423

The C anadian Mine ral o gi stVol. 34, pp.423434 (1996)

RELATIVELY ALUMINOUS ALKALI PYROXENEIN NEPHELINE SYENITES FROM MALAWI:

MINERALOGICAL RESPONSE TO METAMORPHISM IN ALKALINE ROCKS

ALANR. WOOLLEYDepartment of Mineralngy, Natural History Museum, Cromwell Road Inndon SW SBD, U.K.

R. GARTH PLATTDepartnznt of Geology, Inkeheal. IJniversity, Thunder Bay, Ontario P7B sEI

G. NELSON EBYDepartment of Earth Sciences, University of Massachusefts at Lawell, Itwell, Massachusens 01854, U.S.A.

ABSTRACT

The North Nyasa Alkaline hovince (NNAP) of central and nofthern Malawi, eastern central Africa, consists of sevennepheline syenite intrusions which were, to varying degrees, affected by the Mozambique Orogenic event. We have analyzedthe constituent clinopyroxene in rocks of five of these intrusions by electron microprobe. Three groups of pyroxenes can bedistinguished. (a) Alkati pyroxenes range from diopside tbrough aegirine-augite to aegirine. These have total A1 < 0.15 (apfu)anal aie taken to be of primary magmatii origin. @j PyroxeneJcharacterized by Al > 0.15 (aplu), with uAl predominan! forma rim on group-(a) pyroxenes, cut across them and also form, in one intrusion, tiny acicular crystals within nepheline and,|ocaily, feldspm. They comprise aluminian aegirine-augite, aluminian aegirine and omphacite. These pyroxenes are interpretedas metamorphic in origin. (c) fne tma group comprises Al-rich py'roxenes in which NAI is predominant. They are diopside andaluminian diopside exhibiting no trend of alkali enrichment. They are probably of igneous origin, but could also have beenaffected by metamorphism. Although the production ofpurejadeitic pyroxenes probably requires I minimnm pressure ofabout7 kbar, aluminian aegirine with higher contents of Fe3+ can probably be generated at much lower pressures. The preservation ofprimary igneous pyroxenes in the NNAP intrusions probably indicates that neither high pressures nor particularly elevatedlempetatn1es were reached. The formation of the aluminous p'.roxenes was essentially the result of an isochemical event;in general, metamorphism of agpaitic rocks will give rise to aluminian aegirine and ajadeitic pyroxene, whereas miaskitic rockswill contain aluminian aegirine-augite aad an ornphacitic plroxene.

Keywords: nepheline syenite, alkali pyroxene, aegirine, aluminian aegirine, atuminian aegirine-augite, omphacite, Malawi.

SoMM,cB.E

l,a province alcaline du Nyasa m6ridional, dans le secteur cenral et nord du Matawi, partie est-centrale de I'Afrique,est faite de sept massifs intrusifs i sy6nite n6ph6linique qui ont 6t6, b divers degr6s, affectds par I'orogenbse mozambiqu6enne.Nous avons analys6 le clinopyroxbne pr6sent dans cinq des sept massifs par microsonde 6lectronique. Nous distinguonstrois groupes de compositions: (a) pyroxbnes alcalins allant de diopside vers I'aegyrine, en passant par I'augite aegyrinique.Ces pyroibnes possbdent moins de 0.15 datomes d'Al par unit6 formulaire, et seraient dorigine primahe (magmatique)'(b) lrs plroxenes contenant plus de 0.15 alomes d'Al, surtout sous forme de uAl, forment une bordure sur les cristaux dugroupe (i), les recoupen! et apparaissent aussi, dans un cas, sous forme d'aiguilles h f intdrieur de la n6ph6line et du feldspath.Les tompositiotrs sont repr6ientatives d'augite aegyrinique alumineuse, daegyrine alumineuse et d'omphacite, et seraientdorigine mdtamorphique. lrs pyroxbnes du groupe (c) sont aussi alumineux, mais I'aluminium y occupe surlout le sitet6tra6drique. n s'agit <le diopside et de diopside alumineuse sans indication d'enrichissement en alcalins. Ceux-ci seraientprobablement d'origine ignde, mais pounaient bien avoir 6t€ affect6s par m6tamorphisme. Quoique la production de la jad6itepure requiert une pression minimale d'environ 7 kilobars, I'aegyrine alumineuse ayant une teneur 6levde en Fer+ est toutprobablement stable e une pression de beaucoup inf6rieure i cette valeur. La prdservation de pyroxbnes primaires dans cesmassifs t6moigpe probablement des conditions plut6t mod6r6es de t€mpdratue et de pression attsintes lors du m6tamorphisme.La formation de pyroxbnes alumineux rdsulterait essentiellement d'une recristallisation isochimique. En g6n6ral, on peuts'attendre que le m€tamorphisme de roches agpartiques produise une aegyrine alumineuse et un p1'roxbnejad6itique, tandis queles roches miaskitiques dewaient produire une augite aeglninique alumineuse ou un pyroxbne omphacitique.

Ctraduit par la R6claction)

Mots-cl6s: syenite n6phdlinique, pyroxbne alcalin, aegyrine, aegyrine alumineuse, augite aegyrinique alumineuse, omphacite,Malawi.

424 THE CANADIAN MINERALOGIST

hrrnooucnou

The alkaline igneous rocks of central and northernMalawi (Fig. 1) that Bloomfield (1970) included in hisNonh Nyasa Alkaline Proyince (NNAP) consist ofseven intrusions, namely Kasungu, Chipala,Chikangawa, Mphomph4 Telelele, Ulindi and Ilomba.The dominant rock-type of these intrusions isnepheline syenite, with a little syenite and pyroxenite.The only general study of the NNAP is that of

Ftc. l. Location of the major intrusions of the North NyasaAlkaline Province NNAP) in central and northernMalawi. The area containing the Chilwa Alkalineprovince in southern Malawi is also indicated.

Bloomfield (1965), who described the peffology andgeneral field relationships. The Chilwa AlkalineProvince lies in soutlern Malawi, to the south of theNNAP Gie. 1).

The NNAP intrusions in Malawi have, to varyingdegrees, been affected by tectonic, thermal andmetasomatic events that were probably related to thegenesis of the Mozambique orogenic belt. Geo-chronological data for the province are sparse. Rblsrwhole-rock isochron ages of 685 x.62 Ma and 650 *40 Ma have been obtained from nepheline syenites ofIlomba and Chikangawa (Ray 1974, Bloomfield et a/.1981), and Pb-a dating ofzircon from sodalite syeniteat Ilomba gave 655 Ma (Bloomfield 1968). However,K-Ar dating of these intrusions gave 508 t 12 and490 x. 12 Ma for Ilomba @loomfield et al. l98L)and 410 + 16 Ma for Chikangawa (Snelling 1965). Theyounger K-Ar dates led Bloomfield (1970) to concludethat the intrusions had been affected by a lateMozambiquian phase of cataclasis.

Geological evidence within the NNAP intrusions forthese tectonic and metamorphic events is slight. Ray(1975) considered that the most northerly intrusions,Ilomba and Ulindi, comprise part of the SongweSyenile Complex, which forms an igneous unit over30 km long and 2-3 km wide. The Songwe syeniteslocally show signs of deformation, with developmentof a gneissosity indicating emplacement prior to orduring an episode of deformation related to theMozambique Orogeny (Ray 1975).

During field work on all seven of the NNAPintrusions, little evidence for metamorphism ordeformation was observed but. as described in tlenext section, there is microscopic evidence for defor-mation. However, detailed mineralogical work hasrevealed the presence of an unusual suite of pyroxenesthat are interpreted as being products of the meta-morphism.

Gnolocy l,No PgrRocRApHy

The pyroxenes that are the subject of this paperoccur in the Kasungu, Chipala llomba, Mphomphaand Chikangawa intrusions (Frg. l); they are absentfrom the Ulindi and Telelele intrusions, which consistessentially of biotite-nepheline syenite. The followingbrief geological and petrographic notes can besupplemented by the general account of Bloomfi.eld(1965) and, for llomba, by that of Bloomfield er a/.(1e81).

Kasungu

The Kasungu intrusion is approximately circular,has a diameter of 2 km and forms a prominentmountain rising about 400 m above the surroundingplain (Peters 1969). The intrusive rocks, nephelinesyenite, show little mineralogical or textural variation,

?euo,

ol e le le te I o- . t6-

36e

j Ch i lwa

PYRO)GNE IN METAMORPHOSED NEPHELINE SYEMTE. MATAWI 425

though the mafic minerals may develop a markedbanding or form irregular clots. In these coarse-gained, mesocratic rocfts, turbid perthite is generallycracked and strained, and may be traversed by zones offine-grained recrystallized feldspar. Nepheline usuallyforms l0-20%o of most rocls, but otherwise may beminor. A sodic scapotte, which is marialite on thebasis of electron-microprobe data, has been found intwo samples, the first such identffication in a NNAPintrusion.

The clots of mafic minerals are texturally complexand dominated by pyroxene, amFhibole and biotite.The pyroxene typically occurs in the center of clots butmay form at the margins where, usually intergtownwith mica and amphibole, it forms symplectitic inter-growths with feldspar, as well as myriads of small,isolated grains within feldspar; this texture occurstlroughout some rocks. The principal pyroxene formsgreen, subhedral prisms or rounded crystals that maybe zoned and generally contain abundant tinyinclusions. They are usually enveloped by a greenamphibole that also is turbid due to the presence of

inclusions. Titanite, apatrte, zircon and magnetiteor ilmenite are accessory phases.

Although not apparent in hand specimen, the micro-scopic textures and structures of the Kasungu rocksindicate that all have been subjected to some degree ofdeformation and considerable recrystallization. Notextures that are indubitably igneous have beenrecognized. Bloomfield (1965) considered that thenepheline o'shows typical replacement textures" andthat it is the result of nephelinization. However, wecould find no unequivocal textural evidence for such aprocess.

ChipaLa and. Chipala East

The Chipala infiusion forms a small hill 5 kn northof Kusungu, and the smaller Chipala East intrusionoccurs a kilometer to the east @loomfield 1965, Peters1969). To the west of Chipala Hill, a series ofessentially horizontal sheets of nepheline syenite liewithin the foliation of the Precambrian gneisses. Theexact relationship of these sheets to the main intrusion

FIG. 2. Photomicrographs of pyroxenes in NNAP nephelinesyenites. A. Numerous small grains of aluminian aegirine-augite enclosed in a single plate ofalkali feldspar. SmallHill east of Chipala (ME19). Scale bar 0.01 mm.B. Aegirine (da*, high relief) rimmed and penetrated byaluminian aegirine (pale, hieh relief). Ilomba (ME8?l.Scale bar 0.02 mm. C. Myriads of acicular crystals ofaluminian aegirine within nepheline. Ilomba (ME89).Scale bar 0.01 mm.

426 TI{E CANADIAN MINERALOGIST

is not clear, as no contacts are exposed. The ChipalaEast intrusion also is probably a conformably emplacedsheet.

In general, the nepheline syenite of the Chipala andChipala East centers is petrographically identical tothat of Kasungu. Figure 2A illustrates a widespreadtexture in which numerous small grains of pyroxeneare enclosed in feldspar.

Ilomba

The Ilomba complex, located 2 km south of thenorthern border of Malawi with Tanzania @g. 1), is a2 x 1.5 km intrusion of syenite, nepheline and sodalitesyenites and pyroxenite @loomfield et al. l98l). Onthe northwestern side of the intrusion. Bloomfieldobserved a gradation along strike from paragneiss toperthosite. He (Bloomfield et al. I98L) and Ray (1975)noted the close resemblance of the perthosite to thesyenites comprising the whole of the Songwe syenitecomplex.

Alignment of feldspar crystals is apparent through-out the complex and, we believe, an igneous texture.A lineation, which is only apparent in some places,notably in the summit are4 is, however, undoubtedlytectonic and related to structural events of tleMozambique Orogeny (Ray 1974).

The suite of nepheline syenites includes biotite- andpyroxene-bearing varieties. The rocks are generallycoarse-grained but texturally variable; they contain

microcline perthite, abundant nepheline, with somealteration to cancrinite, brown biotite and sparsepyroxene in the former, but abundant deep greenalkali pyroxene, generally zotred, and rare biotite in thelatter. The green alkali pyroxene is in some casesrimmed and partly replaced by a pale green pyroxene(Fig. 2B) that also forms small, isolated patches.Myriads of tiny needles of pyroxene, generally alignedalong two directions, occur in the nepheline (Ftg. 2C)ando to a lesser extent, the feldspar, of the pyroxene-nepheline syenite. Sodalite is locally abundant, andmagnetite, calcite, titanite and zircon are accessoryphases, with eudialyte in the pyroxene-nephelinesyenite.

The main pyroxenite consists of up to 907o diopsidicpyroxene and up to 257o interstitial magnetite. There isminor colorless to pale green calcic amphibole, but anabundance of tiny green crystals of spinel. A small areaof pyroxenite mappedby Bloomfield @loomfield eral.1981) on the western limb of tromba, and occasionallyfound as xenoliths in the nepheline syenite, is variablyaltered, with the pyroxene crystals having a deep greencolor and being marginally altered to amphibole,biotite, or both.

Mphompha

The Mphompha intrusion, referred to as NorthNyika by Bloomfield (1965), comprises part of theRumphi igneous complex (Kemp 1975, Thatcher

TABLE T. REPRESE{TATIVB @MSITIONS OF (IINOFTRO)(E!{ES FROM KASUNGU AND CHIPALA

$io, 50.64 51.84 51.31 fiJ2 50.42 50.96 52.13 51.80 51.47 5r.84 $93 50.07no. 030 037 0.41 l.m 0.11 0.80 1.15 0J0 0.45 0.14 0.84 0.62Ateo, L6 2ffi 2-t7 5.y3 6.98 552 1.6E 1.81 LL8 5.31 5JE 432Fezor* 452 652 &43 7.63 10.90 12.93 253 635 7.92 556 E.8l A6FeO 780 8.O7 851 7.9r 4J0 5,n E 6 1056 U.88 lo.G 7.68 7J0NtrO 0.43 0J8 0J4 0.67 037 052 0J3 0i7 057 0.69 0.65 0,tr'2tt4$ 957 8.31 6& 538 sJL 5.08 11.04 1A 5J9 6.10 5.16 450clo m35 $.7E 16.Et $2n M,'2 13.48 2lJ3 1&33 16J1 1632 15,2 14.42NarO 293 3,n 4.fi 5.16 sJl 6.6 159 323 4.0r 425 531 5.51zrg2 0J5 024 rd- Ld. rd. ed. 4m 0.n 038 0 [d. <0r0Toral 9,35 W9 9E.98 s463 W.42 99.96 l@.4 10055 l@.06 100J5 100.1E 99.EE

Sfic(lml fumla bced m 4 aalim md 6 atm of cyg€n

si 1931 1964 LEn 1.919 1.E96 1918 r.949 1.978 1.v78 1.950 t.vzr 1.916nAr 0.069 0.036 0.028 0.61 0.104 0.@ 0.051 0.w 0fi2, 0.050 0.079 0.084

EAI 0.024 0.056 0.0rl0 0.1&, 0105 0.163 o.@.3 o.gn o.cn 0.185 0.169 0.111Ti 0.0a 0.011 0.012 0.a9 0.006 0.@ 0.032 0013 0.013 0.m4 0.0t4 0.018Fe3'r 0.r0 0.186 024/ 02t7 0.30E 0341 0.071 0.a3 0&3 0.157 0150 0.347Fo& 0ll!9 0216 o.n4 0250 o.t42 0,1@ 0.252 03E2 0,382 0317 0 2 02/)lvlo 0.014 0.019 0.01E 0.@l 0.012 0.017 0.017 0.019 0.019 o.VD 0.@1 0.tr20l4c 0J50 0.469 0.380 0.303 03n 0.A2 0.615 03rA 03n Og2 0290 0 7ca 0352 0.762 0.612 0.616 0J93 0544 0.tr70 0.6E0 0.6E0 0.658 0.615 0J91Na 0.150 0137 0.310 Mn 0.416 O.MI 0.115 0299 02', 0310 0388 0.413k 0.010 0.@4 0.004 0.m7 0.07 0.005 - 0.m1

co@osidc cpcted in rrtTa trddx. . Fep3 elolrred m tte b€si8 of 4 cdm ad 6 alons of cyg@ ed: mtderemdred saryle$ l-{ Kemgs,7-12 (bhsla r-3 md6-.q tlm si6i /t-5 md 1G.12: Al-riclpymm

1l

PYRO)GNE IN METAMORPHOSED NEPHELINE SYENTTE. MALAWI A a 1

L974). It is oval in outline, approximately 8 x 3 kmin are4 and lies within syenites, some of whichare peralkaline, and granites. Pyroxene has beenfound in only about half the samples collected.The main rock type is a nepheline syenite com-posed of nepheline and perthite; there are patchesand layers, which may comprise half the rock, ofamphibole, biotite, pyroxene and abundant titanite.The pyroxene is generally enveloped in amphibole andis clearly partly replaced, small remnatrt pyroxenecores commonly being all that remains of the originalcrystals.

Chilcangawa

The Chikangawa intrusion (Gaskell 1973),originally referred to by Bloomfield (1965) as SouthVipy4 is oval in outline, with a longer axis of 6.5 km.It comprises nepheline-biotite syenites which arecoarse, relatively homogeneous, and of low colorindex. Pyroxene has been identified in only onespecimen, in which it forms sparse clusters of smallgrains enclosed in microcline. This habit is closelysimilar to the fine-grained granular pyroxene found insome rocks of Kasungu and Chipala.

0.3

0.1

tro

Qpo

(I)

o

o Oo g o

ao oO n

+

. ++

1,f +

+Kasungu & Ghipala* igneous group

r + +vrAl-richgroup

r .* o MPhomPha{+ x Chikangawa

+ff-+J:*T'rlF *

+- *.."+ :ff

r ++ '

* ***iJ ** l**e+ r * * r * f

* * f , * * *

o

*rx

tITY

)t(x

.0 -0.8 -0.6 4.4 42 n o.2 4.4 0.6 o,8Na-Mg

* r . +' * t -, . 1 +

t i

+ T ++ {t+. , -+

+ + , + + +

l lombar igneous group+ vrAl-rich groupo vAl-rich group

@@

@o

Chllwa Provlnce nephellna syenftas

Na-Mg

Frc. 3. Plots of Al againstNa-Mg (atoms performula unit) in pyrox-enes of the North NyasaAlkaline Province. A.Kasungu and Chipala,Mphompha and Chikan-gawa. The igneous andAl-rich groups ofKasungu and Chipalaare divided at 0.15 Al.B. Ilomba. uAl-rich,wAI-rich and igneousgroups are distinguished.The arrow shows thegeneral trend of pyrox-enes @ased on approxi-matdy 200 analyses) innepheline syenite of theChilwa Province. Malawi(Woolley & Platt 1986).

428 T}IE CANADIAN MINERALOGBT

Frno:cls

Pyroxenes were analyzed using both a Hitachi52500 scanning elecfion microscope equipped with aLink AN10/55S energy-dispersion X-ray analysissystem, operaled at 15 kV and a current of I nAmeasured on vanadium metal" and a Cameca SX50wavelength-dispersion electron microprobe operated at20 kV and a current of 20 nA measured on a Faradaycage. Representative compositions of pyroxenes aregiven in Tables 1-3.

Kasungu and ChipaLa

The pyroxene population of Kasungu and Chipalanopheliqe lsyenites may be divided on the basis of A1content into two distinct types, Al-poor and Al-rich, thedivision being taken at 0.15 Al (apfu) (Fies.3A, 4A).The Al-poor pyroxene occurs as cores, and the Al-richones as overgtowtls and as small grains in the ground-mass. The Al-poor pyroxene is considered to have anigneous paragenesis, as discussed below; compositionsare distinguished on Figures 31.,4A and later diagrams

+mo

[L6z

+(v)

I6z

Frc. 4. Plots of (Na-Fe3+)against uAl in pyroxenesof the North NyasaAlkaline Province intru-sions. A. Kasungu andChipala; igneous andUAI-rich groups aredistinguished. B. Ilomba;vlAl-rich" ryAl-rich andigneous groups are distin-guished.

Kasungu and Chipala

* igneous group

+ vrAl-rich group

l lomba

* igneous group+ vlAl-rich groupo lvAl-rich group

trChikangawaKasungu/ChipalaMphompha

429

A

+

Frc. 5. Plots of NNAP pyroxenes, excluditrg Al-rich groups (i.e., Al > 0.15), in terms of diopside - hedenbergite - aegirine.A. Kasungu and Chipala, Mphompha and Chikangawa. B. flomba.

as the 'lgneous groupo'. Both types are represented inthe Kasungu and the Chipala intrusions Cfable 1).

The Al-poor suif€ consists essentially of diopsidicpyroxene with a Mg/(Mg + Fe2) value varying from0.46 to 0.74, with most grains having a value greaterthan 0.5. Concomitant increases in Na (L.594.8O trtt%oNa2O) and Fe (2.66-L0.78 vrtVo FqOr) reflect amoderate increase in the aegirine component (Fig. 5A),which is represenled in Figure 34, by the variable(Na-Mg). Compared with observed increases in othernepheline syenile complexes, for example the Chilwaalkaline province of southern Malawi (Woolley & Platt1986), this enrichment is not pronounced, as indicatedby the arrow on Figure 3B. Al contents (1.68-3.06 wtVoAlror) are slightly higher than in pyroxenes ftomfypical nepheline syenite complexes.

The suite of Al-rich pyroxene is distinctly richer inNa (3.9G{.71 wtTo Na2O), Al (3.59-8.71 wt%o N2O),and hence the jadeite component, and uAl (Figs. 34,4A). The amount of Fe (5.73-12.03 wtVo FerOt)partially overlaps the range of the Al-poor pyroxenes,i.e., the extent of solid solution toward aegirine isessentially similar in both suites. Mg/@g + Fe2)values are also similar to those of the Al-poor suite.

The nomenclature of the Al-rich pyroxenes canpartly depend on the method used to reducethe analytical data. Moreover, differences exist in thepublished nomenclatures of the fields defined in termsof the aegirine - jadeite - diopside or hedenbergitemolar proportions (cf Clark & Papike, 1968, Deeret al. L978).In this paper, we use the nomenclature ofMorimoto et al. (1988). The nomenclature suggestedby Curtis & Gittins (1979) for pyroxene of a similslparagenesis to those described here, has not beenadopted, but their classification diagram hasbeen adapted (Frg. 64.) to the IMA-sanctioned systemof Morimoto et al. (L988).

The bulk of the Kasungu and Chipala Al-richpyroxenes correspond to aluminian aegirine-augite,with some classifying as omphacite (Fig. 6.A). TheAl-poor pyroxenes are diopside and aegirine-augite,with rare aluminian diopside. Tie lines on Figure 6.4,join compositions of cores and rims (for clarity only aselection is shown), the latter always proving to be themore aluminous and sodic.

Ilomba

Plnoxene compositions in 17 samples of Ilombanepheline syenite and pyroxenite fall into three distinctgroups, fwo of which have elevated Al contents(Fig. 38, Table 2), but differ in the proportions ofNAI and uAl (Fig. 7). The third group is comparativelypoor in total Al.

The o'low"-Al pyroxene compositions vary ftomdiopside to aegirine (Figs. 5B, 68) and thus, unlike theKasungu and Chipala pyroxenes, show an extensiveevolution typical of clinopyroxene in many types ofatkaline igneous series, for example, the nephelinesyenite suites of the Chilwa hovince, as depicted bythe arrow on Figure 3B. Small excesses of Na over thatrequired for the formation of aegirine componentcombined with uAl, suggest limited solid-solutiontoward jadeite. This pyroxene is generally the mostabundant in the tromba nepheline syenite, but is alsothe principal pyroxene io 1trs rlkali-pyroxene-richpyroxenite occurring near the summit of Ilomba. In thenepheline syenite, this pyroxene commonly has a rimof uAl-rich pyroxene, and in the pyroxenite, a corerich in ryA1. This series will be referred to hereafter asthe "igneous group".

The NAI-rich group @ig. 38) has compositionsranging from diopside to aluminian diopside (Fig. 6B);compared with the uAl-rich group, the group of

, f++R'

+

+

+

- +

$a

0.4

(U

+qtz(Uz

(UC)T(Uz(5z

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1'hl/(oAl+Fe3)

0.1

* - T + + + T i-++ +T r}

* * * * * ' ' + + * 4* * + '* * r+: '* * + f+f,++

tr* * t , +

. Tr +F r

r * *+

l lombar igngous group

+ vAl-rich group

o sAl-rich group

, o

oro o o o 6 o o 9

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1

"Al/('Al+Fe9

Flc. 6. Plots of Na/(Na+Ca) against vlAV(vrAl+Fe3*; latoms per formula unit) for NNAPplroxenes. The pyroxene names, given in A only, follow the nomenclature ofMorimoto et al. (1988). A. Kasungu and Chipala (vlAl-rich and igneous groups aredistinguished), Mphompha and Chikangawa- A selection of Kasungu and Chipalapyroxenes representing individual crystal cores and rims are joined by tielines. In allcases, rims are more sodic and aluminous tllan cores. B. Ilomba. vlAl-rich, rvAl-richand igneous groups are distinguished.

n e

c.7

3.6

1.3

3,aoglrlne

traegirine-augite omphacite

Kasungu & Chipalar igneous group+ u Al-rich group

o Mphomphax Chikangawa

+ ,

+ T+

"fffi7..-cllopside

PYRO)(ENE IN METAMORPHOSED NEPHELINE SYEMIE. MAI.AWI

TABLB e RBRESB{TATM @ITPOSfiONS OF PfROXS.IES FROM trOMBA

431

1I

saTK"Al"o.F%o."F€OthoMSCsONqoTaOzT@1

sirAlFo3r,

nAl

TI&$eFc&ldrMSCaNak

49.4 5tt2 51J9 fiX2 5330 54.% 51.90 5332 52.81 9& 41.940.15 0J0 02i 034 0.r9 0'1 0.19 0.06 0.r'1 0.16 ?.ul1.23 135 2A4 t.l3 5al 7.68 3.93 5.65 4.95 5.16 6.(B

10.75 nS D94 18.@ t3.43 Al3 11.42 1t.6t X.n 2430 L68J,, 3.46 521 6.94 492 ljl 4.08 \98 Ll7 r.@ 3A1oz An 0t9 1.48 039 0.0 052 0.70 031 031 o.fi)630 L1L 1r5 3.49 4:15 0.15 3'a5 221 0.4',1 (L6t 14.6

1738 7.{t 5.68 n:r1 N,42 026 932 6.98 L',ts 3.4 L63fi 93t 10.00 6.v. 7gt ,3rt4 &47 10.18 1238 r2.8 0.75024 0J0 (L16 0.61 0,m 0m 0.1? om 0.01 0.01 0.m

983? 10094 rm.?9 r01r8 101.14 10110 9.65 10.82 1m36 1m.47 10.06

StEaSd Mlsh6edm4 c€dds d6 sns of Gygpn

1.941 1.949 Lget L9ti7 1966 1989 l]W r.980 rgtg l.y'J lnLo.o5? 0.61 0.{,33 0rE1 0.034 0.011 0.o14 0.020 0.921 0.@5 0.n9ilm 002

o.dD 0.010 0.095 0.m 0rl9 03n 0.141 02st ol98 0,02 0.0s30.@ 0.023 0.w, 0.010 0.m5 0.07 0.m5 0.@ o.dE 0.m4 0.G50315 0.671 0.658 oJlE 03n 0.637 0.49t Asn 0.713 0.7@ 0.C00jfto 0.110 0.166 0.a 0.152 0.046 0.1p 0.093 0.w1 0.032 0.101a,w 0.025 0.0,9 0.048 0.012 0.m 0.017 o.gD 0.012 0.012 0.m0J6E 0.131 0.('1 0100 0 l 0.008 0% 0.126 0.w 0.938 0:n4o:t$ 03(b 013,2 0.&, o.4A 0.010 0378 0275 0.110 0.125 0.rt3o2r4 0.693 0rl' 0J6 0.97 096E U@, 0:83 0.899 0A$ o.(El0.m5 0.m9 0.m 0.011 o.qp om4 ilxt3 0.w 0.m 0.m 0.m

ComOUAos r€pctod h*t6 dd* * BezO3 mlmlated o ftebads of 4 cdiG d 6 aM8 of syg€o Sandi& f-a

igpem Eror& 5-10 nAr-df! gru4p, u rAHrjgrurp. t-10 funeedtss in!s&lhe.

Frc. 7. Plot of nomba pyroxenes in terms of wAl and uAl. The uAl-rich, NAl-rich andigneous groups are distinguished.

llomba

* igneous group+ uAl-rich groupo rvAl-rich group

+ -+ +

+ + + +* * ' + + + - i + * - * . ** * , + + ' - F + + ' + ' +- + + + r +

"'Al

432 THE CANADIAN MINERALOGIST

IvAl-rich pyroxenes have lower Fe3+'Ferotal andFe2+:Mg values and higher Ti contents (Table 2). Anyexcess of Na over Fe3+ is small. and the content ofuAl is low. Such pyroxene is the major rock-formingphase of the pyroxenites.

The uAl-rich group lies in the compositional fieldsof aegirine-augite and aegirine, with many warrantingthe modifier aluminian (Fig. 6B). High Fe3+;pstotulratios, lower Ti contents and higher Na values alsodistingulsh this group from the rvAl-rich pyroxenes.The excess of Na over Fe3+ (Fig. 4B), combined withvIAl, indicate a significant trend of increasing solid-solution toward jadeite in this group. This is thepyroxene that occurs as a rim around, and smallpatches adjacent to, the Al-poor pyroxene in thenepheline syenite, and as a rim on IvAl-rich pyroxenein some samples of pyroxenite.

The fine needles of pyroxene that occur withinnepheline (Fig. 2C) and, to a lesser extent, feldspar, in$ome rocks, were found to be generally too narrow(commonly I pm or less wide) to be analyzed with theelectron microprobe, although some slightly largergrains were analyzed. They all proved to belong tothe vlAl-rich group (Table 2, anal. 8-10) and arealuminian aegirine.

The pyroxene sf the alkaline pyroxenite from nearthe summit of tromba is generally Al-poor, 1.e., itbelongs to the igneous series, but it may contain a coreof rvAl-rich pyroxene, suggesting that these rocks

TABII 3. RBPRESEMATIVE @MFOSrIIONS OF PYROX3NES

represent blocks of pyroxenite caught up in andmodified by the nepheline syenite. The correlation ofpyroxene type with rock type at Ilomba can besummarized as follows: (1) rvAl-rich: principalpyroxene of pyroxenite; core in alkali pyroxene ofpyroxenite. (2) uAl-rich: forms a rim on Al-poorpyroxene and discrete grains in nepheline syenite;forms tiny acicular prisms within nepheline innepheline syenite; forms a rim to ryAl-rich p)noxenein pyroxenite. (3) Al-poor: common as a core tovlAJ-rich plroxene in all samples ofpyroxene-bearingnepheline syenite.

Mphompha

Pyroxene grains were aaalyzed in four samples ofnepheline syenite (Table 3). They all proved to bediopside (Frg. 6A) which, apart from those in one rockoare characteized by high Al (>0.15 Al apfu) @ig.3A).Thus they trend to aluminian diopside. However, theAl is essentially tetrahedrally coordinated, and sothe pyroxene is similar to the ruAl-rich group ofIlomba. At Mphompha, the pyroxene has only a smallamount of Na (generally 1-2 wt%o NqO) with noindication of a trend toward more sodic compositions(Frg.6,4.).

Chikangawa

The pyroxene grains in the one pyroxene-bearingnepheline syenite found all proved to be aegirine(Figs. 5.A', 64'), with only moderate amounts of Al(maximum 2.8 wt%o Al2O) (Table 3).

DrscussroN

Pyroxene compositions in five alkaline intrusivecenters of the North Nyasa Alkaline hovince fall intothree distinct groups, one of which, the Al-poor group,ranges from diopside through aegirine-augite toaegirine and thus is typical of pyroxenes in igneousnepheline syenite, some ijolite-series rocks and evenperalkaline granite. The completeness of this series inIlomba reflects the general peralkalinity of the intru-sion and contrasts with the restricted series of alkalipyroxene compositions in Kasungu, Chipala andMphompha. The Al-poor nahre of the aegirine in thesingle pyroxene-bearing rock sampled fromChikangawa is consistent with a primary origin.

In contrast to these igneous pyroxenes, a secondseries has been identified that is rich in vIAl: we areunaware of such alkali-rich plroxenes having beendescribed from pristine magmatic rocks. Thesepyroxenes form a rim on the igneous ones, cut acrossthem and also form fine needles within nepheline atIlomba and small grains in Kasungu and Chipala. Suchaluminous pyroxenes, some of which are emFhacite,are likely to be metamorphic in origin.

sio?Tto2Atro"F%o""F@DtlOt'{goCaONaro7.1O2Tmt

tu97 5tr0024 O828 218

nJ4 30.14037 0.60.14 0210 0.6tzt t.64

rLt2 'a74(L(E 0.@

9&89 9,@

I

51J80.43

4947,6l.m

1138n20.990.00

1m.66

2 3 4 5 - 6

4937 41X9 4194 46,'11 4f.33tjt 1,8 lJ9 ll0 L63.66 4,94 5.6 632 7.18334 7:16 7.43 83t 5.816.6 4gt 'tA &41 6.A050 0.62 0J4 0J5 ojl

10.81 933 9.49 167 8.78a$ 2t3t 2t5s tu% 2t32131 1.86 1t2 LOz rS2,0.16 0,n 0.06 0.a 0.10

99J4 100"50 rm.68 1$l9 1m3t

sr 1.937 lJ65 1.803 l.g)4 1Jl9 r,752 1j84 1.981aAt 0063 0.13J 0.19? 0.196 o2At 0248 0106 0.019

EAr 0.m 0.@t 0.@ 0.0a8 0.043 o,gn oJn o,gDT1 0.012 0.045 0,I/.2 0.044 0.052 0,0t9 0.007 olmFe3b 0.111 0.tO6 0.9t O2r0 02/o 0.165 o.p8 0.t66Fe& oN 0J92 O.tt7 ot6J 0l0i oa3 0.012 offi!,ln 0.033 0.016 OM 0,01? 0!18 0.016 0.6 0.m9MS 0.637 0.@ 0J36 0532 U4X 0.4qr 0.0U 0.(,6ca 0f96 09(5 0& 0.869 08t5 0.r/6 0.62 0ft7Nr Ogn 0.06 0.136 0,133 0.149 0.111 09i9 0J43zt Om 0.006 0.qB o.mr iln2 O,W Omr 0.0fr)

Shrftrsl f@la b6sed @ 4 cari@ dd 6 at@ of 6t!@

ConFocfi@ qtEsr€d h vt% !dd4 i FqO3 €&qh&d @ fre Mr d 4 qd@ dd 6tM8 doqrygo srryrffi l-6 l,bh@h, 7-E (!lk@gm

PYRO)GNE IN METAMORPHOSED NEPHELINE SYENTIE, MALAWI 433

(5

+Gz(!

z

The Al-rich pyroxene needles within nepheline aregenerally oriented in two directions, which areclearly determined by the nepheline structure, and thisorientation would seem to preclude an origin as aliquidus phase. Nepheline generally contains an excessof silica and a small amount of iron, so that an origineither by exsolution, subsolidus growth or throughmetamorphic heating seems feasible. The last isfavored as being more consistent with the rimming andcross-cutting aluminous pyroxene, which is interpretedas being metamorphic.

Aluminous pyroxene has been reported from meta-morphosed alkaline rocks at a number of localities(Fig. 8) including the Monglong nepheline syenitegneiss, Cameroon (Komprobst et al. 1976), schistosenepheline syenite at Norra Kiirr, Sweden (Adamson1944), undersaturated syenite in Japan Qwasaki 1960),nepheline gneiss in Ghana (Holm I97l) andregionally metamorphosed nepheline syenite in theGrenville Province of Ontario (Lumbers 1976).However, the only detailed account of pyroxenes froma metamorphosed alkaline complex is that of Curtis &Gittins (1979). Their pyroxene data for the regionallymetamorphosed agpaitic rocks of the Red Winealkaline complex, in Labrador (Frg. 8) cover much ofthe upper part of the diagram. Also shown on Figure 8are data for the other localities mentioned above. fromwhich it is clear that pyroxenes of wide degrees ofalkalinity and Al-enricbment may be formed as a resultof metamorphism of alkaline rocls. The uAl-richpyroxenes from Ilomba have similarities to some ofthose from Red Wine, but those from Kasungu andChipala have compositions, in terms of Figure 8, rather

Frc. 8. Plot of Na/(Na+Ca)againsr UAY(uAl + Fe3+)(atoms per formula unit)for pyroxenes from meta-morphosed alkaline rocksincluding Red Wine,Labrador (Curtis &Gittins 1,979); Ghana(Holm 1971); Nkonglong,Cameroon (Kornprobstet al. 1976): Norra Kiirr,Sweden (Adanson 19M);Japan (Iwasaki 1960) andOntario (Lumbers 1976).Key to symbols grven onfigure. Pyroxene fields ason Figure 6A.

different from any so far described for this paragenesis.The rvAl-rich pyroxenes from Ilomb4 from textural

and chemical evidence, are interpreted to be essentiallyof primary, igneous origin and are considered to beigneous cumulus minerals. These pyroxenes, and thoseof Mphompha, trend toward aluminian diopsidecompositions, but it is not clear whether this trend is aprimary igneous feature or reflects a metamorphicoverprint.

Jadeitic pyroxenes are generally considered to beindicative of high pressures. However, this is notnecessarily the case. The aluminous pyroxenes in theagpaitic Red Wine rocks were formed underamphibolite-facies conditions; Curtis & Gittins (1979)suggested that the presence ofjadeitic and omphaciticpyroxenes in nepheline-bearing rocks do not indicatethe high pressures normally attributed to silica-saturated or oversaturated rocks that contain theseminerals. Although experimental work indicates thatformation of pure jadeite probably does require aminimum pressure of about 7 kbar, replacement ofAl by Fe3+ will probably lead to the generationof pyroxene at much lower pressures @eer et al.1978). The preservation ofearly igneous pyroxenes inthe NNAP intrusions suggests that the metamorphismwas neither of high-pressure type nor reachedparticularly elevated temperatures. This metamorphismmay have been the "Pan-African tlermal event", whichis widespread over much of the northem half of theAfrican continent, or part of the Mozambique Orogeny'

It is not clear whether the production of meta-morphic pyroxenes in the NNAP intrusions was anisochemical process. Although the aluminian aegirine

g x x x

o 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1

htl(vh+res)

434 T}IE CANADIAN MINERALOGIST

at Ilomba and aluminian aegirine-augite and omphaciteat Kasungu and Chipala are slightly more sodic thanthe coexisting igneous pyroxenes, and the presence ofsodic scapolite in nepheline syenite at Kasungu andChipala is probably indicative of Cl, and possibly Na,metasomatism, clearly the enhanced Na in the meta-morphic pyroxenes may simply reflect re-equilibrationunder metamorphic conditions but within a closedchemical environment. The fact that the more alkalineigneous pyroxenes in the Ilomba rocks coexist withhiqhlV alkaline aluminous pyroxenes, whereas the lessalljaline igneous pyroxenes in the Kasungu and Chipalasuiltes occur with less alkaline aluminous pyroxenes,mdy be evidence that the systems were essentiallyclosed. Thus in this paragenesis, metamorphism ofagpaitic rocls will give rise 16 aluminian aegirine anda jadeitic pyroxene, whereas miaskitic alkaline roclslead to aluminian aegirine-augite and an omphaciticpyroxene.

AcIO{owLEDcElvGNTs

We are most grateful to Drs. Thomas Frisch andMalcolm Ross for detailed cornments on the paper,particularly on pyroxene nomenclature. We also tha::kthe Editor for his painstaking editorial work as well asscientific comments. Staff of the Geological Survey ofMalawi are thanked for their assistance during thisproject. We are gateful to Dr C.T. Williams andJ. Spratt for their help during probe work at the NaturalHistory Museum, London. R.G. Platt acknowledges thesupport of the Natural Sciences and EngineeringResearch Council of Canada. Travel funds wereprovided under the NATO Collaborative ResearchGrants Programme (No. 880038).

RsnnBNcrs

ADAMSoN, O.J. (1944): The petrology of the Norra K[rrdistrict. An ocqurence of alkaline rocks in southemSweden. Geol. Fdren. StocWnbn Fdrh. 66. lI3-255.

Bt ootvrrs.t, L (1965): A reconnaissance survey of alkatinerocks in the Northem and Central reg1ons. Rec. Geol.Surv. Malawi 5,17-63.

(1968): The pre-Kanoo geology of Malawr. Mem-Geol. Surv. Malawi 5,l-166,

(1970): Orogenic and post-orogenic plutonism inMalawi. /n African Magmatism and Tectonics (T.N.Clifford & I.G. Gass, eds.). Oliver & Boyd, Edinburgh,u.K. (119-155).

DEANS, T. & Ws:s, M.K. (1981): The Ilombaalkaline comFlex, northern Malawi and associateduranium-niobium mineralisation, Oyerseas Geol. MineralRes.57, l-21.

Cum, J.R. & Perxs, J.J. (1968): Crystal-chemicalcharacterization of omphacites. A.nz . Mineral. 53,840-868.

CuRrIs, L.W. & Grrrns, J. (1979): Aluminous andtitaniferous clinopyroxenes from regionally metamor-phosed agpaitic rocks in central Labrador. J. Petrol. ?,0,165-186.

Dnun, W.A., Howrs, R,A. & Zussrraar, I. (1978): Rock-Forming Minerals. 2c^. Single-Chain Silicates.Longmans, London, U.K.

GAst{El.;t,, J.L. (L973): The geology of the Mzimba uea'. BuIl.Geol. Surv. Malawi JT, l-30.

HoI-u, R.F. (1971): Sone gamets, pyroxenes, and amphibolesfrom nepheline gneisses in Ghana. Am. Mineral. 56,21tt-21,22.

Iwesarc, M. (1960): Clinopyroxenes intermediate betweenjadeite and aegirine from Suberi-dani, TokusimaPrefecture. J. Geol. Soc. Japan 66,33440.

KErp, J. (1975): The geology of the Uzumuaarea. BuIl. Geol,Surv. Malawi 41, l-56.

KoRMRoBsr, J., CANTAGREL, J.-M, FasRGs, J., LessERRE, M.,Rou-n, M. & Sose, D. (1976): Existence, au Cameroun,d'un -agFatisme slsalin panafricain ou plus ancien, lasy6nite n6ph6linique l mboziite de Nkonglong - com-paraison avec les roches alcalines connues dals la m6merigtron BulI. Soc. G6oI. France\E,1295-1305,

LulvIBERs, S.B. (1976): Omphacite-bearing nepheline syenitein an anorthositic complex, Grenville Province of Ontario.GeoI. Assoc. Can. - Mineral. Assoc, Can,, Program Abstr.1 .73 .

MoRnuoro, N., FesRfs, J., Fnncusolr, A.K., GD{auRG, I.V.,Ross, M., Srmnr, F.A., ZussvaN, J., Aorr, K. &Gorranot, G. (1988): Nomenclature of pyroxenes.M ineral. Mag. 52, 535-550.

Prrrns, E.R. (1969): The geology ofthe Kasungu area. BulI.Geol. Surv. Malawi 25.7-55.

Rav, G.E. (1974): The structural and metamorphic geology ofnorthem Malawi. J. GeoI. Soc, kndonl30.4n440.

(1975): The geology of the Chitipa-Karonga area.Bull. GeoI. Sun. Malawi 42.1,-101.

Sttmwc, N.J. (1965): Age determination uniL OverseasGeol. Surv. Rep. (for 1964),28-38.

THATCHR, E.C. (1974): The geology of the Nyika area, Bull.GeoI. Surv. Malawi 40.'J,-90.

Woourl A.R. & Harr, R.G. (1986): The mineralogy ofnepheline syenite complexes from the northern part oftheChilwa hovince, Malawi. Mineral Mag, 50,597-610.

Received. September 1, 1994, revised manwcript acceptedDecemberT,1994.