Downloaded from http://jgs.lyellcollection.

org/ at University of Virginia on September 10, 2012

Journal of the Geological Society

The IUGS systematics of igneous rocks

M. J. LE BAS and A. L. STRECKEISEN

Journal of the Geological Society 1991, v.148; p825-833.

doi: 10.1144/gsjgs.148.5.0825

Email alerting click here to receive free e-mail alerts when new articles cite this
service article
Permission click here to seek permission to re-use all or part of this article
request
Subscribe click here to subscribe to Journal of the Geological Society or the
Lyell Collection

Notes

© The Geological Society of London 2012

 Downloaded from http://jgs.lyellcollection.org/ at University of Virginia on September 10, 2012
Journal of the Geological Society, London, Vol. 148, 1991, pp. 825-833, 8 figs, 2 tables. Printed in Northern Ireland

The IUGS systematics of igneous rocks

M. J . L E BAS’ & A . L . S T R E C K E I S E N 2
(On behalf of the International Union of Geological Sciences, Subcommission on the Systematics of Igneous
Rocks)
1 Department of Geology, University of Leicester, LE1 7RH, U K

Manuelstrasse 78, Berne, CH-3006, Switzerland

Abstrpct: In order to create a sustainable classification of igneous rocks which all geologists might
use, an international body was set up by the IUGS: the IUGS Subcommission on the Systematics of
Igneous Rocks. In the course of creating the classification, the Subcommission has established ten
principles for its construction and for defining an appropriate nomenclature. The principles are: (1)
use descriptive attributes; (2) use actual properties; (3) ensure suitability for all geologists; (4) use
current terminology; ( 5 ) define boundaries ofrock species; (6) keep itsimple to apply; (7) follow
natural relations; (8) usemodal mineralogy; (9) if mode not feasible, use chemistry; (10) follow
terminology of other IUGS advisory bodies. These principles and their rationale have not previously
been enunciated.
The classification separates and individuallyclassifies the pyroclastic, carbonatitic, melititic,
lamprophyric and charnockitic rocks before entering the main QAPF classification for plutonic and
volcanic rocks which is based on the modal mineral proportions of quartz (Q), alkali feldspar (A) and
plagioclase (P) or of alkali feldspar (A), plagioclase (P) and feldspathoids (F). Rocks withmafic
content >W% have their own classification. If the mineral mode cannot be determined as is often the
case for volcanic rocks, then a chemical classification of total alkalis versus silica (TAS) is used. The
nomenclature for these classifications necessitates only 297 rock names out of the c. 1500 that exist.

200 years ago, the Academy of Sciences of St Petersburg (at This is despite the knowledge that it is not always easy to
present Leningrad) offered a prize for the best essay on the determinethe precise mineral assemblage of many
classification of rocks. Soon after, Kirwan (1794) coined the fine-grained igneous rocks.
phrase ‘igneous rock’. Today, the study of igneous rocks is a Rock names were often given after the type locality for
vast subject, and the task is still to create a systematic and the rock, e.g. gabbro,fromthe village of Gabbronear
sustainable classification of the many differenttypes now Florence in northern Italy; and urtite from Lujavr-Urtin the
recognized. Lovozerocomplex, Kola Peninsula,USSR.Sometimes a
Early attemptsto classify igneousrocksvaried.Some more useful name was derived from the mineral assemblage,
were based on petrography and mineralogy (e.g. von Cotta e.g.
peridotite. Either way, new names proliferated,
1866; Rosenbusch 1887), some utilized a notation of particularly for alkaline rocks. Johannsen sought to
symbols for minerals and textures (e.g. Michel-Levy 1889), systematize the growing nomenclature,and published ‘A
and a few looked to chemistry (e.g. Roth 1861). In the early descriptive petrography of the igneous rocks’ in four
part of this century and well recorded by Johannsen (1939), volumes (1932-1939), which came to dominate the thoughts
Loewinson-Lessing, Lacroix, Niggli and Washingtoninde- of petrologists in the English-speaking world.
pendentlyproduced c. 70 papers which attempted to At much the same time,Niggli (1931) presented a system
systematize igneous rock nomenclature. Loewinson-Lessing for the classification and nomenclature of igneous rocks
emphasized the petrographical and mineralogical as- according to modal mineralcontents. He followed this
semblage as the prime means for defining different rocks. (1936a, b ) with a system of classifymg igneous rocks by their
Lacroix included rock chemistryin his classification and chemical compositions, which system was based on
produced a complex hierarchy of terms with a mathematical molecular numbers (‘Niggli numbers’), systematized as
notation to characterize each rock. Washington used ‘magma types’. These ‘magma types’ were not rock names
chemical analyses to calculate a standard (the
CIPW butadjectival
attributes. The thoroughness of Niggli’s
normative)mineral assemblage which formed the basis of scientific work dominated the minds of petrologists in the
another complexhierarchy of classes, ordersand rangs. German-speaking world. Troger (1935) published a most
These and others like them never found broad acceptance. useful compendium of igneous rock types, listing for each
They were too cumbersome. Igneous classifications based on type the mineral content, rock chemistry, systematic
supposed genesis havealsofailed, usually owing tothe position in the Niggli and other classifications, original type
inadequacies of the genesis proposed, although a truly locality and reference to the original description.
genetic classification may ultimately be constructed. With the establishment of the IUGS in 1961, a greater
Despitethesefailures, the mineralogical and chemical awareness developed of the advantages of international
approaches to rock classification have continued to modern cooperation in science,althoughsome had been possible
times as the basis for differentiatingigneousrocks, with through the meetings every four years of the International
identification of the mineral assemblagebecoming the Geological Congress (IGC). At the IGC meeting in Prague
dominantmeans of distinguishing one rock from another. in 1968, a meeting underthe leadership of Mehnert was
825
 Downloaded from http://jgs.lyellcollection.org/ at University of Virginia on September 10, 2012
826 M . J . LE B A S & A . L. STRECKEISEN

planned to discuss the earlier

and widely circulated can differ, they are not factual. The most common
proposals by Streckeisen (1967) onthe classification and interpretations made by geologists are petrogenetical, but
nomenclature of igneous rocks. Regrettable political events there could also be aesthetic considerations. Even such
prevented any thorough discussion, andthereforeIUGS propertiesas the age of a rock can beuncertain and
created a Subcommission which should deliberate the thereforenotsuitablefor primary rock classification. The
various problems and present definite recommendations to second principle is that classifcation should depend on actual
IUGS. The Subcommission on the Systematics of Igneous attributes and not on interpreted characters.
Rocks began its work in 1969, held a working meeting in Classifications must be intelligible to all who use them.
Berne in 1972, and then presented its first report at the 1972 Inthe presentcase,igneous rock classification must be
IGC in Montreal. Since then, the Subcommission has met at usable by all geologists. A granite identified by a field
Grenoble (1975), Sydney (IGC, 1976), Prague (1977), geologist or mining engineershould also be given that
Padua (1979), Paris (IGC, 1980), Cambridge (1981), name by the specialist petrologist or geochemist. Con-
Granada (1983), Moscow (IGC, 1984), London (1985), versely, the same name should not be given to two different
Freiburg im Breisgau (1986), Copenhagen (1988) and rocks. Such names are bestremovedfromcurrent usage
Washington D C (IGC, 1989). 419 people from 49 countries owing to their ambiguity. Thus the third principle is that the
haveparticipatedin the discussions and recommendations basic or root name given to a rock should be one that is
made at these meetings, and through discussion papers and suitable for all geologists to use.
questionnaires circulated between the meetings. There is considerable logic in the statement that‘if a new
Such sustained discussion brought together the different classification is set up, it should employ new terms, i.e. new
lines of thought that previously existed. International rock names’. The aim would beto avoid ambiguity and
co-operationhad been achieved, andtheoutcome is the equivocal definition in the terminology. However, with
book ‘A Classification of IgneousRocks and Glossary of several hundred igneous rock namescurrently in use, to
Terms: Recommendations of theInternationalUnion of create new ones, which would effectively mean more in at
Geological Sciences Subcommission on the Systematics of least the short term, would lead to unacceptable confusion.
Igneous Rocks’ (R. W. Le Maitre et al. 1989). Although the This point is contentious, and the balance of the argument
internationallanguage of science is English, the book is has been whether the present terminology is so bad that a
being translated in whole or in parttoother leading totally new one is preferable. The Subcommission has taken
languages. the view, but not without lively argument, that radical new
Thebook is based on principles established by the nomenclature is not usually accepted. Ithas beentried
subcommission, and gives recommendations on class- before in igneous rock classification, and the merits of the
ification and nomenclature agreed by the Subcommission. It various systems proposed are evident in some of the terms
also provides a glossary of over 1500 igneous rock names which have been adopted. The normative scheme of Cross,
that have been usedin the past. The Subcommission Iddings, Pirsson, Washington (1902) for realising the
decided that fewerthan 300 igneous rock names were properties of rocks through their chemistry is widely used,
sufficient to characterize all igneous rocks. Fewerthan a as are several of the termsformulated by Shand for his
hundred is all that is required when dealing with the more chemically-based classification (1949). But their class-
common igneous rocks. Unanimity of nomenclature was not ifications never gained wide acceptance. The fourth principle
reached for all igneous terms, and debate continues about is that the terms to be used in any classifcation should follow,
how best to classify the potassic, the lamprophyric and a few as far as possible, those which are currently and widely
other groups of rocks. accepted as being useful terms.
A useful classification necessitates the recognition of
boundariesbetween the different classes of objectsbeing
Principles of classification identified. In contrast, some classifications are based on the
Ten principles have been established over the 20 year period characterization of a typical or averagesample. Onefor
of the deliberations of the Subcommission, but the rationale example, is the classification of Johannsen, quoted above,in
of their development has not previously been stated. which full petrographic descriptions supported by chemical
Objects being classified mustbe well identified. Their analyses are given for each compartment within the
distinction usually results in the object being given a name classification. The more recent two-volume work of
or someidentity tag. Thepurpose of classification is to Andreeva et al. (1983) follows this procedure with even
promote a systematic means of distinguishing oneobject greater emphasis on the chemistry of each rock type. The
from another, using unequivocal terms of identification. For problem here is what to do if a sample is not a close match
igneousrocks, the unequivocal terms available aretheir to any type sample but has similarities to two or three. Since
evident physical and chemical properties. It is also necessary most rocks form a continuum of types rather than falling
that the properties are intrinsic to the rock itself and not to intodiscretegroups,as most biological samples do, it is
the environment in which it occurs. Thus, even an isolated necessary toerect notionalboundaries which divide one
sample obtained from a borehole or from a glacial erratic category from the next. Since the boundaries are not
should be capable of being given a name which conveys to natural, they can be drawn at any convenient place. Thefifth
another geologist the identity of the rock sample. The first principle is that the classifcation must consist of classes which
principle, therefore, is that igneous rock nomenclature should are separated by boundary conditions.
be based on descriptive attributes. Many sophisticated igneous rock classifications have
The descriptive attributes of a rock must be distinguished been created, such as that by Johannseninto families,
from interpretative attributes. The interpreted characters of orders and other subdivisions. But the complexity, like that
a rock are those deduced from a conceptual understanding for the CIPW andShand systems mentioned above, led to it
of it, and they may differ from person to person. Since they never being widely applied. The only classifications that
 Downloaded from http://jgs.lyellcollection.org/ at University of Virginia on September 10, 2012
I U G S SYSTEMATICS O F IGNEOUS R O C K S 827

have been widely used, are those which are simple to apply. analyses, but even with such data, two visibly different rocks
It is a natural and inevitable corollary that geologists have can have identical chemical analyses. In such cases as the
and will create more complex classifications, often devised gabbro-basalt pair, there is no difficulty because the gabbro,
for limited fields of investigation. These have their clear being coarse-grained, would alreadyhavebeen classified
uses. The aim of the Subcommission is to producea modally using the above principles. But with many
classification which will providea sufficiently sound and fine-grained potassic and lamprophyric rocks, for example,
broadly based classification of leastcontroversy, on which there is the difficulty of heteromorphicequivalents. One
the science of igneous petrology can build. Well-established rock may be mica-rich another feldspar-rich, but they can
simplicity is the sixth principle. give indistinguishable chemical compositions, apartfrom
In a pragmatic manner, it is reasonable to discover what water content. This is a problematical area of classification
other considerations cause classifications to be adopted. It is that is still not resolved by the Subcommission.
evident that classifications thatcorrelate with unequivocal The final principle concerns the use of geological terms
inter-relationships, are commonly adopted.The relation- which aremore properly defined by other international
ships may be thoseobserved in the field or they may be bodies rather than by this Subcommission. The correctness
relationships deduced from petrogeneticconsiderations. The of mineralnames is the concern of theInternational
simpler these relationships, the greater the use that has been Mineralogical Association, andthe Subcommission has
made of them. The seventh principle is that any classification endeavoured to follow its recommendations. For example,
of igneousrocksshould follow fundamentalgeological barkevikite is no longer an accepted term, and in any new
relationships. definitions of rocks, such discredited terms are not used. In
The vindication of this last principle is the success of the cases where original definitions of rocks are quoted
division into the plutonic versus volcanic classification which however, a discredited term may have to be used since,
hasbeenaccepted since its formalization by Rosenbusch taking the example of barkevikite, it may notbe known
over 100 years ago. It would not have been acceptable 200 whether the original authormeant kaersuitite or another
years ago in the days of Werner who thought granites were brown amphibole. When the Subcommission came to deal
sediments. Butthe division raises the problem thatthe with the pyroclastic rocks, it wanted to use internationally
terms
plutonic and volcanic are fundamentally inter- accepted divisions of grain s u e forsediments, butnone
pretative. However,their identification with rocks that exist. The recent establishment of a Subcommission on the
have respectively phaneritic or aphanitic textures,
has Systematics of SedimentaryRocks will, nodoubt,create
permitted the terms plutonic and volcanic to be validated. divisions which we will adopt, but in the meantime we have
To most geologists, these two terms imply coarse-grained used those most widely quoted. The tenth principle is that all
and fine-grained characters; forsome,anintermediate terminology should be internationally acceptable.
category is required,andthen hypabyssal and medium- Having established theten principles which shouldbe
grained are employed. The difficulty of thetrio of terms followed in naming and defining rocks and in constructing a
‘plutonic’, ‘hypabyssal’ and ‘volcanic’ is that the boundaries classification, the next step is to determine the methodology
between them are often impossible to establish. Use of grain of the classification.
size would solve the problem, and would be in accord with
the first principle. Thus the sixth principle is not without its
hazards. Tools of the IUGS classification
There is a choice of properties which might be used in The first question is: how are the igneous rocks recognized
describing an igneous rock, but
the
one which has and separated from other rocks? If igneous rocks are those
consistently received the widest use over the last 100 years, which solidify frommoltenmaterial, then very few, apart
is the mineralassemblage.It is not deniedthatother from lavas, can be observed to be igneous although many
characters,
such
as geophysical or geochemical or are so interpreted. Using the term ‘igneous’ violates
mechanical, are of value, but the prime property to beused principle 2, but the certainty of the interpretation is such
in all cases is the mineral assemblage together with the thatthe violation can be overlooked.Nevertheless, there
relative proportions of the minerals in that assemblage. The are cases where the igneous origin is in doubt; for instance,
eighth principle is thatclassificationshouldbebased on some
granites are said to be metasomatic. Thus
modal mineralogy, as far as possible. Rosenbusch’s term ‘Massige Gesteine’ isthe best simple
In many cases, igneous rocks are too fine-grained for the description of an igneous rock because it implies a
minerals to be identified, even under the microscope. Some, crystalline (or glassy) rock which, when observed in
even with identifiable minerals, are so fine-grained that the outcrop, has a uniform texture and massive character unlike
modal proportions cannot be determined accuractely. The metamorphic and sedimentary rocks. By this definition,
latter is also impossible if the rock contains glass. In these charnockites, eclogites and the rocks of the mantle could all
cases, the Submission decided that chemical analysis was the be included with the igneous rocks. Some would say all
next bestdescriptive property to be employed in three are strictly metamorphic, but the Subcommission took
characterizing an igneousrock,but that the mineralogical the pragmatic view that eclogite is metamorphic butthe
approachshould always be applied first. To ensurethat others can be igneous. Dividing igneous rocks into plutonic
chemical analyses may be comparable, all must be and volcanic again violates principle 2, but again is accepted
recalculated anhydrous to 100%. The ninth principle is that as a practicable procedure.
if the modalmineralogy ofan igneousrockcannot be The prime tools of rock identification are the constituent
determined satisfactorily, then chemical analytical parameters minerals (principle g), andthe Subcommission uses the
should be the next property used. feldspars, together with quartz and the feldspathoids, as the
The use of chemical analyses does bringitsproblems. principal component minerals for classification. This follows
Not only is there the difficulty for many to obtain chemical the long established choice of petrologists. It was
 Downloaded from http://jgs.lyellcollection.org/ at University of Virginia on September 10, 2012
828 M. J . L EB A S & A . L . STRECKEISEN

quartz monzodiorite
quartz monzogabbro
quartz alkali feldspar

rnonzodiorite

anorthosite

foid-bearing gabbro
foid-bearing anorthosite

foid-bearing rnonzodiorite
foid-bearing monzogabbro
Fig. 1. Classification and nomenclature
of the plutonic igneousrocks according
foid gabbro to their felsic modal contents when mafic
mineral content is less than 90%. Q,
quartz; A, alkali feldspar;P, plagioclase;
F, feldspathoid (foid). The equivalent
classification for volcanic rocks is given
in Le Maitre et al. (1989, fig. B.lO) Figs.
1-6 in this paper are afterLe Maitre et
F al. (1989).

recommended that the useful abbreviation ‘foid’ for was discussed, butfound to be untenable;it did not fit
feldspathoid be a permissable synonym. current terminological usage. It was necessary to put Q at 20
Classifications based on these felsic minerals tend to be in QAP whilst F must be at 10 in APF.
displayed either on a two-component diagram, such as that
of Nockolds (1954), or on the double three-component plot
of Johannsen which is now formalized as the IUGS QAPF
system for rocks with felsic minerals >10% and mafic
01
dunite -4
minerals <90% by volume (Fig. 1). The system works well
mineralogically, but attemptstodeducethesame class-
ification from chemical data have notbeen satisfactory
owing to the problem of how to divide the albite component
between alkali feldspar and plagioclase. The problem has
not been solved,althoughStreckeisen (1976) showed that lherzolite
with minor adaptation of the CIPW norm, a fair agreement olivine
could be reached between the mineral assemblage deduced orthopyroxenite olivine
from chemical analysis andthemode,for leucocratic, olivine
non-foidal rocks. Streckeisen & Le Maitre (1979) obtained websterite
better agreement between mode and norm by dividing the
feldspar components according to the ratio An/(An Or) +
although ultramafic rocks and foidal rocks still did not fit.
Most of the fields within the QAPF system havelong orthopyroxenite clinopyroxenite
been established, but some names and boundaries required Fig. 2. Classification and nomenclature of the ultramafic rocks
adjustment following principles 4 and 5. An attempt tohave (M > 90%) based on the modal proportionsof olivine (OI),
the same boundariesin both the QAP and APF triangles orthopyroxene (Opx) and clinopyroxene(Cpx).
 Downloaded from http://jgs.lyellcollection.org/ at University of Virginia on September 10, 2012
SYSTEMATICS
IUGS OF
ROCKS
IGNEOUS 829

Table 1. The classification of the pyroclastic rocks based on clast versus pyroclast type was agreed, following principles 4 and
size 6 (Table 1).
The discussions oncarbonatite nomenclature quickly
Pyroclastic deposit dismissed the usage of the plethora of exotic names formerly
Clast
size
Pyroclast
Mainly
unconsolidated:
Mainly
consolidated: favoured by alkaline rockpetrologists,becausecurrent
m in workers prefer straight-forward terms such as calcite-
Bomb,
Agglomerate,
block Agglomerate, carbonatite,dolomite-carbonatite andferrocarbonatiteto
bed of blocks or pyroclastic
breccia characterize the modal properties evidentin thename. If
bomb, block tephra
64
the type of carbonate cannot be determinedoptically, then a
bed
Lapillus
Layer, of tuff
Lapilli chemical classification is used (Fig. 3).
lapilli or Melilite is deemed to be a mafic mineralalthough
lapilli tephra
2 light-coloured. Of the QAPF minerals, it is compatible only
Coarse
grain
ash
Coarse
Coarse
ash(ash
tuff) with the feldspathoids, and any rock containing significant
1/16 melilite usually has less than 10% felsic minerals,thereby
Fine
ash
grain
Fine
ash
(dust)
Fine (ash) tuff
grain) (dust excluding it from theQAPF classification. Nor do these
(After Le Maitre er al. (1989). rocks fit the ultramafic classification, and thus a special
classification was constructed that is based onthe modal
proportions of melilite, clinopyroxene and olivine (Fig. 4).
Atan earlystage in the Subcommission’s discussions,
When the felsic minerals component falls below 10% of charnockites were included because they have
both
the mode, triangular plots involving olivine, pyroxene and igneous-looking textures andan association with igneous
hornblende satisfactorily classify the peridotitic, pyroxenitic rocks. The presence of hypersthene (more accurately now,
and hornblenditic ultramafic rocks (Fig. 2). The prefixes orthopyroxene) in a granite or a granodiorite characterized
leuco- and mela- canbeapplied to all the rocksin the by perthiticfeldspar, was taken to indicate a charnockitic
QAPF fields. The simple division by which leucocratic might rock.
have >SO% light-coloured minerals and melanocratic Whilst the Subcommission has outline agreement on how
<50%, unfortunately does not correspond to current usage. the lamprophyres should be classified modally, it is not clear
Therefore, following principle 4, each field has had to have how lamproites and kimberlites should be treated, and how
its own limits set. For instances, a leuco-granite has 4 %
mafic minerals, and a mela-granite >20% mafic minerals,
but a leuco-gabbro has <35% mafic minerals and a
mela-gabbrohas >65% mafic minerals. The leuco-/mela-
divisions for other rock types are given in Le Maitre et al.
(1989, fig. B.7a, b).
The ultramafic and QAPF classifications work well for
the majority of the plutonic and coarse-grained rocks, but
not for all. The lamprophyres,the charnockites, the
melilite-rich rocks and the carbonatites have each received
entirely independent modal classifications, some more
satisfactory than others. Pyroclastic rocks have alsobeen
classified independently. A Working Group was set up to
consider the pyroclastic rocks and numerous questionnaires
were circulated to more than 150 geologists. Several more
detailed classifications were proposed before the simple size 01 CPX
Plutonic rocks

Me1
CaO

01 CPX
MgO FeO+Fe,O,+MnO Volcanic rocks

Fig. 3. The chemical classificationof carbonatites using weight Fig.4. The classificationof the melilitic igneous rocks based on
percent oxides. modal composition. Mel, melilite; Cpx, clinopyroxene; olivine.
 Downloaded from http://jgs.lyellcollection.org/ at University of Virginia on September 10, 2012
830 M. J . LEBAS & A . L . STRECKEISEN

Table 2. The present classification of the lamprophyres

Light-coloured
constituents minerals mafic Predominant

Amphibole, Melilite,
(barkevikite, biotite,
Biotite, kaersutite),
Hornblende, fTi-augite,
diopsidic augite, diopsidic augite, Ti-augite, folivine,
Feldspar Foid ( folivine) olivine,
(folivine) biotite fcalcite

or > p1 - Vogesite Minette - -

pl > or -Spessaritite Kersantite - -
or > p1 Feld > foid - - Sannaite -
p1 > or Feld > foid - - Camptonite -
- Glass or foid - - Polzenite
Monchiquite
- - - - - Alnoite

Or, alkali feldspar; pl, plagioclase; feld, feldspar; foid, feldspathoid. (After Le Maitre et al. 1989).

they may relatetothe lamprophyres. Inorderthatthe was chosen following principle 6; the point being made that
publication of theIUGSbookon igneous classification the sum of the oxides in whichever parameters were used
should not be held up, the compromise was made to put was an artefact, and therefore the simpler it was, the easier
these three groups of rocks underthe generalheading it was to use (Fig. 5 ) . A proposal to substitute MgO for the
‘lamprophyric rocks’ (Table 2), and that they should receive SiO, parameter was abandoned since MgO is not well tested
attention later. There is now (1991) a Working Group trying for distinguishing volcanic rocks across the whole spectrum
to resolve the several different opinions that exist on these from basic to acid compositions. The Rl-R2 diagram of de
rocks, particularly those expressed by Rock (1990). La Roche et al. (1980) in which R1 = 4Si - 11(Na K) - +
If the mineral mode of a volcanic rock can be 2(Fe+Ti) and R2=Al+2Mg+6Ca cations per mil was
determined, then it is classified by QAPF, following a looked at closely because it had the merit of encompassing
similar procedure to that for the plutonic rocks (Fig. 1). If the use of all the major oxides, which is advantageous when
the mode cannot be determined, thenprinciple 9 is followed compared with the total alkalis-silica system. It was not
and chemical analysis is used. The chemical parameters it adopted, partly because the two cation per mil parameters
was decided to use were silica (SiO,) weight per cent and were difficult to calculate without electronic assistance, and
+
total alkalis (Na,O KzO) wt.% because they appeared to partly because some of the boundaries were thought to be
be the best and were already widely used. The proposal that misplaced. The Subcommission considered it wiser, in such
+
(Na,O 0.7K20) be usedinstead of (Na,O K,O) was + a fundamental issue as classification, to remain closer to a
considered.Using (Na,O + 0.7&0) wt.% brings this scheme that already had a long record of usefulness
parameterinto relativeequivalence with (Na,O + K,O) (principle 4).
molec.% Despite this useful aspect, the simpler parameter Two decisions were madefor determining the bound-

15 -
Phonolite

13 -

S
3
0,
Y
+
0
m”
Z
Foidite

I
1\
rdacite
Rhyolite

Fig. 5. The chemical classification of the

volcanic rocks based on total alkalis

\
Dacita
Andesite versus silica (TAS). The line between
Basanic
andesite \
the fields of foidite and basanite-
.
Basalt tephrite is dashed because other criteria
Picro- \ are necessary to distinguish those rock
I
basalt l types (see Fig. 7). Basanite has
I
normative olivine >10%; tephrite has
normative olivine <10%; trachyte has
I l I Si02 wt% quartz (20% of sum of felsic minerals;
ULTRABASIC 45 BASIC :5 I N T E R M E D I A T E 6: ACID trachydacite has quartz >20% of sum of
l I I felsic minerals.
 Downloaded from http://jgs.lyellcollection.org/ at University of Virginia on September 10, 2012
IUGS SYSTEMATICSOFIGNEOUS ROCKS 831

4 - Na20 + K20 W%
I
1-
,.

:,
.. basanite i.jj
!@rite $ !. I basaltlc : jj andesile
3 foidite .........
..... ........
................... . . ....... .......
.. ....,. andesite
, ,
basalt
l

Fig. 6. The
classification
and
nomencla-
ture of the 'high-Mg' volcanicrocks
I -
M90 > 18% 8 Ti0-s
L
1% meimechite :. i boninite
MgO > 8% 8 TiO, < 0.5%
MgO > 18% 8 TiO, < 1% komatiite Ij.
(picrite, meimechite, komatiite, bonin- 1 I I I 1 I 1 I I 1 1 I I 1
ite) using TAS (thick boundaries)
37 41 45 49 53 57 61 65
together with MgO and TiO, wt% (thin
boundaries). SiO, wt%

aries within thetotal alkali-silica (TAS) system (Fig. 5). discriminate these rock types is because at such low values
First, to use the SiO, categories of ultrabasic, basic, of SiO,, the individual contents of CaO, MgO and A1,03
intermediate and acid established by hewinson-Lessing significantly influence the formation of plagioclase (abun-
(1889) to distinguish the boundaries of the basalts dant in basanite and tephrite but absent in nephelinite) at
(45 < SiO, < 52 wt.%), andesites (52 < SiO, < 63%), dacites the same (Na,O + K,O) contents.
and rhyolites (63 <SiO,%). Second, to utilize the large The details of the individual classifications are not given
computerized data base CLAIR (Le Maitre & Ferguson here, but can be found in Le Maitre et al. (1989).
1978) which contains not only chemical analyses of igneous
rocks but also the rock-name of each analysis. The data base
Nomenclature
could be used to plot each volcanic rock (e.g. all trachytes)
on frequencydistributiondiagrams and so put best-fit An universally acceptable classification must use universally
boundariesbetween adjacent fields. Despite some overlap acceptable nomenclature (principle 3). There exist over 1500
between adjacent frequencydistributionplots, the clear rock names of which only a few hundred are completely
clustering of the points for any one rock typeindicated obsolete.
Somerocks have many names. Some are
where boundaries should be placed according to generally synonyms, such as liparite and rhyolite,dellenite
and
acceptedusage. Ratherthan choosingcurvedboundaries rhyodacite. Many are names given to varieties of common
between the fields, the Subcommission decided to employ rocks, such as domite for biotite trachyte, or cortlandtite for
straight lines between easily determined points to divide a pyroxene olivine hornblendite with a specific poikilitic
them.Thus, basalt is now defined chemically as having texture. Some are now considered to be self-contradictory,
between 45 and 52 wt.% SiO, and <S wt.% (Na,O + K,O). such as nepheline basalt for nephelinite, because basalt by
One of the merits of the TAS system is that the boundaries modern definition has plagioclase but nephelinite does not
are definitive althoughthey could be criticized asover contain plagioclase. Whentheterm nepheline basalt was
simplistic. Definitive boundariesremove the ambiguity in
naming a rock which plots near a boundarybetweentwo
adjacent rock types. The simple boundaries of theTAS
I
system also enable the classification to be constructed in a
few minutes by pencil and ruler.
The TAS system works well for the common rocks but is
inadequate for many low-silica rocks. The high-Mg volcanic Basanite
rocks (picrite, komatiite, meimechite and boninite) form a
groupnot classified by theTAS system. Instead picrite,
komatiite and meimechite are distinguished by >18% MgO Nephelinite
and <52% SiO, and are
further
separated by their
(Na,O + K,O) and TiO, contents. Boninite has (52< ------
SiO, < 63 wt.%), MgO > 8 % and TiOz< 0.5% (Fig. 6).
Likewise the alkaline low-silica rocks also require Melanephelinite
classification apart from TAS. Owing to the small numbers I
of well identified nephelinitic, basanitic and tephritic rocks 0 1 1 1 1 ' 1 1 1 1 ' 1 1 1 1 '
in the geological literature, the seriousness of the overlap of 8 12 16
24 20 28 32
the nephelinitic rocks into the TASfield of the basanites and Normative ne (wt%)
tephrites was not fully appreciated. To overcome this, rocks Fig.7. The CIPW normative ne versus ab classification and
in those TAS fields were found to be better distinguished nomenclature of non-melilitic volcanic rockswith Na,O > K,O that
using CIPW parameters (Fig. 7): normative nepheline and plot in theTAS fields (Fig. 5 ) of foidite and basanite-tephrite (after
normative albite (Le Bas 1989). The failure of TASto Le Bas 1989).
 Downloaded from http://jgs.lyellcollection.org/ at University of Virginia on September 10, 2012
832 M. J . L E B A S & A . L. STRECKEISEN

coined in 1850, the presence of plagioclase was not implied IS the rock
YES Carbonates YES
by that term. + IS it pyroclastic
?
* Use carbonatite
classification
igneous ? ?
> 50’”
Instead of trying to decide which rock names are not
necessary, the Subcommission investigated what names were l AYES I
best for the categories created by theQAPF,TASand
minor classifications. Plutonic and volcanic QAPF require classification
less than 50 root names althoughsome pre-fixing is 1s melilite >IO%
required, such asnephelinesyenite andquartzdiorite,to and M > 90% 7 classification
obtain the full range needed. TAS requires fewer than ten

fi-
further names, but did require three new names to occupy

, j( fl-d
charnockitic
fields otherwise nameless; picrobasalt, potassic trachybasalt classification I

YES Use
and basaltic trachyandesite. Together with the specialized Is it
classifications for ultramafic, charnockitic, pyroclastic, NO lamprophyric ?
* lamprophyric

I
Is it classification
lamprophyric, melilitic and carbonatitic rocks, a total of 297 charnockitic
names andtermsare used in theIUGS classification.
Following principle 4, most are ones already in common use
and most have required only littie alterationfromtheir
plutonic
previous definition. The alterations werenearly all minor
Is it M c 90% ?
changes in the positions of the boundaries tothe fields plutonic
defining them. Somerocktypeshadnot previously been Use

defined in this way, merely characterized (e.g. tonalite). I ultramafic

classification I
Of the 297 names recommended for use by the IUGS,
about 100 are necessary for pre-fixed root-names in order to
define and distinguish such rock types as olivine gabbro, volcanic
possible ?
pyroxene hornblendite,quartz monzonite and analcime volcanic
phonolitefrom ordinarygabbro,hornblendite, monzonite
and phonolite. Another approximately 100 namesarefor
the less commonrocks,many of themalkaline, such as
hauynite,ijolite,italite and kugdite.This leaves less than
100 namesas all thatarerequired by a geologist dealing
with the more common igneous rocks.
Since igneous petrology is an international science, the Use TAS. If it falls in fields F or U1, use
Subcommission wishes all terms to have the same validity in norm ne v. norm ab classification
all countries, allowing for variations arising from problems
of translation. Care was taken, as far as possible, to ensure Fig. 8. Simplified Row chart, beginning at the top left, showing the
that translation would not present misinterpretations. One sequence followed by the IUGS scheme for namingan igneous rock
area where there was variation in the meaning of a standard using the individual classifications described.
term was with the use of the term subalkali basalt. It arose
from a mis-translation and has caused much misunderstand-
ing. Inthe USSR, basalticrockshavebeen divided into (2) Does the rock have more than 50% modal carbonate
‘alkali’ (>S%normne), ‘sub-alkali’ (norm ne <5%) and minerals? If so, use the carbonatite classification.
‘normal’ (norm ne 0%, norm (hy + 01) > 0%). In the rest of (3) If the rock is ultramafic and contains more than 10%
the world, basalt is defined as ‘alkali’ (norm ne >O%) or as modal melilite, use the melititic rocks classification.
+
‘sub-alkali’ (norm ne 0%, norm (hy 01) > O%), as defined (4) Is the rock lamprophyric? This is generally understood
by Iddings in 1895. This has been resolved in the USSR by to mean thatit forms
minor
a intrusion, is strongly
substituting the term ‘mid-alkali’ for USSR ‘sub-alkali’ and porphyritic with only mafic phenocrysts, usually biotite,
thuspermitting ‘sub-alkali’ to be used everywhere for amphibole or pyroxene, and that feldspar, if any, is confined
‘normal’ basalts
in order
conform
to with IUGS tothe groundmass. If so then use the lamprophyric
recommendations. classification.
(5) If the rock is orthopyroxene-bearing,plutonic and
belongs tothe granitic association, use the charnockitic
classification.
Procedure (6) If the rock is plutonic, then use the QAPF classification
The aim of the IUGS classification of igneous rocks is that it for plutonic rocks, noting that if the modal content of mafic
should be capable of beingused in a logical sequence to minerals is more than90%, then theminor classifications for
name any igneous rock. The flow-chart (Fig. 8) gives the ultramafic rocks should be used.
sequence of that logic. The first step is to confirm that the (7) If the rock is volcanic and the mineralmode can be
rock is igneous and therefore suitable for the classification. determined, then use the QAPF classification for volcanic
Then the suitability of the rock for special classification must rocks.
be tested before going on to the main QAPF and possibly (8) If the rock is volcanic and the mineral mode cannot be
TAS classifications. The following questions must be asked determined, then the chemical classifications related to TAS
in sequence. should be used.
(1) Is it a pyroclastic rock? If itis, use the pyroclastic All classifications are regrettablyimperfect,and the
classification, if not, go to the next question. Subcommission continues to seek to improvethem. All
 Downloaded from http://jgs.lyellcollection.org/ at University of Virginia on September 10, 2012
RO
IGCN
KESSOYUSSOTFEIM
UAGT
S ICS 833

comments and contributions for improvement would be 1989. A Classification of Igneous rocks and Glossary of T e r m :
gratefully received. Recommendations of the International Union of Geological Sciences
Subcommission on the Systematics of Igneous Rocks. Blackwell Scientific
Publications, Oxford.
We gladly acknowledge with thanks the unstinting efforts made by LOEWNSON-LESSING, F.1989.Note sur la structure desroches eruptives.
the many members and correspondents who have contributed to the Bulletin de la Societe Be@ de geologie, 3, 393-398.
workof the IUGS Subcommission on the Systematics of Igneous MICHEL-LEVY, A. 1889. Structures et Classification des roches eruptiues.
Rocks. We thank Blackwell ScientificPublications for permission to Baudry, Paris.
reproduce Figs 1-6 from Le Maitre et al. (1989). NIGGLI,1931.
P. Die quantitative mineralogische Klassifikation der
Eruptivgesteine. Schweizerische Mineralogische und Petrographische
Mitteilungen, 11,29&364.
- 1936a. Ueber Molekularnormen zur Gesteinsberechnung.
References Schweizerische Mineralogische und Petrographische Mitteilungen, 16,
295-317.
ANDREEVA, E. D.,BASKINA, V. A., BOGATIKOV, 0. A. ET AL. 1983. - 19366. Die Magmatypen. Schweizerische Mineralogische und Petr-
Magmaticheskie Gornye Porody : Klassifikatziva,
Nomenklatura, ographische Mitteilungen, 16,335-399.
Petrographiua. Moscova, Akademia Nauk USSR (2 volumes,in NOCKOLDS, S . R. 1954. Average chemical compositionof some igneous rocks.
Russian). Geological Society of America Bulletin, 65, 1007-1032.
CROSS,W., IDDINGS, J. P., PIRSSON,L. V. & WASHINGTON, H. S . 19U2. A ROCK,N. M. S. 1990. Lamprophyres. Blackie, Glasgow.
quantitative chemico-mineralogicalclassificationand nomenclature of ROSENBUSCH, H. 1887. Mikroskopische Physiographie der Mineralien und
igneous rocks. Journal of Geology, 10, 555-690. Gesteine. II. Massige Gesteine. Schweizerbart, Stuttgart.
DE LA ROCHE,H., LETERRIER, P.,GRANDCLAUDE, P. & MARCHAL, M. 1980. R o w , J. 1861. DieGesteins-Analysen in tabellariseher Ubersicht und mit
A classification of volcanic and plutonic rocks using the RI-R2 diagram kritischen Erlauterungen. Hertz, Berlin.
and major element analyses. Its relationships with current nomenclature. SHAND, S. J. 1949. Eruptive Rocks. Murby, London.
Chemical Geology, 29, 183-210. STRECKEISEN, A. L. 1%7 Classificationand nomenclature of igneousrocks.
IDDINGS,J. P. 1895. The origin of igneous rocks. Bulletin of the Philosophical (Final report of an inquiry.) Neues Jahrbuch fur Mineralogie,
Society of Washington, U,89-213. Abhandlungen, 107,144-240.
JOHANNSEN, A. 1932,1937,1938,1939. A descriptivepetrography of the - 1976.Classification of the commonigneousrocks by means of their
igneous rocks. 4 volumes, University of Chicago Press. chemical composition: a provisional attempt. Neues Jahrbuch fur
KIRWAN, R. 1794. Elements of mineralogy. Volume 1, 2nd edition, Elmsly, Mineralogie, Monatshefte, 1976, H.l, 1-15.
London. - & LE MAITRE,R. W.1979. A chemical approximation to the modal
LE BAS,M. J. 1989. Nephelinitic and basanitic rocks. Journal of Petrology, QAPF classification of the igneous rocks. Neues Jahrbuch fur
M,1299-1312. Mineralogie, Abhandlungen, 136, 169-206.
LE MAITRE,R.W. & FERGUSON, A. K.1978. The CLAIR data system. TROGER,W. E. 1935. Spezielle Petrographie der Eruptivgesteine: ein
Computers and Geoscience, 4, 65-76. Nomenklatur-Kompendium. Schweizerbart, Stuttgart.
LE MAITRE,R. W. (editor), BATEMAN, P,,DUDEK, A.,KELLER,J. Er AL. VON COTTA, B. 1886.Rocks classified and described. Longmans, London.

Received 30 March 1991; revised typescript accepted 9 May 1991.