Canadian Mineralogist
Vol. 21, pp. 655-660(1983)
lr lN KEWEENAWAN
BASALTSoF THEMAMA|NSEpotNT FoRMATtoN,oNTARto
N.W.D. MASSEY*, J.H. CROCKET aNn A. KABIR
Departmentof Geology,McMasterUniversity,Hamilton, Ontario LgS4Ml
ABSTRACT
Olivine-phyriclavasof the Mamainsepoint Formation
(Lake Superior,Ontario) show an averageIr contentof0.4g
ppb. This is about ten timesthat of overllng plagioclasephyric laval, which average0.041ppb Ir: Simih;rhough
smallerdifferencesin Cr and Ni contentsof the two rock
types appear to result from the incorporation of earlier_
formed cumulate olivine and chromiie into the olivine_
phyric lavas.Ir tenorsof plagioclase-phlriclavasmay thus
be more representativeof this proteioioic sequencelthey
are comparableto those found in modern ocean_floor
basaltsand ancient ophiolitic suites.
Keywords: geochemistry,basalt, proterozoiq, Mamainse
Point Formation, Ontario, iridium, chromium,
nickel, cumulates.
SoN,rMelns
Les lavesA phdnocristauxd'olivine de la formation de
Mamainse,Point(lac Supdrieur,Onmrio) contiennent,en
moyenne,0.48ppb d,iridium, soit environdix fois plusoue
Iescoul6esA phdnocristauxde plagioclase
qui lesrecouwint
(0.041ppb). Desdiff6rencesmoins frappantesdansla con_
centration de Cr et de Ni dans ces deux types de basaltes
rdsulteraientde l'incorporation de cumulaiJd'olivine et de
chromite dansleslavesd olivine. La teneuren iridium des
basaltesi plagioclaseseraitdonc la plus repr6sentativede
cettesequence
prot6rozoique;cesrochessont comparables,
a cepolnt <tevue,aux basaltesdesfondsoc6aniques
modernese[ dessuitesophiolitiquesanciennes.
(Traduit par la R6daction)
Mots-clds: g5ochtmie,basalte,prot6rozoique, formation de
Mamainse Point, Ontario, iridium, chrome, nickel,
cumulats.
INTRoDUCTIoN
The Mamainse Point - Alona Bay area lies at the
eastern end of Lake Superior, about g0 km north
of Sault Ste Marie (Fig. l). The late proterozoic Ma_
mainse,Point Formation, of Keweenawan age, rests
unconformably on the Archean granite-greenstone
terrane ofthe Superior province. A thickness of4ffi
to 5000 m of the formation is exposed; it is
dominated by mafic flows, with subordinate shallow_
EPresentaddress:PrecambrianSection,Ontario
Geological
Survey,77 Grenville Street,Toronto, Ontario M5S iB3.
level felsic intrusive bodiesand interlayeredcoarse
clastic sediments(Annells 1973,Thomson 1953).
The mafic flows are mainly ophitic or melaphyric
in lithology, although occasionalcoarsergabbroic
and glomeroporphyritic typesare found. Sparse(5go)
phenocrystsof olivine or plagioclase(or both) are
developedin many flows, with olivine-phyric flows
at the baseof the sequencegiving way to plagioclasephyric flows higher up. Someof the olivine-phyric
flows may contain up to 200/oolivine phenocrysts,
suggesting the incorporation of early-formed
cumulate olivine in theseflows.
Low-gradeburial metamorphismin the zeoliteand
prehnite-pumpellyite facieshasresultedin alteration
of the original mineralogy, with only augite remaining relatively fresh. The effectsof alteration are most
severein vesicularzonesat the tops and bottoms of
flows. Consequently,the data presentedhere have
been determined on samplescollected from flow
centres.
Fortunately, a large group of elements(Al, Fe,
Mg, Ti, P,Y; Zr, Nb, Ni, Cr, Co, REE,Th, Ta,
Hf and U) appear to have been immobile or only
moderatelyaffected by geochemicalreadjustments
accompanyingburial metamorphism(Massey1980).
Stratigraphicvariation in theseimmobile elements
revealstrends with stepwisecyclicity that allow the
MamainsePoint sequence
to be subdividedinto five
successiveseries (Massey 1980). The Alona Bay
volcanic rocks are believedto be stratigraphically
equivalentto the series-Ilavasof Mamainsepoint;
both are olivine-phyric.The othersseries(II-V) are
plagioclase-phyric.CIPW norms and discriminant
diagrams based on immobile elementsallow the
volcanic rocks to be classifiedas tholeiites.
The petrogenesisof the Mamainsepoint volcanic
rocks wasmodeledfrom major and immobiletraceelementabundances,including rare-earthelements,
by Massey(1980).Thepreferredmodel for magma
generation,which relies heavily on rare-earthpatterns,suggests
that parentalmagmasweregenerated
by dynamic melting within a rising diapir of garnet
lherzolitemantle. Incompatibleelementratios, particularly ThlU, imply that volcanic rocks of series
I, II and III were formed from one diapir, whereas
rocks of series.IVand V were probably generated
in a seconddiapir labeledwith a slightly differenr
Th,/U ratio. Variation trendsin AlrOr, FeO,,MgO,
TiO2, Ni and Cr in the five volcanic series'suggest
655
656
THE CANADIAN MINERALOGIST
tfe
rod
ff:d''n / I
iif,
F------1
I
nb
ma
'%),li\""
.-';2'r-qliri
'6ii€;4"^
pro. 1. 6glglalized geologicalmap of the KeweenawanRift, Lake Superior arca.a:
Pre- and Lower Keweenawansediments,b: Lower and Middle Keweenawan
volcanicrocks, c: intrusiverocks, d: Oronto Group and equivalents,e: BayfieldJacobsvilleGroup. Strata are shownshadedwhereexposed,unshadedwhereburied
beneaththe Palaeozoiccover (boundariesshownwith dashedline). Major faults
are indicatedby heavydashedlines. MP MamainsePoint, MBA Michigan Basin
gravity anomaly, MGA Midcontinent GeophysicalAnomaly, GF Grenville Front.
that shallow-levelfractional crystallizationof olivine,
plagioclaseand very small amountsof chromite has
occurred.
Ir Rssut-rs eNo DrscusstoN
Iridium was determinedby radiochemicalneutroilactivation analysisusing, with minor modification,
the procedureof Crocket et al. (1968).Samplesof
250 mg were irradiated in the McMaster reactor at
a neutron flux of approximately l0r3
neutrons/cmzlsfor 3 days.After a decayperiod of
3 weeks,Ir was separatedradiochemicallyand the
74-day 1e2lrcounted on an intrinsic germanium
detector. The sensitivitylimit for Ir is 0.005 ppb.
Ir Rssut-ts ANDDIScussIoN
Ir contents,along with Ni and Cr, are listed for
individual samplesin Table l. Averages for the
various seriesaregivenin Table 2, togetherwith the
averagesfor the olivine-phyric and plagioclase-phyric
flows. The major feature of the data is the clear
distinction betweenolivine-phyric and plagioclasephyric lavas. Ir values differ by an order of
magnitude,0.48 ppb for olivine-phyric flows comparedto 0.041ppb for the plagioclase-phyricflows.
Similar differencesare also found in the Cr and Ni
contents,which arehigherin olivine-phyricflows by
factors of 5.5 and 3, respectively.There are no
significant differencesin averageIr content of the
Ir-poor plagioclase-phyricflows of seriesII-V (Iable
2). Differencesin averageNi and Cr content ofthe
plagioclase-phyricflows are more distinct. Plots of
Ir againstCr and againstNi are shown in Figures
2 and 3, respectively.
For the total population of 54 samples,there is
a strong positivecorrelationbetweenIr, Ni and Cr.
Ir IN KEWEENAWAN BASALTS, ONTARIO
TABLEl.
Ir,
Ni AND Cr C0NTENTS
0F BASATTSFR0M'IHEI,IAilAINSE
POINT FORI'IATION
Ir.
Nl
Cr
ppD ppn ppn
Sample
No.
6
d
gr
]3
I
: :
,- 9'
r5
19
22
23
-28
p, 29
-,2'
Jt
47
5l
52
54
rsu
Ji60
63
176 721
I . 3 5 0 835 2353
0.2M 168 733
0 . 3 I 171
'173 479
0.319
699
0 . 4 0 0 ]53 624
0.804 5 5 2 l l l 3
0 . 4 6 6 210 7&
0 . 4 5 6 360 6 t8
0 . 0 3 6 r20
' 1 6 6 241
0.053
190
0.033 l J t
Jt5
0.023 80 153
0.029 106 300
0.020 6 1 1 9 1
n n?o ' t 2 7 1 0 4
't02
0 . 0 3 1 l?8
0.038 ]38
98
0 . 0 4 0 83 406
't72
0.036 l8l
't64
U . U b 5 144
0 . 0 6 6 168 146
0 . 1 4 0 ]65 164
0.016 Jt
co
0 . 0 2 5 37 t56
Sample Ir
Ni Cr
ppb ppm ppm
No.
66
67
69
7t
72
74
75
76
77
78
79
81
87
0.085 44 86
0.044 73 128
0.040 45 126
0.036 42 106
0.059 19 98
0.036 40 186
0.054 Il3 150
L
0.061 176 tl6
0.031 1l 24
0.054 I 37 136
<0.005 <5 18
0.023 28 86
0.015 41 64
0.079 162 248
4
0.047 173 273
82
0.015 281 274
AA
0.019 254 388
>:;
0.019 271 292
d;
o,
0.066 202 245
0.026 139 253
3sd
0.020 88 173
97
0.043 t96 266
'100
0.048 147 2?4
112 0.030 37 201
0.41p 398 996
0.s74 372 1058
0 . 2 9 1 3 5 11 0 5 8
i 1 1 3 0.409 351 1222
0.327 369 1260
657
The Ir-Ni-Cr correlations for the nine series-I
rocks are 0.97 and 0.96 for Ir versusNi and Ir yer"sasCr, respectively;thesecorrelationsare significant
at >0.190 confidencelevel. Thesestrong positive
correlations, the olivine-phyric character, and the
high Ir, Ni and Cr contentsof theserocks suggest
a strong associationof all threemetalswith olivine.
The strong partition of nickel into olivine from basic
magma is well known. Massey(1980) found that
precipitation of small amountsof chromite was required to accountfor the Cr content of the various
volcanic seriesusing the fractional crystallization
flo^
^..i
F 133
^
" .gr't]11..
o
ig
.ot
l.
TABLE2. AVEMGEIr, Ni ANDCr CONTENTS
OF BASALT
SERIES,IiIAMAINSE
POINTFORI.4ATION
Local i ty
Series
TT
ppb
AlonaBay
Mamainse
Point
0l -phyri c
(AlonaBay+I)
Plag-phyric
(rr-v)
I
II
III
rll
v
Ni
ppm
ppm
0.40
3 6 6 1 1l 9
u.az
311
903
0 .0 3 2 ! 0 . 02i 112t38 232!65
0.05010.035120t53 157195
0.04210.021 59!52 102148
0 . 0 3 9 i 0 . 0 2 1177!75 258155
0.48
331
980
0. 041
I l3
0.o
0l
lr ppb
FIG.2. Cr versuslr for MamainsePoint basalts.Samples
are divided into series(seeTable l): opentrianglesAlona
Bay volcanic suite, closedtrianglesseriesI, opensquares
seriesII, opencirclesseriesIII, closedsquaresseriesIV,
closedcircles seriesV.
178
Note: Seriesare numbered
fron baseuoward.
The relevantcorrelation-coefficients,
r(Ir versrzs
Ni)
and r(Ir versusCr), are 0.86 and0.92, respectively,
and are significant at a 0.lVo level. However, with
few exceptions,theseelementsare not slrongly correlated wiqhin the subpopulationsrepresentedby the
various series.Except for the olivine-phyric,Ir-rich
series-Ibasaltsand the b verws Ni correlation in
series-Ilrocks, no correlationssignificant to better
than a 590confidencelevelwere found for either Ir
yersrs Ni or b versusCr in any of the various rockseries.The strong positive correlations found for the
total population arise in part from the essentially
bimodal distribution of data, in which the olivinephyric rocks representa high Ir-Ni-Cr subpopulation and the plagioclase-phyricrocks constitutea low
Ir-Ni-Cr subpopulation.
lp araP
.f'.'t{'
A
o
^aA^
:"%
t
om
o.f
0.1
lr ppo
t0
FIc. 3. Ni versuslr for MamainsePoint basalts.Symbols
as in Fig. 2.
6s8
THE CANADIAN MINERALOGIST
modeling-techniquesof Ale$ e et al. (1977), It is suggestedthat all theseobservationsare reasonablyaccountedfor by proposingthat the olivine grains of
series-Irocks carry minute inclusionsof chromite,
which in turn concentrateIr. The concentrationof
Ir by chromite has been demonstratedin several
studiesof mineral separatesfrom sulfur-poor mafic
and ultramaficrocks(Gijbelset al. 1974,1976,Oshin
& Crocket 1982).
Concentration of Ir with olivine in maficultramafic rocks hasbeensuggested
in other studies.
Keays& Davison (1976)found an enrichmentof Ir
in sulfur-poor dunitic komatiites from Western
Australia.Keays(1982)arguedthat Ir is concentrated
with olivine in dunitic komatiites and that the concentration mechanismis probably an early precipitation of an iridium-osmium alloy alongwith olivine.
It was also noted that Ir enrichmentin sulfur-poor
olivine-rich rocks is characteristicof thoserocks that
crystallizein a plutonic environment,being usually
absent in extrusive komatiites. Whereasthe Mamainse Point volcanic rocks are tholeiitic and
generallymuch poorer in Ir than komatiitic rocks,
the tendencyfor olivine of plutonic rocks to show
Ir enrichment is compatible with the suggestion
presentedhere for the MamainsePoint olivine-phyric
volcanic rocks. As olivine, along with plagioclaseand
chromite, are inferred to be on the liquidus at extrusion, it is probable that a significant fraction of
thesemineralsprecipitated as cumulatesin high-level
magma chambersand that the Ir enrichmenttook
place in this environment. Whereas the precise
mechanism by which the Ir-olivine association
becomesestablishedis uncertain, a strong casecan
be made for chromite as the Ir concentratoron a
basisof high Cr contentand high positiveCr-Ir correlation in olivine-phyricrocks, and model calculations that requirethe precipitationof smallamounts
of chromite, possibly as small inclusions within
olivine.
Another possibleexplanationof the high Ir content of the olivine-phyric rocks is that they carry
magmatic sulfides. Whereasthe Ir-Ni correlation
might ariseby concentrationof the metalsin sulfide,
the Ir-Cr correlation is not explainedby a sulfide
control, and this mechanismis discountedas a likely means of concentratingIr in the olivine-phyric
volcanic rocks. Phenocrysts are the distinctive
characteristicof theselavasand can explainthe high
Ir, Ni and Cr if the olivine is interpretedas being
concentratedin earlier-formedcumulatesthat have
been variably incorporated into the lavas. This
mechanismwould be expectedto haveprofound effectson the compatibleelements,suchas Ir, but little effect on mildly compatible or incompatible
elements. To fully explain the Ir and Cr data,
however,somechromitemust also be incorporated,
possiblyas smallinclusionswithin the olivinegrains.
(ppb) 0F SULFUR-P0oR
Ir C0N]tNTS
MFIC R0CK5
TABLE
3. AVERAGE
n
ocean-floor basalt
(malnly Mid-Atl anti c Ridge)
oceanlc intraplate and
off-rldqe island basalts
( m a l n t yu o t e r l r e s J
Continental plateau basalts
(nalnly tholel ltes from
Parana, Karroo, Deccan
and columbia River)
Diabase dykes, North and
south Carolina
Gr€at Lake dolerite,
Tasmania
o p h i o l i t l c M g - r l . c hb a s a l t
ophiolitlc Fe-rich basalt
Ir,
ppb
References
< 0 . 0 5 7a , d , f , h , i , J
27
0.36
a,b,e,f
0.09?
b,c
0.27
f
2i
0.082
S
I
l0
0.056
0.0t9
k
k
R e f e r e n c e s : a B a e d e c k e re t a l . ( 1 9 7 1 ) , b c r o c k e t u n p u b l .
data, c crccket & skippen-(1966), d crccket & Teruta
(1977\, e Crccket et al. (1973), f Gottfried & Greenland
( 1 9 7 2 \ . a G r e e n l a n i l ( f 9 7 ] ) . h H e r t o q e ne t a l . ( 1 9 8 0 ) , I
t i e a y si 5 c o t t ( 1 9 7 6 ) ; j L ; u l e t a l . - ( l 9 Z - ) ; _ l o s h i n ( l e 8 l ) .
n: nunber of samples
CoMPARTSoN
wrrH OrHERAREAS
Although only a relatively small data-baseis
available,comparisonof Ir contentsin the Mamainse
Point Formationvolcanicrockswith thoseof basalts
from other areasis useful. Table 3 includesbasalts,
mainly tholeiites,from severalsettings:ocean-floor
basalts,oceanicand continentalintraplate basalts.
Although thereis somevariation within eachgroup
(dueto different degreesof fractionation, for example) that is not apparentin the tabulation, there is
a significant difference in averageIr contents in
basalts from different geologicalsettings.In particular, ocean-floorand ophiolitic basaltsshowlow
Ir contents(0.056ppb), whereasintraplate basalts
are much richer in Ir (0.36 ppb). The continental
plateau basalts are also higher than ocean-floor
basaltsand comparemore closelywith the oceanic
within-plate basalts.
A cursory comparison might suggestthat the
olivine-phyricvolcanicrocks of the MamainsePoint
Formation are comparableto the oceanicintraplate
and continental plateau basalts, whereas the
plagioclase-phyricrocks are comparablewith the
ocean-floorbasalts.However, if the higher Ir contents of the olivine-phlric lavas result from inclusion of earlier-formed cumulate olivine and
chromite, their useful comparisonwith other rock
that the low
suitesmay be precluded.It is suggested
Ir content of the plagioclase-phyricrocks is more
representativeof the parentalmagma. Data on immobile incompatible elementsfor the Mamainse
Point Formationtholeiiteshasalreadyled to the suggestionthat they comparevery closelywith oceanfloor basalts(Massey1980).Ir data seemcompatible with this conclusion.
IT IN KEWEENAWAN BASALTS. ONTARIO
659
This comparison i$ interestingbecausethe MaGrrnus, R., HrNornsoN, P. & Zzrs, J. (1976):
mainsePoint Formation, like other Keweenawan
Geochemistryof sometrace elementsin mineral
sequences, is subaerial and believed to have acseparates
from Rhum,InnerHebrides,with special
emphasison iridium. Econ. Geol. 71, t36+1370,
cumulatedin an elongaterift (King & Zietz 1971,
Chase& Gilmer 1973).The Ir data presentedhere
along with geophysical(Halls l97g) ana otnei --._'__ Mlttl,*:;. H., DpssoroucH. G' &
geochemical
oatilirassey
le80),
lead
totrreinterpretation of the KeweenawanRift as an immature
proto-oceanicrift.
ACKNOWLEDGEMENTS
The financial supportofthe Natural Sciencesand
EngineeringResearchCouncil of Canadais gratefully acknowledged.Scholarshipsupport to the senior
author from the Natural Sciencesand Engineering
Research Council of Canada and the Ontario
GraduateScholarshipProgram is appreciated.The
thoughtful commentsof the two refereescontributed
significantly to the improvement of the final
manuscript.
RrrsRENcrs
ALLEcRE,
C.J., TREUIL,M., Mrrsren,J.-F., MrtsrEn
B. & Al-senEpn,F. (1977):Systematicuseof trace
elements in igneous processes.I. Fractional
crystallizationprocesses
in volcanicsuites.Corlr.
Mineral. Petrology60, 57-75.
ANNeLLS,
R.N. (1973):Proterozoicflood basaltsof
easternLake Superior:the Keweenawan
volcanic
rocksof the MamainsePoint area,Ontario.Geol.
Sum. Can.Pap.72-10,
Beeorcrrn, P.A., ScueuDy, R., Etzln, J.L.,
KrMrEnLrN,
J. & Wnrson,J.T. (1971):Traceelementstudiesof rocksand soilsfrom OceanusProcellarumand Mare Tranquillitatis.Proc. 2nd Lunm
Sci. Conf. 2, 1037-1061.
f
ii;f;"ltiJ?,|3?I'ffi:i,,,1":lllhllilti#
;;;;;;;hr.iJ
Complex,SouthAfrica. Geochim.
Cosmochim.Acta 3E. 3lg_337.
GorrFRrEo,D, & GnsBNr-ANn,L.P.(1972):Variation of
iridium and gold in oceanicand continental basalts.
Int. Geol. Cong. (24th) 10, 135-144.
GREEN'I-aNo,
L.P. (1971): Variation of iridium in a differentiated tholeiitic dolerite. Geochim. Cosmochim.
Acta 35,319-322.
Harrs, H.C. (1978): The Late Precambrian Central
North American Rift System - a survey of recent
geological and geophysical investigations. In Tectonics and Geophysics of Continental Rifts (I.8.
Ramberg & E.-R. Neumann, eds.). D. Reidel, Dordrecht, The Netherlands.
HrnrooeN, J., JaNssENs,
M.-J. & Pallre, H. (1980):
Trace elementsin ocean ridge basalt glasses:implications for fractionation during mantle evolution and
petrogenesis. Geochim. Cosmochim. Acta 44,
2125-2143.
KeAys, R.R. (1982): Palladium and iridium in
komatiites and associated rocks: application to
petrogenetic problems. .In Komatiites (N.T. Arndt
& E.G. Nisbet, eds.). G. Allen and Unwin, London.
& Davrsor.r,R.M. (1976): Palladium, iridium
and gold in the oresand hostrocks of nickel sulfide
depositsin WesternAustralia. Econ. Geol. 71,
1214-t228.
& Scorr, R.B. (1976):Preciousmetalsin ocean
ridgebasalts:implicationsfor basaltsassourcerocks
for gold mineralization.Econ. Geol. 71, 705-720.
CHasr,C.G. & GrLrvrnn,
T.H. (1973):Precambrian
platetectonics:the Midcontinent gravitybtgJr,..
Eorth KrNc,E.R. &Zrcrz,I. (1971):Aeromagnetic
studyof
Planet.'Sci,Lett. 21, 7U78.
the MidcontinentGravity High of centralUnited
States.Geol. Soc.Amer. Bull. E2,2187-2208.
CRocrcr, J.H., Krevs, R.R. & HslrH, S. (1968):
Determinationof somepreciousmetalsby neutron Laut, J.C., Kravs,R.R., GanelerHy,R., AtnEns,
E.
activationanalysis.J. Radioanol.Chem.t,487-507.
& MonceN,J.W. (1972)lChemicalfractionations
in meteorites.V. Volatileand siderophileelements
MacDoucerL,J.D. & Hannrss,R.C, (1973):
-,
in achondritesand oceanridge basalts.Geochim,
Gold, palladiumand iridium in marinesediments.
Cosmochim.Acta 36, 329-345.
Geochim.Cosmochim.Acta 37. 2547-2556.
Massrv, N.W.D. (1980\ The Geochemistry
of some
& SnrnrN, G.B. (1966): Radioactivation deterKeweenawanMetabasitesfrom MamainsePoint,
mination of palladium in basaltic and ultrabasic
Ontario.Ph.D. thesis,McMasterUniv., Hamilton.
rocks, Geochim. Cosmochim. Acta 30. 129-141.
Ontario.
& Ttnura, Y. (1977): Palladium, iridium and
gold contents of mafic and ultramafic rocks drilled
from the Mid-Atlantic Ridge, Leg 37, Deep Sea
Drilling Project. Can. J. Earth Sci. 14,777-784.
OsnrN, I.O. (1981): The Abundances and
Geochemistryof some Noble Metals in Thetfurd
Mines Ophiolites,P.O. Ph.D. thesis,McMaster
Univ., Hamilton, Ontario.
660
THE CANADIAN MINERALOGIST
- & Cnocrrr, J.H. (1982):Noblemetalsin Thetford Minesophiolites,Quebec,Canada.I. Distribution of gold, iridium, platinum, and palladium in
the ultramafic and gabbroic rocks. Econ. Geol.71,
1556-1570.
THorursoN,
J.E. (1953): Geology of the Mamainse Point
copper area. Ont. Dept. Mines Ann. Rep. 62(4).
Received May 14, 1982, revised manuscript accepted
March 3, 1983.