t75

The C anndian M ine ralo gi s tVol, 34, pp. 175-186 (1996)

CLASSIFICATION OF LAMPROPHYRES, LAMPROITES, KIMBERLITES'AND THE KALSILITIG, MELILITIC, AND LEUCITIC ROCKSX

ALAN R. WOOLLEY

Departuent of Mineralogy, Natural History Museum, Crornwell RoaL bndon SW SBD' U.K

STEYEN C. BERGMAN

ARCO Resources Technnlogy, Exploration and Production Research, 2j00 West PIarc Parkway, Plano, Texas 75075' U'S'A'

ALAN D. EDGAR

Deparxwnt of Earth Sciences, University of Westem Ontario, LondaU Ontario N6A 587

MICHAELJ. LEBAS

Department of Geology, University of leicester, Isicester IEI 7RH' U'K'

ROGER H. MITCIIELL

Department of Geotogy, Inkehzad lJniversity, Thunder Bay, Ontario P7B 5EI

NICHOLAS M.S. ROCKT

Department of Geology, University of Westem Australia. Nedlanh 6009, Westem Austalin" Australia

BARBARA H. SCOTT SMITH

Scox-Smith Petrology, 2555 Edgemount Boulevar4 NorthVancoaver, Brttish Columbin'WR 2M9

ABSIRAC"I

The nomenclature and classification of lamprophyres, lamproites, kimberlites and the kalsilitic, melilitic and leucitic rocls

are inaaequately detined. The recommendations of three working groups establghed by t{e, IUGS-Subcommi-ssion on the

Systematics of igneous Rocks to resolve aspects of tlle problem are presented for discussion. New definitions of the above are

given, variouslyln mineralogical and geochemical terms, and a reviied sequence for the systematic classification of the rocks

is provided which integrates with the existing IUGS hierarchical system.

Kewords: lamprophlre, lamproite, kimberlite, kalsilitic rocks, melilitic rocks, leucitic rocks, classification.

Sovnrtarnn

l,a nomenclature et la classification du clan dgs tamFrophyres, lamproites, kimberlites, et des roches a kahilite, m€lilite,

et leucite n'ont jamais 6t6 d6finies de fagon satisfaisante. Nous presentons i"i p9* 1* de discussion les recommandations de

trois groupes de travail 6tablis par la sous-commission charg6e dL la syst6matisation des roches igrdes de I'Union Intemationale

Ces S"cientes G6ologiques. Nous pr6sentons des d6finitionJnouvellei des membres de ce clan, en tennes soit min€ralogiques'

soit gdochimiques, et une s6quende nouvelle de classification syst6matique, int6gr6e au sch6ma hidrarchique du systbme existant

de ITUGS' (fraduit par la R6daction)

Mots-cl6s: lamprophyre, lampro'ite, kimberlite, roches i kalsilite, roches i m6lilite, roches l leucite, plassification'

* Recommendatrons of the Intemational Union of Geological Sciences (IUGS), Subcommission on the Systematics of Igneous

Rocks.t Deceased.

176

IlqrnonucttoN

In the book "A Classificarton of lgneous Roclcs andGlossary of Terms: Recommendations of theInternational Union of Geological SiiencesSubcornmission on the Systematics of lgneous Rocks.(LelVlaite et al. 1989),lamprophyres, lamproites andkimberlites were amalgamated under the hearting"lamprophyric rocksoo. This was regarded as aprovisional expedient until, as noted in the heface ofthe book, a more satisfactory classification could beestablished. In the book "Lamproplryres", Rock (1991)went even further and placed lamproites andkimberlites in his "lamprophyre clan',. However, tomany petrologists, inclusion of lamproite andkimberlite with the lamprophyres is either inappro-priate or incorrect because the terrn tampropSe isdevalued, whereas the classification of lamproites andkimberlites is not furthered.

During a meeting of the Subcommission at theIntemational Geological Congress in Washington in1989, three working groups were established toreconsider the classification of these rocfts, togetherwith certain feldspathoidat and leucitic, melilitii, andkalsilitic rocks, which are also inadequately defined inLe Maitre et al. (1989). The working groups includedmembers of the Subcommission, and severalspecialists, particularly in the areas of kimberlitic andlamproitic rocls, were invited to participate. Thiscontribution brings together the findings of theworking groups and the Subcommission on tlesetopics, and is presented as the best compromisecurrently achievable. The recommendations arepresented for discussion and should not be regarded asthe definitive statement on the topic.

The objective of the Subcommission is not to crearca highly detailed ttxonomy, either mineralogical orchemical, but to provide a broadly based classificationto be _used in a logical manner by any geologist.The- -classification suggested follows tle principtesestablished by Le Bas & Streckeisen (1991).-

The classification of these exotic alkaline rocks isnot a trivial tash and there will probably never be, atleast in our lifetimes, a truly practical and workablescheme. Advances have been made in the past twodecades in kimberlite and lamproite nomenclature,prinarily because of their economic importance. Incontrast, the vast array of lamprophyres has notreceived the same amount of attention, and there willbe a slow natural evolution of our ideas regardingtheir classification as a consequence of subsequentinvestigatious.

INTEGRAIoN wrrr{ TIIEIUGS l:lrrnnncucAr, Sysru\4

Any classification should be capable of integrationwith the IUGS hierarchical system described by

TI{E CANADTAN MINERAIOGIST

Le Maite et al. (1989) and Le Bas & Streckeisen(1991, Fig. 8). In the published hierarchical sysrem,lamprophyres, lamproites and kimberlites werecombined under "lamprophyric rockso', but if thisportmanteau term is to be abandoned, then it must bereplaced by a logical hierarchy of classification forthese three groups of rocks that integrates them withthe melilitic and leucitic rocks.

Lamprophyres do not lend themselves to eithermineralogical or chemical classification on account oftheir very variable HrO and CO, contents, which areprimary in some instances and secondary in others, andwhich are not taken into account in most proposedclassifications. Nevertheless, the members of theworking groups and the Subcommission agreed almostunanimously that the following proposals were a stepin the direction of establishing a nomenclature withminimum ambiguify and maximum usefulness.

It must be sfiessed, however, that the hierarchicalsystem presented here is not definitive and has minimalgenetic conlent. Any classification of a rock as alamprophyre, kimberliie or lamproite using this systemshould be considered provisional until further investi-gations have been undertaken using the specializedliterature as a key to correct classification.

h.oposED Svsrura FoR TIrE Ctassmcetowor Msr,trffIc, Kersrrnc, KrvrsFRlmc. Latwnomc

AND LEUCmc Rocrs aNo Lavpnopnynrs

Classification of melilitic rocla

The melilite-bearing rock classification is used forrocks with >lUVo modal melilite. Triangular plots arepresented for plutonic (melilitolite) (Fig. 1) andvolcanic (melilitite) rocks @ig. 2), and Table I isintended for melilitic rocks containing kalsilite.

If the mode cannot be determined" then one shouldapply the total alkalis versus silica CIAS) chemicalclassification, as follows: (a) The rock should plot inthe foidite field. (b) If the rock does not containkalsilite but has larnite in the norm, then one shouldapply Figure 3. (c) If nomrative lamite is greater thanl07o and KrO is less than NqO (wt7o), then it is amelilitite or olivine melilitite. (d) If K2O is greaterthanNarO and KrO exceeds 2 wt%o, then it is a potassicmelilitite orpotassic olivine melilitite. The latter can betermed a lcanngife, which mineralogically is a kalsilite- leucite - olivine melilitite. (e) If normative larnite isless than lUVo,then the rock is a melilite nephelinite ora melilite leucitite according to the nature of thedominant feldspathoid mineral.

Classification of lealsilitic rocks

The principal minerals of the katsilitic roctrs includeclinopyroxene, kalsilite, leucite, meliliteo olivine andphlogopite (Iable 2). These rocks cannot be called

pyroxenemelilitolite

pyroxeneolivine

melilitolite

melilile-bearing peridotite and pyroxenite

CI.-ASSIFICATION OF LAMPROPTIYRES 177

Mel

Ol CpxPlutonic rocks

Frc. 1. Classification of the plutonic (melilitolite) meliliticrocks with modal melilite >10Vo O-e Maifre er al, 1989Fig. B.3).

pyroxenite because that lerm is reserved for plutonicrocls. The rock types maftrite and katungite, togetherwith the closely associated leucitic rock ugandite(which is excluded from Table 1. as it does not containkalsilite and is more logrcally classified as an olivineleucitite), constitute the kamafugitic series of Sahama

Ol CpxVolcanic rocks '

Ftc. 2. Classification of the volcanic (meilitite) meliliticrocks with modal melilite >10Vo (Le Maitre er al. 1989,Fig.8.3).

TABIfi 1. SUGGESTED NOMENCLAN'RE FOR

KAMAFUGITIC ROCKS

Rocknane MineralAssemblage

MafuriteKafilngite

Venarzite

Coppaeute

Olivine-pyroxene kalsilititeKalsilite-leuote-olivine melilitite

Kalsitt+,pblogopite-olivine-leucite nelilitite

Kalsilite-phlogopit€ melilitite

(1974). From the point of view of the IUGS system ofclassification, the presence of essential melilite orleucite (or both) indicates that either the classificationthat deals with melilitic or leucitic rocks should beapplied. However, the presence of kalsilite aud leuciteis considered petrogenetically so distinctive andimportant that the accepted term, kamafugite, shouldbe retained for this consanguineous series of rocks.Table I indicates their nomenclature as a function ofmineral assemblage.

Plutonic kalsilitic rocks of the Aldan and NorthBaikal petrological provinces of Russi4 which are notkamafugitic, may be distinguished by the prefixo'kalsilite". Thus, synnyrite becomes kalsilite syenite,and yakutite becomes kalsilite-biotite pyroxenite.

Clas sffi c ation of kimb e rlite s

Kimberlites are currently divided into Group I andGroup tr (Smith et al. 1985, Skinner 1989). Group-Ikimberlites correspond to archetypal rocks fromKimberley, South Afric4 which were formerly termed"basaltic kimberlites" by Wagner (1914). Group-trkimberlites correspond to the micaceous or lampro-phyric kimberlites of Wagner (1914).

Petrologists actively studying kimberlites haveconcluded that there are significant petrologicaldifferences between the two groups, although opinionis divided as to the extent of the revisions required totheir nomenclature. Some wish to retain the sntus qw(Ski:rner 1989), whereas others (Mitchell & Bergman1991, Mtchell 1994b) believe that the terminologyshould be completely revised (see below). Regardless,the working group is unanimous in agreeing that asingle definition cannot be used to describe both rocktypes. Because of the mineralogical complexity of therocks, a simple succinct definition cannot be devised.Following a concept originally developed by Dawson(1980), the rocks may be recognized ns containing acharacteristic assemblage of minerals.

The following characterization of Group-Ikimberlites is a.fter Mitchell (1995), which is basedessentially on that of Mirchell (1986, 1994b), and

Mel

melalite-bearing ultramaflc volcanics

178 T}IE CANADIAN MINERALOGIST

A

AAAA

F'

evolved from earlier "definitions" given by Clementet al. (1984) and Mtchell (1979).

Group-I kimberlites consist of volatile-rich (domi-nantly COr) potassic ultrabasic rocks commonlyexhibiting a distinctive inequigranular texture resul :ngfrom the presence of macrocrysts (a general term forlarge crystals, fypically 0.5-10 mm in diameter) and,

TABII 2. IVID].IERAL ASSEMBLAGES OF (ALSII,ITTC RocKs

Phl Cpx Lct Kal Mel Ol

M a f t r i t e - x - x xK a t u n g r t e x x x x x

V e n a D z i t o r r x x r r

C o p p a e l i t e x x - x x -

A MELIL IT ITE

X MELIL ITE NEPHELINITE

rxJ - -{ rr l tr f

II

in some cases, megacrysts Qarger crystals, t,?ically1-20 cm) set in a fine-grained matrix. The assemblageof macrocrysts and megacrysts, at least some of whichare xenocrystic, includes anhedral crystals of olivine,magnesian ilmenite, plrope, diopside (in some casessubcalcic), phlogopite, enstatite and Ti-poor chromite.Olivine macrocrysts are a characteristic and dominantconstituent in all but fractionated kimberlites. Thematrix contains a second generation of primaryeuhedral to subhedral olivine, which occurs togetherwith one or more of the following primary minerals:monticellite, phlogopite, perovskite, spinel (magnesianulviispinel - magnesiochromite - ulvdspinel -magnetite solid solutions), apatite, carbonate andserpentine. Many kimberlites contain a late-stagepoikilitic mica belonging to the barian phlogopite -kinoshitalite series. Nickeliferous sffides and rutileare common accessory minerals. The replacement ofearlier-formed olivine, phlogopite, monticellite andapatite by deuteric serpentine and calcite is common.Evolved members of the Group may be poor in,

20 A A K 6 4

A A L A

Aur 18F

2 rcE

J 1 4ul

? tz= 1 0(Eoz 8

A

2 4 6 8 ' t o 1 2 1 4 1 6 1 8 2 0 2 2 2 4 2 6 2 8 3 . O

NORMATIVE NEPHELINEFlc. 3. Plot of normative nepheline versus normatle lamite for melilitic lavas showing

the proposed boundary (dashed line) between the melilitites and melilite nephelinites.The data plotted are taken from eastem, western atd southem Afric4 Europe, theformer USSR, Australia, and U.S.A., including Hawaii, and are variously described asolivine melilitite, melilitite or melilite nephelinite (see Le Bas 1989, Table 2).Also plotted are typical potassic melilitic volcanic rocks: A, melilite ankmatrite;K, katungite; L, leucite melilitite; M, potassic olivine melilitite; V, venanzite ftalsilite- phlogopite - olivine - leucite melilitite).

oo {io"oo

4 '+ ^"a'ffi:bf{

Symbolg: Phf phbgopite, Cpc olimpyroxrc, kt ltuire, Msl ndflita, OL olivine,Gb: glas. x prcem, -: ab&d. Afrrr Mjtc,hel & Bsgoln (t991, Table 2.3).

CI/,SSIFICATION OF I.AMPROPHYRES 179

or devoid of, macrocrysts, and composed essentially ofsecond-generation olivine, calcite, serpentine andmagnetite, together with minor phlogopite, apatiteand perovskite.

It is evident that kimberlites are complex hybridrocks in which the problem of distinguishing theprimary constituents from the entrained xenocrystsprecludes simple definition. The above characteriza-tion at0empts to recognize that the composition andmineralogy of kimberlites are not entirely derived froma parent magma, and the nongenetic terms macrocrystand megacryst are used to describe minerals ofcryptogenic, i.e., unknown, origin. Macrocrysts includeforsteritic olivine, chromian pyrope, almandine-pyrope, chromian diopside, magnesian ilnenite andphlogopite crystals, that are now generally believed tooriginate by the disaggregation of mantle-derivedlherzolite, harzburgite, eclogite and metasomatizedperidotite xenoliths. In most cases, diamond, which isexcluded from the above "definition", belongs to thissuite of minerals but is much less common. Megacrystsare dominated by -agnesian ilnenite, titanian pyrope,diopside, olivine and enstatile that have relativelyCr-poor compositions (<2 vttvo CrrOr). The origin ofthe megacrysts is still being debated (e.9., Mirchell1986), and some petrologists believe that they may becoguate. Both of these suites of minerals are includedin the characterization because of their commonpresence in kimberlites.

It can be debated whether reference to these charac-teristic constituents should be removed from theoodefinition" of kimberlite. Strictly, minerals that areknown to be xenocrysts should not be included in apehological definition, as they have not crystallizedfrom the parental magma. Smaller grains of both themacrocryst- and megacryst-suite minerals also occurobut may be easily distinguished on the basis of theircompositions. In this respect, it is important todistinguisl' pseudoprimary groundmass diopside frommacrocrystic or megacrystic clinopyroxene. Group-Ikimberlites do not usually contain the former except asa product of crystallization induced by the assimilationof siliceous xenoliths (Scott Smith et al. 1983). T'heprimary nature of groundmass serpophitic serpentinewas originally recognizedby Mitchell & Putnis (1988).

Recent studies (Smith et aI. 1985, Skinner 1989,Mitchell l994b,lggs,Tainton & Browning 1991) havedemonstrated that Group I and Group tr kimberlites aremineralogically different and petrogenetically separaterock-types. A definition of Group-tr kimberlites hasnot yet been agreed upon, as they have been insuffl-ciently studied. Mitchell (1.994b, 1995) has suggesredthat these rocks are not kimberlitic at all. and should betermed o'orangeite', in recognition of their distinctcharacter and unique occturence in southern Africa.Wagner (1928) previously suggested that the rockswhich he initially termed micaceous kimberlite(Wagner L91.4) be renamed "orangite" (src,). The

following characterization of the rocks currentlydescribed as Group-II kimberlites or micaceouskimberlites follows that of Mitchell (1995).

Group-tr kimberlites (or orangeites) belong to a clanof ultrapotassic, peralkaline rocks rich in volatiles(dominantly H2O), characterized by phlogopitemacrocrysts and microphenocrysts, together withgroundmass micas that vary in composition fromphlogopite to'lefiafeniphlogopite". Rounded Inacro-crysts of olivine and euhedral primary crystals ofolivine are common, but are not invariably majorconstituents. Characteristic primary phases in thegroundmass include: diopside, commonly zoned to,and mantled by, titanian aegirine; spinels rangingin composition from Mg-bearing chromite toTi-bearing magnetite; Sr- and kEE-ich perovskite;Sr-rich apatite; REE-ich phosphates (monazite,daqingshanite); potassian barian titanates belonging tothe hollandite $oup; potassium triskaidecatitanates(K.{!g'O2); M-bearing rutile and Mn-bearingilmenite. These are set in a mesostasis that may containcalcite, dolomite, ancylite and other rare-earthcarbonates, witherite, norsethite and serpentine.Evolved members of the group contain groundmasssanidine and potassium richterite. Zirconium silicates(wadeite, zircon, kimzeyitic garnet, Ca-Z-silicate)may occur as late-stage groundmass minerals. Barite isa common deuteric secondary mineral.

Note that these rocks have a greater mineralogicalaffinity to lamproites than to Group-I kimberlites.However, there are significant differences in thecompositions and.overall assemblage of minerals, asdetailed above, that permit their discrimination fromlamproites (Mitchell L994b, 1995).

Clas s ffi cation of lamp roites

The classification system of lamproites describedby Mitchell & Bergman (1991) is recommended; itinvolves mineralogical and geochemical criteria, asfollows:

Lamproites are characterized by the presence ofwidely varying amounts (5-90 vol.Vo) of the followingprimary phases: (1) titanian (2-10 wt%o TiO),aluminum-poor (5-12 wtVo Al2O) phenocrysticphlogopite, (2) titanian (5-10 wt%o TiO) groundmasspoikilitic "tetraferriphlogopite", (3) titanian (3-5 vtt%oTiO) potassium (4-6 wt%o K2O) richterite, (4)forsteritic olivine, (5) aluminum-poor (<l t*,t%o N2O3),sodium-poor (<l wtvo Na2O) diopside, (6) non-stoichiometric iron-rich (14 wt%o FqOr) leucite, and(7) iron-rich sanidine (typicafly L-5 vttvo Fe2Or). Thepresence ofall the above phases is not required in orderto classi$ a rock as a lamproite. Any one mineral maybe dominant, and this, together with the two or tlreeother major minerals present, suffices to determine thepetrographic name.

Minor and common accessory phases include

180 TIIE CANADIAN MINERALOGIST

priderite, wadeite, apatite, perovskite, magnesio-chromite, titanian magnesiochromite, and magnesiantitaniferous magnetite; less commonly, but characteris-tically, jeppeite, armalcolile, shcherbakovite, ilmeniteand enstatite also are present.

The presence of the following minerals precludes arock from being classified as a lamproite: primaryplagioclase, melilite, monticellite, kalsilite, nepheline,Na-rich alkali feldspar, sodalite, nosean, hauyne,melanite, schorlomite or kimzsyite.

Lamproites conform to the following chemicalcharacteristics: molar K2O/1.{a2O > 3, i.e., ulfiapotas-sic, (2) molar KrO/Al2O3 > 0.8 and commonly > L,(3) molar (KrO + NqO)/Al2O3 typically > | i.e.,peralkaline, (4) typically <1,0 wtVo each of FeO andCaO, TiO, l-:7 v,ttVo, >2000 and commonly >5000 ppmBa, >500 ppmzx, >1000 ppm Sr and >200 ppm La.

The subdivision of the lamproites should follow thescheme of Mitchell & Bergman (1991), in whichthe historical terminology is discarded in favorof compound names based on the predominance ofphlogopite, richterite, olivine, diopside, sanidine andleucite, as given in Table 3. It should be noted that theterm "madupitid' in Table 3 indicates that the rockcontains poikilitic groundmass phlogopite, as opposedto phlogopite lamproite, in which phlogopite occurs asphenocrysts.

The complex compositional and mineralogicalcriteria required to define lamproites result from thediverse conditions involved in their genesis, comparedwith those of rocks that can be readily classified usingthe IUGS system. The main petrogenetic factorscontributing to the complexity of composition andmineralogy of lamproites are the variable nature oftheir metasomatized source-regions in the mantle,depth and extent of partial melting, coupled with theirconrmon extensive differentiation.

TABLE 3. NOMENCT-ATURB OF LAMPROITtsS

llislorical name Revisod name

Classification of volcanic leucitic roclcs

The leucitic rocks. after elimination of thelamproites and kamafugites, should be namedaccording to the QAPF (volcanic) diagram with theprefix '1eucite" or ooleucite-bearing" as appropriate.Rocks containing liftle or no feldspar, i.e., falling infield 15 (foidite) of QAPF 0-e Maitre et al. 1989,Fig. B.10), are leucitites, which are divisible into threesubfields: (a) QAPF subfield 154 phonolitic leucitite:Foids 60-907o of light-colored constituents, and alkalifeldspar > plagioclase; (b) QAPF subfield l5b,tephritic leucitite: Foids 60-907o of light-coloredconstituents, and plagioclase > alkali feldspar; (c)QAPF field 15c, leucitite sensu stricto: Foids 9G-1007oof light-colored constituents, and leucite practically thesole feldspathoid.

The essential mineralogy of the principal groups ofleucitic rocls is given in Table 4. No unambiguouschemical criteria have been found to distinguish thisgroup of rocks. In terms of the TAS plot, leucititesextend significantly beyond the foidite field intoadjacent fields @g. 4). They are better distinguishedfrom lamproites by other compositional parameters,although even here some overlap occurs. The chemicalcharacteristics of the potassic rocks, and attempts atdistinguishing lamproites from certain leucitic rocks,using a variety ofcriteria, are explored byFoley et al.(1987) and Mitchell & Bergman (1991).

Cl as s ifi c ati on of Lamp r o p hy r e s

Lamprophyres are mesocratic to melanocraticigneous rocks, usually hypabyssal, with a pan:idio-morphic texture and abundant mafic phenocrysts ofdark mica or amphibole (or both) with or withoutpyroxene, with or without olivine, set in a matrix of thesame minerals, and with feldspar (usually alkalifeldspar) restricted to the groundmass. For the generalclassffication of these rocks, see I* Mal:tre et aI.(1989), and for detailed descriptions, see Rock (1991).

TABLE 4. MINERAIOGY OF TI{E PRINCIPAL ROUPSOFLEUCIT1CROCKS

cpx

Wyoniagite

Orendite

Madupite

Cedricite

Masilite

Wolgidit€

Fihoyilo

Verite

Jumi[ite

Fortnite

Cancalitc

diopside-leucite-pblogopite lsmproite

diopside-sanidine,phlogopite lamproile

diopsftle nafupitic lamproite

diopsftte.leucite lamproite

leucite-richtedte lamproite

diopside-lercite-richerie madupitic lamproite

lerrite-phlogopite lamproite

hyaloolivinediopsidepblogopite laoproite

olivine.diopside-richterite madupitic laqtroite

hyalo-enstatite-phlogopie lamFroite

ensutite-sanidin+'phlogopite laryroite

I*ucitite x

Tephritic leucitite r

Phonolitic leucitite rkucitct@hdte x

L4cito basanite rLeucitephonolit€ x

x

x

I

x

x

x

x > x

x < r

x -

x -

>10%

<to%>to%

$llnbols Cpc dinopyrorog I,a: lruite Pt plagioolase, Ss: midiao (lroducn olits emoluion), OL olivine. x prs€il, -: abs€d. A[ ths roc&s nay @dab somon€pbslirc.after Mitchell & Bergoan (1991)

CT.ASSIFICATION OF I.AMPROPFIYRSS

37 41 45 49 53 s7 61 65 69 73SiOz wt%

FIG. 4. Percentage frequency disribution diagram for 112 samples of leucitite plottedin the TAS diagram, showing that only about 507o plot in the foidite fiel{ with majorconcentrations in the basanite-tephrite and phonotephrite fields. Reproduced fromFig. 23 oflr Bas et al. (1,992), Star indicates the peak ofthe frequency distribution.

181

;eEtoqY8o

AI

to

The Subcommission no longer endones the terms'lamprophyric rocks", or'lamprophyre clan", as usedby Le Maitre et al. (1989) and Rock (1991) toencompass lamFrophyres, lamproites and kimberlites,because lamproites and kimberttes are best consideredindependently of lamprophyres.

Rsusm SseuBrflAL Sysrsr,r ronCrassn'yNc Iovnous Rocrs

The revised hierarchy, which modifies that of thewall chart accompanying Le Maitre et al. (1989) anidFigure 8 of Le Bas & Sneckeisen (L991), is givenbelow @igs. 5, 6). Each statement is a precondition forthe next in the sequence.

1. If the rock is fragmental, the rock should beclassified as aPYROCLASTIC ROCK.

2. If the rock contains more than 50Vo pimarycarbonate. the rock should be classified as aCARBONATITB.

3. If the rock contains >107a modal melilite andM > 90Vo. the rock should be classified as aMELILITIC ROCK.

4. If the rock contains kalsilite. the rock should beclassified as a KALSILITIC ROCK.

5. If the rock is fine-grained or glassy and has larnitein the norm. the rock should be classified as aMELILITICROCK.

6. Ifthe volcanic rock contains essential leucite" withor without phlogopite (biotite), or is from a minorintrusion vdth mafic phenocrysts only (generally micaor amphibole, or both), apply the following criteriasequentially:(a) If the rock is free of leucite but rich in olivine(typically 35-55 modal % including macrocrysts,xenocrysts and phenocrysts), and one or morgdominaal primary minerals in the groundmass aremonticellite, phlogopite, carbonate, serpentine ordiopside, the rock should be classified as aKIMBERLITE.@) If the rock contains titanian phlogopi0e as Al-poorphenocrysts or groundmass grains (or both), togetherwith common Fe-rich leucite or forsteritic olivine(or both), as well as one or more of titanian potassiumrichterite, Al-Na-poor diopside, Fe-rich sanidine,accessory wadeite and priderite in the groundinass,then the rock should be classified as a LAMPROITE.(c) Any rcrnnining leucitic rocks should be classifiedusing the terms provided in the classification ofVOLCANIC LEUCMC ROCKS.(d) Apart from certain pyroxene- and olivine-bearingrocks, e.g., ankaramite and oceanite, which are notlamprophyres (for tlese rocks proceed to the

Use pyroclastic rockclassification

Use charnockiteclassification

Use ultramaficclassification

Is ithigh-Mg ?

Use TAS. If it falls in fields F or Ul.use nonn ne r. norn ab classification

182 THE CANADIAN MINERALOGIST

Ftc. 5. Flow chaxt (after Le Bas & Streckeisen 1991) forthe classification of igneous rocks following the IUGSscheme Q.e Maitre er al. 1,989), but modified for theproposals presented in this report paper (dashed box).The 'horm ze versus norm ab classification" is that sivenby Le Bas (1989).

QAPFITAS systems), any remaining rocks containingphlogopite (biotite) or amphibole (or both) and thoseftom minor intrusions with only mafic phenocrysts, goto the LAMPROPIIYRE classification (Le Maitreet al.1989).

7. If the rock is glassy or the mode is not otherwisedeterminable, the rock may be lamproitic, if molarK2O/NqO is greater than 3, molar K2O/N2O3 greaterthan 0.8 and molar (KrO + NqO)/AlrO, greater than I(peralkaline); then go to the LAMPROITE classi-fication.

8. The classification ofthe charnockitic, plutonic andrcmaining volcanic rocks proceeds as described inte Maihe et al. (1989).

TnE PRoBLEMS oF CI-A.SSFICATIoN

Many schemes of classification for igneous rocks,such as that based on the total alkali versr.s silicamethod (TAS), have a major petrogenetic component,and ultimately all taxonony of igneous rocks v/illincorporate genetic factors. For no other igneous rocksis the petrogenetic component of classification moreimportant than for lamprophyres, lamproites andkimberlites. One criticism of the IUGS classification ofrocks, discussed by Le Bas & Streckeisen (1991), isthat it is deficient in a strong genetic component. Thisis essentially an historical artifact of the pragmaticapproach originally adopted by Streckeisen to attain aconsensus. Because of the plethora of classificationsystems that have been suggested and applied in thepast, the Subcommission followed a consensualapproach. The widespread adoption of the IUGSsystem would seem to havejustified this approach formany igneous rocks. However, although much of theIUGS system undoubtedly has some petrogeneticsignificance, and is used in genetic discussion, forexample the TAS system, purely descriptive terms mayhave to be applied where there is disagteement as tointerrelationships of rock suites. This would seemto apply at present to the lamprophyres, for whichdifficulties remain in erecting a petrogenetic classi-fication.

The main problems of the classification of lampro-phyres, lamproites, kimberlites, feldspathoidal,kalsilitic, melilitic, and leucitic rocks are brieflyoutlined below.

Inrnprophyres

The lamprophyres are a complex gtoup of rocksthat have mineralogical similarities to some kimberlitesand lamproites. Lamprophyres are difificult to classiSunambiguously using existing criteria. They are notamenable to classification according to modalproportions, such as the system QAPF, norcompositional discrimination diagrams, such asTAS (Le Mure et al. 1989). It seems unlikely thata simple taxonomic system viill be found unlessappropriate genetic criteria are applied, that is, unlessthe classification takes into account the genesis ofthe rocks.

The term "lamprophyre" was introduced by vonGumbel in 1874 for a group of dark rocls that formminor intrusions, contain phenocrystal brown mica andhornblende, but lack feldspar phenocrysts. Followingits introduction, the term was used by Rosenbusch(1877) to encompass a wide variety of hypabyssalrocks containing ferromagnesian phenocrysts, e.9.,minette, kersantite, camptonite and vogesite.Eventually, spessartite, monchiquite and alnoite alsowere included in the group. Thus, the group became arepository for any mafic-phenocryst-rich rock that was

use melilitic rocks classificationmelilite > 1070

contains kalsilite

leucite orminor intrusion with

only mafic phenocrysts + l c t + o l +

K2O/Na2O > 3.0molar K2O/AI2O3> 0.8

and peralkaline

CLASSIFICATION OF LAMPROPIIYRES 183

difficult to categorize. Subsequently, Middlemost(1986) and Rock (1986, 1991) extended the definitionfurther to include kimberlites, lamproites and evenrocks containing feldspar and leucite phenocrysts.Reviews of the history of lamprophyre nomenclaturecan be found in Rock (1991), Mitchel & Bergman(1991) and MtcheU (1994a).

Rock (1991) used 'lamprophyre" synonymouslywith "lamprophyte clan',, a term that includedlamprophyres, lamproites and kimberlites. To manypetrologists working on lamproites and kimberlites,"lamprophyre" is an inappropriate general term.Consequently, the working group and Subcommission6vsrwfuelmiagly rejected the use of the term"lamprophyre clan" to encompass lamprophyres (sensuRosenbusch), kimberlites and lamproites.

Mitchell (L994a) discussed at some length theinadequacies of the concept of the 'lamprophyre clan"and proposed adoption of the term "lamprophyrefacies" to convey the concept that some members of apetrological clan crystallized under different conditionsthan other members of the clan. Mitchell's approach tothe problem is determined by a conviction that systemsof classification should be genetic in character.

The working groups found it impossible to devisea definition of'lamprophyric rocks" that is not sobroad as to be almost petrologically meaningless.For the lamprophyres (sezsa Rosenbusch or Rock),the working group could not draft a satisfactorydefinition, in part because Rock (1991) included anumber of rock types that differ mineralogically from

Fto. 6. Flow chart for themelilitic. kalsilitic andleucitic rocks and thekimberlites, lamproitesand lamprophyres (dashedbox ofFig. 5). It is enteredafter the "carbonates>50Vo" box and exits tothe "charnockitic" box ofFig. 5. The symbols usedfollow those of Kretz(1983) where possible.

the generally accepted characteristics of lamprophyres.However, of greater significance is the realizationthat members of the lamprophyre group havedistinctly different origtns, and thus it is unwise todescribe or group together rocks that are geneticallydifferent.

Among the rocks included in the lnmprophyre groupby Rock (1991) are alnoite and polzenite. Thesecontain more than L0 vol.Vo melilite, and thus are nowconsidered as varieties of melilitic rocks. $imililly, ftsabundant carbonate in aillikite suggests that it may beconsidered a variety of silicocarbonatite.

Minette is a biotite-rich lamprophyre. However,certain mica-rich rocks that usually occur in minorintrusions should not be called minette because themica is phlogopite (commonly titaniferous), andthe rock is alkaline. Such rocks might be beftertermed 'alkali minette" rather than "glimmerite" or"phlogopitite", as it is the alkalinity rather than theFe/IVIg ratio that is their characteristic feature.

The widespread usage of the term lamprophyre inEnglish, French and German petrological literature isin marked contrast to its infrequency in the extensiveRussian literature. There the term is usually reservedfor alkaline rocks. Specific rock types, such askersantite" are considered as varieties of diorite.Similarly, camptonite and alnoite are regarded asvariants of gabbro and melilitite, respectively(Atdreeva et a\.1.985, Kononova 1984). This approachto lamprophyre nomenclature 15 5imil6 to the faciesconcept proposed by Mitchell (L99ad.

184 TIIE CANADIAN MINERALOGIST

Lamproites

Lamproites have always been a difficult groupof rocks in terms of their identification andnomenclafure. Recent interest in them has beenprompted by the discovery of economically viablediamond-bearing varieties, which has led to adetailed re-examination of the group and completerevision of the nomenclature (Scott Smith &Skinner 1983, Mitchell & Bergman 1991). Theserevisions have resulted in the reclassification aslamproites of some rocks previously regarded askimberlites (e.g., Prairie Creek, Arkansas, andMajhgawan, India).

It is only in the last few decades that lamproiteshave been considered to crysrallize from a distincttype of magma. Formerly, the presence of leucite,the similarity of some olivine-bearing lamproiteto kimberlite, and the presence of some"lamprophyrico' characteristics, e,g., abundant pheno-crysts of pblogopite, led not only to a plethora ofnames, but to an ill-defined place in petrologicaltaxonomy. The problem was further exacerbated bytheir geochemical characteristics, with the magrnachemistry stabitzing a large number of unusualminerals (K-Ba titanates and silicates, K-Z silicates),some of which can be used as discriminants for theserocks. For most examples ef lamproite, the presence ofthese minerals reflects derivation from a sourseenriched in incompatible and large-ion-lithophileelements; this enrichment distinguishes lamproiticrocks from Group-I kimberlites 6d lam.Frophyres.A full discussion of the nomenclature of lamproites,their relationships to other potassic and ultrapotassicrocks, in both mineralogical and chemical terms,together with a suggested revised nomenclature, isfound in Mitchell & Bergman (1991). TheSubcommission essentially accepted the detailed workof Mitchell & Bergman (1991) and Scott Smith &Skimer (1983), and needed only to integrate theserocks into the IUGS hierarchical svstem.

Kimberlites

The Subcommission considered it inappropriateto re-investigate the nomenclature and definition ofkimberliles in detail, because this has been doneextensively in the last few years by the manyspecialists of kimberlites (Skinner & Clement 1979,Dawson 1980, Clement et al. 1984, Mitchell 1986).Nevertheless, a clear definition of kimberliteshould be formulated, particularly for purposes ofdist''rguishing these rocks from olivine lamproite,and for placing kimberlites in the hierarchicalclassification system. Currently, the classification ofkimberlite is undergoing revision, and the nomen-clature advanced by Mtchell (1994b) has not yet beentully explored.

Melilitic roclcs

The main problems here concern volcanic meliliticrocks. In Le Maitre et al. (1989), the coarser-grainedmelilitolite is classified on the basis of modalproportions, but a completely satisfactory classificationfor the finer-grained rocks was not attained. Adefinition based on rock chemistry is desirable, butunfortunately these rocks cannot be distinguishedadequately ftom other volcanic rocks in the TASsystem. However, the presence of melilite in more thantrace modal amounts results in the formation of lamite(or calcium orthosilicate) in the CIPW norm, and thiscan be used as a potential discriminant. Althoughlamite may appear in the norm of some melilite-freenephelinitic rocks containing clinopyroxene rich in theTschermaks component (H.S. Yoder, Jr., pers. comm.),we find that the lafier is typically expressed in the normas anorthite.

A further problem remains, that there is a continuousseries from melilitite, tlrough melilite nephelinite tonephelinite. Investigation of this problem indicates thata reasonably clear discriminant between melilitite andmelilite nephelinite is normative larnite. In Figure 3,samples of melilitite and nephelinite are plotted interrns of normative lamite versus notmatle nepheline.The best division appears to be at LlVo larrtrte.

When classilying the melilitic rocks, the followingshould be taken into consideration: (a) The presentclassification for melilitic rocks in I* Mao.ne et al.(1989, p. 12) is based on the presence of modal meliliteexceeding l0 vol.Vo in either plutonic (melilitolite) orvolcanic (melilitite) occlurences, in combination withM > 9AVo. O) In the IUGS scheme (see flow chartaccompanying Le Maitre et al. 1989), their identifica-tion is made after excluding the lamprophyres butbefore entering QAPF. It is now considered preferableto identify melilitic rocks before lamproites,kimberlites and lamprophyres. (c) Even in fine-gainedrocks, melilite can usually be idenffied in thin sectionwhere it occurs in essential proportions, i.e.,>10 voLvo. This assumes the rock is not altered; if it is,melilite is usually carbonated. (d) Some fine-grainedmelilitic rocks are strongly potassic, e.g. katungite, thepotassic character usually being reflected in thepresence of modal leucite or kalsilite (or both). (e)Melilitite is characterized by the presence of meliliteand perovskite and contains less than 38 wtVo SiO2 andgreater than 13 WVo CaO.

Kalsilitic rocl<s

Kalsittic rocls have not previously been consideredby the Subcommission. These fall into two groups: thekamafugrtic series of Sahama (1974), and the kalsilite-bearing syenites and pyroxenites, e.9., synnyrile andyakutite, occurring in the Aldan and North Baikalpetrological provinces ofRussia (Kogarko et al.1995,

CLASSIFICATION OF T.AMPROPTIYRES 185

Kostyuk et al. 1990). Some of the kamatugitic rockscontain leucite or melilite (or both) and might beconsidered feldspathoidal or melilitic rocks. However,the presence of kalsilite is considered so important thatit requires assignment ofthese rocks to a special group.

Foiditic and leucitic rocl<s

The problems posed by these rocks are restricted tothe fine-grained members. The coarser-grained rocksof equivalent composition containing nepheline orleucite (or both) and, despite heteromorphism in somecases (Yoder 1986), can be satisfactorily classifiedusing the QAPF and other mineralogical systems(Le Maitre et al. L989).

The boundaries between the feldspathoidal field andthe basanite+ephrite, phonotephrite, tephdphonoliteand phonolite fields in the TAS system are not whollysatisfactory, as they do not provide an acceptableboundary for the nephelinitic and leucitic rocls. Theproblem with regard to the leucitic rocks is illustratedby Figure 4.

It is evident that leucitic rocks cannot be distin-guished chemically on the TAS diagram. However, asleucite is, with few exceptions, a phenocryst phase, orforms small but identifiable crystals, a modal systemshould be feasible. This approach was not adopted byLe Maitre et al. (L989). The distinction betweennephelinite and basanite has been considered by Le Bas(1989).

AcIC{OWLEDGENmNIS

The authors thank their colleagues in the workinggroups and the Subcommission for their constructiveapproach to the controversial problems encountered.The members of the working groups and otherswho contributed include D.S. Barker, J.L. Brandle,A. Cundari, S.V. Efremovq J. Keller, S. Kravchenko,E. Lazko, A.N. Mcl.aurin, H.O.A. Meyer, E.A.K.Middlemost, T.F.D. Nielsen, A. Streckeisen,E.J. Visentin, K.Yag1, H.S. Yoder, Jr. and B. Zanettin.We acknowledge the numerous useful suggestions oftwo referees, particularly Frank Dudas, and of theEditor for his major "tidying" operations on themanuscript.

Sadly, Nick Rock died in February of 1992, beforehe saw the final dmft of these recommendations. He isincluded as a coauthor here not only because of hissignificant participation in the discussions of theworking groups, but also because of his immensecontribution to the investigations of lamprophyres andrelated rocks and the enthusiasm with which he studiedthem. Although he may not have agreed with all theideas and conclusions presented in this reporl he sawmost of its components, supporled them, and wasactive in their discussion almost to the final stage,including the meeting in Brazil (1991), during the Fffi

International Kimberlite Conference, when agreementwas reached on some of the major problems.

RSFERB{cEs

ANDREVA, E.D., Basrore, V.A. & BocArtr<ov, O.A. (1985):Magmatic Rocks: Classification, Nomenclature andPetrography. Nauka, Moscow, Russia (2 volumes).

Crrrranrn, C.R., Srnwnn, E.M.W. & Scom Surru, B.H.(1984): Kimberlite redefined. J. Geol. 92,223-228.

DAwsoN, J.B. (1980): Kimberlites and Their Xenoliths.Springer-Verlag, Berlin, Germany.

FoLEy, S.F., VENTURET, G., Gnm{, D.H. & ToscaNr, L.(1987): The ultrapotassic rocks: characteristics, classifica-tion, and constraints for petrogenetic models Earth-Sci.Rev. ?4.81-134.

VoN GuMBEL, C.W. (1874): Die PaleolithischenEruptivgesteine des Fichtelgeberges. Ftanz, Munich'Germany.

Kocamo, L.N., KoNoNova, V.A., ORLovA, M.P. &Woorrrv, A.R. (1995): Alkaline Rocl<s and' Carbonatitesof the WorM. 2. Former USSI?. Chapman & Hall, London,U.K.

KoNoNovA, V.A., ed. (1984): Alkaline Igneous Rocks. Nauka"Moscow. Russia.

KosryuK V.P., Par,rNe, L.I., Zmrov, A.Y., ORtovA, M.P.& Bea.ova, T.Y. (1990): Potassic Alkaline Magmatismof the Baikal-Stanovoy Rifting System. Nauka'Novosibirslg Russia.

Knsz, R. (1983): Symbols for rock-forming minerals.Am. MineraL 68"n7-n9.

Ls Bas, M.J. (1989): Nephelinitic and basanitic rocks.J. P etrol. 30, 1299 -L3 12.

Ls MAFRE, R.W. & Woor.trv, A.R. (1992): Theconstruction of the total alkali-silica chemical classifica-tion of volcanic rocks. MiruraL Petrol 46,l'22.

- & SrRrcrctsw, A.L. (1991): The IUGS systematicsof igneous rocks. ./. Geol. Soc. Inndon 14{i, 825-833.

Le MArrRE, R.W., Berruar, P., DUDEK, A., IfurER, J.,LATEYRE, J., Ln Bas, M.J., SABNE, P.A., Sctnao, R.,SOnr.NssN, H., Smscr<srssx, A., Woor,lrt, A.R. &Zawerrw, B. (1989): A Classificatian of lgrcow Rocksand Glossary of Terms: Recommendations of theInternational Union of Geological SciencesSubcommission on the Systelnatics of lgneous Rocks.Blackwell Scientific Publications, Odord, U.K.

MDDlio\losr, E.A.K. (1986): The nomenclature and origin ofthe noncumulate ultramafic rocks and the systematicposition of kimberlites. /z Extended Abstr., Fourth Int.Kimberlite Conf. (Perth). Geol. Soc, Aust. 16'12-74.

186 TIIE CANADIAN MINERALOGIST

Mncrru,, R.H. (1979): The alleged kimberlite-carbonatirerelationship: additional contrary mineralogical evidence.Am. J. Sci. U9, 570-589.

(1,986): Kimberlites: Mineralogy, Geochemistryanl Petrology. Plenum Press, New York, N,Y.

(1994a): The lamprophyre facies. Mineral. petrol.51.137-t46.

(L994b): Suggestions for revisions ro theterminology of kimberlites and lamprophyres from agenetic viewpoint, .fn hoc. Fifth Int Kimberlite Conf. 1.Kimberlites and Related Rocks and Mantle Xenoliths(H.O.A. Meyer & O.H. konardos, eds.). Companhia dePesquisa de Recursos Minerais (Brasilia), Spec. Publ.ua. 15-26.

- (1995): Kimberlites, Orangeites and Related Rocks.Plenum hess, New York, N.Y.

- & BERcuan, S.C. (1.991): Petrology of lamproites.Plenum Press, New York, N.Y.

& PurNrs, A. (1988): Polygonal serpentine insegregation-textured kimberlite, Can. Mineral. 26,991,-997.

Roctg N.M.S. (1986): The nature and origin of ultramaficlamprophyres, alniiites and allied rocks. J. Petrol. 27,155-196.

- (1991): Lamprophyres. Blackie, Glasgow, U.K.

Rosarnuscn, H. (1877): Mikroslapischz Physiographie dzrMineralien und Gesteine, Schweizerbart'sche Verlag-handlung, Stuttgart, Germany.

Sanaua, T.G. (1974): Potassium-rich alkaline rocks. lz TheAlkaline Rocks (H. Sgrensen, ed.). Wiley, London, U.K.(96-109).

Scorr Slfin{, B.H., Dencnn, R.V., Hannrs, J.W. & Srnecrc,K.J. (1983): Kimberttes near Orroroo, South Australia.Proc. Third Int. Kinberlite Conf, | (J. Komprobst, ed.).Elsevier, New York, N.Y. (121-142).

& SrrNNER, E.M.W. (1983): A new look at PrairieCreek" Arkansas. Proc. Third Int, Kimberlbe Conf. I(J. Komprobsg ed.). Elsevier, New York" N.Y. (255-283/.

Sxwrm., E.M.W. (1989): Contrasting group-l and group2kimberlite petrology: towards a genetic model forkimberlites. /z Proc. Fourth Int. Kimberlic Conf. (Perth).Geol. Soc. Awt, Spec. Publ. 14,528-544.

& CI-nmlr, C.R. (1979): Mineralogical classi-fication of southern African kimberlites. /n Proc. SecondInt. Kimberlite Conf. 1. Kimberlites. Diatremes andDiamonds: their Geology, Petrology and Geochemisfy@.R. Boyd & H.O.A. Meyer, eds.). Am. Geophys. Union,Washington, D.C. (129 - 139).

SMrn{, C.8., Gunr:er, J.J., Srnnwn, E.M,W., Cmmn, C.R.& EsRAHn4, N. (1985): Geochemical character ofsouthern African kimberlites: a new approach based uponisotopic constraints. Trans. Geol, Soc. S. .lfrica EE,267-280.

TArNroN, K.M. & Bnowrswc, P. (1991): The group-2kimberlite - lamproite connection: some constraints fromthe Barkly-West District, northern Cape Province, SouthAfica. In Extended Abstr.. Fifth Int. Kimbedite Conf.(Araxa). CPfuM Spec. PubL A9l, Bra*ilia (405407).

WAcNER, P.A. (1914): The Diamond Fields of Southem\frica Tmrsvaal l-eader, Johannesburg, South Africa.

(1928): The evidence of kimberlite pipes on theconstitution ofthe outer part ofthe Earth. S. \fr, J. Sci,25,ln-r48.

YoDm, H.S. (1986): Poassium-rich rocks: phase analysis andheteromorphic relations. J. Petrol. n, nl5-L228.

Received September 27, 194 , revised mamtscript acceptedJanz 15. 1995.