Canadian MineralogistYol.26, pp. 97-108 (1988)

MINERALOGY, PETROLOGY AND GEOCHEMISTRYOF THE ALKALINE MALPEOUE BAY SILL, PRINCE EDWARD ISLAND

JOHN D. GREENOUGH*Department of Earth Sciences, Memorial University of Newfoundland, St. John's, Newfoundland, AIB 3X5

AKIO HAYATSUDepartment oJ Geophysics, University of Western Ontaio, London, Onturto, N6A 587

V. STEPHEN PAPEZIK+Deryrtment of Eorth Sciences, Mernorial University ol Newfoundland, St. John's, Newfoundland, AIB 3Xs

AssrRAcr

A lamprophyric sill exposed at Malpeque Bay, PrinceEdward Island, formed circa 2.47 Ma ago (Late Permian)in response to early tectonism that preceded opening oftheAtlantic Ocean. The high Ca and Ti contents ofclinopyrox-ene and extremely high Ti concentrations in phlogopiteattest to the sill's alkaline nature. Extremely low Rb andK concenfiations 5rrggest that phlogopite or other K-liearingmantle phases had only a minor effect on formation theMalpeque Bay magma. Other elements such as Sr, Ba, andthe REE show more substantial metasomatic enrichmentbut the concentrations of Hf, Ta, U, Tb, Zr and Mincreased the most. The metasomatizing agent (fluid ormagma) was probably CO2-rich and either incapable oftransporting K and Rb or carried these elements throughthe source region leaving behind elements such as Zr andNb. Whole-rock Fe-Mg-Ni relationships, the high Mg con-tent of olivine phenocrysts, and the presence of spinel lher-zolite nodule indicate a,primary magma. The low equilibra-tion temperatures of the xenoliths (-970"C) suggest thatthey cannot represent the magma source but are typical ofthe recrystallized and re-equilibrated xenolitls associatedwith alkali basalts and nephelinites.

Keywords: lamprophyre, olivine nephelinite, primarymagma, spinel lherzolitA, sill, Late Permian, rifting,Prince Edward Island, geochemistry, major elements,trace elements, rare-earth elements, me&Nomatism,petrogenesis.

Sola\aalns

Un filon-couche lamprophyrique a 616 mis en place prbsde la baie de Malpeque, sur I'lle du Prince-Edouard, il ya environ 7A7 Ma @ermian tardif), suite aru( cotrtraintestectoniques qui ont pr€cede l'ouverture de l'oc6an Atlan-tique. Les concentrations 6lev6es de Ca et Ti dans leclinopyroxdne et tris 6lev6es de Ti dans la phlogopite mani-festent son caractdre alcalin. Les teneurs trds basses en Rb

et K font penser que phlogopite ou autre phase porteusede potassium dans le manteau n'a pas exerce d'influenceimportante sur la formation du magma. Les 6l6ments Sr,Ba et les terres rares montrent un enrichissement m€taso-matique plus important, mais ce sont les concentrations deHf, Ta, U, Th, Zr et M qui ont le plus augment€. L'agentdu m6tasomatisme (fluide ou magma) 6tait tout probable-ment enrichi en CO2, et soit incapable de transporter le Ket le Rb, soit trds efficace i le faire, Laissant i la sourceune concentration relative des 6l6ments tels le Zr et le Nb.Les relations Fe-Mg-Ni parmi les roches globale, le carac-t0re magn6sien des phdnocristaux d'olivine, et la presenced'enclaves de lherzolite i spinelle seraient les indicationsd'un magma primaire. Une temp6rature d'dquilibratiorr desenclaves anormalement basse (enwion 970'C) montrentqu'il ne s'agit pas d'€chantillons repr6sentatifs de la sourcedu magma, mais plutOt des populations d'enclaves typiquesdes basaltes alcalins et des n6phdlinites.

(Traduit par la Rddaction)

Mots-cl4s: lamprophyre, nephdlinite a olvine, magma pri-maire, lherzolite i spinelle, filon-couche, Permian tar-dif, tectonique d'extension, ile du Prince-Edouard, chi-misme, 6l6ments traces, terres rares, mdtasomatisme,p6trogenbse.

INTRoDU TIoN

The only igneous rocks exposed in the provinceof Prince Edward Island occur on George Island inMalpeque Bay (Fig. l). Larochelle (1967) suggestedthat the unmetamorphosed mafic sill may be UpperPermian, but the importance of this age and thepetrologic and tectonic significance of the intrusionhave been recognized only recently. A compilationof dates for volcanic rocks in eastern North Ameri-ca, that are considered related to opening of the At-Iantic Ocean, shows that most of the magmatic ac-tivity took place between 170 and 200 Ma, with apeak at 190 Ma (McHone & Butler 1984). Most ofthese continental basaltic rocks show a tholeiitic af-finity, with many trace-element characteristics similarto those of ocean-floor basalts, but the continentalbasalts also exhibit the effects of crustal contami-

tPresent address, Department of Geology, Mount Allison. University, Sackville, New Brunswick EOA 3C0r-Passed away summer 1984.

97

98 THE CANADIAN MINERALOGIST

nation (e.9., Smith et al. 1975, Dostal & Dupuy 1984,Greenough & Papezik 1986a).

This study of the Malpeque Bay sill was under-taken to determine its petrologic and tectonic sig-nificance. We discuss (l) the age of the sill, (2) itsmineralogical composition, (3) whole-rock geo-chemistry, (4) the composition of minerals in as-sociated spinel lherzolite xenoliths, and (5) the sill'spetrogenesis.

GnNBnar- Georocv aNn Acs op rHs SnL

The sill was first described by Milligan (1949) asa dyke approximately 3 m thick which cut red sand-stones. Larochelle (1967) pointed out that the unitis a sill and that it probably intruded unconsolidat-ed (penecontemporaneous) sediments, as indicatedby the convoluted nature of some of the contacts.Prest (1972) gave thorough field and petrographicdescriptions of the rocks and also presented twochemical analyses indicating low SiO2 contents (40w. q0). Deposition of the Prince Edward Island redbeds may have taken place as early as late Pennsyl-vanian or early Permian, but the age of the upper-

TABLE 1. RESULTS OF RADIOI1ETRIC DATINO OF THE I,IALPEQUE BAY SILL

K ( % ) a o A r

r a d l o g e n i ct c m ' J

Age (ila)

mo$t beds, intruded by the sill, remains in doubt dueto a lack of fossils (Langston 1963). Paleomagneticdata obtained by Larochelle (1967) indicate that in-trusion occurred during the late Permian.

Table 1 shows the results of K-Ar dating of fourwhole-rock samples. The conventional ages ofthreesamples (P83-43,45 and 47) agtee with each othervery well in spite of differences in their potassiumcontents, whereas the apparent age of P83-48 is 3Volower than the other three. The discrepancy in agesis probably due to either argon loss or low initialqAr/3eAr ratio but not to analytical error becausea second analysis on another specimen from P83-48yielded the same result. Therefore, the most likelyage of the sill,247 Ma (average of samples P83-43,45 and 47), confirms Larochelle's (1967) late Per-mian estimate of the age of intrusion and resolvesthe discrepancy with K-Ar ages (207 t 8 Ma) re-ported by Poole el al. (1968).

PsTnocRApHY AND SAMPLING LOCATIONS

In outcrop the sill varies from a vesicular brec-ciated rock containing baked sediment and lherzo-lite xenoliths to a dark fine-grained rock, with oli-vine phenocrysts, that also contains xenoliths.Sampling locations of the nonvesicular and nonbrec-ciated specimens appear in Figure 1. All samples con-tain 5-1590 euhedral olivine microphenocrysts (aver-age 0.2 mm) set in a very fin+prained (clinopyroxenegrains <0.02 mm) matrix. The matrix containsclinopyroxene (5090) and olivine (590) with an inter-

PB3-43P83-45P83-47P83-48APB3-488

0.40720.87881.0440.47t20 .4773

4.189 x0 .050 x

10.733 x4 .703 x4 .761 x

246.9247 .I246.7240.0239.9

1 0 .

10-;ro-il 0 - -

Frc. l. Location map showing the Study area on George Island, within the provinceof Prince Edward Island, and detailed sampling locations.

P 8 3 - 4 5 , 4 7 .

Gult of St . LawrenceH o g l s l a n d

George ls

P r i n c e E d w a r d l s l a n d

. MALPEQUE BAY SILL, P.E.I.

IABLE 2. REPRESENIATIVE COMPOSITIONS OF SILL CLINOPYROXENE, OLIVINE AND PHLOGOPITE

99

P83-43 PB3-45 P83-47cPx cPx cPx

P83-48 P83-43cPx 01

P83-45 P83-47 P83-4101 01 Ph l

si02 47."14 47.5LT i 0 2 3 . 2 6 3 . 1 3At203 5 .87 5 .45Cr203 O.02 0 .05Feo* 7 .1 ,4 7 .22Mno 0 .09 0 .10M90 12.52 12.28N i o 0 . 0 0 0 . 0 3Cao 22.07 22 .98Na20 0 .68 0 .53KzO 0 .00 0 .00T0TAL 99.09 gg .28

s i 1 . 7 7 7 7 . 7 9 2l_yAl 0.223 0.208v r A l 0 . 0 3 7 0 . 0 3 3r i 0 .092 0 .088Cr 0 .000 0 .001Fe 0.225 0.227Mn 0 .002 0 .002Mg 0.729 0.689Nr 0 .000 0 .000Ca 0 .891 0 .929Na 0 .050 0 .038K 0 .000 0 .000

Ca 48.5 50 .3M9 39.2 37.3FerPn 1,2.4 I2.4Fo (ml . %)

47.68 49 .342 . 8 1 2 . 5 t6 . 4 8 4 . 7 50 . 1 8 0 . 5 26 . 8 3 6 . 6 10 . 0 6 0 , 1 2

t2 .28 1 ,4 .170.00 0 .00

21.53 2 t .490.72 0 .450 . 0 5 0 . 0 0

98.62 99.96

1 . 7 9 8 1 . 8 3 10,202 0 .1690.086 0 .0380.080 0 ,0690.004 0 .0150.214 0 .2050.001 0 .0030,689 0 .7830.000 0 .0000.869 0 .B540.052 0 .0310.002 0 .000

4 9 . 0 4 6 . 338.9 42 .41 ? 1 1 1 i

38.42 38.890 . 0 7 0 . 0 40 . 0 4 0 . 1 30 . 0 0 0 . 0 2

1 9 . 0 9 1 5 . 7 00 . 3 2 0 . 1 9

40.91 43 .340 . 1 5 0 . 1 90 . 5 4 0 . 3 9

99.54 98 .89

0.991 0 .9920.000 0 .003

0.001 0 .0010.000 0 .0000.412 0 .3340.006 0 .0031.573 1 .6490.003 0 .003orotn 0.010

79.2 83.2

39.11 37 .640 . 0 3 8 . 6 30 . 0 3 1 1 . 4 10 . 0 4

1 4 . 1 3 i 3 . 1 80.1 ,2 0 .08

45,36 13 .90o . 2 90.27 0 .78- 0 . 6 9_ / . o J

,,.38 'L' 'A

0 .987 5 .6420.000 2 .015

0.000 0 .9860.0010.297 1 .6520.002 0 .010r .707 3 .1050.0050.006 0 .125- 0 .200_ 1 .459

85.2

MaJor-elerent oxideis in wt. g, Hith total Fe as FeO.Number of lons on the basis of 6 oxygen atoms for clinopyroxene, 4 for ollvlne and A2 forph l ogop i te .A l l A l s h o m a s r v A l f o r o l i v i n e a n d p h t o g o p i t e .

sertal mesostasis of orange-brown mica ( < 5 go) andfiner-grained devitrified glass and felsic minerals(plagioclase? + nepheline?). Minor,,iddingsite" andserpentine appear on otherwise unaltered olivinegrains and constitute the only obvious alteration inthe rocks.

Ultramafic xenoliths are associated with samplesP83-47 and 48, but comprise less than 5go of therocks; the xenoliths consist of equigranular, I mmolivine, orthopyroxene, and clinopyroxene, and 0.5mm brown spinel grains in approximate proportions60:23:12:5. The lherzolite nodules show no signs ofshearing or deformation.

MTNERAI- CHEMISTRY

Chemical compositions of minerals from the silland from an associated lherzolite xenolith appear inTables 2 and 3. Analyses were carried out at theMemorial University of Newfoundland on a JEOLJXA-50A (wavelength-dispersion) electron-microprobe analyzer with Krisel sotfware usingminerals of known composition as standards.

Minerals in the sill

Clinopyroxene from the sill (Table 2, Fig. 2) con-tains more than45t/o Wo and averages 48go Wo. Incomparison with pyroxene from basaltic rocks, Ti(- 2.8 wt.9o TiO), Al (- 5.? wt.9o AlrOr) and Ca

(-22wt.o/o CaO) show relatively high concentrations(Le Bas 1962) as well as a positive correlation withone another. Such characteristics are typical of

TASLE 3. REPRESENTATIVE COMPOSITIONS OF CLINOPYROXENE'ORTHOPYROXENE, SPII IEL AND OLII INE IN THE XENOLITHS

01

5 i02Ti0zA1203C1203Fe0r!ln0itg0Nt0Ca0Na20(zo

TOIAL

stI vAlvl AlTI

Fel'lnl'19NILANaK

52.2 '1 53 .08 54 .86 54 .A7 0 .27 0 .350.39 0 .40 0 .08 0 . r0 0 .09 0 .065.23 4 .80 3 .25 3 .50 58 .25 58 .340.59 0 .44 0 .23 0 .18 7 .72 7 .482.74 2 .66 6 .89 7 .L4 12 .73 L2 .A70.04 0 .1 ,2 0 .13 0 .20 0 .07 0 .09

14.96 16 .82 33 .37 33 .49 19 .6B t9 .710 . 0 8 0 . 0 4 0 . 0 0 0 . 0 7

2r .41 20 .43 0 .4s 0 .411 . 1 8 1 . 0 7 0 . 0 9 0 . 0 5

40.07o.o20 . 0 10.00

10.840 . 1 3

48.050 . 3 30 . 0 4

0 . 0 1 0 . 0 1 0 . 0 2 0 . 0 0 . - - -98.90 99.87 99.37 100.01 98,81 98.96 99.49

1.914 1 .917 1 .913 1 .904 0 .056 0 .066 0 .9930.086 0 .083 0 .087 0 .096 0 .0000.140 0 .120 0 .046 0 .046 14 .372 14 .3680.011 0 .010 0 .002. 0 .002 0 .012 0 .008 0 .0000.022 0 .016 0 .008 0 .006 1 .276 L .216 0 .0000,083 0 .079 0 .200 0 .207 2 .224 2-224 0 .2250.001 0 .003 0 .003 0 .005 0 .012 0 .012 0 .0020.816 0 .905 1 .735 1 .732 6 .136 6 .158 1 .7750.002 0 .001 0 .000 0 .001 0 .0060.839 0 .790 0 .016 0 .015 0 .0000.083 0 .074 0 .005 0 .0030.000 0 .000 0 .000 0 .000

ca 48.2l'{g 46.9Fe+liin 4.9Cr/(cr+Al )F o ( m l . X )

44 .5 0 .8 0 .850.9 88 .8 88 .44 . 6 1 0 . 4 1 0 . 8

0.082 0.078BB. B

l{aJor-elenent oxldes ln wt. %, ulth total Fe as Feo.Nunber of ions on ihe basis of 6 oxygen atons for pyroxene, 32 forsp lne l and 4 fo r .9 l i v ine .A i l A1 snown os IVA1 fo r o l i v lne and VIAI fo r so lne l .

. MAGMAI XENOLITH

4

r (9) 21)

100

Frc. 2. Plot of the En (Mg), Wo (Ca) and Fs (Fe+Mn)contents of pyroxene from the lamprophyre andassociated lherzolite xenolith.

pyroxene compositions from strongly undersatuated(nephelinitic) magmas.

Olivine compositions range from Fors to Fortwith a mean of Fo3a (Table 2, Fig. 3). Grains withless than 86 mol.9o Fo have calcium contents gtreaterthan 0.25 wt.Vo CaO, typical of hypabyssal intru-sions (Simkin & Smith 1970). The core of a fewgrains with higher Fo and lower Ca contents (0.02wt.9o CaO) may have crystallized at a greater depthand may have been carried along with the magmato crustal levels. These grains could representxenolith debris; however, they a.re much finer grainedthan olivine grains in the associated lherzolite anddisplay euhedral forms.

Titanium in phlogopite is considered "high" if itexceeds 4 wt.Vo TiO2, but grains from the MalpequeBay sill contain 8-9 wt.9o TiO2. Strong & Harris

8 4

Fo o/o

Frc. 3. Histogram showing the Fo contents of olivine fromthe lamprophyre and associated lherzolite xenolith.

THE CANADIAN MINERALOGIST

(1974) noted relatively high TiO2 concentrationsO (6.5-7.5 wt.9o) in phlogopite from a lamprophyre

in Newfoundland, and Wagner & Velde (1986)reported up to 11.2 wt.9o TiO2 in phlogopite froma lamproite.

Minerals in the xenoliths

The compositions of mineral grains in a xenolithassociated with sample P83-48 appear in Table 3.Spinel in the xenolith is characterized by Crl(Cr + Al)value less than 0.08, i.e., it is Al-spinel according toCarswell (1980). Both olivine (Fig. 3) andorthopyroxene (Fig. 2) show constant compositionsof Fo6e and Enee, respectively. These values falltoward the Fe-rich end of the narrow distributionof compositions normally observed in Al-spinel lher-zolites (Carswell 1980). The clinopyroxene is typi-cal of "Cr-diopside" found in spinel lherzolites. Itdisplays a greater range of compositions (particularlyCaO contents) than other minerals in the xenolith(Fig. 2) as is commonly reported for such xenoliths(e.9., Yarne 1977),

WHoLE-RocK GEocmMIsrRY

Samples were prepared for analysis by removingall weathered surfaces and ensuring the absence ofxenolith material. Whole-rock chemical compositions (Table 4) were determined by G. Andrews(Memorial University of Newfoundland) using

IABLE 4. iIA,JOR-ELEI.IENT AND NORI.IATIVE COMPOSIIION OF T}IE I.IALPEQUEBAY SILL

PB3-43 P83-45 P83-47 P83-48

5102Tl02A1203Fez03Fe0l'ln0l,ls0Ca0Na20KzoPzos1 . 0 . I .TOTAL

ttl9,

a0rAbAnNEDiHy01MtI lAp

9 28 0

aoo

6c

o

ctz

41.s 42 ,42 .74 2 .679.60 9 .085.25 5 ,037.00 7 .320 . 1 7 0 . 1 7

13.58 15.0810.70 9 .60

t 1 ? 1 t 6

0.55 r .o70,92 0.84J . 0 0 a . J o

,&r4 ,r. rt

42.5 4t.72 . 7 7 2 . 7 79,24 9 .424.24 4 .948.03 7 .400 . 1 7 0 . L 7

14.43 14. 159.98 10 .43

1.06 0 .650.88 0 .862.57 3 .41

99.40 99.56

0,70 0 .72 0 .71

BARTH NIGGLI MOLECULAR NORIiS

0.70

8.579.75

r3 .6031.08

25-.702.013.942.O0

6.407.806.62

13.6728.68

2g-.2r2 .003.781.78

o.:s s.go7.22 6 .896.30 7 .s2

14.94 15.9630.36 31 .64

27-,06 26--302.00 z ,oL3.90 3 .941, .87 1 .84

XENOLITH

MAGMA

ilaJor elerent concentratlons ln oxide rt. %.ils' - Mg/(mg r 0.9 Fe) atomlc.i lolecular nom calculated assm'lng Fe203/Fe0 E 0.185.

MALPEQUE BAY SILL, P.E.I. 101

P R I N C E E D W A R D I S L A N D

NORTH AMERICA

O C E A N I C I S L A N D S

N E W Z E A L A N D

AAaraV

TABLE 5. TRACE-ELEiIENT CONCENTRATIONS IN THE ilALPEOUE BAY SILL

P83-43 P8!45 p83-47 p83-48 I-l (o) NBS1632an . 7 n = 4 ( o )

4A

.AR b 6 9c s i . 6 0 . 8Sr 1,257 1017Ba 706 616T h 7 . t 6 . 4u 1 . 8 2 . 3Zr 234 217H f 5 . 5 5 . 0Nb 101 94T a 5 . 6 5 . 1f 2 5 2 7

B 1 7 2 2 . 6 ( 2 . r )0 . 6 i . 8 2 . 7 ( 0 . 2 )

917 822 r72 (3)639 623 r83 (10)

6 . 4 7 . 3 4 . 6 ( 0 . 1 )i . 6 t . 7 1 . 4 ( 0 . 1 )

225 23r 87.3 (1.7)5 . 8 5 . 9 r . 9 ( 0 . 1 )

95 t02 7 .3 (1 .5 )5 . 1 5 . 7 < 0 . 5

25 29 2s .1 (3 .1 )

G R E E N L A N D

EUROPE

A.* D

I. t J t

t 1

R bFrc.4. Plot of Rb and Sr concentrations (ppm) in relatively

unfractionated (Mg'>0.65, or Ni>200 ppm) lampro-phyres and related rocks. Sources of data: PrinceEdward Island, tlis study; North America, Phelps etol. (1983), Luhr & Carmichael (1981); Oceanic Islands,Kay & Gast (1973), Greenough (1979), Wilkinson &Stolz (1983); New Zealand; Briggs & Goles (1984);Greenland, Hansen (1980, l98l), Clarke et a/. (1983);Europe, Alibert et al. (1983), Downes (1984).

atomic-absorption spectrophotometry. Concentra-tions of trace elements determined by D. Press (alsoat M.U.N.) using X-ray fluorescence methods maybe judged for precision and accuracy from replicate

A PRINCE EDWARD ISLAND

4 0 0

3 0 0

Zr

200

1 0 0

O OCEANIC ISLANDS

6 . 6 ( 0 . 1 )263 (5 )e 8 . 6 ( 2 . s )

6 . 5 ( 0 . 5 )78 .7 (2 .8 )

8 3 . 0 ( 0 . 2 )

r Concentrations in ppm except for Au, in ppb.Elements detemlned by XRF rere run rlth the standard l,i-l and thoseby instrunental neutron-actlvation, with NBS 1632 a.

analy$es of U.S.G.S. standard W-l (Table 5). Traceelements determined by instrumental neutron-activation analysis at Nuclear Activation ServicesLtd. were run with standard NBS 1632 A.

Mg'values [Mg' :Mg/(Mg+0.9Fe) atomic,wlere Fe : total ironl are very high (-0.70) indicat-ing that the rocks are relatively primitive. They canbe classi{Ied as olivine nephelinites as all samples are

aa

art

vq

VV

V Ya

1Yt

LA

NdSmEuTbYbLu

52 48 4890 85 8840 39 409 . 0 8 . 5 8 . 92 . 7 2 . 6 2 . 80 . 9 1 . 0 r . 01 . 5 1 . 4 1 . 30 .20 0 .19 0 .19

5410146

2 . 81 . 01 . 60 . 2 L

1 s . 3 ( 0 . 2 )2 9 . 3 ( 0 . 5 )

2 . 5 ( 0 . 1 )0 .49 (0 .04)0 . 3 4 ( 0 . 0 2 )1 . 1 3 ( 0 . 0 7 )0 . 1 8 ( 0 . 0 1 )

<5<2

9e.o (2.2) -

95 .7 ( 1 .7 )16.2 (2.7)

Sc 23 22 22 24\ 276 259 268 27ICr 296 373 370 353Co 79 75 68 72Ni 332 463 4t6 383Au' 23 45 36 36M 0 1 0 7 7 L 0Cu 68 70 6'1 68Zn 108 106 fl7 111Ga 15 15 16 1B5e 4 .5 5 .4 6 .5 4 .

I NEW ZEALAND

O eneexunHoY EUROPE

F

5 0 r 0 0 1 5 0Nb

Ftc. 5. Plot of Nb yersz.s Zr (ppm) in relatively unfractionated (Mg'>0.65 orNi >200 ppm) lamprophlres and related rocks, Linc denote constant M/Zr ratios.Sources as in Figure 4,

t02

200

1 0 0

highly nepheline-normative (- 14 mol.9o Ne, Table4). They also show some basaltic characteristics inthat Ti (-2.7 wt.Vo TiO2, Fig. 4), Ca (- 10 wt.9oCaO), and K (-0.8 wt.9o K2O) concentrationsoverlap with those found in primitive alkali basalts(cf. Kay & Gast 1973, Hansen 1980, 1981, BasalticVolcanism Study Project (BVSP) 1981).

Concentrations of some large-ion lithophile (LIL)elements (e.g., RB-10 ppm, Sr-1000 ppm,

aUJF 5 0Enzoxo 2 0

utJg 1 0

@6

THE CANADIAN MINERALOGIST

P 8 3 - 4 3P 8 3 - 4 5P83 -47P 8 3 - 4 8

LA Ce Nd SmEu Tb Yb LuFIc. 6. Chondrite-normalized R.EE patterns for the Malpeque Bay lamprophyre.

Values for normalization from Taylor & Gorton (1977).

Ba-650 ppm) are high in comparison with mostbasalts (c/ BVSP l98l) but are low (especially Rb)in comparison with most unfractionated lampro-phyres @ig. 4). In contrast, the high-valenry cationstend to show very high concentrations (e.g., Nb - 100ppm, Zr -225 ppm) and values of ratios such asNb/Y ( - 3.7) and Nb/Zr ( - 0.5) only compare withthose in lamprophyres (Fig. 5).

Concentrations of the light rare-earth elements

Eo-o-

z

2 0 0 0

1 5 0 0

1 0 0 0

5 0 0

0

MgO, Weigh t %Frc.7. Plot of Ni (ppm) versus MgO (ft. q0) concentrations in potential primary

magrras (solid dots, as reviewed in BVSP 1981, p. 424) and, the Malpeque Baylamprophyre (open triangles). Solid lines show constant values of the Ni,/MgO ratio.

4 03 02 01 0

MALPEQUE BAY SILL, P.E.I. 103

(R.E@ in the Malpeque Bay lamprophyre are higherthan in most basaltic rocks but are similar to thelower concentrations recorded for some lampro-phyres (c1,, Kay & Gast 1973, Moller et ql. 1980,Alibert et ol. 1983, Clarke et al. 1983). The heavyREE are only about 6 times chondritic levels (Fig.6) but are comparable to those in the lamprophyresnoted above as well as to unfractionated basalticrocks from Hawaii (BVSP 1981, p. 172). Thechondrite-normalized REE patterns define relativelystraight lines with stsep slopes typical of alkalinerocks (Fig. 6).

Relatively high Ni and Cr concentrations (-400and 350 ppm, respectively) seem representative ofunfractionated magmas (e.g., BVSP 1981, pp.422-424). Comparison of concentrations of Sc, V,Co, Mo, Cu and Znwiththose in other unfractio-nated lamprophyres is hampered by the scarcity ofdata. Most of these elements, however, show onlysmall variations (by much less than a factor of l0)in mafic magmas in general; the concentrationsreported here fall within the ranges normallyobserved (cl, BVSP 1981). Gold concentrations inthe lamprophyre are higher than those in earlyMesozoic continental tholeiites from the Atlanticprovinces (<20 ppb versw -35 ppb, J.D.G. unpub.data).

DISCUSSIoN

Primary nature af the lamprophyre

Most of the evidence from lherzolites suggests thatmantle olivine is approximately Fo* in composition(Carswell 1980). The Mg' values of magmas formedin equilibrium with such olivine should be between0.72and 0.70 (Roeder & Emslie 1970, Green l97l).It follows that the Malpeque Bay lamprophyre withMg' values between 0.70 and 0.72 qualifies as aprimary magma. Estimates of values for thetemperature-dependent partition coefficientK" : (Ni/MgO)o6u1o",/(Ni/MgO)*"1, also are usefulin identifying primary magmas. Given that Kovalues should fall between 1.65 and 2.72 for mosrmafic magmas (Hart & Davis 1978, Leeman & Lind-strom 1978) and assuming a lherzolite sourse withFoes olivine containing 3200 ppm Ni, primarymagmas will have Ni/MgO values between 23 and39 (reviewed in BVSP 1981, pp. 414-426). Figure 7illu$trates that the Malpeque Bay lamprophyre showsNi/MgO values expected for a primary magma. Fur-ther, the composition of the most Mg-rich olivinephenocrysts @oe1) indicates that fractionation hasnot taken place.

Concentrations of the heavy REE in the lampro-phyre are as low as normally observed in maficmagmas devoid of substantial accumulations ofphenocrysts. As removal of low-pressure phases such

IABLE 6. ESTII4ATES OF EOUILIBRATION TEI'IPERATURE OF XENOLITHSBASED ON VARIOUS GEOTHER'.IOI''IETERS

Reference lilethod Range of I'leanTenperatures ( 'C)

l e l l s (1977)l, lori (1977)lbrl & Green (1978)ilorl (1977)llorl (1977)Fabrles (1979)ilorJ A 6reen (1978)ilori & 0reen (1978)Roeder d t a l . (1979)l.lysen (1976)

AYerage Tenperature

Sol vus 880-1009solvus 1002-1203Sol vus 911-1058Cpx-o]-Sp (Eq. 2) 980-9940px-01-Sp (Eq. 3) 924-943Sp-01 900Fe/Mg opx-Cpx 820-992Fellilg 01-Cpx 886-1003gl-sp 1030- 1044YrA l /c ropx-cpx 854-1076

Range of tenpefatures calculated from al l possible conblnat lons0f analyses shorn ln Table 3, the mean represents the averageof these tenperatures (usual ly 4),

as olivine, pyroxene or plagioclase would raise theseconcentrations (Frey et ol. 1974), the REE are con-sistent with the proposal that the rocks are primitive.

Further evidence that the lamprophyre representsa primary magma comes from the fact that it con-tains lherzolite xenoliths. Estimates of the ascentvelocities necessary to carry these nodules into theupper crust (1-50 cm/s) preclude substantial frac-tionation en route (Spera 1980).

Spinel lherzolite as a source maleriol

The Fo content of olivine in the spinel lherzolitexenoliths @ose) is similar to that expected in themantle source for the magma itself @VSP 1981,424-426). High values of the Allcr ratio show thatthe spinel was not precipitated from a magma atdepth, but the values are representative of mantlesource-material (Carswell 1980). To test the hypothe-sis that the spinel lherzolite represents the sourcematerial for the lamprophyre, equilibration temper-atures for the former were estimated using variousgeothermometers (Table 6).

Geothermometers using clinopyroxene yielded awide range of temperatures, in part due to the vari-ations in composition shown in Table 3. This is par-ticularly true of methods based on the solvus.Nevertheless, averages of temperatures estimatedfrom all possible combinations of mineral grains inTable 3 show similar results for all geothermome-ters except the solvus-based method of Mori (1977).Carswell (1980) noted that Mori's equation for thesolvus yields higher temperatures than the solvus-based methods of Wells (1977) or Mori & Green(1978) but that temperatures from the Mori equa-tion agree with results from methods not based onthe solvus (e.9., Mysen 1976). However, whenapplied to the Malpeque Bay pyroxene the Mysenequation based on the solvus gives higher tempera-tures than all other geothermometers (Table 6),whereas the Wells (1977) and Mori & Green (1978)methods tend to agree with other methods. Thedifferences in results may be related to slightly lower

9941103944987934900904945

1032960

104 THE CANADIAN MINERALOGIST

Al concentrations (Mori 1977)in the Malpeque Baypyroxenes relative to the pyroxenes used in calcula-tions by Carswell (1980). However, Al concentrationsin the Malpeque Bay pyroxenes are high enough thatthe Lindsley & Andersen (1983) two-pyroxenegeothermometer is not applicable. An average of themean temperatures from all methods reported inTable 6 yields a temperature of 969"C. All temper-atures fall within the range normally reported forspinel lherzolites (900-1150'C) and demonstrate tharthe lherzolite cannot represent the source for the lam-prophyre (which must have been hotter) or phasesprecipitated from the magma at depth (Carswell1980). The lherzolites apparently represent recrystal-lized and re-equilibrated mantle material overlyingthe magma's source area, and were included in themagma on its way to the surface.

Petrogenesis

Any attempt to characterize the source for theMalpeque Bay lamprophyre must account for the fol-lowing characteristics: l) low Rb and Ti concentra-tions (for example) within the range of some alkalibasalts, 2)hght-REE nncentrations at the lower endof the range shown by lamprophyres, and 3) highZr and Nb concentrations and low Nb,/Zr ratioscharacteristic of lamprophyres from highly"enriched" source$.

It has long been known that K and Rb concentra-tions in alkaline magmas are much more variable andirregular than (for example) Sr and REE concentra-tions. This variability has in some cases beenattributed to the presence or absence of phlogopiteduring mehing events (e.g., Kay & Gast 1973). Morerecent studies of mantle xenoliths have shown thatthe nodules best representing the sources for alka-line magmas (because of enrichment in incompati-ble elements) usually contain phlogopite (Jagou|z etal. 1979, Gurney & Harte 1980). Similarly, meltingexperiments bearing on the origin of alkaline magmashave consistently pointed to the importancq ofphlogopite in magma genesis (Modrcki & Boettcher

TABLE 7. PARMETERS FOR THE I'IELIINO I{ODEL

Sr mdel

Modal prcp. 55 25||lelt prcp. 5 5Dsr 0 .016 0 .016

Rb model

1973, Arima & Edgar 1983, Llyod et ol. 1985).TheMalpeque Bay lamprophyre clearly showsanomalously low Rb (Fig. a) and K, given its highSr-concentrations (1000 ppm), indicating thatphlogopite playe( a subordinate role in its genesis(in comparison with other nepheline-normativemagmas).

Early students of alkaline magmas commonly sug-gested that small percentages of melting of a garnet-rich mantle source can explain their enrichment inincompatible elements (e.g., Gast 1968, Kay & Gast1973). More recently it has been recognized thatmodels using a primitive undepleted source, thelowest recorded distribution coefficients, andextremely small percentages of melting cannot dupli-cate observed incompatible-element concentrations(Sun & Hanson 1975, Cullers et al. 1985). This is truefor the Malpeque Bay lamprophyre as well. To illus-tratE this, Shaw's (190) non-modal batch-meltingequation was used to model magmatic Sr concentra-tions. A garnet lherzolite souice was assumed (toaccount for the low heavy-R.EE concentrations: Kay& Gast 1973) together with reasonable values for thepartitioning coefficients (D values), modal percen-tages of minerals and melting proportions (Table 7).Assuming primitive to slightly elevated Sr values(e.g.,28 ppm: Jagoutz et al, 1979), the lowest Sr con-centrations observed in the sill cannot be duplicatedeven when the percentages of melting approach zero(model predicts -600 ppm). Furthermore, it is notclear how extremely small proportions of melt couldbe removed from the mantle. If minor phases (suchas phlogopite) exist, then small percentages of melt-ing make it more difficult to explain, for example,the high Rb concentrations common in many alka-line magmas because the minor phases make Rb amore compatible element (DRb in phlogopite : 3.1:Philpotts & Schnetzler 1970).

High-pressure fractionation of eclogite has beensuggested as a means for elevating incompatible-element concentrations in alkaline magmas.However, as noted by Mitchell & Bell (1976), pro-hibitively high percentages are required. A fractio-nation model seemslo be at odds with Fe-Mg-Nirelationships that support a primary origin for theMalpeque Bay lamprophyre.

Most petrologists have concluded that the sourceregions for alkaline rocks have been metasomaticallyenriched (Fitton & Upton 1985). Suggested compo-sitions for the metasomatizing fluid vary fromassorted combinations of H2O, CO2, F and Clcharged fluid (e.g., Bachinski & Scott 1979, Molleret ol. 1980, Phelps et ql. 1983, Cullers et al. 1985)to basanite magma itself (e.9., Alibert et al. 1983,Menzies 1985).

Unless partition coefficients are much lower thanthose given for Rb and Sr in Table 7, the "basanitemelt" hypothesis would have to be a two-stage

1045

0.140

1 0 5L7 67

o . o 2 3 . 1

Modal prop 53 22Melt !.0D. 1 3o s o . o l o . o l

10

10

Sources:

D v a l u e s : F r e y e t a l . ( 1 9 7 8 ) , P h l l p o t t s & s c h n e t z l e r ( 1 9 7 0 ) .l ' le l t lng proport ions: Kay & Gast (1973), l lodreskl & Boettcher (1973).Phase proport lons: carswel l (1980), curney & Harte (1980),

MALPEQUE BAY SILL, P.E.I. 105

process where the metasomatized source is remeltedin a second step. The maximum amount ofincompatible+lement enrichment possible is gven bya single-pass zone-melting model (Kay 1979) in whichthe melt percolates through a large volume ofdepleted to primitive mantle-material. The final con-centrations in this melt equal those given by thebatch-melting equation (Shaw 1970) as the percen-tage of melting nears zero (Alibert et al. 1983), Wealready have suggested that such a model cannorexplain the observed Sr concentrations unless Dvalues are much lower than in Table 7. If phlogo-pite is assumed to be present in the mantle materialtraversed by the melt, then Rb concentrations wouldremain very low (approximately 2 ppm assuming 0.3ppm Rb in the source, and melting parameters as inTable 7). To achieve higher concentrations of theincompatible elements, the liquid must be concen-trated (pooled) and then undergo fractionation, orpartial melting (after crystallizatiot) at some latertime.

The other possibilities, that the metasomatizingsolution was a H2O, CO2, F, or Cl rich fluid ormagma (with lower D values for most incompatibleelements), are difficult to test. Recent experimentaldata on the solubilities of various elements in pro-posed mantle metasomatizing fluids indicate thatvery large fluid/rock ratios are required, thus favor-ing a magma as the metasomatizing agent (Schneider& Eegler 1986). Nevertheless, the association of highvolatile concentrations with alkaline magmas is wellestablished (Bailey 1982). H2O- and possibly F-richfluids result in silica-saturated melts such as minettesand lamproites with very high K and Rb concentra-tions (Bachinski & Scott 1979, Jaques et al. 1984).More undersaturated melts such as kimberlites andcarbonatites appear related to high CO2 and lowH2O metasomatizing fluids, which also result inenrichment of K and Rb as well as the R.EE and high-valency cations. Conflicting with these observationsare experiments by McGetchin & Besancon (1973)showing that carbonate inclusions from mantlepyrope contain very low concentrations of K (alsosee Moller et sl, 1980). However, Wendlandt &Harrison (1979) found that at upper-mantle pressuresthe light REE are strongly fractionated into aCO2-rich vapor phase. trf COt-rich metasomatizingfluids do result in only minimal K and Rb enrich-ment, but otherwise leave all of the other meltcharacteristics typical of kimberlites and carbona-tiles (i.e., Si-undersaturated melts), then such fluidsmight produce source characteristics requisite of theMalpeque Bay lamprophyre. Menzies et ol. (1985)proposed that chemical variations in a suite of spinellherzolite xenoliths were due to metasomatic enrich-ment fronts in a contact-metamorphic aureole sur-rounding apophyses of basanitic melt. In a wet zoneclose to the interface of the magma and wallrock,

the precipitation of mica or amphibole (or both)results in enrichment of most incsmpatible elements.To explain lherzolites enriched only in the lieht REEand not in large-ion lithophile elements, they sug-gested a dry zone of metasomatism farther from thecontact that was affected by CO2-rich metasomaticfluids. Possibly this process, acting on a larger scale,created the source for the Malpeque Bay lampro-phyre.

Greenough & Papezik (1986b) recently showedthat movement of volatile complexes in a mafic sillcan explain enrichment in elements such as K, Rb'Zr, and Nb. In some parts of the sill where volatileswere apparently incapable of escaping, all incompat-ible elements were endched. In granophyric pegma-tites in the upper half of the sill only Zr and Nb wereenriched through the formation of minor phases,whereas K, Rb, Sr, and Ba were removed from thesill system by escaping volatiles. Such a scenario inthe upper mantle could explain the origin of the lowK and Rb but high Zr and Nb concentrations oftheMalpeque Bay lamprophyre in the case where aCO2-rich fluid was capable of carrying all elements.The source was not enriched in K and Rb becausethe metasomatizing fluids did not leave these ele-ments behind.

A final hypothesis sugge$ts that the CO2-richlamprophyre had very high concentrations of K andRb, but the loss of volatiles during ascent and cool-ing removed these elements. This hypothesis does notexplain why Zr and Nb remained enriched especiallybecause the fine-grained lamprophyre (onceemplaced) cooled rapidly, allowing little time forseparative processes (such as envisaged for the New-foundland sill of Greenough & Papezik 1986b) tooccur.

Tectonic implications

The247 Ma darc on the Malpeque Bay sill sets itapart from most other postmetamorphic igneousrocks in the nothern Appalachians. These show a fre-quency peak at around 190 Ma (McHone & Butler1984) and are typically related to opening of theAtlantic Ocean. Rocks of similar age to the Mal-peque Bay sill include those of the coastal NewEngland igneous province (210-20 Ma syenites andalkali grarrite: McHone & Butler 1984, Foland & Faul1977), the mildly alkaline Seabrook diabase dykesin New Hampshire Ql2-236 Ma: Bellini et al.1982),and tholeiitic diabase dykes recovered in the North-umberland Strait, CanadaQl2-239 Ma: Pe-Piper &Jansa 1980. With the exception of the Nofihumber-land Strait dykes, these rocks show alkaline compo-sitions typical of early-rift igneous activity (cf. LePichon & Sibuet l98l). Their timing and composi-tion relate them to an early cycle of North Atlanticcrustal attenuation and magmatism. The next cycleof magmatism, 40 Ma later, is more closely associa-

106 THE CANADIAN MINERALOCIST

ted with actual opening of the Atlantic Ocean @e-Piper & Jansa 1986).

Suuvany

The age (247 Ma) and alkaline geochemistry of theMalpeque Bay sill relate it to an early stage of crustalattenuation that preceded opening of the AtlanticOcean. The lamprophyre has an olivine nephelinitecomposition and, in comparison with other lampro-phyres, displays relatively low concentrationi ofsomeZ,IZ elements (e.g., K and Rb) but elevated con-tents of high-valency cations (e. g., Zr and Nb). Con-straints from trace-element modeling suggest thatthese characteristics can be rehtJd io- mantlemetasomatism. Analogy with other types of alkalinerocks indicates that the metasomatizing material wasa-CO2-rich fluid or magma that either was incapa-ble of carrying substantial K and Rb, or that car-ried these elements through the source system leav-ing behind only the high-valency cations. phlogopiteor other K-bearing phases had little effect on thecomposition of the final magma. This is in markedcontrast to the sources for many lamprophyres wherelarger amounts of these minerals probably help con-trol Rb and K concentrations in the melt. Spinel lher-zolites associated with the lamprophyre cannotrepresent its sourca as they show low temperaturesof equilibration. Olivine compositions, major- andtrace-element data, and the presence of lherzolitexenoliths indicate that the magma came to the sur-face rapidly and constitutes a primary melt.

AcKNoWLEDGEMENTS

This research was supported by an NSERC oper-ating grant to Dr. V.S. Papezik. Mrs. G. Andrewscarried out the major-element analyses and Mr. D.Press the trace-element analyses (XRF). Dr. H.Longerich provided assistance in the acquisition ofelectron-microprobe data. J.D.G. extends hisappreciation to Mrs. Papezik for her cooperation inthe completion of the project afler her husband,sdeath.

RsrsnrNcss

ALBEnr, C., MrcHano, A. & Arsanste, F. (1983): Thetransition from alkali basalts to kimberlites: isotopeand trace element evidence from melilitites. Contr.Mineral, Petrologt 82, 176-186.

Antua, M. & Eocaa, A.D. (1983): A hieh pressure

BAcurNsKr, S.W. & Scorr, R.B. (1979): Rare-earth andother trace element contents and the oriein ofminettes (mica-lamprophyres). Geochim. Cos-mochim, Acta 43, 93-100.

Banley, D.K. (1982): Mantle metasomatism - continu-ing chemical change within the earth. Nature 296.525-530.

Basarrrc VorceNnu Sruny PnorBcr (BVSp) (1981):Basaltic Volcanism On the Terrestrial planets. per-gamon Press, New York.

Bnrmu, F.X., Conrurvr, D.H. & Srrwanr, A.J. (1982):Geology of foundation excavations at Seabrook Sta-4on, Seabrook, New Hampshire. 1z Geotechnologyin Massachusetts (O.C. Farquhar, ed.). proc. Conf.March 1980, Univ. Mass., Amherst, Mass.

Buccs, R.M. & Goros, G.G. (1984). Petrological andtrace element geochemical features of the Okete vol-canics, western North Island, New Zaland. Contr.Mineral. Petrology 86, 77 -88.

Canswerr, D.A. (1980): Mantle derived lherzolitenodules associated with kimberlite, carbonatite, andbasalt magmatism: a review. Lithos 13, l2l-138.

Clanrr, D.B., Murcrr, G.K. & Pe-Pnnn, G. (1983):The lamprophyres of Ubekendt Ejland, West Green-land: products of renewed partial melting or extremedifferentiations? Contr. Minerol. Petrology 83,117-127.

CuLLeRs, R.L., RauerusHNAN, S., BrRENDsBN, p. &GnrrrrN, T. (1985): Geochemistry and petrogenesisof lamproites, Late Cretaceous age, WoodqonCounty, Kansas, U,S.A. Geochim. Cosmochim.Acta 49,1383-1402.

Dosrar, J. & Durw; C. (198a): Geochemistry of theNorth Mountain basalts (Nova Scotia, Canada).Chem, Geol. 45, 245-261.

Dowms, H. (1984): Sr and Nd isotope geochemistry ofco-existing alkaline magma series, Cantal, MassifCentral, France. Earth Planet. Sci. Lett.69.32l-334.

FenRrEs, J. (1979): Spinel-olivine geothermometry inperidotites from ultramafic complexes. Contr.Mineral. Petrology 69, 329-336.

Ftr.tol.t, J.c. & Uerox, B.C.J. (1985): Alkaline igne-ous rocks: a review symposium. J. Geol. Soc. Lond.142, 697-708.

FoLaNo, K.A. & Faur, H. (1977): Ages of the WhiteMountain intrusives - New Hampshire, Vermont,and Maine, USA, Amer. J. Sci.277,888-904.

experimental study on a magnesian-rich leucite- Fnrv, F.A., BnvaN, w.B. & THoursoN, G. (1974):lamproite from the West Kimberly area, Australia: Atlantic Ocean floor: geochemistry and petiologypetrqgenetic implications. Contr. Mineral. Petrol- of basalts from Legs 2 and 3 of theDeep Sea Drill-ogy E4,228-234. ing project. J, Geophys. Res. 79, 5501-5527.

MALPEQUE BAY SILL, P.E. I . t07

-, GnEEN, D.H. & Rov, S.D. (1978): Integratedmodels of basalt petrogenesis: a study of quartztholeiites to olivine melilitites from South EasternAustralia utilizing geochemical and experimentalpetrological data. J. Petrology 19,463-513.

Gesr, P.W. (1968): Trace-element fractionation andthe origin of tholeiitic and alkaline magma types.Geochim, Cosmochim. Acta 32, 1057-1086.

Gnner, D.H. (1971): Composition of basaltic magmasas indicators of conditions of origin: application tooceanic volcanism. Phil. Trans. Roy. Soc. Lond.A268.'707-725.

GnrnwoucH, J.D. (1979): The Geochemistry ofHawaiian travas. M.Sc. thesis, Carleton Univ.,Ottawa, Ontario.

& Parezrr, V.S. (1986a): Petrology andgeochemistry of the early Mesozoic Caraquet dyke,New Brunswick, Canada. Can. J. Earth Sci. 23,193-201.

& - (1986b): Volatile control of differen'tiation in sills from the Avalon Peninsula, New-foundland, Canada. Chem. Geol. 54,217-236.

GunNey, J.J. & Henrs, B. (1980): Chemical variationsin upper mantle nodules from southern African kim-berlites. Phil. Trans. Roy. Soc. Lond. 4297,273-293.

HaNsEN, K. (1980): Lamprophyres and carbonatiticlamprophyres related to rifting in the Labrador Sea.Lithos 13. 145-152.

- (1981): Systematic Sr-isotopic variation in alka-line rocks from West Greenland. Lithos 14, 183-188.

Henr, S.R. & Devrs, K.E. (1978): Nickel partitioningbetween olivine and silicate melt. Earth Planet. Sci.Lett. 40,203-219.

Jacou'rz, 8., Patue, H., BaoonnHeuseN, H., BLUM,K., CSNDAL$, M., Dneraus, G., Srnrru, B.,Lonrrz, V. & WaNrr, H, (1979): The abundancesof major, minor and trace elements in the earth'smantle as derived from primitive ultramafic nodules.Proc. Lunar Sci. Conf.10th, 2031-2050.

Jaeurs, A.L., LnwIs, J.D., SuttH, C.8., Guconv,G.P., FrncusoN, J., Cnarerl, B.W. & McCut--LocH, M.T. (1984): The diamond-bearingultrapotassic (lamproitic) rocks of West Kimberlyregion, Western Australia. Iz Kimberlites l: Kim-berlites and Related Rocks (J. Kornprobst, ed.).Elsevier, Amsterdam.

Kav, R.W. (1979): Zone refining at the base oflithospheric plates: a model for a steady-stateasthenosphere. Tectonophys. 55, l-9.

and origin of alkalirich basalts. J. Geol. El,6s3-682.

LeNcsroN, W. (1963): Fossil vertebrates and the latePaleozoic red beds of Prince Edward Island. Nal.Mus. Can. Bull, 187.

LenocurLrn, 4..(1967): Paleomagnetic directions of abasic sill in Prince Edward Island. Geol Sum. Can.P. ap. 67-39, l-6,

Lr Bas, M.J. (1962): The role of aluminum in igneousclinopyroxenes with relations to their parentage.Amer. J. Sci. 260. 267-288,

LEruar, W.P. & LIt.rostnou, D.J. (1978): Partitioningof Ni2+ between basaltic and synthetic melts andolivines - an experimental study. Geochim. Cos-mochim. Acta 42, 801-816.

Le PrcHolr, X. & Stsuar, J.-C. (1981): Passivemargins: a model of formation. J. Geophys. Res.86,3708-3720.

LrNosr-ey, D.E. & ANoEnsEN, D.J. (1983): A two-pyroxene thermometer. Proceedings of the Thir-teenth Lunar and Planetary Science Conference,Pafi 2. J. Geophys. Res. 88, A887-A906.

Lrovo, F.E., Antue, M. & Eocen, A.D. (1985): Par-tial melting of a phlogopite-clinopyroxenite nodulefrom south-west Uganda: an experimental studybearing on the origin of highly potassic continentalrift volcanics. Contr.\ Mineral. Petrology 9l'32t-329.

LuHn, J.F. & Carutcresl, I.S.E. (1981): The Colimavolcanic complex Mexico: Part II. Late Quaternarycinder cones. Contr. Mineral. Petrologt T6' 127-147 .

McGercHN, T.R. & BEsaNcoN, J.R. (1973)t Carbonateinclusions in mantle-derived pyropes. Earth Planet.Sci. Lett. lE, 408-410.

McHoxn, J.G. & Brnren, J.R. (1984): Mesozoic igne-ous provinces of New England and the opening ofthe North Atlantic Ocan. Geol. Soc. Amer. Bull.95,757-765.

MBNzres, M. (1985): Alkaline magmas: a window onthe earth's mantle. J. Geol. Soc. Lond. 142,70+705(abstr.).

KruproN, P. & DuwcaN, M. (1985):Interaction of continental lithosphere andasthenospheric melts below the Geronimo volcanicfield, Arizona, U.S.A. J. Petrology 26' 663-693.

MnrrcaN, G.C. (1949): Geological Survey of PrinceEdward lsland. Prince Edward Island Dep. Indus'try Nat. Res.

Mrrcsell, R.tI. & Brrr, K. (1976): Rare earth elementgeochemistry ofpotassic lavas from the Birunga and& Gesr, P.W. (1973): The rare earth content

108 TT{E CANADIAN MINERALOCIST

Toro-Ankole regions of Uganda, Africa. Contr.Mineral. Petrology 5E, 293-303.

Moonesn, P.J. & BorrrcHrn, A.L. (1973): Phase rela-tionships of phlogopite in the system K2O-MgO-CaO-AI2O3-SiO2-H2O to 35 kilobars: a bettermodel for micas in the interior of the earth. Amer.J. Sci. 273,385-414.

MorLen, P., MoRrEANr, G. & Scurev, F. (1980): Dis-cussion of REE distribution patterns of carbonatitesand alkalic rocks. Lithos 13, l7l-179.

Moru, T. (1977): Geothermometry of spinel lherzolites.Contr. Mineral, Petrology 59, 261-277.

& GnEmt, D.H. (1978): Laboratory duplicationof phase equilibria observed in natural garnet lher-zolites. J. Geol.86, 83-97.

MysEN, B.O. (1976): Experimental determination ofsome geochemical parameters relating to conditionsof equilibration of peridotite in the upper mantle.Amer. Mineral, 67, 677 -683.

PE-Prern, G. & JeNsa, L.F. (1986): Triassic olivine-normative diabase from Northumberland Strait,eastern Canada: implications for continental rifting.Can. J. Earth Sci.23. 1013-1021.

Puurs, D.W., Gusr, D.A. & Wooonx, J.L. (1983):Petrogenais of the mafic feldspathoidal lavas of theRaton-Clayton volcanic field, New Mextco. Contr.Mineral. Petrology 84, 182-190.

PHrrrorrs, J.A. & ScHNrrzrEn, C.C. (1970):Phenocryst-matrix partition coefficients for K, Rb,Sr and Ba with applications to anorthosite and basaltgenesis. Geochim. Cosmochim. Acta 34, 307 -322.

Poole, W.H., Sem.ono,8.V., Wru-revs, H. &Knlrv,D.G. (1968): Geology of southeastern Canada. InGeology and Economic Minerals of Canada (R.J.W.Douglas ed.). Geol. Surv. Can., 227-304.

Pnrsr, V.K. (1972): Geology of Malpeque-Summerside area, Prince Edward Island. Geol. Sum.Can. Pap.7145.

RoEnER, P.L. & Enaslrc, R.F. (1970): Olivine-liquidequilibrium. Contr. Mineral. Petrologlt 59, n 5-289.

-, CawrerL, I.H. & JaurnsoN, H.E. (1979): Are-evaluation of the olivine-spinel geothermometer.Contr. Mineral. Petrology 6t, 325-334.

Scnmrorn, M.E. & Eccun, D.H. (1986): Fluids inequilibrium with peridotite minerals: irhplicationsfor mantle metasomatism. Geochim. Cosmochim.Acta 50.711-7U.

SHew, D.M. (1970): Trace element fractionation dur-ing anatexis. Geochim, Cosmochim. Acta 34.237-243.

Srurn, T. & Srrarru, J.V. (1970): Minor element dis-tribution in olivine. J. Geol. 78.304-325.

Surrn, R.C., Rose, A.W. & Lerxnc, R.M. (1975):Geology and geochemistry of Triassic diabase inPennsylvania. Geol. Soc. Amer. Bull. 86, 943-955.

Srena, F.J. (1980): Aspects of magma transport. .IzPhysics of Magmatic Processes (R.8. Hargraves,ed.). Princeton University Press, Princeton, NewJersey,265-323.

SrnoNc, D.F. & Hannrs, A. (1974): The petrology ofMesozoic alkaline intrusives of central Newfound-land. Can, J. Earth Sci. 11, 1208-1219.

SuN, SHrwe SooN & HersoN, G.N. (1975): Origin ofRoss Island basanitoids and limitations upon the her-erogeneity of mantle sources for alkali basalts andnephelinites. Contr. M i nera l. Petro lo gt 52, 7 7 -lM.

TAyLoR, S.R. & Gonrox, M.P. (1977): Geochemicalapplication of spark source mass spectrography -III. Element sensitivity, precision and accuracy.Geochim. Cosmochim, Acta 41. 1375-1380.

Vanxr, R. (1977): On the origin of spinel lherzoliteinclusions in basaltic rocks from Tasmania and else-where. "/. Petrology 18, l-23.

WacNER, C. & Veror, D. (1986): The mineralogy ofK-richterite-beaxing lamproites. Amer. Mineral. 71,t7-37.

WeLLs, P.R.A. (1977): Pyroxene thermometry in sim-ple and complex systems. Contr. Mineral. Petrol-ogy 62, 129-139,

WENnANDT, R. & Hannrsox, W.J. (1979): Rare earthpartitioning belween immiscible carbonate and sili-cate liquids and CO2 vapour: results and implica-tions for the formation of light REE enriched rocks.Contr. Mineral. Petrology 69, 409-419.

Wnrrxsoll, J.F.G. & Srorz, A.J. (1983): Low pressurefractionation of strongly undersaturated alkalineultrabasic magma: the olivine - melilite - nepheliniteat Moiliili, Oahu, Hawaii. Contr. Mineral. Petrol-ogy 83,363-374.

Received August 12, 1986; revised n anuscript acceptedFebruary 26, 1987.