Vorwort
burg 19II-1919. " tiber die Entstehung der Runenzeichen" vennehrte Aufiage.
JACOBSEN L. - MOLTKE E., Danmarks runeindskrifter. Under med­ 13erlin 1854.
virkning af A. BJEKSTIID og K. M. NIELSEN. 2 vols. (Text, Adal!). KIRCHNER J., Germanistische Handschriftenpraxis. Munchen 1950.
Kabenhavn 1941-1942. KLAEBER F., Beowulf and the Fight at Finnsburg. Third Edition with
JAFFE G., Geschichte der AuJJassungen von'der Herkunft. Bedeutung Supplement. Boston, etc. 1941.
und dem Alter der Runen im I6., I7. und I8. JahThundert. Diss. KLUGE F., Etymologisches WlJrterbuch der deutschen Sprache. II. Auf~
Gottingen 1936. lage. Mit UnteIStiitzung durch W. KRAUSE bearbeitet von
-, G eschichte der Runenforschung. Geistesgeschichtliche Betrachtung A. ·GOTZE. Berlin-Leipzig 1934.
der AuJJassungen im I6.-I8. Jahrhundert. Berlin 1937. KOHLER J. J. : cf. GREIN C. W. M.
JAMES M. R.. The Ancient Libraries of Canterbury and Dover. Cam­ KOHLER T., Die altengiischen Namen in Baedas Historia Ecclesiastica
.bridge 19Q3. . und auf den altnordhumbrischen Miinzen. Diss. Berlin 1908.
JELLINEK. M. H. : cf. STRI!ITBERG W. KRAPP G. P. DOBBIE E. V. K., The Anglo-Saxon Poetic Records.

440 441
'",
 I: The Junius Manuscript. II: The Vercelli Book. III: The LllXER M., Mittelhochdeutsches HandwOrterbuch. 3 vols. Leipzig

Exeter Book. V: The Paris Psalter and the Meters of Boethius. 1872-1878.

VI: The Anglo-Saxon Minor Poems. New York 1931-1942. VON DER LEYEN F., Die germanischen Runenreihen und ihre Namen.

KRAUSE W., Beitrage zur Runenforschung (I)-II. (Schriften der Z. f. Vkde., N. F. 2 (1930), 170-182.

Konigsberger Gelehrten Gesellschaft 1932, "Heft 2; 1934, Heft I). VON LILIENCRON R., Zur Runenlehre. Zwei Abhandlungen von

Halle/Saale 1932-1934. R. VON LILIENCRON & K. M ULLENHOFF. Allgemeine Monats­

-, Runeninschriften im alteren Futhark. (= Schriften der Konigsberger schrift fUr Wissenschaft und Literatur. Halle (1852), 16<;~193;

Gelehrten Gesellschaft. Geisteswissenschaftliche Klasse / 13. 3 10-348.

Jahr / Heft 4)· Halle/Saale 1937. LILJEGREN J. G., Run-Uira. Stockholm: 1832.
- , Was man in Runen Titllte. 2. verbesserte AuHage. Halle 1943. LINDQUIST I., En skinnhandskrift fran Flatey som beskriver Rokstensrunor.

Untersuchungen zu den Runennamen. I. Die Lauch-Rune. In: Minneskrift utgiven av Filologiska Samfundet i Goteborg

II. Runennamen und Gotlerwelt. Nachrichten der Akademie der

(1920), II4-121 = Goteborgs Hagskolas Arsskrift 24, 1920.

Wissenschaften in Gottingen. Philologisch-historische Klasse,

-, Religiosa runtexter. I. Sigtuna-galdern. Runimkriften pa en

1946/ 1947, 60-63; 1948, 93-108.

amulettfunnen i Sigtuna I93I. Ett tydningsforslag. (Skrifter utg.

: cf. PLASSMANN I. O.
av Vet.-soc. i Lund IS). Lund 1932.

LAISTNER M. L. W., Thought and Letters in Western Europe A. D. -, Trolldomsrunorna fran Sigtuna. Fomvlinnen 1936, 29-46.
500-900. London 1931. -, Religiosa runtexter II. Sparliisastenen. Ettsvenskt runmonument

(with the collaboration of H. H. KING), A Hand-List of Bede fran Karl den Stores tid. Ett tydningsfOrslag. (Skrifter utg. av

Manuscripts. Ithaca (N. Y.) 1943. Vet.-soc. i Lund 24). Lund 1940.

VAN LANGENHOVE G. C., Brusseuche Runen 1. De lsitWri Runen. Aca­ LINDQVlST N., En isliindskroartkonstbokfran I$OO-talet. Uppsala 1921,
derie Royale de Belgique, Bulletin de la Classe des Lettres 192 3, LINDROTH H., Bidrag till Rokstensinskriftens tolkning. Stud. i nord.
21 4-2 38 . filologi 2 (1911), Nr. 8. "
Aligemeines zur Runenlehre. In: Beitrage zur Runenkunde, 7- 2 4. LINDSAY W. M., Early Irish Minuscule Script. (St. Andrews University
LXNGIN T., Altalemannische Sprachquellen aus der &ichenau, in: Publications VI). Oxford 1910. 'I
"

K. BEYERLE et al., Die Kultur, etc., 699 ff. -, l sidcri Hispalensi Episcopi Etymologiarum sive Originum Libri XX.

LAUTH F. j., Das germanische Runen-Fudark, aus den Quelkn kritisch 2 vols. (Scriptorum Classicorum Bibliotheca Oxoniensis).

"erschlossen und nebst einigen Denkmalern zum ersten Male erkllirt. Omnii (19II).

Ein sprachwissenschaftlicher Beitrag zur iiltesten Cultur-Geschichte UiFFLm K., Die Sankt Galler Schreibschule in der 2. Hiilfte des 8. Jahr­
des europiiischen Central-Volkes. Munchen 1857. kunderts. (palaeographia Latina VI St. Andrews University

LAZIUS W., De gentium aliquot migrationihus, sedibus ... /ibri XI I Publications XXVIII). London, etc. 1929.

(Basilere 1557). -, Die Sankt Galler Schreibschule in der ersten HiilJte des 9. Jahr­

LEFFLER (LAFFLER) L. F., Om Rokstenen. Antiqvarisk tidskrift for hunderts. Neue Heidelberger Jahrbiicher. Neue Folge, 1937,

Sverige 6 (1880), 2. 28-54·

-, Bidrag till tolkningen av Rokstensinskriften. In: Nordiska Stud:ier LoSCH F., Die Berner Runenalphabete. Germania 30 (1885), 287-306.

tiIHign. A. NOREEN, Uppsala (1904), 191-216. -, Zu den Berner Runenalphabeten. Germania 31 (1886), 1I8 ff.

LEHMANN P. et. al., Mittelalterliche Bibliothekskataloge Deutschlands LoTH j., Vocabulaire vieux-breton. Paris 1884.

und der Schweiz, herausgegeben von der Bayerischen Akademie LOWE E. A., Codices Latini Antiquiores. A Palaeographical Guide.

der Wissenschaften. MUnchen 1918 ff. Oxford 1934 fT.

LEVISON W., England and the Continent in the Eighth Century. Oxford LUFf W., Studien zu den iiltesten germanischen 'Alphabeten. Gutersloh

(1949)· 1 898.

442 443
-.
LUICK K. et. aI., Hiftorische Grammatik tier Englischen Sprache. Leipzig Language to the End of I949. Mediaeval Studies 12 (1950),

i921 if. 237-3 24.

MAAssEN F., Geschichte tier Quellen und tier Literatur des catwnischen MUCH R., Die Germania des Tacitus. (Gennanische Bibliothek,

Rechts im Abendlo.nde his %um Ausgang des Mittelalters I. Gratz, I. Abteilung, V. Reihe, 3. Band). Heidelberg 1937.
1871. MiiLLENHOFF K. - SCHERER W., Denkmlikr deutscher Poesie und Prosa

MACALISTER R. A. S., The Secret Languages of Ireland. Cambridge aus dem VIII.-XII. Jahrhu.ndert. Dritte Ausgabe von E. Stein­

1937· meyer. 2 vols. Berlin 1&j2.

MANGEART J., Catalogue descriptif et raisonne des manuscrits de la MULLENHOFF K. : cf. VON LILlENCRON R.

BibliothCque de Valencienfles. Paris-Valenciennes 186o. MYNORS R. A. B., Cassiodori Senatoris Institutiones. Oxford 1937.

MANmus M., G"schichte tier lateinischen Literatur des Mittelalters. NAPIER A. S., Old English Glosses. Chiefly Unpublished. Oxford 19°0.

(Handbuch der classischen Altertumswisseuschaft). MUnchen -, The Franhs Casket. In: An English Miscellany Presented to

191I if. DR. FURNIVALL, Oxford (1901), 362-381.

MARSTRANDER C. J. S., Om rUn/me og nmenaunenes oprindelse. N. T. S. NIELSEN K. M. : cr. JACOBSEN L. - MOLTKB E.

1 (1928), 85-188. NORDEN A., Magiska runinskrifter. Arkiv 53 (1937), 147-189.

-, De gotiske runeminnesmerker. N. T. S. 3 (1929), 25-157.

MAssMANN H. F., Neue Runen. Anzeiger fur Kunde des deutschen
NOREEN A.: Altnordische Grammatik j : Altisliindische und altnorwegische

Grammatik. Vierte, vollstiindig umgearbeitete Auflage (Samm­

Mittelalters I (1832), col. 27-32. lung kurzer Grammatiken gennanischer Dialekte IV). Halle!.
-, Runen aus Rom und Wien. Germania 16 (1871), 253-258. Saale 1923.
: cf. GRAFF E. G. OLSBN, M., Om troldruner. Edda 5 (1916) 2, 225-245.
MAURER K., Ober Rrmenhandschriften. Gennania 23 (1878),1°4-109' -, Til Rok-indskrifter. Arkiv 37 (1921), 201-232.
MENNER R. J., The Poetical Dialogues 0/
Solomon and Saturn. (The Kingigtdrsoak-stenen og sproget i de griinlandske runeinnskr~fter. ,
Modem Language Association of America. Monograph Series. N. T. S. 5 (193 2), IB9-257.

XIII). New York-London MDCCCCXLI. -, Rrlnm' er ristu rynastir menno N. T. S: 5 (1932), 167-188.

MERONEY H., Early Irish Letter-Names. Speculum 24 (1949), 19-43. -, Sigtuna-amuktten. Nogen tolkningsbidrag. Avhandlinger utgitt

MICHELI G., L'enluminure du haut moyen f1.ge et ks influences irla7Ulaises. av Det NQrske Videnskaps-Akademi i Oslo. II. Hist.-Filos.

Bruxelles 1939. Klasse (1940) 3, 1-48.

MICHELS V. : cf. STREITBERG W. et. ai., Norges innskrifter med de yngre runer. I. (Norges indskrifter

MIGNE J. P., Patrologia Latina. Parisiis, 1844 if. indtil Refonnationen. Md. 2. 35)~ Oslo 1941.
MOLTKB E. : cf. JACOBSEN L. -, Grimhilds og Gudruns nmei:nnskrifter. Avhandlinger utgitt av

MONE F. J., Quellen und Forschungen %Ur Geschichte der teutschen Det Norske Videnskaps-Akademi i Oslo. II. Hist.-Filos. Klasse

Literatur I!J Sprache. I. Aachen-Leipzig 1830. (1943) I, I-5 6•

DE MONTFAUCON B., Palmographia grax.a, sive de orm et progressu : cf. BUGGE S.

litterarum grmcarum et de vanis omnium smculorum scriptWnis grmcm : cf. CAllEN M.

generihus. Parisus MDCCVIIl. : cf.VON. FlUESEN O.

MossE F., Manuel de la langue gotique (Bibliotheque de philologie PAUBS A. C., The Name of the letter 3. M. L. R. 6 (19 I l), 441-454.

gennanique II). Paris MCMXLII. -, Runes and Manuscripts. The Cambridge History of English

-, Manuel de L'anglais du Moyen Age des origines au XlVe ~ck Literature I (1949)0 7-18.

1. Vieil Angw (Bibliotheque dephilologie germanique VIII). PAUL H., Geschichte tier germanischen Philologie. In: Grundrisz der

2 Vols. Paris MCMXLV. germanischen Philologie hg. von H. PAUL, Zweite ... Auflage. I.

- , Bihliographia Gotica. A Bibliography of Writings on the Gothic Strassburg 1901, 9 if.

444
~,
#s
PEDERSEN H., L'origine des runes. Mem. Soc. Roy. Ant. Nord SIEVERS E., Runen und Runenituchtiften2 • In: Grundrisz der germa­
1920-1924, 88- 136. nischen Philologie hg. von H. PAUL, Zweite ... Auflage. I. Strass­
PERTZ G. H., ItaliiiniscJu Reise oom Nooemher I8aI his August I823. burg 1901, 248-262.
Archiv der GeselIschaft rur iiltere deutsche Geschichtskunde 5 SIEVERS E. - BRUNNER K., Altenglische Grammatik nach der angel­
(1824), 462 f., 5I1. sitchsischen Gramrno.tik. Zweite revidierte Auflage der Neube­
PETERSEN C. S., Den datuke Litteratur fra Folkevandringstiden indtill arbeitung (Sammlung kW"zer Grammatiken gennanischer Dialekte.
Holberg. Kebenhavn MCMXXIX. A. Hauptreihe. 3). Haile/Saale 1951.
PHILIPPSON E. A., Germanisches Heidentum bei den Angelsachsen. SIEVERS E. ; cf. STEINMEYER E.
(Kainer Anglistische Arbeiten IV). Leipzig 1929. SILVESTRE J. B., PaUographie universelle. 4 vols. Paris 1839-1841.
PLASSMANN' I. O. - KRAuSE W., Die Hrabanische Runenreike. Ger­ SINGER C. : d. GRATTAN J. H. G.
manien 1943, 17 1- 188• SISAM K., Cynewu.lf and His Poetry. Proceedings of the British
POOLE R. S. : cf. KEARY C. F. Academy 17 (193 2), 303-331.
PROKOSCH E., A Comparative Germanic Grammar. (William Dwight SMITH T., Catalogus Librorum manuscriptorum Bibliotheca! Cottoniana!.
Whi1ney Linguistic Series). Philadelphia 1939. Oxonii MDCXCVI.
RAUCQ E., Die Rrmen des Briisseler Codex No 9565-9566. Mede­ SPEHR H., Der Ursprung tIer isliindischen Schrift und ihre Weiterbildung
deelingen van de Koninklijke Vlaam.'IC~e Academie voor Weten­ his ZUT Mitte des I3. JllS. Halle 1929.
schappen, Letteren en Schoone Kunsten van Belgie. Klasse der STAMMLER W. et al., Deutsche Philologie im Aufriss. 1. (Berlin-Bielefeld
Letteren, Jaargang III, NO.4. Brussel 1941. 1952 ).
VON RAUMER R., Geschichte der germmiischen Philologie vorzugsweise in STBPFENS F., Lateinische Paliiographie. Zweite Auflage. Freiburg 1929.
Deutschland. (Geschichte der Wissenschaften in Deutschland. STEINMEYER E. SIEVERS E., Althochdeutsche Glossen. 5 vols.Berlin
Neuere Zeit. Neunter Band). MUnchen 1870. 1879-1922. ,
REICHARDT K., Rrmenkunde. Jena 1936. STEINMEYER E., Die kleineren althochdeutschen Sprachdenkmiiler.
"

ROGER M., L'enseignement des lettres classiques d'Ausone a Alcuin. Berlin 1916. .
Introduction a l'histoire des leoles carolingiennes. Paris 1905. -: cf. MULLENHOFF K. - SCHERER W.
ROSENFELD H., Buch, Schnft und lateinische Sprackkenntnis bei den STENTON F. M., Anglo-Saxon England. Second edition. (The Oxford
Germanen vor der christlichen Mission. Rheinisches Museum ftir History of England II). Oxford 1950.
Philologie. Neue Folge, XCV (1952), 193-209. STEPHENS G., Handbook of the Old-Northern Runic Monuments of
Ross A. S. C. : cf. DICKINS B. Scandinavia and England, now First Collected and Deciphered.
SALIN B., Die altgermanische Thierornamentik. ""Obersetzt von J. London-Copenhagen 1884.
MEsTORF. Stockholm 1904. -, The Old-Northern Runic Monuments of Scandi'lUl'lJio. and England,
SCHATZ J., Altbairiscke Gramrno.tik. (Grammatiken der althoch­ now First Collected and Dec,iphered. 4 vols. Cheapingbaven
deutschen Dialekte I). Gattingen 1907. 1866 - Lund 1901.
SCHERER W. : cf. K." MULLENHOFP. STOKES W., The Anglosaxon Prose and Glosses in Rome. Beitrage rur
SCHERRER G., Verzeichniss der Handschriften der Stiftsbibliothek von Kunde der indogermanischen Sprachen 17 (189i), 144 ff.
St. Gallen. Halle 1875. STORMS G., Anglo-Saxon Magic. The Hague 1948.
ScHLtnTER O. B., Aldhelm's runic alphabet and that of the Codex Regius. STORNA}OLO C., Codices Urbinates Latini Bibliothecae Vaticanae.
J. E. G. Ph. 2 (1898), 29 f. 3 vols. Romae 1902~1921.
ScHWARZ E., Deutsche und gerrno.nische Philologie. Heidelberg 1951. STREITBERG W., Die gotische Bibel.· 2 vols. (Germanische Bibliothek.
SHIP D. A., Norsk sprdkhistorie til omkring I370. Oslo 1931. Zweite Abteilung, Dritter Band). Heidelberg 1919-1928.
SHETELIG H. FALK H., ScandinavUm Archaeology. Translated by -, Gotisches Elementarbuch. 5. und 6. Auflage. (Germanische
E. V. GOtIDON. Oxford 1937. Bibliothek. I. Reihe, 2. Band). Heidelberg 1920.
44 6
•. 447

 STREITBERG W. - MICHELS V. JELLINEK M. H., Geschi;hte der und Glossar. (Niederdeutsche Sprachdenkmiiler VI). Norden­
indogermanisclum Sprachwissenschaft. II. Die Erforschung der Leipzig I~.
indngermo:nischen Spraclum. 2. Germanisch. (Grundrisz der indo­ - , The Clermont Runic Casket. (Skrifter utg. af K. Hum. Vet.-Samf.
gennanischen Sprach- und Altertumskunde II. 2). Berlin­ i Uppsala VI, Nr 7). Uppsala 1900.
Leipzig 1927 ff. - , Zum Clermonter Runenkiistclum. Z. f. d. Ph. 34 (1902), 127.
STROH F., Handhuch der germanisclum Philologie. Berlin 1952. WANLEY H.: d. G. HICKES.
STROM H., Old English Personal Names in Bede's History. (Lund
WBBSTER A. B. : cf. BROWN G. B.
Studies in English VIII). Lund-London-Copenhagen (1939).
WESSEN E. : cf. BRATE E.
Suss W., Ueber antike Geheimschreibemethoden und ihr Nachlebtn.
WILMART A., Codices Reginenses latini. 2 vols. Romae 1937-1945.
Philologus 78 (1923), 1'1-2-175. WIMMERL. F. A., Die Runenschrift. Vom Verfasser umgearbeitete
Sveriges runinskrifter utg. af KungL Vitterh. Hist. och Antiq;'it. und vermehrte Ausgabe. Aus dem Danischen iibersetzt von
Akademien. Stockholm 1900 ff. F. Holthausen. Berlin 1887.
SWEET H., The Oldest English Texts. (Early English Text Society, WORM 0., [RUNER] seu Danica Literatura Antiquissima, vulgo Gothica
Original Series 83). London MDCCCLXXXV. dicta. Hafnire MDCXXXVI.
Tabulae Codicvm Manv Scriptorvm praeter graecos et orientales in WRENN C. L., Late Old English Rune-Names. Medium lEvum I
Bibliotheca Palatina Vindohunen.si asservatorvm. Edidit Academia (193 2), 24-34·
Caesarea Vindobonensis. Vindobonae MDCCCLVIII . ff. WRIGHT E. C., A Postscript to •Late Old English Rune-Names' (M. lEo
THORNDIKE L., A History of Magic and Experimental Science During I 24-34). Medium lEvum 5 (193 6), 149-151.
the First Thirteen Centuries of Our Era. 2 vols. London 1923. - , Robert Talbot and Domitian A IX. Medium lEvum 6 (1937),
THORSEN P. G., am Rllnernes Brug til Skrijt udenfor det Monumentale. 170 f.
Felger med Udgaven af Codex Runicus 80m Tillreg. Kjeben­ WULCKER R. P., Grundriss ftUr Geschichte der angelsiichsischen Literatur.
ham 1877. Mit einer Obersicht der angelslichsischen Sprachwissenschaft. ", .

THURNEYSEN R., Zum Ogom. Beitriige 61 (1937), 188-208. Leipzig 1885.

- , A Grammar of Old Irish. Translated by D. A. BINCHY and WUTI'KE H., Die Kosmographie des Istner Aithikos im lateinischen
O. BERGIN. Dublin 1946. Auszuge des Hieronymus. Aus einer Leipziger Handschrift zum
TOLLER T. N. : cf. BOSWORTH J. erstenmale besonders herausgegeben. Zweite vermehrte Ausgabe.
TRAUBE L., aRoma Nohilis. Philologische Untersuchungen aus dem Leipzig 1854.
Mittelalter; Abhandlungen der Philosophisch-philologischen Classe WySS U., Libellus valde doctus, elegans, fsf utilis, multa fsf uaria scriben­
der Koniglich Bayerischen Akademie der Wissenschaften 19 diJ:rum titerarum genera complectens. Facsimile edition by
(1892), 299 ff. H. KIENZLE. Basel 1927.
TRAUTMANN E. : cf. ALTHEIM F.
ZACHER J., Das gothische Alphabet Vulfilas und das Runenalphabet.
TRAUTMANN-NEHRING E. : cf. ALTHEIM F.
Eine sprachwissenschaftliche Untersuchung. Leipzig 1855.
VIETOR W., Die northumbrischen Runensteine. Beitriige.:rur Textkritik.
ZEISS H. : cf. ARNTZ H.
Grammatik und GlossaT'. Marburg 1895.
- , Das angelstJchsische Runenkiistchen aus Auzon bei Clermont­
Ferrand. Fiinf Tafeln in Lichtdruck mit erkliirendem Text.
Heft I : Tafeln; Heft 2 : Text. Marburg 1901•
VIGFUSSON G.. cf. CLEASBY R.
VAN DE VIJVER A., Les tEu/ores inimtes d'Abbon de Fleury. Revue
Benedictine.47 (1935), 125-169.
WADSTEIN E., Kleinere altsiichsische Sprachdenkm.ii1er mit Anmerkungen.
448 449
~.
 PERIODICALS QUOTED BY ABBREVIATED TITLES INDICES

(Italics refer to the pages with figures and to the plates)

Aarb0ger for nordisk Oldkyndighed og Historie.
A. f. d. A. = Anzeiger fur deutsches Altertum.
A. Ph. S. = Acta Philologica Scandinavica.
A. - MANuSCRIPTS

Archiv fUr das Studium der neueren Sprachen.

Avr:anches 107: 32, 2.j.4, 275 ff.

Arkiv for nordisk filologi.

Bamberg Class. 6 H J IV 1 I : 98 f.

Beitrage zur Geschichte der deutschen Sprache und Literatur.

Msc. hist. nat. 7; lviii.

D. St. Danske Studier. Msc. patr. 130/2 B IV 29: 275 f., 277 f., 284, 326 ff., 345 ff.,

E. St. Englische Studien. 359 ff. 3 60 •

Basle FIll 15a (. Basler Rezepte t) : 389. 421.

G. R. M. = Germanisch-romanische Monatsschrift.
FIll 15c: 27 I f., 42I.

J. E. G. Ph. The Journal of English and Germanic Philology. Berne 118: 96, 407.

Kgl. humanistiska vetenskapssamfundet i Lund. Arsberattelse. 207: xl. xlviii, 130, 174 ff., 205. 217, 274.279,285,287,371.380 f.• 412,

Kgl. vitter:hets, historie och antiqvitets akademiens handlingar. 414. I78, I82, I88, I89. I9I.
Memoires de la Societe royale des Antiquaires du Nord. 36 3: 155·

762; 275.

M. L. N. = Modern Language Notes. Brussels 931I-9319: xl, xlvi, xlviii, !iv, 2, 3, 60, 63 ff., 189. 205, 236, 245. 266,

M. L. R. = Modern Language Review. 314, 421, 428. 68, [Vb.

,i. ­
M. o. M. = Maal og Minne. 9565"9566; x:.a.iv. I, liii.lv. 3. 89, 93, 95 ff., 103,105. II2 f., II7, I19.

120 ff., 152, 15.4, 160 f., 256. no, V.

M. S. L. = Memoires de la Societe linguistique de Paris. 10420-10425 ; 276. 276.

Neues Archiv der Gesellschaft fur altere deutsche Geschichts­ Cambridge CCC 41 (S. 2) : 8, 401, 420 f.

kunde. CCC 326; 421.

N. T. S. = Norsk tidsskrift for sprogvidenskap. CCC 422: 419 f.

TC RI4.34: Iviii.

P. M. L. A. = Publications of the Modern Language Associa­ Univ. Library Ii. ii. I I : 394.

tion of America.
Chartres 214: 263.

Skrifter utgivna av Vetenskaps-societeten i Lund. Copenhagen AM 28, 80 (Cod. Runicus, Wonnianus) : xxiii, xlii, lvii.

Sprakvetenskapliga sallskapets i Uppsala forhandlingar. Ny KgI. Sm!. 167b: 400:

Studier i nordisk filologi. 1867, 4°: 284.

Dresden Cod. Boemerianus: 155.

T. N. T. L. = Tijdschrift vbor Nederlands(ch)e Taal en Let­

terkunde. Dublin Book of Ballymote: lviii, 148, 203. 275.

Trinity College 60: 272.

Z. f. d. A. = Zeitschrift rur deutsches Altertum. Durham Collectar (A. iv. 19): 401 f.
Z. f. d. Ph. Zeitschrift fur deutsche Philologie. Exeter 3501: 387, 394 ff., 417 ff.

Z. f. vgl. Sprachf. = Zeitschrift fur vergleichende Sprachfor­ 3507: 81, 122, 123 f., 130 f., 212, 219 ff., 224 ff., 229 ff., 245, 7;5 0 , 7;55.

schung. 266 f., 269, 7;71, 403. 405 f. 230, z34, 235.

Florence S. Marco 604; 33. 284, 343 ff., 345 ff., 359 ff.

Z. f. Vkde = Zeitschrift fur Volkskunde.

945: 343·
45 0 4-51

-.
Fulda .A;a 2: 200. Otho B 10: XlOCvii, 2, 8, 16 ff., 4:3, 47, 70, 89, 124, 390,421.

C I I : aoo.
illS. II.
Cod. BoniE. I : 401.
Tiberius D 18 :336.
St. Gall 48: ISS. Titus D 18 : xli, xliv, xlvii, I, Iii, 31, 157 ff., 190, 2 03, 279,

127: 41 I C. 4II, VIIIb.

281, 283, 335 ff., 345, 346 ff., 349 ff., 359 ff.

176: 34, 13S, 162 C. IS8,360.

199: 94, 96.

Vespasian A 18 :Iviii.

270: XlOCix, xliii, I, liii, Iv, 3, 89, 90 ff., 95, 99, 101 C., lOS, II2 C.,
. Vitellius A I2: 8r, 122, 124, 131, 219, 222 ff., 229 ff., 267,

II6 C., 1I9, 120 ff., ISO ff., 160 f., 164, 195,217 ff., 256, 282, 411.
271, 403, 405. 230 , 234, 23S·

IIlO. Vitellius A 15: 399 f.

273: 94,96. Harley 1772: 266, 271, 403 f. 404, 40S.

295: 300.
2399: lvili.

85S: 94, 96.

3017: xl, lIS, 212 ff., 260, 269, 272. 2IS·

876: 284, 288, 290 ff., 346 ff., 349 ff., 377, 422.
Royal 7 C XII: 272.

878: xx:rix, 3, 66, 73 ff., 186, 211 C., 216, 233, 236, 255, 257, 262 f.,
12 D XVII: 417.

271, 290, 365, 376 f. 78, 80.

Stowe 57: lviii.

899: 163, 200, 300.

Munich Clm. 276: lviii, 368.

904: 147, ISS f. 3731: 42 2 •

911 : 422.
6229: 422.

Ghent 92: 286, 382 f.

6250: xli, 407 ff. po.

306: 3, 83 ff., 97, 201. 85.

6272: xli, 407 ft, 413. 4IO, VIIh.

Heidelberg Salem 9.39: xl, 279, 283, 288, 305 ff., 317, 322, 335, 345 ff.,
6291 : xli, 402, 407, 409 f. 4 IO •

359 ff. 360, Vllla.

8lIZ: 198.

13067: xli, xlvii, 414 ff.

Hildebrandslied: cf. Kassel Theal. F 54.

14179: 402..

Karlsruhe Aug. 163: Iviii.

14436 : XlOevili, xli, xlvii, lviii, 14, 32., 41 f., 50, 80 f., 117, 122.,

J76: 268, 284, 322, 323 ff., 346 ff., 359 ff., 414. 360.
124 f., 12.8, 130, 181, 190, 2.03 ff., 211 f., 216, 2.18 f.,
"

254: 284, 320 ff., 346 ff., 359 ff. 360.

2.31, 234, 2.36, 244 f., 250, 2.51 ff., 269,2.74,276 ff., 285,

Rastatt 22: 198. 32.9,342., 368, 373, 4II, 416.254,257, VI.

Kassel Theol. F. 54 (Hildebrandslied) : 388 ff. 14561 : 402. f.

F. 65: 270 fr., 414 f. 1l7I. 14684: 2.75·

Lauderdale MS: cf. Tollemach~ MS. 14725: 275.

Leyden Voss. lat. F 12 8: 13, 60, 115, 122, 130, 180, 192 ff., 217, 237,244,
14737: 200.

266, 269, 272, 411. I94.

19410: xxxix, xlviii, 81, 204 f., 206 ff., 216, 230 f., 233 f., 2.36,

Q 33: 193·
2.45, 2.55, 2.57 f., 260, 2.62., 279, 2.85, 411. 208.

Q 67: ISS.
·22053: 389 f.

Q 83 : xliii, Ivii f., 30, 231, 365.

Hochatift Freising Lit. 3a: 412. f. 4I3.

A. Weinmiiller: I, 283, 31Z, 346 ff., 349 ff., 359 ff. 360.

London Addit. 4783: lvili, 148, 203, 275.

Arund. 393 :389 f.

Niimberg 1966 (Rl. 69Iz): 186, 190, zR4, 308 C., 345 ff., 349 ff., 359 ff. 360 •

Cotton Caligula A IS : lvili.

Oxford Bod!. 572.: lvili, 165, 168.

Domitian A 9: xlv f., lviii, 2, 3 ff., 18, 21 ff., 37 f., 42 f., 67,
Auct. F.4.3Z: 32, 38, 157 ff., 340, 343. IS8.

71, So, 124, 130, ISo, 216, 236, 244, 251, 258,
Eng. tho e. I : 18.

26 3 f., 274 f., 371, 429. 9, I.

Junius Orrmulum: lviii.

Galba A -: 34 C. Junius Psalter: 40 I.

Galba A 2: XlOCVii, lviii, 2, 9, IS, 22, 24, 31 f., 34 fr., 40, 42 f.,
St. John's College 17: Iviii,:iI. II, 13, IS, 22.,24,2.6 ff., 37, 38 ff., 46 f' J

45 ff., 48 ff., 67, 70, 83, 89, 122, 157 ff., 185,258,
48 ff., 61 f., 81, 83, 89, 116, iZ2 C., 130, 149,

264ff., 343, 368. 46, 48, IS8, 1116S.

154, 157 ff., 163, ISo, 196, z05, 2.16, 219, 260,

Galtla A 3: 34 fr.
2.64 ff., 2.69, 2.74, 331, 343, 368. 39, 48, ISS,

Nero C 2: 27.
26s, lll.

45Z 453
~.
Paris grec 375 : xxxviii, lviii. Brunswick casket: So.
~

Mortain: So.
lat. 943 : 219. Chamay: xvii, xxvii, 251. Mos: xiv.
52 39: xxiix, I, Iii, 33; 69, 190, 231, 259, 279, 282 f., 329 fr., 333 fr., St. Cuthbert's coffin: xxi f., II, 41, Ma:lifell: 143 fr., 168. I43.
342, 345 fr., 349 fr., 359 fr. 360. 85,272. Norum: 143, 145. I43.
7520: 175." Dahmsdorf: xiv, 269, 412. Nydam: 139.
9666: Iviii. Dover: xxi f., 60. Rak: 139, 143 fr., 149 fr., 430. I43.
Arsenal II69: 124, 195, 216, 218, 248 fr., 256, 277 f., 405, 416. 249. Fl0ksand: 259. Rotbrunna: 143, 145. I43.
Franks casket: 138, 413. Ruthwell: :xxi C., So, 60, 81, 87.
Phillipps 3715 : xl, 81, 122, 124 f., 204, 209, 219, 227 fr., 267, 285, 405. 230,
Fmihov: xiv. Redven: 143, 145, 168. I43.
, 234, 235.
Grumpan: xvii f., 139. Scanomodu-solidus: 86.
Salzburg a IX 32: xl, xliii, 3, 89, IIO fr., II3 fr., 120 fr., 160 f., 192,217. G0r1ev: 30. Sigtuna: 145, 152.
Schlettstadt I 153bis: 229. Hackness: 140 fr., 149 fr., 153 C., 156, SlSderkoping: 145.
Strasbourg 326: 33, 69, 284, 329 f., 332 fr.; 342, 346 fr., 349 fr., 359 fr. 360, 160 f., 430. I4I. Thames scramasax: xxii, xxvii, I, I I •
VIla. Hantum: 86. 13 C., 41 C., 47, 180, 186, 272. I_
Tollemache MS. Orosius : 400 f. Hartlepool: SO.
Thornhill III : xxi,431.
Ja:11inge (grester stone): xxxii.
Torsbja:rg: xiv.
Trier RIll 13 (61): xliii, 3, 14,89, 101, 102 fr., II2 f., II7 fr., 120 fr., 160 f., Kingigtorsoak: 143 C. I43.
Urswick: xxi.
244, 255 f., 260. I20. Karlin: 140.
Vadstena: xvii C., xxvii, 139.
lUppsala Cod. Argenteus (DG. I): 318. Kowel: xiv.
Vi: xiv.
V.alenciennes 59: 179, 406 fr. Kylver: xvii, xxxviii, 139. Vilsta: 143, 145. I43.
4II: 200. Lorn VII : 143 fr. I43. Westeremden: xxi, xxii (B), 47 (B).
843: 200 f. Lunda: 143 fr. I43. ISo.
'Vatican lat. 266: 202 f., 261, 275. Maeshowe VIII: 143 fr., 149. I43. 0vre Stabu: xiv.
Regin. lat. 294: 284, 313 fr., 345 fr., 349 fr., 359 fr. 360. Maeshowe XVIII: 143 fr., 149, 168.
33 8 : xxxix f., xlvii, 32, 61, 128 f" 196 f., 205, 209, 212, I43.
23 6 , 237 fr., 255, 261, 269, 274, 285. 242, 248.
421 : 164, 302.
Urbin. lat. 290: xliii, 3, 89, 101, 103, 106 fr., lI8 f., 120 fr., 160 f.,
187, 218, 244, 255 f., 261, 279 f., 284, 317 fr., 320,
346 fr., 359 fr. I20, 360.
Vercelli CXVII: 391 fr., 3¢.
Vienna 751: xxxix, 70, 84, 97, 197 fr., 209, 231, 245, 260, 274, 285. 20I.
795 : xxvii, xxxix, xlii, Ii, 2, 22, 52 fr., 72 f., 86, II6, 128 f., 180, 18 5,
20 5, 245, 280, 318, 429.59, IVa.
1010: 261, 284, 309 fr., 346 fr., 349 fr., 359 fr. 360.
1224: 403.
1609 : xxxix, 163, 186, 279, 283, 295 fr., 303, 346 fr., 349 fr., 359 fr. 360.
1761 : xxxix, 163, 186,266,267 fr., 279, 283, 286, 288, 297, 299 fr., 307,
322, 325, 346 fr., 349 fr., 359 fr., 412. 268, 360.
2732: 300.
Wolfenbiittel Cod. "Carolinus (4148) : 58, 318.

"B. - INSCRIPTIONS

Arguel: 251. Bmcteates nos. 21, 22, 61 : 140.

Arum: 140. Breza: xvii.
Bewcastle: 50,183. Bridekirk: 45.
Bingley: So. Britsum: 269, 403, 412.

454 THE ST-CATHERINE PRFSS LTD, BRUGES, BELGIUM.

~.
rtl~~ ~ < xt Hi- J ~1~'Y~ t~ MMt<>D4~
fupo.rkgw hnijepzs tbemltjdo
(a) Germanic Fu}>ark (cf. p. xvii if.).

Danish rl'l~ ~ ~r
Swed" Norw. rl~ ~ ~r
: *
,I t
i l-t '1 : tB 'Yt!
~I ~ ·1 ttl
I ~
f u P~ r k h ni as ~ bm 1 R
(b) Norse FUl?l!rks (cE. p. xix f.).

1 t 3 4 5 (, 7 8 , 10 1111 1314151' 1718 17 10 2.122 23 14­

rtl~~khXJ> Mi-l +J'~Ta, t~ M"t~M~
fuporc.9w hnij3pxs tbemll)d~

r~ tiT~*. ~
a z yeak 9 k J s~
*
25 U 17 18 2' 30 31 U 33 1ta 1&.. fOb 24...

+11 r t t2 B8
33C1.

(c) English Ful?orc (cf. p. xx if.).

 I', \ II' III
PLA'I E II

I'" '1fi !tl re u~ F""h E COlA

Jf\.~tll 1111 1: byp J!
~ -J- ~~ F r ~
b B btq'/c by f~ U hUJ' by~ea .'t"'~"'"
~ rrr-..-L...
~r ~ "'.~
T.~:.\',~,
..... ,. .
',1.1',,1' ~.~!l<: ..~"
cl. . til ~ l-,'>$ol'tl byp '6 ~ ...... n... n~". ~. .,'1:..
....J='.. ,Ge t,.. ·, l' ~, t - S rr 'l'I. ..... '

e:- M
1, ' " ...
by; f r mal ',n . . .1
·t·
'1l';'(' .~'..('
"~!l:'~'
~'JoI:.'),
of:. ~ - ~
tV .H'I,.
'J-J' ~ •...)~
b' ~e;s
r'O' t.. _:
hyp
J' ;.

d )ost';d.an 0 R'I' c ,,~, ,0)$,'1,';;:'.-1'-

+:,. ,.:
I'f I, Ii
IJ!, t'! ,,'" :'l.! .} f:
li. H.,.......

nl J} Yy ( I" eop _1\,';'

·r
L,,~
... lit"",,· :J.... q:.". 'l' ,,v. .r. 1_ ..~
f\. (el ' fO] (t R tw
) by; c : i: .•I.~ O:>r'·· \1' ! . f..": ~. .e'i,... f,. ,
L ~ -~ byp 11 , ufa • -1-" !
;11''>~
1-. ..,.... .;:
II~
f • ",
r,t
11 r ~
.", ':" ,: ......
,(: 1;. ~.l_
X' VlU! rre C h tClt
byp c
!
• .,..'
.Is ' f 1_.
Sr~. ~~~,. '1; . .~. H
i... '1... i t ~
<tl_ h,~"
~ 1.-...
~fU
t' <.1
X i'i
1 P, ,"
WI prer a hi . 5 ..•,' ~~~,~ <i'.~\., ~, ~!. ~ '.,t..,;.
stt~C~ 3")(.. ume ~ 'of'· i I!~ h . , ,· ¥ r N~ * .&,", ,',),
~"L.J..
ot
~
11 by;;Sel( ,ttl I' PP
b
ant
ne bPb
,&
,,:,.
!
'1','
l'"
.l.~_.".
'wJirJ:I:,...t. ,h ..· ' 6. \1' t i\.
,II<'.......R., j ~.r~ -.It
u· to
.1'" ~N..I.
\1.1-:";'
0luJ<

'b<f!; on tL U
"l
<::; __~'l1:".t~~·;t~ -nd
)11. 'Af N
!' 'f; - : ~
IY'f
tlf'.:4);.r..u.;.
v""""'U -
(\
V:c
~
yto
b'

byp 0
l
h H; H"~L byp h
.
1

,pa.
.>-g~ e~ Ii M..,.
" ;... , I:p 0
'i. L:Z ~! :D:,W",
...; .• i\ '. r
....

I, 1".,..
·t. ~oI"
'YAI: ..J:.. , ,.\
,H. ,~.
III I· ~.
A. d­
'#
F. \:'
',,\\
1 tl
...
'·.of Il:iiIIMII"'H':~

;~

rc('ul..,Ot .'
., • .y,.:A '{~~J,!..

~ A'fc i Ity~ byt~

.... )j<
bre] n of ,'.'. t<. ,). "f ,~'-'~ 0/" ~. ~.~"'(f.':
f T I'll 1. io... . 4: ~ •..::, 1.f . ~:,-:~,

\.le" ~. blp of' 1 I Ir byp ( \; 1~' ,:·ik.d.1.1t

il ~ 1:, U'!.. HI.J
: If{'' " "I
!( a. 1lI-1'
. ..!.."

t rl.-'­ rn.
Y 1"\ YJt reo cUT .,f. It> '~. f.~~
.. :.4: ihl n "J;'.J'~

m.o byp re )'$ ~!o/t byp ; : ~ b All!"- ~li'il.< ~ ... .,.

~ • q ~~ 14 <twfC f ' ~.. IYI.
:".l'.,l ..~
'tli.;'';'"'' I.'

l 0- ~ tar- he ( ]' k beo t.~"tV~J.>;

~ lJ •••! 1'- ~.. .,~ ,. 'H...l.: ';;:i .J'l;H.... \~.

~'l~~'
,'r byp e. tp t- -rtf) by; \ t·~ u" n..
'.;, "ll I\.~
JX-8lI(."t
~f N. ""I _
A" ~,,:
p~ ........... ~.
..;, 'l' ~
/.

" .;'~ /·r~./.!I''''I

eo,llrVfVC'or pret oJ ...J.... un: .", .•• ~S;I.S~I-'';''.",
1\ l><C<'T
~l:"
c eq?r L byp e~ p TF"l'1 byp ~ .;t,'~.,~:.:IJo!1(··~·~rI
:,:,'f ,1'_.
..' .. - ' .
r \tJ' ceol H bh :'(tX~_
'i I X 'ft".>I!J..y reccal I"'~'t
.rk;~
...
-~~'

.' .

1
m
)tall [
fl: 't
'91.\J
I' r "1f1~
.. ~'fI'
re
bpc
m
1111
.. ",~£, j:!'2.-.J"_
:.~

.~ t~ "r'W'h(riri<~~'l "'.' r..., ~ tho"'!'Jr

fY. '.>C" ... "e:. i" '

• pa:t'pc bcpoppcn. ~a:Jl he pynnum leop]>:.

by pft;lc eopla t;chpy leun.~onn frertlJce fla:r c ?nt;mnc p, hpapcoltan hpura Oxford,:::it. John\ Coll.-g<'. :'lIS. Ii, 1'. . 1. s' (p. JO ,1'.,
ceoran blac 'Co l)ebebban bleb;) l)rbpCOrap. pymu ljfp1'l:;ap pepa iScrptcap

ILn M
~'tkl
1/(15 /lU7 ra aCreJ r
Ji" N i 't' P J:I \I addHr;z fe ftinan r
r

1'1'- 3t' '

.rI1I.!l,},:1 1..!lcJr/ Iflterprdarzda' reft'llqUO.

L I1 Er.a,
(·.,lIon ,IS. (lIb .. II 10 (fwllI (;. IIi<-kt". 1'111'.<(//11'11.,. (;n//Ill/lflli,'(/ ,·lllld,,·
\':",·".,;1'" " t· ,.( 1"'1 '0 fT \
 - 4'5 .~ ·~q-f$

, ~~-.,t:-;s ~r7i )<'4 7""~' 'rT"I'7'".Mr:;><~~:

1:', r '!.
l {,1 l i ~a. t ~,t14
J
c ::-: ~~;>-~'-~~ 77, ?Z~:.v~ ~""~l""'?-::Zl :x.,.~•. ~-\'
.. c ~ 'i , ~ til f:. fIt- ~. ~
~

:::..() "C. !. -r ~- ~
-G-,. C SI-s CI- (lit y

. ' .

(h) Bru~,,<·b. Koninklijkt· Bihliotlwck. :\IS. ()JII-<)J"J. foJ. .1' (I'. h:; fL)

o
o
 i:.
..-:­
.,
....;
~~
]
..c.~
:
'2
:'l.
a
. .:
.....
:::

"<:..
S
or
-
...'?
r.
'E,:
~ --:
,
"~.
~t
5:;::-.
~ ~
'T
~ §:
;.
"r . "r.
'"f­ '":I.
:I.
,.­
~l
§=­
C
'~
--~ c
~
s:
'-'
.~
-

-­..
.. :

"':l
w
0
::­
;::;
~t"11
C c::
Z
.....
<
-z<
tTl
t'rJ ~
~
r::n
.....
~
-

(J)

~
><
~
-
t'"'
t::I
~
>
>­~ ~
~