Journal of Archaeological Science (1999) 26, 1089–1092

Article No. jasc.1999.0403, available online at http://www.idealibrary.com on

The Use of Sulphur in Hollow Ancient Gold Objects

H. Kutzke
Mineralogisch-Petrologisches Institut, Universität Bonn, Poppelsdorfer Schloß, D-53115, Bonn, Germany

G. Wagner
Landesamt für Archäologie, Japanisches Palais, D-01097, Dresden, Germany

Archaeologists interpreted residues inside hollow gold rings found in Roman burials of girls in the Rhineland as a
‘‘cement for chasing’’. The aim of this study was to analyse this material and to compare it with Mediaeval and modern
specimens made from pitch with inorganic filler and organic additives. The cement inside the Roman rings was
identified as crystalline sulphur, which has occasionally been reported in the analytical literature as a filler in hollow
gold objects from Hellenistic to Merovingian times. As was shown by practical experiments, the consistency of solidified
sulphur can easily be adjusted to allow embossing and chasing of filled objects. Therefore, sulphur might act not only
as a filler in such items, as was sometimes suggested, but also play a decisive role in the manufacturing of them.
 1999 Academic Press

Keywords: SULPHUR, GOLD OBJECTS, FILLER, CEMENT FOR CHASING.

Introduction of Roman objects and can even be found in some

D
uring all periods gold has been a very expen- so-called golden Hercules clubs of the 3rd and 4th
sive material which people tried to save. The century . Gold objects from the Thetford Treasure
ancient goldsmiths were masters in producing contain sulphur as a filler and a cement for gemstones
hollow objects from thin gold sheet. Because of the (Cowell, La Niece & Meeks, 1983). There is even an
ductile character of gold such items must contain a interesting passage in classical literature. In his In-
filling material to prevent them from deformation terpretation of Dreams, Artemidorus (2nd century 
during use. (trans., 1975) says (Book II, chapter 5), while discuss-
The occurence of sulphur in hollow ancient gold ing dreams of golden finger-rings:
objects has so far not been studied systematically. An ‘‘For those that are hollow and filled with brimstone
on-line search for the words ‘‘sulphur’’ (or ‘‘sulfur’’) signify deceit and treachery, since they contain something
and ‘‘gold’’ occurring together in titles, keywords, and that is hidden. They also signify expectations that exceed
texts of the abstracts of the Conservation Information the actual benefits, since their mass is greater than their
Network (CIN, including, among others, the Art & weight.’’
Archaeology Technical Abstracts, AATA) produced
no relevant hit. By manual browsing through the This entry in an ancient encyclopaedia of dreams and
literature on ancient gold objects only a few passages, their supposed meaning proves that sulphur was at that
which mention sulphur could be found, starting with time commonly known to people as a filler in gold
Henkel’s (1913: 278) book on Roman finger-rings. jewellery.
Deppert-Lippitz (1985: 5) reports sulphur in addition The use of sulphur did not end with the Roman
to gypsum, clay, and wax in the gold jewellery of the period. Merovingian garnet jewellery (Arrhenius,
Römisch-Germanisches Zentralmuseum in Mainz. By 1985), the great gold buckle from a 7th century Anglo-
employing EDX-analysis, virtually pure sulphur was Saxon ship burial (Cowell, La Niece & Meeks, 1983),
found in a Roman armring from Eauze la Gars (Eluère the head of a 4th century Early Alamannic needle
& Raub, 1991). According to Ogden’s (1982) discus- (Raub, 1981), and the 6/7th century lions’ heads termi-
sion of filling materials, sulphur occurs in the majority nals and their connection of a Byzantine torc in the
1089
0305–4403/99/081089+04 $30.00/0  1999 Academic Press
1090 G. Eggert et al.

Figure 1. The pair of gold rings from Bonn, Josefstraße (Haberey, 1961). Scale 1:1.

British Museum (Hockey, 1989) were all filled with which he interpreted as cement for chasing. Using
brimstone. X-ray diffraction, we found plain (orthorhombic) sul-
Although sulphur occurs quite frequently in volcanic phur (JCPDS 8–247) in both cases (Ring i: RLMB
areas and is associated with gypsum deposits, it must Lab. 8905; Ring h: RLMB Lab. 8906). In a micro-
have gone out of use at some time. No further objects test-tube the material could be easily melted and
could be found in the analytical literature. Theo- evaporated without residue condensing at colder parts
philus’s detailed 12th century treatise on goldsmithing of the glass, so there is no hint of the presence of any
and later works do not mention this use of sulphur; other material except sulphur.
modern goldsmiths are seemingly unaware of it. The sulphur sample taken from Ring h had a sooty
It should be noted that because of the reactivity of black material on top. After extraction of the sulphur
molten sulphur, its use in metal objects is restricted to with carbon disulphide, copper (II) sulphide (Covellite,
the noble metal gold (and, theoretically, platinum and JCPDS 6–464) and some copper sulphate pentahydrate
iridium) and its not too debased alloys (Ogden, 1993). (Chalcanthite, JCPDS 11–646) were identified.
Silver and copper alloys would be destroyed by the Chalcanthite is known as an oxidation product in the
exothermic reactions to the corresponding sulphides. oxidation zone of covellite ores. The covellite must
have formed by the reaction of copper out of the gold
alloy with the sulphur. There were even some green
The Material Inside Three Roman particles inside the ring which contain copper (tested
microchemically), but could not be identified by XRD.
Gold Rings The loose, powdery, and sooty appearance of the black
This study started with the analysis of the ma- material resembles that of the copper sulphide occuring
terial inside three gold rings in the Rheinisches sometimes in black spots on copper alloys (Hjelm-
Landesmuseum Bonn (RLMB) which were found in Hansen, 1984; Eggert & Sobottka-Braun, 1999).
the graves of Roman girls in the Rhineland. Two were Ogden’s (1982: 40) observation that ‘‘Sometimes there
excavated in Bonn, Josefstraße, in a girl’s grave of the is a noticeable admixture of charcoal [in the sulphur
second half of the 3rd century (Figure 1). They were filling of gold items, G.E.], but there seems no reason
each made from two thin (0·2 mm) sheets of gold, why this should have been deliberate’’, might very
which were rolled into tubes and soldered. The two well be a misinterpretation of the appearance of such
tubes were then bent into semi-circles which were copper sulphides in one case or another.
soldered together to create the full ring. This soldered Sulphur was also found by XRD (RLMB Lab. 8555)
joint is not visible even under the microscope and could in an armring of the 4th century from Zülpich-Enzen
only be detected by X-ray radiography (RLMB 4193, (Follmann-Schulz, 1989). The sample was totally
B. v. Zelewski). Finally, the torsion grooves had to be soluble in carbon sulphide (no long chain sulphur
chased with punches. molecules present). Other rings of the Rheinisches
The rings were possibly worn in the ear, although the Landesmuseum Bonn could not be sampled, because
catch does not look well-suited for this purpose. There the opening for pouring in the filling material is
is visible wear on one side only. Haberey (1961) covered by gemstones in situ.
noticed, through the holes near the catches, loose Typically, in all three cases much of the now brittle
pieces of a light brown crystalline material inside, sulphur is obviously missing. Only loose, rustling
 The Use of Sulphur in Hollow Ancient Gold Objects 1091

sulphur particles are left, which might easily get lost (or 95·6C). Së is transparent, light yellow and fluid. Above
sampled) through the openings. Further loss of elemen- 159C the viscosity increases rapidly, owing mainly to
tal sulphur may not only be due to chemical reaction the formation of long sulphur chains with 103–106
with alloying metals, which may end in water soluble atoms (Sì). Above the viscosity maximum, at 187C,
products like chalcanthite, but also to bacterial action the average chain length in the dark red brown viscous
forming either SO2 or H2S gas depending on the melt decreases again. If such a melt is quenched in cold
(an)aerobic situation. In the course of the millennia, water, a plastic and rubber elastic modification is
even the low vapour pressure of orthorhombic sulphur formed (‘‘plastic sulphur’’). The transformation of Sì
(2·310
6 hPa at 23C) must be taken into consid- into crystalline Sá is extremely low at room tempera-
eration. There might be cases of now empty gold items ture in pure sulphur, but can be catalysed by impuri-
where the original sulphur has totally disappeared ties. Therefore, one cannot conclude from the absence
today. of Sì (full solubility in carbon disulphide, proven
for the ring from Enzen) in objects today that this
modification was never formed during production.
To check if the right consistency for chasing is
Sulphur—A Cement for Chasing? achievable practically, the forming of the torsion
When chasing was the last step in the manufacture of a grooves of the rings from the Josefstraße by punches
gold object, is it reasonable to assume that the cement was taken as an example. A smooth ring with open
for chasing was then exchanged with sulphur as a filler? ends was manufactured from a gold sheet using
Is sulphur, which can even be produced in a plastic wooden models. Filling in Së-melt yielded a much too
variety, suitable for chasing, as was postulated by brittle crystalline material inside. On the other hand,
Hockey (1989), Ogden (1993), Haberey (1961) and plastic sulphur formed in situ in the ring is much too
Follmann-Schulz (1989); the latter two not knowing soft and elastic for use as a cement for chasing. If the
the chemical nature of the material inside? To answer melt is heated still higher and filled into the ring, the
this question, one has to look for the required proper- sulphur tends to fracture after cooling like cracknel
ties of such a material and then perform experiments. during punching. The best results were obtained when
Theophilus (trans. 1979) gives the following recipe the red brown melt was slowly cooled in the air, until it
for ‘‘The Composition Called Chaser’s Pitch’’ (Book became fluid again. This was filled into the gold ring,
III, chapter 59): ‘‘Grind a piece of brick or tile very which was cooled by water. Because of the volume
small and melt some pitch in an earthenware dish and shrinkage on solidifying, sulphur was poured in as long
add a little wax.’’ As opposed to recipes for other as possible to prevent the formation of cavities in the
materials, he gives no definite amounts for the compo- interior, which would collapse during chasing. If the
nents (‘‘some’’ pitch, a ‘‘little’’ wax). A compositional sulphur stands overnight, it becomes too brittle again
adaptibility is characteristic for such a material. Cellini because the S8-rings, which act as plasticizer in the
(trans., 1967: 15), the 16th century artist and crafts- polymer Sì, recrystallize. In that case the sulphur could
man, used the same recipe and said that the compos- be removed from the ring by reheating: the molten
ition has to vary with the season. Even today chaser’s sulphur gushes out of the ring and residues burn away
pitches are composed similarly and during winter more with a blue flame. Removal of the sulphur was also
oil or fat is used to make the mixture softer, while necessary if the metal had to be heated red-hot and
during the hot summer more inorganic filler is used to quenched to restore its flexibility. On the other hand,
make it more rigid. Resin, turpentine, and tallow are the later hardening of the rubbery sulphur makes this
now in use as a modifier, and grinding bricks can be material better suited as a firm filling material (Ogden,
avoided by the use of gypsum powder as an inorganic 1993).
filler. Every workshop has its own recipe and the When worked directly after filling in the sulphur,
consistency must be adjusted to the material and the good results which replicate the original rings were
form which is to be chased. The chaser’s pitch must not obtained. The preparation method uses easily observ-
be too hard (it must give way during punching) and not able properties of the melt (colour and viscosity)
too soft (otherwise it will result in indistinct indenta- and, therefore, requires no measurements of time and
tions instead of a sharp relief). The chaser can easily temperature and no high degree of skill.
test the right consistency with a hammer after the
mixture has cooled (Brepohl, 1987: 222).
Sulphur melts easily and can then be filled into
hollow items without problem. The consistency of the
Conclusion
solidified melt can vary considerably depending on As in the three Roman girls’ rings from the Rhineland
temperature and time (Hollemann & Wiberg, 1985). At discussed here, sulphur is frequently found in hollow
a temperature marginally above the melting point of ancient gold objects (and only in golden ones!). It can
119·6C, the melt (called Së) consists mainly of eight- easily be melted and used to fill the insides of the
membered rings of S-atoms, as in the solid phase objects to protect the thin walls from damage. As was
(orthorhombic Sá below and monoclinic Sâ above shown by reproduction experiments, sulphur can be
1092 G. Eggert et al.

adjusted by simple means to the right consistency to Deppert-Lippitz, B. (1985). Goldschmuck der Römerzeit im Römisch-
allow the chasing of filled gold objects. Such a function Germanischen Zentralmuseum. Bonn: Habelt, p.5.
Eggert, G. & Sobottka-Braun, U. (1999). Black spots on bronzes and
seems reasonable, because the exchange of another elemental sulphur. In (ICOM Committee for Conservation, Ed.)
chaser’s pitch for sulphur as a filler does not make 12th Triennial Meeting Lyon 29.8–4.9.1999, Preprints. London:
sense for the sake of parsimony. James & James, in press.
Because of the frequent occurence of elemental sul- Eluère, C. & Raub, C. J. (1991). Schwefel in frühem Schmuck.
Lecture at Jahrestagung Archäometrie, Berlin 1991.
phur in nature and the possible contamination during Follmann-Schulz, A.-B. (1989). Ein römischer Grabfund des 4.
its ancient use and the long time of burial, trace Jahrhunderts n. Chr. aus Zülpich-Enzen, Rheinland. Kölner
element analysis of sulphur for provenancing does not Jahrbuch für Vor- und Frühgeschichte 22, 49–68.
seem very promising. Nevertheless, measuring the iso- Haberey, W. (1961). Ein Mädchengrab römischer Zeit aus der
topic ratio of the sulphur might lead to a distinction Josefstraße in Bonn. Bonner Jahrbücher 161, 319–332.
Henkel, F. (1913). Die römischen Fingerringe der Rheinlande. Berlin:
between different kinds of geological sulphur deposits Reimer, p. 278.
(Hoefs, 1980). Hjelm-Hansen, N. (1984). Cleaning and stabilization of sulphide-
corroded bronzes. Studies in Conservation 29, 17–20.
Hockey, M. (1981). The composition and structure of a Byzantine
Acknowledgements torc. Jewellery Studies 3, 33–39.
Hoefs, J. (1980). Stable Isotope Geochemistry. Berlin: Springer-
We thank Clive Bridger-Kraus, Anna-Barbara Verlag.
Follmann-Schulz, Christoph J. Raub, and Frank Hollemann, A. F. & Wiberg, E. (1985). Lehrbuch der Anorganischen
Willer for their advice and help. Chemie. Berlin: W. de Gruyter.
Noll, R. (1984). Zwei römerzeitliche Grabfunde aus Rumänien in
der Wiener Antikensammlung. Mit einem Exkurs: Goldene
Herkuleskeulen. Jahrbuch des Römisch-Germanischen Zentral-
References museums 31, 435–454, note 57.
Arrhenius, B. (1985). Merovingian Garnet Jewellery. Stockholm: Noll, R. (1988). Eine neue goldene ‘‘Herkuleskeule’’ mit Inschrift aus
Almquist & Wiksell. Carnuntum. Carnuntum Jahrbuch 1987, 115–118.
Artemidorus, (R. J. White, Trans., 1975). The Interpretation of Ogden, J. (1982). Jewellery of the Ancient World. London: Trefoil
Dreams. Park Ridge: Noyes Press. Books, p. 40.
Brepohl, E. (1987). Theorie und Praxis des Goldschmieds. Leipzig: Ogden, J. (1993). Aesthetic and technical considerations regarding
Fachbuchverlag. the colour and texture of ancient goldwork. In (S. La Niece &
Cellini, B. (C. R. Ashbee, Trans., 1967). The Treatises of Benvenuto P. Craddock, Eds) Metal Plating and Patination. Oxford:
Cellini on Goldsmithing and Sculpture. New York: Dover. Butterworth-Heinemann, pp. 39–49.
Cowell, M. R., La Niece, S. & Meeks, N. D. (1983). The scientific Raub, C. J. (1981). Ergebnisse der chemischen Untersuchung der
examination of the Thetford treasure. In (C. Johns & T. Potter, Nadel. Fundberichte aus Baden-Württemberg 6, 661–665.
Eds), The Thetford Treasure. London: British Museum, Theophilus (J. G. Hawthorne & C. S. Smith, Trans., 1979) On Divers
pp. 56–61. Arts. New York: Dover.