American Mineralogist, Volume 70, pages 1087-1 100, 1985
Geochemistryand REE mineralsof nephelinesyenites from the Motzfeldt Centre,South Greenland
A. P. JoNrsr Diuisionof Geologicaland Planetary Sciences CalifurniaInstitute of Technology,Pasadena, California 91125
.rNp L. M. LensrN GrsnlandsGeologiske U ndersogelse @sterVoldgade 10, DK-1350,Copenhagen, Denmark
Abctract Whole-rock analysesincluding the REE are given for nineteensamples from four of the six overlappingnepheline syenite intrusions which constitutethe Proterozoic(- 1310ma) Motz- feldt Centre. Larvikite margins of one unit (SM5) are apparently related to the nepheline syenites by early fractionation of Ol + Pl followed by fractionation of Ol + Pl + Cpx t Amph at current levels of exposure.Positive Eu-anomaliesin the larvikites might be 875r/865r explained by large-scaleincorporation of earlier Pl-rich cumulates.Low initial ratios imply a mantle source,and the most basic magma was a late alkali gabbro giant dike with plagioclasemegacrysts. REE- and incompatibleelement-rich peralkaline lujavrites with large negativeEu-anomalies are late stageproducts of extremefractionation. Electron probe analysesof REE in eudialyteconfirm this mineral as the major host to Zr and REE in the lujavrites. Other REE-bearingminerals in the nephelinesyenites are Ca-Na{e-Ti-Zr-Nb- silicatesof the rinkite-mssandrite and l6venite series,with up to 22 wt.% RE-oxides;some may be related to metasomaticfluids which had geochemicalsignatures similar to the lujav- rites.
Introduction sphene.Electron probe analysesincluding the REE in sev- discussed. The Motzfeldt Centre is the older of four major centers eral of theseminerals are reportedand becausea of igneousactivity which make up the Igaliko Complexof The Motdeldt Centre is especiallyinteresting was developed. nephelinesyenites (Emeleus and Harry, 1970).It is situated late stage peralkaline magma, lujavrite, bulk of silica- at the eastern end of the Proterozoic alkaline igneous This provides a link between the and the un- Gardar Province of southern Greenland (Emeleus and undersaturatedsyenite plutons in the Gardar (Ferguson,1964, Upton, 1976).The Motzfeldt Centre has a determined usual peralkalineIlimaussaq Intrusion that Rb/Sr (whole-rock)age of 1310+31 ma and an initial 1970;Bailey et al., 1981;Sorenson, 1969). We suggest 8?5r/865rratio near 0.702 (Blaxland et al., 1978).The incorporation of earlier feldspar-richcumulates in addition played im- MotzfeldtCentre covers an areaof approximately350 km2 to extensivecrystal fractionation could have an and is well exposedas a steeply dissectedand glaciated portant role in the overall petrogenesisof the syenites. 2000 m plateau. The mineralogy of the mafic silicatesto- getherwith a briefoutlineofthe geologyhas been given by Geology Jones(1984). The Motzfeldt Centre was formed by a seriesof overlap- In this paper we present new whole-rock geochemical ping intrusions emplacedby a combination of block stop- data for major and trace elementsincluding the rare earth ing and 5-10 km radius ring fractures.The country rock is elements(REE). In particular the REE are contained in a mostly gneiss,known as Julianehib granite (Allart, 1973) variety of minerals,some of which show evidenceof meta- and a sequenceof supracrustalrocks including metasedi- somatism. These minerals are principally eudialyte and ments, agglomerates and lavas collectively called the alkali-zircono-silicates of the rinkite-mssandrite and Eriksfiord Formation (Emeleusand Upton, 1976).Meta- livenite series,and to a lesserextent, apatite, zircon and somatizedand alteredrelics of the supracrustalrocks form large founderedraftlike xenoliths within the syenitesand measure up to several kilometers in length and a few I Current address, School of Geological Sciences,Kingston hundred meters in thickness.Adjacent lavas in the supra- Polytechnic, Penrhyn Road, Kingston Upon Thames, England crustal volcanic rocks, including the altered xenoliths, KTl 2EE. range from alkali basalts to phonolites and are possibly 0003-004x/85/l l 12-1087$02.00 1087 JONESAND LARSEN:NEPHELINE SYENI?ES FROM THE MOTZFELDT CENTRE
Fig. l. Geology of the Motzfeldt Centre and location of samples used for whole-rock analysesin this study. Map after Jones (1984) with revised eastward extension to the major E-W fault after Tukiainen et al (1984). AGGD is alkali gabbro giant dike; EM : East Motzfeldt : syenites; NM North Motzfeldt syenites. Ornamentation, solid : supracrustal rocks. NQ : North Q6roq Centre (Chambers, 1976),SQ : South Q6roq Centre (Stephenson,1976).
related to the Motzfeldt Centre itself (Jones, 1980). The structurally high levels. Pegmatites are sometimes associ- center is cut by a major transcurrent E-W fault (The ated with contacts between xenoliths and host syenites. Flink's Dal Fault, Emeleus and Harry, 1970) which was Several pegmatites, microsyenite sills and altered limonite- shown to have a left lateral displacementof severalkilome- rich shear zones in the center are enriched in radioactive ters (Jones, 1980) measured as 6 km to the east of the minerals (Tukiainen et al., 1984). inland lake, Moztfeldt So (Tukiainen et al., 1984).The Geo- The simplified geological map in Figure 1 shows the logical Survey of Greenland is currently engaged in de- location of analyzed whole-rock samples (XRF) selected for tailed investigations of several syenites and of radioactive additional REE determination by Instrumental Neutron mineralization in and around the center. There are six Activation Analysis (INAA). The XRF analyses were ob- major syenite units, prefixed "SM" and numbered I to 6 in tained by Jones (1980) from which the following summary order of intrusion as indicated on the summary geological is largely extracted. The ranges of zoned mafic minerals, map in Figure 1. Future refinements to the geological especially Fe/(Fe + Mg) and Na/(Na + Ca) in pyroxene, boundaries of the syenite units are unlikely to affect the are related to the calculated normative differentiation index major conclusions of this paper, but might provide answers (D.L) of the host rock. The D.L was calculated after Thorn- to some of the outstanding problems. A unit of late-stage ton and Tuttle (1960) with the addition of normative lujavrites (SM6) and a sub-unit of cancrinite-rich roof-zone acmite (NaFe3Si2O6), which is important in the more frac- syenites (HY) were recognized subsequent to the work of tionated syenites.Whole rock P and Ti decreasewith in- Emeleus and Harry (1970) by Jones (1980). Unequivocal creasing D.I. and increasing Fe/Mg, consistent with p€tro- time relationships between units SM5 and SM6 have not graphic textures which indicate early crystallization ofapa. been established, since they are separated by the Flink's tite and Ti-magnetite. Many of the syenites show cumulus Dal Fault. The center is cut by a regional swarm of ENE- textures in thin section and classic mineral banding, al- trending trachytic dikes and a prominent alkali gabbro though rare, does occur in the center (Emeleus and Harry, giant dike (AGGD) intruded in similar orientation. A few 1970; Jones, 1980). Mineral compositions show normal near-vertical carbonatitic and ultrabasic breccia pipes up zoning and combined trends which are the expectod conse- to 200 m across also cut the center, but are not shown on quences of crystallization from increasingly fractionated oy- Figure 1. The minor intrusions remain an attractive goal enite magmas. In addition to continuous zonation, the for future study and it is not known how they relate to the mafic minerals also show discontinuous zonation of the syenites.Thin discontinuous fluorite- and calcite-rich veins type pyroxene-amphibole-pyroxene (Jones, 1984). Bulk are found sporadically throughout the center and may pro- feldspar probe analyses are difficult to obtain because of vide clues to late stage metasomatic processesseen in some perthitic textures and show a wide range in Ca-poor com- JONESAND LARSEN:NEPHELINE SYENITES FROM THE MOTZFELDT CENTRE 1089
dike relatedto the marginsof unit SM5 (SM5{')contains well defined plagioclasecores with irregular outlines and lamellar twinning abruptly terminatedby mantlesof crypt- operthitic alkali feldspar.Three reconnaissancemicroprobe analysesof the plagioclaseshow a range in composition from Ab.rAnrrOr, to AbroAnrrOr,n and are plotted in the feldspar ternary in Figure 2. Similar, though slightly lesscalcic plagioclasecores (
Table l. Rock samples analyzed
Unit Sample Rock type Special features Mafic mi4erals anal yzed"
st41 63703 syen1 te AMPN AM4 nepheline syenite Zr-r'ich aegirine 63760 syen i te Nb-rich astrophyllite AMPH 58007. nepheline syenite 54163D
SM4 58039c nepheline syenite o'l i vi ne-beani ng AMPH 58065 nepheline syenite AMPH,AENIG AM65 nepheline syenite i ntercumulus amphibole
SM5 58036 larvik'ite ling dyke plagioclase relics 0L larvikite plagioclase 58037 'larvikite ring dyke rel ics MICA AM54 ring dyke plagioclase rel ics AM39 larvikitic margin ol ivine-r'ich CPX.0L AM38 larvikitic margin olivine-rich CPX.0L AMl1O nepheline syenite CPX At482 nepheline syenite cumuluseudialyte(rare) AENIG
SM6 Al"l139 white lujavrite 46261 dark lujavrite eudialyte phenocrysts At4137 white lujavrite eudialyte-poor AI1140 white Iujav[ite AMPN
-Mafic silicate analyses given in Jones (1984) for AMPH= amphibole.0L = olivine, CPX = clinopyroxene. AENIG= aenigmatite and lt4ICAas ind'icated, Dwhole rock sample not analysed for major elements but believed sirnilar to 58007. cRb and Sr data given in Blaxland et al (1978) Samplesfrom each unit arrangedwith increasing D.I. downwards(see Table 2). 5-digit numbersare GGU(Greenland Geological Survey)s AM-prefixes are from A,P.Jones.
Table 2. Whole-rockmajor elementanalyses (wt.%)
Unit SIt[ Sl'14 s!15 s},t5 Sdrple rc: 63703 Al44 63760 58007 58039 58066 a,r65 58o36a58o3fAM54a lr',r39a a,r38 Ar{Lo Ar.482 AI'1I3946251 AI"1I37 AXt40 si-o2 60.98 56.51 63.69s8.27 54-'1754.85 56.2'7 s1.r0 sr.87 53.44 49.7355.95 57-24 5'1.93 54.0455.97 55.59 58.0? Tio2 0.88 0.36 0.22 0.3I 0.91 0.88 0.09 2.4'1 2.30 2.26 I.58 1.58 0.34 0.34 0.37 0.02 0.16 0.07 N20: 16.5520.sl 15.48 20.14 18.99 19.30 21.55 15.8215.96 t5.'?'t 16.I4 r'1.9120.22 20.66 16.6818.72 19.47 2L.29 F"203 3.lo o.oo 0.28 0.48 Lo.257.42 5.c3 '1.93 FeO* 4.4'1 5.06 3.05 4.09 6.54 2.6'1 10.87 9.s4 9.46I2.3r 6.70 4.30 2.9L 1.48 1.39 I.26 2.3r IlrO 0.19 0.]-8 0.2L O.L7 0.30 0.28 0.r7 0.29 0.25 0.25 0.51 0.21 0.20 0.14 0.37 0.17 0.27 0.30 r@ 0.67 0.60 0.08 0.29 0.87 0.93 0.04 2.'t2 2.50 L.73 2.43 1.21 0.44 0.30 0.48 0.03 0.03 0.18 4,32 I.45 0.57 1.10 2.t4 2.L8 I.L2 5.65 6.40 s.26 4.r5 3.43 r.O3 0.92 0.99 0.?5 0.92 0.35 '7.'72 Na20 s.s8 8.06 6.41 8.00 7.35 r0.50 4.38 4.s3 5.24 5.99 6.00 9.I1 8.74 10.9210.5s 9.40 r0.70 K2o 6.27 5.94 5.s0 6.28 5.46 5.51 6.24 3.92 3.10 4.34 4.sr 4.67 5.80 6.28 3.61 3.96 5.15 4.85 0.24 0.lr 0.02 0.16 0.25 0.62 0.o7 1.85 2.01 1.30 r.43 1.20 O,14 0.13 0.04 0.02 0.tI 0.03 H20 0.36 0.74 0.85 0.47 0.94 0.05 2.51
SUM l00.ls 98.78 99.69 98.81 98.9'7 98.87 98.72 99.08 99.05 99.79 99.92 98.92 98.82 99.30 100.17100.06 100.00 98.15
Na,/(Na+K) 0.57 0.67 0.64 0.65 0.6'1 0.68 0.72 0.53 0.65 0.65 0.67 0.55 0.70 0.68 0.82 0.80 0.73 0.77 (Na+K)/AI 0.96 0.96 1.00 0.99 o.94 0.97 r-r2 0.72 0.72 0.84 0.91 0.83 1.05 1.02 1.30 1.16 I.08 I.07 I.PRMNe(Qz) I.3 19.1 (4.3) 16.6 16.0 I8.4 30.0 1.5 1.0 5.1 l3.s 3.6 22.8 22.5 r6-s 20.2 16.9 24.3 D.r. 82.8 88.I 91.5 91.7 80.8 83.5 89.0 s9.4 59.8 66.2 66.6 76.0 8't.9 90.7 73.9 82.2 90.8 93.0
Note I - *totil ircn wtse Fe2o3 rct detenhed. ArEIyss trrarged in ircrsirlg order of differotiatj-on index (D.I.) for eadl mit. a*nples frm ring dyke ard rerqiDl syenite @ntainirg plagioctase reli6. Al-I arElyses bV XXF of pressed pcdder pellets at Durhm UniveFiq', q€Pt H2O ard Fe2O3 bV titraLion. Avsage Fe3Ae2 values for iFFM €ldlatioN obtairEd for eadr uit rerer SMt = I.8, 944 = 0.66, SM5 = 0.05 (58036, 58037. AI,154)ild 0.30 (ttE Gt), St'16= 5.6. JONES AND LARSEN:NEPHELINE SYENITES FROM THE MOTZFELDT CENTRE 1091
Table 3. Whole-rock trace element analyses (ppm)
Unit Sarple 63703 AI'14 63760 58007 54163 58039 58066 AM65 58035 58037 AI454 A1439 4438 Al4I10 AI\,182 A![39 46261 AI!'1137AI{140
Ba 2L42 568 ff) /ud 5& 1254 898 57 4618 5143 2635 1702 3766 433 878 190 40 250 263 Rb 482 542 '169107 22't 422 25L 200 232 185 280 41 37 89 t22 84 222 201 313 357 't3 sr 280 71 284 131 4t3 412 55 1219 160r 872 841 1992 24r 428 50 26 56 ra 53.3 86.3 195 I29 r20 r19 167 2IL 72-2 60.4 82.7 L96 104 83.1 6?.1 270 352 339 253 Ce I33 199 4I4 274 273 290 397 444 151 ]-22 181 430 2L5 I73 L42 577 460 Nd 48.7 69.9 r20 80.4 98.s t1l 137 163 82.7 63.7 63.0 19s 99.1 59.2 48.3 2t9 2E 199 I84 Sn 8.35 11.0 17.3 12.8 15.4 15.I 17-5 26-4 15.0 11.5 L4.2 28-8 15.0 9.95 8.37 a.4 45-r 2'1.2 24.3 Eu 5.09 I.60 1.27 1.53 1.58 2.16 5.00 2.15 6.84 6.60 8-33 2.45 2.90 r.95 2.52 r.87 1.82 rb o.77 L.23 lqqta?tat L.77 2.06 3.L2 r.70 0.82 l-68 2.75 1.28 1.35 0.96 3.19 8.15 3.30 4.78 Yb t.6r 3.24 o-zJ J.zt z.)l 4-73 5.01 8.68 2.r2 I.9l 3-44 5.52 z.LL ).J) Z-OL 8.13 34.5 (3.6) 12.3 Lu 0.25 0.48 0.91 0.48 C.38 0.75 0.86 1.17 o.32 0,27 0.55 0.80 o.29 0.78 0.41 r.22 4-74 l.0r 1.68 v2433 58 30 42 52 98 33 29 32 37 26 86 25r 108 L2'l rh 3.4 10.9 22.6 L3.L 9.7 L7-2 lr9.6 25.9 2.0 2.3 5.r 10.1 3.6 7.9 10.0 ?a? qq 186 9r.4 u 4.0 2.6 113 8.9 7.3 6.0 4.0 6.9 2.7 7.O 2.O 8.2 5.7 r0.0 4.5 27 2r s8 2r 4230 1040 2449 zr 308-t.4 870 558 634 406 808 748 519 L76 180 448 783 282 1324 70'l 20c5 Hf 20 28^ 18 10 24 2I T4 5.0 5-2 13 23 7.3 4I 16 53 r47 26 50 Nb 73 205 1589' 176 267 r'17 L77 46 42 102 224 82 2'tI r54 608 478 688 595 ^--a . ^ Ta 4.8 12 ZJ' LZ t6 t3 6.6 3.6 2.4 7.5 L6 5.5 26 l0 2'1 79 L2 34 to Sc 11 I.2 1.6 1.5 8.0 7.5 0.6 25 35 51 9.s 6.6 3.6 0.3 0.9-189 0.3 zn L22 123 454 r58 94.3 263 176 91.5 29'1 337 287 r47 107 84.5 682 Pb 23 10 58 19 t3 24 2233414 vl 27 51 48 57 58
Ey'E\r* 2.33 0.52 0.28 0.44 0.36 1.61 1.49 z-L5 U-6L L.ZZ o.24 0.L7 c.24 0.22 0.50 0.99'19 0.28 2.38 0.91 ce/Yb 82.6 6I.4 ob. f dJ.6 tub. z 61.3 .2 5r.2 75.9 63.9 52.6 77.9 101.932.3 54.4 74.2 L9.3 zr4tt 41.6 43.5 23.9 35.2 40.6 33.6 35.6 37.1 35.2 34.6 34.5 34.0 3A.6 32.3 44.2 37.8 28.8 40.0 40.8 Hf/"ra 1.54 I.67 0.12 1.50 1.28 r.50 L.60 2.I2 1 ?O 2 It L.73 r.44 f. rz r.t6 f.ou 1.96 1.86 2.L7 L.76 NbAa 15.2 I7.1 5.8 r4.7 16.7 2I-3 17.7 12.8 I7.5 13.6 t 4.0 12.6 r0.4 15.4 zz.> o.L
Note: - Ba Rb Sr Y Pb zr ard Nb ffi1fred by )(RF at Drrtrdn University, tlE rmirder weie mlyzed at Risd (t)eIrwk) by INAt s€ tdt for details. IIq staJdard rsults for GPG-BR versus international referc in brad detector gave a resolution of approximately 4'arc, or 5 eV for the This samplehas beenhydrothermally alteredand contains region of interest, but several overlapping lines could not be neither free quartz nor nepheline.The peralkalinity index avoided, despite selection of La or Lp lines, and an interference increaseswith D.I. such that the fractionated samplesof matrix was established by experiment. Thus, interferen@s among each unit are peralkalinewith (Na + K)/Al > 1, as are all the REE were removed iteratively and finally conventional ZAF samplesof unit SM6 (lujavrites).The ratio Na/(Na + K) corrections were applied to the complete analysis. Errors become varieslittle within eachunit, and the total rangefor all of particularly serious at the low levels of heavy REE (especiallyHo, the unitsis from 0.57to 0.82. Tm, Lu). Subsequent inspection of the results showed residual plot of the major oxidesversus SiOr (not illustrated) interference from Mn on Dy, which was discarded on Mn-rich A the larger data base minerals, and Eu represents a maximum since a small contri- showsthe samefeatures exhibited by bution from Pr or Sm at the EuLp peak used could not be avoid- (90 additionalsamples) of Jones(1980). Such Harker vari ed. ation diagrams are of limited use, due to the restricted rangeof SiO, in mostunits, except for SM5.The latterunit Whole rock geochemistry showssimple trends of decreasingMgO, FeO, PrOr and TiO, with increasingSiOr; CaO decreasesafter a hiatus Major elements near52 wt.% SiOr, and AlrO. remainsnearly constant to Whole rock major element analyses are given in Table 2 approximately53 wt.ok SiOr, above which it increases. and include FerO. and HrO for several samples. Calcu- Apart from the two silica-rich samplesof SMl, all the lated differentiation indices (D.L) from the CIPW norm are major oxidesof the other syenitescluster loosely around based on the D.I. of Thornton and Tuttle (1960) plus nor- the SM5 data in the range54-58 wt.% SiO2. mative acmite. D.I. values range from 59 to 93 (Table 2) and on this basis samples can be conveniently referred to Minor elements as more- or less-fractionated.The syenitesall contain nor- Table 3 givesanalyses for 8 REE and 13 other minor mative nepheline (Ne) which, apart from the low values, elementsfor the 19 samples.Blank spacesin Table 3 are corresponds more or less with modal proportions esti- values below detection limits and the precision for each mated from hand specimens and thin sections. One obvi- elementcan be approximatelyinferred from the number of ous exception is sample 63760 of unit SMI which has nor- significant figures quoted. Ba, Sr and Sc decreaseby ap- mative quartz (Qz) at the Fe3*/(Fe2* + Fe3*; value used. proximately two orders of magnitude from their highest tw2 JONESAND LARSEN:NEPHELINE STENITES FROM THE MOTZFELDT CENTRE UJ F E o z O lo I O Ne-syenif jovrite Y e SM4 Lu SM6 O O E [! ul E Loc. Loc. s#u YbLu srEu YbLu Fig. 3. Whole-rock REE normalized to chondrite abundances (Nakamura, 1974) plotted on log scale. values in the least-fractionated samples and have their expressednumerically as Eu/Eu* where Eu* is the expected lowest values in the lujavrites of SM6. Zn decreaseswith Eu interpolated from the two adjacent REE (in this case increasing D.I. in each unit but is highest again in unit Sm and Tb) assuming no Eu-anomaly. The least frac- SM6 and the hydrothermally altered sample of unit SMI tionated sample grouped with unit SMI (63703) has the (63760\. Zr, Hf, Nb and Ta are highest in unit SM6, al- lowest REE and a prominent positive Eu-anomaly. This is though these elements do not increase systematically with very different from other samples of SMI and is now D.I. or any other obvious element.The averageZrlHf ratio thought to belong to an earlier syenite that predates SMI is 37 (+4.2,2o) lor the nineteen samples,excluding 63760 (Tukiainen, pers. comm. 1984) and is excluded from Figure which has ZrlHf :23.9. The ratios Hf/Ta and Nb/Ta are 3. All samples from SMI have a negative Eu-anomaly reasonably constant at approximately 1.5 and l5 respec- which shows some correlation between its size (Eu/ tively, for most samples except 63760 which contains Nb- Eu* < l) and the level of total REE. The most fractionated rich astrophyllite. The Nb/Ta ratio is also particularly vari- sample of SMI contains the most REE, except for Eu, and able in SM6 which contains pyrochlore. is the silica-rich hydrothermally altered 63760. The three samples of unit SM4 have similar REE levels (Fig. 3) de- REE spite their large geographical separation (Fig. l) and mea- The REE whole-rock data are given in Table 3 and were sured range in major element chemistry (Table 2). One normalized using the customary chondrite abundances of sample has a negligible Eu anomaly and two have substan- Nakamura (1974\.The normalized REE (REEN) are plotted tial negative Eu-anomalies.By a narrow margin, unit SM5 for each syenite unit in Figure 3. Recent revisions of has the overall lowest REE levels and five of the seven chondritic abundances,such as those employed by Davis et samples have positive Eu-anomalies of varying size. The al. (1982)would increasethe REEI levels overall, but cause latter include three samples from the earliest marginal lar- only minor differencesin their relative distributions. For vikite (SM5* partial ring dike, Table l) which contain descriptive purposes, a distinction is made between least plagioclase cores to alkali feldspars. There is a suggestion fractionated and fractionated samples of each unit, based that the three samples of SM5 with small Eu-anomalies on D.I. This does not require that all samples are related (AM82, AMll0, AM54) have a kinked REE pattern, which by crystal fractionation, but is a useful mechanism for dis- is steep from La to Sm but flatter in the heavy REE from cussl0n. Tb to Lu; the same is true for two samples of SMl. This As illustrated in Figure 3, all of the samples are steeply really requires a closer spread of analyzed heavy REE be- light REE-enriched, with high positive values of Ce/Yb and tween Tb and Lu to analyze further. Unit SM6 lujavrites most have a Eu-anomaly. The Eu-anomaly may be nega- have the highest REE levels, and these become less steep tive or positive and varies greatly in relative size. It can be (CelYb from 162 to 19, Table 3) with increasing total REE JONESAND LARSEN:NEPHELINE SYENI?ES FROM THE MOTZFELDT CENTRE r093 usually has a negative Eu-anomaly (Watson and Green, 1981)is common in the early and mafic syenites,such as the larvikites of SM5, but is largely absentfrom the more fractionatedsyenites. This is reflectedin whole-rock PrOt values in Table 2, which decreasewith increasing D.I. A Zircon, which has a heavy REE-dominated pattern (e.g., o Exley, 1980) is common in many of the nepheline-poor A syenites.Euhedral zircon forms up to 10% of the mode in rare mafic bands with melanite garnet and euhedral py- o8o in unit (Jones,1980). Zirconium also occurs 63760 rochlore SMI in solid solution in late-stageaegirine, with up to 7.0 wt.o/o 200 400 m0 8m 1m0 Dm 1400 1600 18m 2m ZrO2 in syenitesof unit SM3 (Jonesand Peckett, 1980).In Sr pprn all analyzedaegirine, however, REE levelswere below de- Fig. 4. Eu/Eu* versusSr (whole-rock).Eu* is straight line in- tection (50 to 100 pprn etrement)even in the most Zr-rich terpolated value betweenSm and Tb, assumingno Eu-anomaly. aegirine.More irnportantly" Zr occurs in the mineral eu- The data show good correlation betweendecreasing and Eu/Eu* dialyte, Naa(Ce,Ce,Fe)tZrSiuOtt(OH,F,Cl)2in the most Sr. Sample 63760 indicated for SMI is hydrothermally altered syenite. fractionated syenitesand this mineral contains up to 7.0 wt.% RE-oxides(RE2O3, Table 4). Eudialyte occursas an essentialmineral in the lujavrites (SM6) where its habit of phenocrystor interstitial texture (Fig. 5) distinguishesdark lujavrite from white lujavrite (Jones,1980). Eudialyte from abundances.The one dark lujavrite (46261)typified by eu- unit SM4 (AM49, Fie. 6) is light REE-enrichedand eu- dialytephenocrysts (Fig. 5) hasthe highestREE measured, dialytefrom white lujavrite(Table 4, AMl59) containsthe and the pattern is more obviouslykinked. Comparison of highest RE2O3. It should be emphasizedthat the heavy all the patternsin Figure 3 showsthat of the threesyenite REE measuredby electron probe are close to theoretical unitsanalyzed, SMI and SM4 havethe closestsimilarity to detection limits and have serious limitations in precision unit SM6; this is important when considering the pet- (circa * l0oh or worse). Nevertheless,the analyzed eu- rogenesisof the lujavrites.Thus, there is a reasonableover- dialyteshave a kinked REE pattern which is steepfrom La lap betweenthe highestsamples of SMI (63760)and SM4 to Nd and flatter from Sm to Yb (Fig. 6). Y generally (AM65) and the lowest REE sample of the lujavrites behaveslike Ho, a heavy REE (e.g.,Jones and Ekam- (AM140). baram, 1985) and measuredvalues for the two eudialyte There is an interesting relationship between the Eu- grainsin Table 4 havechondrite-normalized abundances in anomaly and Sr for all the syenites,shown in Figure 4. the range 3-7 x 103, in agreementwith the heavy REE Eu/Eu* decreasesuniformly with decreasing Sr, with shown in Figure 6. This kinked pattern for eudialytecom- Eu/Eu* > I in Sr-richlarvikites to Eu/Eu* < I (i.e.,nega- paresfavorably with the whole-rock REE for dark lujavrite tive Eu-anomaly)in Sr-poorernepheline syenites of SM5. (46261')which is similarly kinked (Fig. 3) and shown for A similar relationshipis shownby the other syenites,with comparison in Figure 6. REE in the dark lujavrite might lowest values of Eu/Eu* representingthe largest negative reasonably be expected to reflect the REE-bearing eu- Eu-anomaliesfound in Sr-deficientlujavrites of SM6. Simi- dialyte. In detail, the ratios ZrlCe and ZrlYb for eudialyte lar correlationsare found in analogousplots of Eu/Eu* and whole rock do not agreevery closely.This is probably versusBa and Ca (not shown).This is important sinceBa, becauseof additional Zr and REE mineral hosts in the Sr, Ca and P all decreasewith increasingD.I. ahd SiOr, lujavrites including alteration products of eudialyte.Whole and both apatite and plagioclasefeldspar are common in rock Zr abundancesfor the dark lujavrites (Table 3) agree the early larvikitesof SM5. Their relativeroles might be with modal estimatesof circa 5% eudialyte phenocrysts expectedto directly influence Eu-anomaliesand are dis- (i.e.,5o/o x L2.2wt.%o ZrOr:4500 ppm Zr rock).Conse- cussedin a latersection. quently, the REE are expectedto be 2-3 times lower in unanalyzedinterstitial eudialytefrom white lujawite, which REE-bearing minerals has similar amounts of eudialytebut lower whole rock Zr. The REE are presentin a number of different minor and Much of the eudialyte in the lujavrites has been replaced accessoryminerals in the Motzfeldt syenitesand some of by secondary hydrous minerals, including catapleiite these increasedor diminished in importance possibly in NarZrSi3Oe'2HzO (Fig. 5). The samelujavrites show re- responseto changingcompositions of host magmas.Some placement of nepheline by cancrinite, fine-grained white may have beenreplaced during circulation of metasomatic mica (?Li-rich) and sodalite; coexistingfeldspars of albite fluids. A number of theseminerals have been analyzedby and microcline are uneffected.The evidencesuggests eome electron microprobe and representativeanalyses are given alteration by alkali-rich, Cl-bearing hydrous fluids, which in Table 4. Examples of their textural relationships are might also have modified the REE, ZrlCe and ZrlYb rock illustrated in Figure 5. Apatite, which is characteristically ratios. light REE-rich (e.g.,Henderson, 1980; Larsen, 1979) and Anhedral sphene(CaTiSiOr) is locally common as an tw4 .IONESAND LARSEN:NEPHELINE SYEN/TES FROM THE MOTZFELDT CENTRE Fig. 5. Photomicrographs of typical accessory phases in the Motzfeldt syenites: all have width near 2 mm, all plane polarized light except (c) which has partly crossed polars. (a) Clear, well-cleaved (and parallel twinned, not shown) rinkite in center is interstitial to turbid alkali feldspar, clear nepheline and dark amphibole (AM7, unit SM4). (b) High relief and irregularly fractured phenocrysts of eudialyte (center and above center) with aegirine, nepheline, albite and microcline (various shades) in dark lujavrite (AM164, unit SM6). (c) High relief interstitial and partly altered eudialyte (just right of and below center) with aegirine blades (dark) in white lujavrites (AM138, unit SM6). (d) Moderate relief interstitial Mn-pectolite with good cleavage(near center) with albite (clear),microcline (turbid) and bladed aegirine (dark) in white lujavrite (AM139, unit SM6). accessory mineral in the Motzfeldt syenites; although common in the adjacent North Q6roq Centre, where based on typical analytical totals near to IOO%, it is not a Chambers (1976) established a correlation between the major repository of REE. Sphene can be light REE- abundance of rinkite and the degree of contamination of enriched in felsic rocks (e.g.,Exley, 1980; Fleischer, 1978; the syenites by metasomatized country rocks. These min- Henderson, 1980). The majority of the Motzfeldt syenites erals are usually interstitial, or less commonly euhedral, contain no sphene but usually have accessory minerals and colorless to pale yellow;they are rarely visible in hand of a variety of Ca-Na-Ce-Ti-Zr-Nb-silicates of the specimen. They often have anomalous birefringence and rinkite-mssandrite and livenite series;these contain up to can have lamellar twinning superficially resembling plagio- -22 wt.Yo RE2O3 (Table 4). The same minerals are also clase. Rinkite can be represented by the formula JONES AND LARSEN:NEPHELINE SYENIT'S FROM THE MOTZFELDT CENTRE 1095 Table 4. Electron probe analyses of accessory minerals (wt.%) 23* 4* 5c* 5r* 6 7* 8r sphene zircon 1av rink rink rink lav eud eud Mn-pect Nb2o5 0 .90 2.63 2 .66 3.0r 1.61 13.50 L.74 2.47 n.a. sio2 29.42 32.29 30.30 30,45 31.19 30.04 30.24 47.33 46.51 55.95 TiO2 31.41 0. I5 o.74 10,41 I0.34 6.69 r..01 o.15 0. 11 0.o2 zro2 0.20 66.l3 I7.13 I. 47 6.26 9.15 15.09 72.52 t2 .27 Hfo 1.03 2 AI2O3 0.64 0.05 0.01 5.03 1.38 0.19 0.08 0.r8 0 .'13 0. r1. Y ^O. 2.7'1 2.65 1.90 0.9- 1.5- La2O3 0.17 3 .46 1.53 0.4r 0.49 1.40 Ce203 0.49 10.49 f .fJ t-)z t.29 2.87 Pt203 0.09 I. U] 0.53 0.29 o.22 0.39 Nd203 0 .26 4.36 2.20 0.73 0.41 0.16 Sm203 0. 18 o.77 0.51 0.24 0.15 0.26 Gd203 0. 15 0.68 0.48 0.25 0.10 0.28 Dy^o. b 0.14 0.63 0.36 b -" 0.13 .27 0.20 2'3 0 0.28 0.24 o.12 Yb203 0.07 0.15 0.25 0.2r 0.06 0.25 FeO 1. 35 0.27 1.53 6.63 0.97 0.82 0.78 6.9',7 6.73 1.07 MnO 0. 13 0.22 I.37 0 .09 0.79 0.91 0.71 2.r9 4.59 15.56 MgO 0.19 0.08 0.05 0.03 n.d. n.d. 0.1I n.d. n.d. 0.16 CaO 27.83 0.04 30.78 10.18 20.30 28.29 30.54 10.40 b . z) f o .4{ Na2O 0.84 0.05 6 .49 I.96 3.r2 8.r2 2.L6 12.13 r0.79 8.82 K2O 0.03 0.02 0.02 u. a5 0.2L 0.08 n.d. 0.I7 0,t2 0.08 99.50 100.33 94.56 93.2L 92.26 92.06 94.22 91.53 98.43 98.2L Note: - *REE patterns illustrated for these minerals in Fig. 5. I = AlaSl interstitial sphene in coarse-grained ne-syenite (SIa5). 2 = 63717 zircon core of euhedral grain (0.2 cm) j.n mafic banded syenite (SM3). 3 = AM5l ?lavenite interstitial grain in foyaite (SM4). 4 = AM81 yellow euhedral ?rinkite (0,3 cm) in coarse-qrained ne-syenite (sM5). 5 = AM84 core (c) and rim (r) of yellow zoned ?rinkite euhadral grain (0.5 cm) in coarse-grained ne-syenite (SM5). 6 = 63725 Nb-rich ?lavenite interstitial grain (sM4). 7 - AM49 interstitiaf colorless eudialyte in foyaite (sI44). 8 = AM159 interstitial pale yel1ow eudialyte in white lujavrite (SM6). 9 = AM138 colorless to paLe yellow/ brown interstitial Mn-pectolite in white lujavrite (SM5). aestimated bserious from peak counts only, inteference from Mn. n.d. = rDtdetected.. (Ti,Nb,Zr{Na,Ca)r(Ca,Ce)o- (SirO?)r(O,F)u (Galli and Al- 6. The larger crystals ( > I mm) are zoned from REE-rich berti, 1971)and a typical example from Motzfeldt is shown cores to relatively REE-poor rims (Table 4). in Figure 5. A more general formula for this family would be XYZTOF(SirO?) and a lew representative (partial) Interpretation analysesare given in Table 4. Recasting the cation propor- tions of seven such analysesassuming Si : 2.00 gave totals Petrogenesis i for (X + Y + Z) from 3.01 to 4.05, with cation totals vary- The systematic variations shown by major elements be- ing from 7.91 to 9.25, indicating more than one mineral tween early mafic larvikites and evolved nepheline syenites type. There is a chemical similarity between the of unit SM5 can be related either by fractional crys- rinkite/lavenite minerals and sphene (Table 4) considering tallization or partial melting. Using the most basic mineral the nearly constant SiO, and the common substitutions of compositions analyzed from the larvikites (Jones, 1984) the Na. Nb and Zr for Ca and Ti seen in other minerals. trends have been modelled (not shown) by early extraction Sphene and rinkite do in fact share some structural ele- of olivine (Ol) and plagioclase(Pl) in approximately equal ments (P. B. Moore, pers. comm., 1984) and chemically proportions, followed by extraction of Ol + Pl with ad- rinkite representsan enriched variety of sphene.The not- ditional clinopyroxene (Cpx) with or without amphibole ably fresh appearance of rinkite in some altered syenites (Amph). Alternatively, successivelyderived partial melts with clouded feldspars and general petrographic textures would be in equilibrium with residues of Ol + Pl + Cpx hint at a possible metasomatic origin. Detailed electron t Amph and Ol + Pl respectively.Olivine and clinopyrox- probe analysesof the REE in rinkite are plotted in Figure ene occur as phenocrysts or early formed crystals, and 1096 JONES AND LARSEN:NEPHELINE SYENITES FROM THE MOTZFELDT CENTRE subtraction of Cpx + Amph above approximatelySiO2 : 54 wtJ/o SiO, coincideswith the transition from larvikites to nephelinesyenites. This might be interpretedas a change from Ol * Pl controlled magmas in the ring dike to Ol + Pl * Cpx * Amph in the main chamber at current exposurelevels. We believethat the magmasare derivedby crystalfractionation processes. By analogy with unit SM5, the other syeniteunits of the Motzfeldt Centre are taken to have similar origins. This is supported by overlapping whole-rock and mineralogical o trends(Jones, 1980, 1984) and field relations,although only L SM5 is clearly associatedwith larvikitic magmas.Time intervalsbetween overlapping intrusions were at least long o (-)-c. enoughto allow solidification.Rather similar fractionation schemeswere deduced by Upton and Thomas(1980) for an ul lu important suite of transitionalolivine basaltsin the Late GE Gardar Younger Giant Dike Complex at Tugtutdq (YGDC). The Tugtutdq YGDC was related by relatively low pressure(crustal, < 10 kbar) Ol + Pl fractionation,and an even deeper level evolution apparently dominated by pyroxeneand garnet.The Tugtutdq YGDC encompassesa larger compositionalrange than 8 preliminary analysesof the Motzfeldt AGGD, but their ultimate petrogenetichis- tories are undeniably similar. Even plagioclasemegacrysts in the Motzfeldt AGGD and anorthositeinclusions in the Tugtut6q YGDC have similar compositions (Anrr, Anrr_rr). These giant dikes tapped a magma source that may have beenavailable in large quantitiesthroughout the Lo Ce ft Nd Sm GdTb Dy Er '/b Lu Gardar Period. Indeed they may have beendirectly related Fig. 6. Chondrite-normalizedREE data for accessoryminerals. to the intrusive centers by dilation and subsequentring (a) RinkiteJivenite types including core and rim of zoned grain; fracture,as envisagedby Bridgwaterand Harry (1968).The (b) eudialytefrom unit SM4 (AM49) and unit SM6 (AM159) consistentlylow initial E7Sr786Srratios are taken to imply showing kinked REE patterns(steep La to Sm, flatter Gd to Yb) source(Blaxland et al., 1978;Patchett et al., 1976) which compare well with the whole-rock data for a lujavrite a mantle sample,which has eudialytephenocrysts (46261, similar to illustra- and preclude large scale involvement of the Julianehib tion in Fig. 5 (b)). Granite (uppercrust). Eu-anomalies The most fractionatedsyenites have the largestnegative Eu-anomalies,there is a simple relationship for Eu/Eu* plagioclaseis found in coresof alkali feldsparsin the larvi- versus Sr (Fig. a), and Sr decreaseswith increasingfrac- kites; amphibole only occurs as an interstitial mineral. tionation. Thus, large positive anomaliesexist in the least- Field relationsshow that the larvikitic marginsgradually fractionated Sr-, Ba- and P-rich larvikites of SM5, which mergeinto nephelinesyenites and occasionalbanding with are consideredto be relatedto the evolvednepheline syen- cumulustextures is found in the nephelinesyenites of SM5 ites by crystal fractionation (section above).Similar posi- (Emeleusand Harry, 1970).Plagioclase compositions used tive Eu-anomaliesexist in the Oslo region alkaline rocks, were constrained by relations between CaO, AlrO, and in most of the lardalites (nepheline-richlarvikites) and SiO, to be at least as An-rich as the most calcic cores ih about half of the larvikites (Neumann, 1980),and their the larvikites( by Larsen(1979) and theseare remarkablysimilar to the dark lujavrite (46261)of SM6 from Motzfeldt, as shown in Figure 7. The euhedraloutlines and petrographictextures llimaussaq of eudialyte in fractionated syenitesfrom the central re- gions of SM5 (Jones,1980) suggest that eudialytejust achievedthe statusof a cumulusmineral at Motzfeldt.It seemsreasonable that extrapolationof this processcould have led to the formation of layered cumulate eudialyte- syenitessimilar to thoseso well displayedin the Ilimaussaq Intrusion (Ferguson,1970). The lujavrites at Ilimaussaqare late-stageproducts, just as in Motzfeldt, and future com- parative work betweenthe two occurrencesmight be very useful,although it is beyondthe scope ofthis paper. 46261 Metasomatism Evidencefor metasomaticfluids is seenin the lujavrites in the form of replacementof eudialyte by catapleiiteand La Lu other minerals.On a more widespreadbasis, syenites from Fig. 7 Chondrite-normalizedREE distributions for the lujav- elsewherein the centershow nephelinereplaced by sodalite rites (unit SM6, 46261)from Motzfeldt the hypo- comparedwith and cancrinite,with the developmentof albitic rims to per- thetical parental magma for the llimaussaqintrusion (Larsen, with experimentally-derivedstability r979). thites. By analogy limits of sodalitein aqueoussyenite systems (Wellman, 1970)these replacement textures are interpretedto have resultedfrom metasomatismby aqueousfluids. These fluids may have aflectedwhole rock ZrlCe and ZrlYb ratios, and kinked patterndue to someheavy REE. It is possiblethat might evenhave beenresponsible for somegrowth of rink- most of the REE which would have beenpresent in zircon, ite minerals.The fairly common replacementof nepheline are instead accomodatedin eudialyte in the more frac- by natrolitemay alsobe relatedto the samefluids. tionatedsyenites. Thus, there is a strongand expectedcor- Given that metasomaticfluids were widely availablein relationbetween the probe-measuredREE in eudialyteand the Motzfeldt syenites,especially in the roof zones,then the the INAA-measuredwhole-rock analyses of the lujavrites sampleof alteredSMI syenite(63760) becomes particularly (Fig.6), whose primary REE host is eudialyte.Similar com- interesting.This sampleis enrichedin many of the same parisons between probe analyses of dominant REE- elementsfound in the lujavrites,notably Zn and Pb (Table mineralsand host rock REE existbetween perovskite and 3), and is evenmore enrichedin Nb and Ta, due to the kimberlite(Jones and Wyllie,1984). presenceof Nb-astrophyllite.The REE,including the nega- tive Eu-anomaly,are also quite similar to the lujavrites D euelopment of luj aurite (Fig.3). This suggeststhat the metasomaticfluids, although Field evidenceshows that the lujavritesintruded earlier not necessarilycontemporaneous, have apparently en- syenitesand behavedas magmas(Jones, 1980). The sim- richedsome of the normal syeniteswith the sameelements plest interpretationis that the lujavrites are late-stage that are characteristicof the lujavrites. The origin of the productsderived by extensivefractionation of the syenite lujavriteswas not metasomatic,as outlined above,but it magmasof unit SM 1 or SM4, and they wereemplaced at a should be noted that the processesof late stageigneous high structural level.Their remarkablyhigh Zr contents enrichmentand metasomaticfluids are not mutuallyexclu- areconsistent with experimentalresults and aretaken to be sive,and may of coursebe related. due to complexingof dissolvedZra* with "free" alkalies not associatedwith Al in the melt; a natural consequence of peralkalinity (Watson, 1979; Watson and Harrison, Summary 1983).In terms of their chemistry,mineralogy and oc- The following petrogenesisof the centerhas beenpro- currence,the lujavites (SM6) from Motzfeldt are com- posedabove: parable with lujavrites and other agpaitic rocks in the (1) Sr isotopic evidencesuggests an ultimate mantle IlimaussaqIntrusion, S. Greenland(Ferguson, 1964; Bailey source and the most basic magma available was alkali et al., 1978)and in Lovozero,U.S.S.R. (Vlasov et al., 1966; gabbro (Ti-rich olivine-basalt)seen in the late giant dike Gerasimovskyand Kusnetsova,1967). The SM6 lujavrites (AGGD). Alkali gabbro,larvikites and nephelinesyenites are somewhatcoarser-grained than their counterpartsin can be relatedby progressiveremoval of Ol + Pl at deeper Ilimaussaq,and have geochemicalaflinities with kakorto- levelsiollowed by Ol + Pl + Cpx t Amph at currentlevels kites from the latter intrusion (Ferguson,1964). Likely of exposure. rangesfor the REE content of the initial magma to the (2) PositiveEu-anomalies in the larvikiteshave been ex- IlimaussaqIntrusion (?olivine-phonolite)were estimated plained by incorporation of earlier plagioclase-richcumu- JONESAND LARSEN:NEPHELTNE SYEN/TES FROM THE MOTZFELDT CENTRE 1099 latesand are seenas plagioclaserelics. This could be tested land, 186, also Bulletin Grsnlands geologiske Under- since plagioclaseseparates from the larvikites, and prob- sogelse,85. ably from the AGGD shouldhave substantial positive Eu- Emeleus,C. H. and Upton, B. G. J. (1976)The Gardar periodin anomalies. southern Greenland.In A. Escher and W. S. Watt, Eds., The Geologyof Greenland,p. 152-181.Geological Survey of Green- (3) The lujavritesare late-stageproducts derived by ex- land,Copenhagen. tensive fractionation of nephelinesyenites, either of unit Exley,R. A. (1980)Microprobe studiesof REE-rich accessorymin- SMI or SM4. erals: implications for Skye granite petrogenesisand REE mo- (4) Metasomatic fluids apparently affectedmany of the bility in hydrothermal systems.Earth and Planetary Science syenitesand have geochemicalsignatures similar to the L€tters,48, 97-l 10. late-stagelujavrites. A relatedorigin is implied but more Ferguson,J. (1964)Geology of the llimaussaqalkaline intrusion, work is needed. South Greenland.Meddelelser om Grsnland, 172,also Bulletin This schemeis in generalagreement with the widely ac- Grsnlands geologiskeUndersogelse, 89. ceptedview of parental augite syenite(similar to SM5 lar- Ferguson,J. (1970)The differentiation of agpaitic magmas: the vikites)to many of the Gardar centers,and emphasizesthe IlimaussaqIntrusion, South Greenland.Canadian Mineralogist, importanceof giant dikes such as those of transitional oli- 10,335-349. Fleischer,M. (1978)Relation of the relativeconcentrations of lan- vine basaltfrom Tugtut6q (Upton and Thomas,1980), in thanides in titanite to type of host rocks. American Mineral- the Gardar province. The unexpectedlyhigh positive Eu- ogist,63, 869-873. anomaliesand the apparentsignificance of plagioclase lead Galli, E. and Alberti, A. (1971)The crystal structure of rinkite. back to the ideasof Bridgwaterand Harry (1968),which Acta Crystallographica,B.27, 1277-1284. seemto warrant renewedinterest. Gerasimovsky,V. I. and Kusnetsova,S. Ya. (1967)On the pet- rochemistryof the Ilimaussaqintrusion, South Greenland.Geo- Acknowledgments chemistryInternational, 4, 236-246. Giradi, F., Guzzi, G. and Pauly, J. (1965) Reactor neutron ac- The Danish Natural Science Research Council are thanked for tivation analysisby the single comparator method. Analytical grant no. 1l-O42'1.J. G. Holland and A. Peckett gave useful assist- Chemistry,37, 1085-1092. ance with XRF and electron probe techniques at the University of Hamilton, D. L. (1961)Nephelines as crystallisationtemperature Durham, and P. B. Moore (Chicago) is thanked for help with indicators.Journal of Geology,69, 321-329. some REE-minerals. The Geological Survey of Greenland provid- Henderson,P. (1980) Rare earth element partitioning between ed generous support at all times and this paper is published with sphene,apatite and other coexistingminerals of the Kangerd- the Director's permission. Iugssuaqintrusion, E. Greenland.Contributions to Mineralogy and Petrology,72, 8l-85. References Jones,A. P. (1980)The petrology and sructure of the Motdeldt Allart, J. (1973)Geological map of Greenland1: 100,000.Juliane- Centre, Igaliko, South Greenland.Unpubl. Ph.D. thesis,Uni- haab 60V.2Nord descriptive text. Grsnlands geologiske versityof Durham. Undersogelse,Copenhagen. Jones,A. P. (1984)Mafic silicatesfrom the nephelinesyenites of Bailey,J. C., Larsen,L. M. and Sorenson,H. (1981)Introduction the Motzfeldt Centre, South Greenland.Mineralogical Maga- to the Ilimaussaqintrusion with a summaryof the reported zine,48,l-12. investigations. Rapport Grsnlands geologiske Undesrogelse, Jones,A. P. and Ekambaram,V. (1985)New INAA analysisof a 103,5-17. mantle-derivedtitanate mineral of the crichtonite series,with Bailey,J. C., Gwozdz,R., Rose-Hansen, J. and Sorenson,H. (1978) particular referenceto the rare earth elements.American Miner- Preliminary geochemicalwork on the Ilimaussaqalkaline intru- alogist,70,4l4_4l& sion, South Greenland. Rapport Grsnlands geologiske Jones,A. P. and Peckett, A. (1980)Zirconium-bearing aegirines Undersogelse,90,'l t19. from the Motzfeldt Centre.South Greenland.Contributions to Blaxland,A. B., van Breeman,O., Emeleus,C. H. and Anderson,J. Mineralogy and Petrology,7 5, 251-255. G. (1978)Age and origin of the major syenitecentres in the Jones,A. P. and Wyllie, P. J. (1984)Minor elementsin perovskite Gardarprovince of SouthGreenland: Rb-Sr studies.Bulletin of from kimberlitesand distribution of the rare earth elements:an the GeologicalSociety of America,89, 231-244. electron probe study. Earth and Planetary ScienceLetters, 69, Bridgwater,D. and Harry, W. T. (1968)Anorthosite xenoliths and 128-140. plagioclase megacrystsin Precambrian intrusions of South Larsen,L. M. (1979)Distribution of REE and other traceelements Greenland.Meddelelser om Gronland,185, l-66. betweenphenocrysts and peralkaline undersaturatedmagmas, Chambers, A. D. (1976) The petrology and mineralogy of the exemplifiedby rocks from the Gardar igrreousprovince, south North Qdroq Centre, Igaliko Complex, South Greenland. Greenland.Lithos, 12,303-315. Unpubl.Ph.D. thesis, University of Durham. Larsen, L. M. (1981) Chemistry of feldsparsin the ilimaussaq Cox, K. G., Bell,J. D. and Pankhurst,R. J. (1979)The interpreta- augite syenite with additional data on some other minerals. tion of igneousrocks. George Allen and Unwin, London. Rapport Grsnlands geologiskeUndersogelse, lO3, 3l-3'1. Davis,A. M., Tanaka,T., Grossman,L., Lee,T. and Wasserburg, Nakamura, N. (1974)Determination of REE, Ba, Fe, Mg Na and G. J. (1982)Chemical composition of HAL, an isotopicallyun- K in carbonaceouschondrites. Geochimica et Cosmochimica usual Allende inclusion.Geochimica et CosmochimicaActa, 46, Acta, 38, 757-775. 1627-165t. Neumann, E.-R. (1980)Petrogenesis of the Oslo region larvikites Emeleus,C. H. and Harry, W. T. (1970)The Igaliko nepheline and associatedrocks. Journal of Petrology,21, 499-531. syenite complex. General Description. Meddelelserom Grsn- Patchett,P. J.,Hutchinson, J., Blaxland, A. B. and Upton, B. G. J. JONES AND LARSEN:NEPHELINE SYENjTES FROM THE MOTZFELDT CENTRE (1976) Origin of anorthosites, gabbros and potassic ultramafic Undersogelse,48; also Meddelelserom Gronland, 169. rocks from the Gardar province, south Greenland: initial Upton, B. G. J. and Thomas, J. E. (1980) The Tugtut6q Younger ItSr786Sr studies Bulletin of Geological Society of Denmark, Giant Dyke Complex, South Greenland: fractional crys- 25.79-84. tallization of transitional olivine basalt magrna. Journal of Pet- Powell, M. and Powell, R. (1977) A nepheline-alkali feldspar geo- rology,2l, 167-198. thermometer. Contributions to Mineralogy and Petrology, 62, Vlasov, K. A., Kuz'menko, M. V. and Eskova, E. M. (1966) The t9t-204. Lovozdro alkaline massif. Oliver and Boyd, Edinburgh and Ssrenson, H. (1969) Rhythmic igneous layering in peralkaline in- London. trusions. An essay review on Ilimaussaq (Greenland) and Lovo- Watson, E. B. (1979) Some experimentally determined zir- zero (Kola, USSR) Lithos, 2, 261-283. con/liquid partition coeflicients for the rare earth elements. Ge- Stephenson,D. (1976) The South Q6roq Centre nepheline syen- ochimica et Cosmochimica Acta, 44,895-897 . ites, South Greenland: petrology, felsic mineralogy and pet- Watson, E. B. and Green, T. H. (1981) Apatite/iquid partition rogenesis. Bulletin Grsnlands geologiske Undersogelse, 118. coellicients for the rare earth elements and strontium. Earth and Sun, S-S. and Hanson, G. N. (1975) Rare earth element evidence Planetary Science Letters, 56, 405421. for differentiation of McMurdo volcanics, Ross Island, Antarc- Watson, E. B. and Harrison, T. M. (1983) Zircon saturation re- tica Contributions to Mineralogy and Petrology, 54, 139-155. visited: temperature and composition effects in a variety of Thornton, C. P. and Tuttle, O. F. (1960) Chemistry of igneous crustal magma types. Earth and Planetary Science Letters, 64, rocks: 1, differentiation index. American Journal of Science,252, 295-304. 6'.7-75. Wellman, T. R. (1970) The stability of sodalite in a synethetic Tukiainen, T., Bradshaw, B. and Emeleus,C. H. (1984)Geological syenite plus aqueous chloride fluid system. Journal of Petrology, and radiometric mapping of the Motzfeldt Centre of the Igaliko rr,49-'lr. Complex, South Greenland. Rapport Grsnlands geologiske fJndersogelse,I 20, 78-83. Upton, B. G J (1964)The geology ofTugtut6q and neighbouring Manuscript receioed,December 3, 1984; islands, South Greenland. Bulletin Grsnlands geologiske acceptedfor publication, J uly 30, 1985.