.111 .1 1.11;. ' 1111‘ r X Ho ‘11 ’11:: ‘1 3111'. 1‘ "1 3 l 1 3 ' 1 111 1 "1' {.1 1‘ I! 1' 111 1‘ ' 1" ‘ I ' o .1 I. 1 ,1.' I '1 “I 1 . ‘ 1... ‘1.- .1 1 11111 .. 1 l 1 31‘ “‘3‘ 11111 11111 11 "111111 1.11 . 1.11111 11311 1113111 1'3 3 . 1'11 .1 ' 1131133 11111.. 11' 1 .1. 1.11.111 _ 11 .. 11 “111.111.1111 11. 1111 117 '1": 111." 11 1111111111 1111111111 1 11 1.11: 1 11111111111131 3 11 11 1 -~—m-—_ 1111 1 '11: i931 .1. ‘11111‘ 11 11.1111111'1. 11,1; 11111111111" 11111 11111111 111-.1 1;. 1,111,111 11’ 133- 1' 1‘, 1.11- 11 .111 . .13 1 1 1111" 1 1.11.. 11111111 11 11.111 1 _ A 4 -.— 4 - . .7 1 A . - ‘ A‘ ._ - — _ - ‘ — - .. V - V- . ‘ 1 1 ‘:; - A. .. _ _ w' __.v j _._..:‘_<_ .-4 A < -. . ‘ 1' _ _ A 1 . .m ____,~_ M, m .w I-a—o - 1 1111311 111 1.11.1111 1 1111 11111 111111133111 1 Z..— _-.___. 1.,- “—— ——..—~ ~— ——v-.——.__::—___ tm‘r%_ .._4- _ N$————— HA —. w. ._ ‘——.—' I 1 1 1 1 o I H’— .__._:‘____A______ '_ ' 1 11 11 1 w 2.1:: My? m -‘_. —o. ’.G=. —_.. —fi 4; ‘4‘ n‘:A-H‘;AT——L W WJ-fbfl—T fig”. “a: __ éfl—m gm J’f‘LM ___ , 11 .—_—x"—. 1 1 11111331 1 1'1111111111111111.1511 1 1 1. 1 111 1311111 1111 1' 1111111111111 3111 I111 1 1111 1 1 1 11 111 1 1111 1111.111111111 11 3' 1 113 1. .1... 13.11 1 3W1 3'31 131 1 1 .. 1 1 11331 1111111 1111 11 131111111 1 1 1111111 ~.-.1 1’1 1', .1 . 1 .11, 11 I11111111111 1 1.11 1 1 .11111111 ‘ 1 :11“ "“31 11115511113 . 111 .. 11 1 111111 11111 1 111 11 111 113331 1311111111 1 111111 1111111111 11111 1.111 ““113. Date l l um; will“ LII! Im II Illll 12» ll l all; II 006 This is to certify that the thesis entitled Models of Basalt Petrogenesis: A study of Lower Keweenawan Diabase Dikes and Middle Keweenawan Portage Lake Lavas, Michigan presented by Pipob Wasuwanich has been accepted towards fulfillment of the requirements for Masters degreein Geology \- (\k LLCK Majorprofessor 8/1/79 07639 Michigan State University LIBRARY OVERDUE FINES ARE 25¢ PER DAY PER ITEM Return to book drop to remove this checkout from your record. AP $5402 3320!: MJDELS OF BASALT PETROGENESIS: A STUDY OF LOWER KEWEENAWAN DIABASE DIKBS AND MIDDLE KEWEENAWAN PORTAGE LAKE LAVAS, MICHIGAN By Pipob wasuwanich A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF SCIENCE Department of Geology 1979 ACKNOWLEDGMENTS I would like to thank Dr. J.T. Wilband for suggestion of the problem, and for his patience, support, and constructive criticism during the development of the thesis. I would like to thank Dr.Tr.A. ngel, and Dr. H. F. Bennett for their constructive criticism of the thesis. I would also like to thank Mr. D.W. Snider fer providing maps of the area surveyed. I would also like to thank the Thai government for scholarship support and financial assistance fbr this work. Finally, I would like to thank Sudhira, Mom and Dad for their support. ii ABSTRACT MODEL or BASALT PETROGENESIS: A sum 01: LOWER KEWEENAWAN DIABASE DIKES AND MIDDLE KEWEENAWAN PORTAGE LAKE LAVAS, MICHIGAN By Pipob Wasuwanich Diabase dikes and lava flows can both be subdivided into two chemically distinct groups: a low TiOZ-PZO5 group with a higher A1203 content and Mg ratio, and lower total REE abundances than a high TiOZ--PZOS grOUp. Both groups are enriched in the LREE relative to l-IREE, and have similar REE patterns. The Mg ratios of the lavas and the diabase dikes are too low to be considered as primary melts. The systematic variations of the major elements and REE are consistent with a model of fractional crystallization of the observed phenocryst phases, olivine and plagioclase. This model requires that the primary liquid had an REE abundance pattern which is a suppressed replica of the fractionated rocks. Thus, the mantle sources of differing depth and RE abtmdances are required to produce the TiOz-PZOS subgroupings. TABLE OF CONTENTS LIST OF TABLES LIST OF FIGURES . INTRODUCTION . General Geology Paleotectonics . . . . . Location and Sample Description PETROGRAPHY MAJOR ELEMENT CHEMISTRY RARE EAREH ELEMENT CHEMISTRY RARE EARTH ELEMENI MOBILITY . FRACTIONAL CRYSTALLIZATION MODELLING . PARTIAL‘MELTINGiMODELLING. DISCUSSION CONCLUSIONS APPENDICES A. Analytical Methods B. Whole Rock Compositions REFERENCES iii Page iv H 00 bLNN 10 19 30 33 43 48 56 58 61 65 Table 10. LIST OF TABLES Average, range, and standard deviation of major element whole rock data for 7 samples of high T102— P205 basalts, and 11 samples of low TiOz- PZOS basalts . . Average, range and standard deviation of major element whole rock data for 12 samples of high TiOz- P205 diabase dikes, and 3 samples of low TiOZ- P205 diabase dikes . . Average, range and standard deviation of normallized REE whole rock data for 7 samples of high TiOZ- P205 basalts, and 11 samples of low TiOz- P205 basalts . Average, range, and standard deviation of normallized REE whole rock data for 12 samples of high TiOz- P205 diabase dikes, and 3 samples of low TiOZ- P205 diabase dikes . . Rare earth element concentration in several continental basaltic rock suites Mg ratios in several basaltic rock suites Partition coefficients for fractional crystallization calculations Original modes used in melting calculations Melting modes used in melting calculations . Partition coefficients for melting calculations iv Page 12 13 20 21 28 36 39 44 44 44 LIST OF FIGURES Figure 1. 10. 11. Keweenawan igneous rocks of the Lake Superior region (after Halls, 1966) with locations of Winona basalts and Baraga-Mhrquette diabase dikes . . Baraga-Marquette county map showing sampled diabase dikes . . .A1203- TiOz- P205 ternary plots of (a) basalts, (b) diabase dikes . AFM diagram of several Keweenawan basaltic rock suites showing their general relation- ships AFM diagram of (a) basalts, (b) diabase dikes . variation of key major element chemistry as a fUnction of stratigraphic position of basalts . REE distribution patterns of the sampled basalts of (a) high TiOz- P205 basalts and (b) low TiOz- -P2 O5 basalts . . . REE distribution patterns of average REE concentration of (a) basalts, (b) dikes . (a) La/Sm vs. La and (b) La/Sm vs. Yb plots of basalts , REE distribution patterns of the sampled diabase dikes of (a) high TiOz- P205 diabase dikes, (b) low TiOZ- P205 diabase d1kes (a) La/Sm vs. La and (b) La/Sm vs. Yb plots Page 11 15 16 17 22 24 25 26 27 Figure 12. 13. 14. 15. 16. A.plot of percent olivine fractionation vs. ng ratio showing the covariation of these two parameters in Rahleigh equation calculations . . . . A1203-(La + Ce)4Mg ratio ternary plots of (a) basalts, (b) diabase dikes . Variation diagram of La/Sm, an index of overall REE enrichment REE abundance patterns as a fUnction of degree of fusion of some hypothetical assemblages . TiO /P20 vs. TiOz plots of basalts and dia ase aikes . . . . . . vi Page 35 38 40 47 52 INTRODUCTION The purpose of this study is to use the whole rock distribution of major elements and selected REE in Winona basalts and Baraga-Marquette diabase dikes to evaluate several petro- genetic models and relate them to the Keweenawan rift system. In most genetic modeling schemes the rare earth elements (REE's) in volcanic and intrusive rocks have been widely used to evaluate such processes as partial melting and crystal fractiona- tion and to infer compositional features of the mantle source region (e.g. Gast, 1968; Kay and Gast, 1973; Shilling, 1975; Frey, et. al., 1978). The REE are very important since they show distinctive relative abundance patterns among natural basalts. These differences in REE abundance reflect the characteristic and very different solid/liquid partition relationships for major minerals. It is possible to model REE distributions by partial melting of various mantle mineral assemblages according to the equations of Shaw (1970) and Hertogen and Gubels (1976). Frac- tional crystallization models can also be used to show the effect of crystallization of mineral phases, such as olivine, pyroxene, garnet, spinel, and plagioclase upon the REE distribution. General Geology The Keweenawan rocks consist of two separated successions of lava flows: the southeast succession onMichigan and'Wisconsin and.the northwest succession of Minnesota. These two successions are locally separated and overlain stratigraphically by sedi- mentary rocks. The flows are dominantly mafic but include inter- mediate and felsic varieties and are mainly subaerial. They outcrop in a linear pattern on both rims of the rift system. The sedimentary rocks occupy synclines developed over the lava flows along the edge of the system and fbrm local thick wedges that flank the flows . In Michigan, the Keweenawan Lavas are distinguished on the basis of remnant magnetic polarity into three main groups: the lowest Keweenawan lavas (Siemens Creek Formation, Hubbard, 1975), the lower Keweenawan lavas (South Range Traps, Irving, 1883; Powder Mill Group, Hubbard, 1975) and the middle Keweenawan lavas (the well-known Portage Lake Lavas). The less extensive Siemens Creek Formation (120m thick) in Ironwood County confbrmably overlies a cratonic quartz arenite (Hubbard, 1975) and shows normal magnetic polarity (Books, 1968, 1972). The 6100m thick sequence lower Keweenawan lavas (Hubbard, 1975) extends more than 160 km from southern Houghton county, Michigan, westward to Grandview, Wisconsin and shows reverse magnetic polarity (Books, 1968, 1972). The upper part of the group is composed predomi— nantly of basaltic lavas but both andesitic and basaltic lavas are the major constituents of the lower part. The 6000m thick sequence of middle Keweenawan lavas (Hall, 1966) outcrops the entire length of Keweenawan peninsula and shows normal magnetic polarity (Books, 1972). Olivine Tholeiite, with quartz tholeiite and alkali olivine basalt are the main rock type in this group. .A system of lower Keweenawan unmetamorphosed diabasic dikes occurs in a 820 km.long and 70 km wide belt in Gogebic, Dickinson, Baraga, Ironwood and Marquette counties, Michigan. The dikes are poorly exposed but they have easily been traced from 7aeromagnetic data due to typical reversed.magnetic polarity. The l/t/Iéakes invariably trend east west with steep dips. Nest of them 37/;;/ have/been rpbrted cutting the pre- -Keweenawan rocks. Few of them // /}htruded er Keweenawan lavas (Hubbard, 1975) in Gogebic county. I / .ZENer of tz:: / ~47" \ // have been found cutting middle Keweenawan lavas. Paleotectonics .A system of linear Bouguer anomalies extends 1300 km from Lake Superior region southwestward into Kansas (King and Zeitz, 1971). The gravity anomaly is well-known as the'Midcontinent Gravity High. Several authors (King and Zeitz, 1971; Chase and Gilmer, 1974; weiblen and Morey, 1975) interpreted this feature as an aborted continental rift system. The rift system may have been opened in Keweenawan time approximately 1.1 bybp (Green, 1977). .According to isotopic age studies, Green (1977) estimated the active rifting extended from 20 to 40nwu Petro et. al. (1979) have shown that the suite of Keweenawan rocks fits into the extensional tectonic setting based on a chemical type comparison 'with known extensional rift zones. Another gravity high that extends southwestward under Michigan basin (Hinze et. al., 1972) is believed to be an extension of Keweenawan rift zone. Burke and Dewey (1973) cited this as an example of their plume generated triple junction with the Midcontinent Gravity High, the Michigan basin anomaly, and the Michipicoten area as the three arms of short-lived Keweenawan spreading center. Hubbard (1975), Green (1972), and White (1972) have deter- mined from geologic and geophysical data that volcanic rocks surrounding Lake Superior were deposited in separate tectonic basins, rather than accumulating during one single event. Case and Gair (1965) believed that the lower Keweenawan dikes of Michigan were emplaced along a major zone of longitudinal tension fractures of Keweenawan age. Several investigators (WOod, 1962; HUbbard, 1975; Morris, 1977) have suggested that the dikes may be the feeders fer lower Keweenawan lavas on the basis of stratigraphic position, reverse magnetic polarity, chemical and petrological similarities. Location and Sample Description Samples were taken from several basalt outcrops located in the central part of the Portage Lake Lavas in the western half of the Winona quadrangle and were sampled on the boundary line of Heughton and Ontonagon counties (Figure 1). Eighteen basalt samples were selected for the whole rock REE analyses. The whole rock major element analysis except fer P205 were carried out by weis (1974). Both major elements and REE analyses are presented in.Appendix B. ‘Mbst of the samples are located in the bottom half of the pile since the upper portion of the pile is only poorly exposed. In the Winona Quadrangle, the lavas trend NNE-NE. Dips gradually decrease from the bottom to the top of the pile, ranging from 50-60 degrees with locally 70 degrees (Weis, 1974). The estimated thickness of the whole pile is about 3100m. Baraga-Marquette dike swarms were selected to represent the lower Keweenawan igneous rocks. Sixteen diabase dikes (includ- ing six sampled by Mbrris, 1977) were sampled and.analyzed for the whole rock major element and REE. The dikes invariably trend east-west with nearly vertical dips. The width of the dikes vary from less than a meter to ten meters. Sample locations are presented in Figure 2. ht on a % «223 o .3er omwnwflo opposgmzémmpmm pom 33mg v.85: mo $8383 53 803 £me 3&8 :3on Houoosm 9:3 05 mo 8H8." macaw“ 555563314 953m “. II E o /// . ow. _ . J _ . L :w ”/0. 1 . q _ . a . / o a 3.... ._ 447% a o uS>>w¥ w>.m3~:2_ m<>>0._. DZ< mm:_u. ZO...> and 'N 817']. PETROGRAPHY The basalts exhibit a variety of textures, predominantly ophitic to subophitic. .A few flows have a diabasic texture. In a few flows relict olivine grains are recognized mostly as chloritic pseudomorphs throughout the groundmass, or more rarely as microphenocrysts. Surprisingly, no relict olivine was observed associated with ophitic pyroxene. Plagioclase has a similar distribution to olivine where, apart from its ophitic and subophi- tic relationship to pyroxene, it occurs as microphenocrysts and rarely glomerophenocrysts. Clinopyroxene phenocrysts were not observed. The basalts underwent low rank metamorphism (prehnite- pumpellyite facies; weis, 1974), which mainly affected olivine and sUbordinately plagioclase and Fe-Ti oxide; whereas clinopyroxene, especially ophitic pyroxene, survived virtually unaltered. The basalts are composed mainly of pyroxene and.p1agio- clase, which together comprise from 60 to 80% of the basalts. Chlorite, Fe-Ti oxide nfinerals, epidote, calcite and pumpellyite make up the remainder of the rock. Pyroxene, dominantly clino— pyroxene, occurs as either subhedral to anhedral grains or as large patched ophitically enclosing euhedral to subhedral laths of plagioclase. Chlorite and epidote are the main secondary minerals; they totally replaced all olivine grains in some specimens, and partially replaced plagioclase. Chlorite is the chief mineral in amygdules. Magnetite occurs primarily as euhedral or equant grains. Secondary magnetite occurs as medium to large skeletal crystals associated with iddingsite and Chlorite with relict olivine grains. Hematite is finely dissiminated through some highly altered flows. Leucoxene usually outlines other mineral grains. Pumpellyite is generally closely associated with chlorite. The Lower Keweenawan diabase dikes are dominantly coarse grained with subophitic to diabasic textures. However, samples JW 3, JW 6 and JW 8 are porphyritic with medium-grained sub- ophitic to diabasic groundmass. Phenocrysts are dominantly plagioclase, rarely olivine. The dikes are composed mainly of plagioclase (50-70%) 'with subordinate pyroxene, Fe-Ti oxide and small amounts of euhedral olivine in some samples. Plagioclase occurs as laths in the groundmass and as coarse equant euhedral normally zoned phenocrysts or glomerophenocrysts. Pyroxene, dominantly clino- pyroxene, occurs as small equal euhedral grains. IMagnetite occurs as euhedral and subhedral crystals. Minor amounts of apatite were observed in most specimens. The diabase dikes are only slightly affected by altera- tion. The main effects are chloritization, serpentinization in olivine and sericitization in plagioclase. iMAJOR ELEMENT CHEMISTRY The major element chemdstry of the basalts were determined (weis, 1974) by atomic absorption spectroscopy except fer ferrous iron which was determined by titrametric methods and for P205 which was determined (this study) by x-ray fluorescence analysis. The major element oxides fer the diabase dikes were determined by x-ray fluorescence analysis, except fer NaZO and fer ferrous iron which were determined by INAA and titrametric methods, respectively. The average, range, and standard deviation of whole rock composi- tion in designated subgroups of the basalts and the diabase dikes are given in Table 1 and 2. The chemical composition fer each sample is given in Appendix B. The Winona basalts can be easily partitioned on the basis of T102, P205, A1203 and 100 Mg/Mg + Fe2+ ratio (Mg ratio) into 'two distinct groups which will be referred to as the low T102- P205 and high TiO2 - P205 groups (Table 1). Apart from their low TiOz-PZO5 content this group has a higherAIZO3 content and a greaterng ratio than the high TiOZ-PZO5 basalts. The relative immobile elements, A1203, TiO2 and P205 (e.g. Hart, 1970; Cann, 1970; WOod, 1976; Floyd and Winchester, 1977) have been plotted (Figure 3) to test the significance of above classification and to show the relative variations among the 10 11 RL203 P205 X 100 RLZOS T102 X 10 P205 X 100 Figure 3.-—A1 O3-TiOEjP205 ternary plots of (a) basalts, (b) diabase di es. ( ) h1gh TiOZ-PZO5 rocks; GC)) low TiOz-PZO5 rocks. 12 com mm on flapoe Ame mm.~ + Noah u memom mm noumfizoamu +momu ofluma +mom + w2\wz oofl mo ofipmn oasoum ma oflumn w: flay . I . . . . . m N m ov um mm 5H oo NN mm ma I we HH mm HH 0 ax OMB mm.HH I w5.oH 5m.HH 5m.vH I NH.mH ww.v~ Room 5m.oc I Hm.om oo.Ho mm.vm I Hm.5v mm.om oflumm m2 mo. mo.ooH oH.H wm.ooH HmuOH H5. Nv.v I Ho.~ 5v.m oc.~ O5.v I o~.~ mm.m Ho; Ho. mm. I «a. 5H. mo. mm. I ma. mm. 0:: No. mo. I vs. 5o. mo. mm. I 5H. ma. moNa mo. H5.H I ov.H vm.~ mH. He.m I mo.m om.m mofih mm. mH.H I Hm. mm. mm. 55.H I M5. vo.~ Omx Hm. ~m.~ I mN.~ ma.~ QN. oo.m I oe.~ H5.N onz No. H5.oH I mm.w wo.oH Ho.H mo.oH I mm.5 No.m omu m5. mN.m I cw.o mw.5 so.H 05.5 - 05.4 Na.o om: om.H ~m.w I m5.m V5.m wo.~ oo.HH I HH.v w5.5 com am.H mm.w I mv.e Hm.o Hm.m vo.oH I 55.m m5.o memo; mm. Ho.5~ I vm.mH mv.oH om. O5.mH I oo.v~ ww.v~ mo~H< 5N.H oH.5v I NH.mv mv.mv mm.m vm.om I m5.Hv mm.ov moflm Gawumfi>oo omcmm ommno>< :ofiumfi>oo omcmm omnuo>< pnmpcmpm pumpcmum mu m m IN m N IN Hammm o m Cab :04 muflmmmm o a ofie :mfi: .mufimmmm memaImlo 304 mo moanemm HH cam .mpapmmm memmI~0wh saw: mo moamamm 5 now mama xuom oHonz ucosofim acnmz mo cowumfl>oo chateaum one .owcmm .ommno>oo owcma owmuo>< :00000>oa omen“ ommno>< unmozmum numpempm 00000 0000000 0000-0000 300 0000000 0000-0000 000: 0 0 0 .00000 0000000 0000-0000 300 00 0000000 0 000 .00000 0000000 o 0- 000 0000 00 000mawm N0 new 0909 zoom @0003 ucoson Menu: mo :60000>oo vumvempm 0:0 omemx .om000>om mo Emhmm0w 2m<--.0 ohswwm no: owx+owmz “on. whacnco clog—a .0. >4491 an. ova—.oaguzzx ~0N~ zuuea aw. vac; 09—. nra~.u44uzz¢ 0h“. xu—xuaoxn nun. 44:32:99 gnu. cum—aha gnu. cuuaz EIE)‘1 + X O'Q-D 16 FEB NHZO+KZO HGO FEO NRZO+K20 H00 Figure 5.--AFM diagram of (a) basalts, (b) diabase dikes. ([3) high TiOz-PZO5 rocks; (O) low TiOZ-PZO5 rocks. 17 1:02 9205 A1203 MG RATIO l 2 3 O .l .2 13 IS 17 4O 50 60 70 l l l l l J Li l L] l I L l METERS ' ' * 3CXDC>- L—9Ic.n3 , o q q I, [I \ ‘\ I ' ‘\ \ I I \ -1 I, I, “ ‘ I ,’ \\ “ I I \ \ I I \ \ I l ‘\ \ I I \ \ I I \ \ I ,’ \ \ I I \ \ 2000— ; ,I \ \ I-7o f + * § 1 | I ' ‘ I I i ‘ | I I 1 I I I I I , I l l . ' I I I ' d | l I l O I '-42 1-64 #112663 ~43 lOOO—T 3739: . .390 h'3:I.1 I-4.2. I—IOJZO .4 O_I Figure 6.--Variation of key major element chemistry as a function of stratigraphic position of basalts. The bottom of the pile (Keweenawan fault) is at 0 meters. (Small numbers are sample numbers.) 18 Thus, the accumulated crystalline phases should contain lower P205 than in the associated liquid becasue of its partitioning characteristics (Anderson and Greenland, 1969). The AlZOS-TiOz-PZOS diagram for the diabase dikes (Figure 3) also indicates two recognizably distinct clusters, although more scattered. The low TiOz-PZO5 diabase dikes are less differentiated relative to the high TiOZ-PZOS diabase dikes. RARE EARTH ELEMENT CHEMISTRY The REE (La, Ce, Sm, Eu, Yb and Lu) of the basalts and the diabase dikes were determined by neutron activation analysis. For simplicity and easy comparison the REE throughout this paper will be reported and discussed as normalized values relative to the ' average chondrite concentration (Haskin, et. al., 1968). The average, range and standard deviation of whole rock normalized REE concentrations for the subgrouped basalts and the diabase dikes are given in Table 3 and 4. The absolute REE concentrations for the basalts and the diabase dikes are given in Appendix B. Throughout this paper, La/Sm and La/Yb ratios have been used as indices of relative fractionation of the REE and La and Yb contents have been used as indices of the enrichment of REE as a whole (e.g. Frey et. al., 1974; Shilling, 1975). It is apparent from the chondrite normalized REE plots (Figure 7) that the REE abundance of the basalts are enriched relative to the average chondrite (Haskin, et. al., 1968) with greater enrichment of light REE (LREE) relative to heavy REE (HREE). There is no significant Eu depletion anomaly. The high TiO z-PZO5 basalts have higher REE concentration (La 2 60; Yb 2 18) than the low TiOz-PZOS basalts (La 2 38; Yb 2 12) but the relative abundance of LREE to HREE are quite similar (La/Yb = 3.1) as 19 20 mo. 00. I on. ma. co. cm. I 05. m5. mesm cm. 00.0 I ne.~ mo.m mm. 05.0 I mo.~ 0~.n 0>500 Hm. 5N.N I 5n.H 00.0 00. mm.” I 00.0 No.0 sm\mq n~.~ 00.00 I cw.m 00.00 mm.0 mm.n~ I mm.mH 0m.mH :0 00.0 m5.n~ I ca.on Nm.~H 50.0 me.- I 00.50 00.00 0» mm. 0m.a~ I oo.mn No.00 0m.~ 5N.Hn I ~c.m~ 5~.mN 3m on.” oo.0~ I 50.00 an.n~ no.0 00.00 I 00.00 00.0n Em 00.n nm.~n I mo.m~ mm.5~ 00.0 00.mm I 00.00 00.00 cu 00.0 00.00 I 00.00 00.00 00.00 00.00 I 05.00 05.0m 00 00000050 800001000 pumpemgm omemm ouwuo>< vumuemum omcmm ow000>< muammmm mom Imcwh_3oq muammmm mo~ NO0H :mwm 00.0 0000000 00 .nwomav .00 .00 :0xmmm_eoym 000 000 0- 00.0. 200 00 0000000 00 000 .0000000 00 0-0000. 0000 00 0000000 mmm ouwhwoonu ommpo>m 5 00m 000p 0000 000:3 mmm 0000000500e mo co0um0>mw 00000000 0:0 omcwp .ommho>\00 50. 0o.~ I 00.0 00.0 00. 00.0 I 00.0 00.0 smx00 00.0 00.00 I 00.0 00.00 m0.~ 00.00 I 00.00 05.00 :0 00.0 00.5 I 00.0 00.0 05.0 00.50 I 00.00 00.50 0» 00.0 00.00 I 00.5 00.00 05.0 00.00 I -.m~ 00.00 0m 00.0 00.50 I 00.0 50.00 00.0 00.00 I 00.50 00.00 00 00.0 00.00 I 00.5 00.00 00.0 00.00 I 00.00 00.00 00 00.00 00.00 I 50.5 05.00 50.00 00.05 I 00.00 00.00 00 000000>00 . 000000>0a 00000000 00000 00000>< 00000000 00000 00000>< 00000 0000000 000 -0000 300 00000 0000000 000 -0000 000: .000000 .00 .00 00000: 0000 0000 mmm 0000 0000 00000>0 009 .00000 0000000 0000-0000 300 00 0000000 0 000 .00000 0000000 000-0000 0000 00 0000000 N0 000 0000 0000 00003 mmm 00000000000 00 000000>00 00000000 000 .00000 .00000> ommpm>m map .muammmn mon-NOMH 30H any Ea 3:93 33-8: ng 3 mo 3:93 Bfiam 2: Us 9:8me 832:5me mmm--..\. 95mm mmmtsz u~20p¢ mmmznz watchc Nb 2. or mm mm hm mm mm vw mm Nm _m cm mm mm hm mm Nb :. or mm mm hm mm mm vm mm mm _m cm mm mm hm on p h p h h b b p p b P b u h b I- p h h P r p P p p b lb p p h p nw :4 m» am no mu a4 001 :4 m» :m :m mu c4 v v v IIIHEJ 0" " aauuoundu 03211uuaou BI] oiv‘r' ' ‘ aauuoundu oazxwuuson EJIIEm v "'0. I 1“." v W Y r v 23 indicated by the similar REE patterns (Figure 8). However, two distinct trends for the basalt groups are recognized in plots of La/Sm vs. La and La/Sm vs. Yb (Figure 9). From.these two diagrams, it is clear that the low TiOz-PZOS basalt population has a higher slope and thus higher La/Sm ratios than the high TiOz—PZO5 basalts fbr similar values of La and Yb. The diabase dikes also show two trends on La/Sm vs. La or Yb plots (Figure 11), and.have similar REE abundance patterns (Figure 8) as the basalts when subdivided into high and low TiOz-PZOS populations. The unusual low REE concentration and irregular pattern in low TiOz-PZOS diabase dikes may be related to the high plagioclase phenocryst content in sample JW 3 and.JW 6 since plagioclase does not accomodate most REE, Ti or P in its structure (e.g. Schnetzler and Philpotts, 1970; Anderson and Greenland, 1969). The similar REE patterns between the basalts and the diabase dikes may reflect similar events during magma generation and/or similar source rOCR in Lower and Middle Keweenawan time. This possibility will be examined in a following section. REE concentration data for several continental basaltic rock suites are summarized and presented for comparison in Table 5. The REE concentration of the Keweenawan rocks are within the range of most of the continental basaltic rock suites (except two Chilled margin rocks from Bushveld and Stillwater intrusions). The wide range in REE concentration (Yb, 4.4 - 20), the small range in 24 . m 830." o .mofiw Sp $338. new mo 8332850 mmm omega... mo magnum 8352233 mmm--.w 0.53m N N mmmZDZ UHEOFG NP :. on me mm hm mm mm vm mm mm #0 ow mm mm hm mm 0 — 34 m» 3m 2m mu c4 0 n “ fi ‘ N m 9 r w 7 .fim N O m. B N mg mu . N mm B v B a B . u ”I 0; o; 33 5o "938 momabofl. EB an: mwmzsz quopa Nb :. or mm we no mm mm vw mm mm "m om mm mm hm mm r m 34 m» 3m 2m mu co 0 r N 0 a v u w H m 11.0 08 OHMU m m w J H“ OM .QU . D i av _ R e . H a e . m a u a fi h C. H zox 25 80 (Q1 <3 A 6:" O 8 C) 2: “d [I] ‘Q 0 CE 8 (T) [I] ‘1 " a film 00 8 C) “‘1 O D c'2u.oo 4b 00 6b.oo ab.oo 1bo.oo 1bo.oo [.9 D ‘9 N“ ‘3 8 6'14 0 8 5. °"‘ m \ 0 CI: 3 CI!) [1] 0 E] (D O (D a.) o O "3 cb.00 1E.oo 1b.oo 25.00 23.00 2b.oo ‘(8 Figure 9.--(a) La/Sm vs. La and (b) La/Sm vs. Yb plots of basalts. ([3) high TiOz-PZOS basalts; 8C)) low TiOz-PZO5 basalts. 26 NH. mmmzaz UHEOHm .mova ommanp ; or mm mm ho mm mm vm mm mm Hm om mm mm hm mm r P _ b p h b h r b p b r F p 34 m» 3m 2w mu a4 Y V V V V V ' 0 O OY"‘ ' - f aawuoundu OJZITUUUON NH. m NAH N on oHH 30H ADV. mmmzaz mowa mmmano mon-N oHH :wH; hwy mo mowa mmmanw onoEmm map mo mchmuumo :oHuanHume mmm--. oH mpszu quoH¢ 2. or mm mm hm mm mm cm mm Nw LP » p F p p p p b » Hm ow mm mm hm mm b p b p b :4 m» 2m 2w EEIIIDHEI mu cg EJIII EJEJ EJEII EJEDE] fi— T: V f 0‘... ' aauuowndu oazlwuuaow 0 O zu: 27 9. 03‘ A 0 E] O 9‘ ID 2 ,3-1 I: E] U) \ mm CE 8 E] E] g m [I] [I] 5’. “6.005 20 00 40 00 30 00 00 00 100 00 [.9 53 é“ a (D [D 8 [I] 2 '4 Z "‘ [I] U) \ m [5 CI: .J :1 a) E] g} [I] i m [I] E] 5’. “ifgb 0100 15.00 13.00 2b.00 4453.00 ‘(8 Figure 11.--(a) La/Sm vs. La and (b) La/Sm vs. Yb plots. (D) high TiOZ-PZOS diabase dikes; ((D) low TiOZ-PZO5 diabase dikes. 28 TABLE 5.-—Rare earth element concentration in several continental basaltic rock suites. (La/Sm and'Yb reported in normalized value) Localitz La/Sm_ _Y_b_ Snake River Plain McKinney Basalt pillow glass 1.9 19 low REE tholeiite 1.4 11.5 high REE tholeiite 1.9 20 Columbia River Plateau lower Picture Gorge 1.4 12.5 upper Picture Gorge 1.7 14.5 high-Mg lower Yakima 2.5 14 low Mg-lower Yakima 2.4 19 middle Yakima (Frenchman Springs) 2.3 17 Prineville 1.8 23 upper Yakima (Ice Harbor) 2.3 21 Steen Mountain avg. of 52 , .0 16.5 avg. groups 0 and 8 1.6 12 Deccan Mahabaleshwar (avg. of 10) 1.7 13 lfiscellaneous (range of 18) 1.8-2.5 12-20 Keweenawan Duluth Gabbro 3.0 4.4 North Shore Vblcanics basalts 1.6-3.1 8-20 (range of 12) Gabbroic intrusions-chilled.margin Skaergaard 1.4 5.8 Stillwater 0.5 1.5 Bushveld 0.6 4.3 Palisade 2.0 10 W-l 1.9 10 Keweenawan (this research) LOW TiOz’pzos basalts 1.62 18.92 High TiOZ-PZOS basalts 1.64 12.52 Low TiOz-PZOS diabase dikes 1.53 17.39 High TiOz-PZOS diabase dikes 1.55 6.56 After Leeman (1977) except the rocks in this research. 29 relative REE abundance (La/Sm, 1.4 — 3.0) and similar REE patterns are surprisingly Characteristic of the continental basaltic rock suites. Leeman (1977, p. 160) stated that the similarity in (La/Sm) between continental tholeiites is consistent with their derivation from similar source materials, whereas the variations in this ratio and Yb within individual magmatic suites, may reflect differences in the degrees of partial melting of these similar source rocks which have chondritic REE relative abundances. RARE EARTH ELEMENT MOBILITY It is of great significance in this investigation to determine the effect of low grade greenschist facies metamorphism on the REE's. Unfortunately, there have been no systematic studies of REE mObility during metamorphic processes. Haskin et. al. (1966) observed that fluctuations in the REE profile of basalts may be due to metasomatic removal of LREE, especially La. 1M0re recent studies, involving zeolite and green- schist facies lavas, partially support this observation. In greenschist facies alteration, various authors (Frey et. al., 1974; Herrman, 1974; Hellman et. al., 1977; Sun and Nesbitt, 1978; Humphris et. al., 1978) concluded that low grade metamorphism did not significantly change the REE profile. Hellman et. al. (1977) demonstrated, within a single outcrop of basaltic rock within the prehnite-pumpellyite facies, that there are some variations in La and a consistent negative Ce/Ce* (Ce* is interpolation value in normalized REE plot) anomaly. Frey et. a1. (1974) have suggested that altered glass can show a marked decrease in REE (except Ce) while altered crystalline basalt sample may show an increase of LREE and a decrease of HREE. Herrman (1974) suggested that minor fluctuations may exist but in no case could LREE enriched lavas be changed to LREE depleted lavas. 30 31 Humphris et. al. (1978) found a very small increase in all REE progressively from the least altered to the most altered part of the flow. Sun and Nesbitt (1978) suggested that REE normalized patterns, particularly LREE and Eu can be affected by metamorphism, but they argued that the consistency of pattern enables recogni- tion of primary magmatic patterns. In this study, the altered rim.and amygdule lavas were avoided but in some samples alteration is extensive, especially sample 91C. When the basalts were segregated into fresh and altered groups based on the ferric/ferrous ratio plus H20, as an indication of alteration, no correlation or systematic variations were feund.between the REE to indicate REE mObility. Although minor abundance fluctuations fer Ce occur relative to the other REE, these depletions do not correlate with chemical or petrographic alteration indicators. Ce depletion may be the result of a primary magmatic condition (Fleet et. al., 1976), whereby Ce is oxidized to the Ce4+ state. However, the afore- nemtioned La variation as a function of alteration makes it difficult to assess whether a rock shows a La enrichment or a Ce depletion. In sunmary, the lack of correlation between petrographic or chemical alternation criteria and the REE, the fact that the variations within the suite subgrouping are relatively minor, and the overall similarity of REE patterns will probably enable recognition of primary magmatic processes upon REE abundance. 32 In sumary, the lack of correlation between petrographic or chemical alternation criteria and the REE, the fact that the variations within the suite subgrouping are relatively minor, and the overall similarity of REE patterns will probably enable recognition of primary magmatic processes upon REE abundance. FRACTIONAL CRYSTALLI ZATION NDDELLING The Mg ratio (atomic ratio of 100 Mg/Mg + Fe2+ ratio where Fe2+ calculated as FeZO3 = TiO2 + 1.5) of basaltic rocks is used as a measure of the degree of crystal fractionation of ferromag- nesian phases from ascending magma and as a criterion to recognize basaltic rocks that generate directly by partial melting of the mantle (Green et. al., 1974; Frey et. al., 1978). Green et. a1. (1974) noted that between 20-30 percent partial melting of a peridotite mantle having a‘Mg ratio of 90 will generate a liquid 2 ‘with a ratio of 70i . Frey et. a1. (1978) calculated a Mg ratio of 68-75 for primary magma (where K(6§§€§Eiq) = 0.3 from.Roeder and Emslie, 1970) derived from 20-30 percent melting of the source rock having a,Mg ratio of 88—89. Frey et. al. (1978) concluded that Mg ratio of primary basalt magma ranges from 68 to 75. Following this criterion, the Mg ratios of the Winona basalts and the Lower Keweenawan diabase dikes are too low to be considered as primary melts. Because olivine is the only ferromagnesian mineral observed as phenocrysts, it will be assumed to be the phase which primarily reduced the Mg ratio through fractional crystal- lization. The Mg ratio can therefbre be used to calculate the percent olivine which was fractionated from the rocks. 33 34 The partition coefficient of Fe/Mg between co-existing liquid and olivine (K (0iicygiq)) was feund to be independent of temperature (Roeder and Emslie, 1970). Assuming the original Mg ratio in primary liquid equals 70 and using K (é§:6y§1q) = 0.3 (Roeder and Emslie, 1970), the change of Mg ratio with olivine fractionation can be calculated by the Rahleigh equation. Figure 12 is a plot of olivine fractionation as a function of the Mg ratio. Note that the low TiOz-PZO5 subgrouped basalts and diabase dikes indicate 38 and 42 percent olivine fractionation, and the high TiOZ-PZO5 basalts and diabase dikes have olivine fractionation values of 66 and 74 percent. However, the use of Mg ratio for estimating the crystal fractionation of olivine is limited by uncertainty of the Feb/Fe2+ + Fe3+ ratio (Sun and.Hanson, 1975). Thus, the aim.of this calculation is only to show the relative olivine fractionation between groups of the basalts and the diabase dikes. Mg ratios of several basaltic rock suites are presented for comparison in Table 6. ‘Mg ratios of the Keweenawan rocks are similar to most of the continental basaltic rock suites. The low Mg ratios of continental basalts may indicate a high degree of ferromagnesian mineral fractionation. On the other hand, Mg ratios of oceanic basaltic rock suites are equal to or close to 70. This may indicate that oceanic basalts are primary liquids and/or have small degrees of ferromagnesian mineral fractionation. An alternative interpretation is that both suites have low amounts of ferromagnesian fractionation and that the mantle beneath 3S 100- 80- 74-J1.' EH. .1192; mPLA£A¥_E'§§ z 9 I— - I - < ,6 -mor- 0239532112-“-.. 2 I 9 6° ' I " I- I 2 - I . == 5 . “' 42-.l_°1V.I.'97_-££5.2|AILAEE_|?J'S£-1-1--- m 4° ' 3.--t2v_v.".°2:!z<>.r..s.e.s.eua--__2.4---- ‘ Z 1 I I. _ I I g, ->- ' i E II . 8 I I I: N 20 - g : :5 H i I :I I i I} " ' I i! - I I I' I I I: O '- '19 ' 60.49 ‘ 50.85 61.60 I J l 1 1 1 l I O 20 4O 60 70 MG RATIO Figure 12.--A plot of percent olivine fractionation vs. Mg ratio showing the covariation of these two parameters in Rahleigh equation calculations. Tie lines show the Mg ratio values in observed basalts and diabase dikes sub- groupings with percent olivine fractionation. 36 TABLE 6.--Mg ratios in several basaltic rock suites. Locality ‘Mg ratio _ References CONTINENTAL BASALTS Columbia River basalts High Mg Picture Gorge S9 (1) Low Mg Picture Gorge 47 (1) High Mg Lower Yakima 44 (1) Low Mg Lower Yakima 38 (1) Middle Yakima 43 (1) Paineville 50 (l) McKinney basalts (Snake River Plain) 48 (2) Deccan Traps Lower Deccan Traps 46 (3) Upper Deccan Traps 44 (3) OCEANIC BASALTS Average oceanic tholeiite 70 (4) Average Atlantic ridge basalts 67 (5) (glasses) Indian oceanic basalts 66 (6) Indian oceanic basalts (glasses) 64 (5) East Pacific rise (glasses) 60 (S) KEWEENAWAN BASALTS (this research) Low TiOz-ons basalts 51 High TiOz-ons basalts 62 Low TiOz-PzOS diabase dikes 46 High TiOz-PZO5 diabase dikes 60 References: (1) Nathan and Fruchter (1974), (2) Thompson (1975), (3) Sukheswala and Poldervaart (1969), (4) Engel et. a1. (1965), (5) Bryan et. a1. (1976), (6) Fleet et. al. (1976). 37 continents has a lower Mg ratio than the mantle beneath oceanic areas. A possible way to trace fractional crystallization trends to include plagioclase, the only other observed phenocryst phase, and olivine on the REE may be demonstrated in.A1203-(La + Ce)-Mg ratio ternary diagrams (Figure 13). In this plot A1203 is assumed to best represent the plagioclase content (CaO and alkali oxides are more effected by alteration; Joly and Smith, 1974), and the Mg ratios reports the olivine fractionation. The progressive increase of (La + Ce) with decreasing Mg ratio and lower (La + Ce) with higher Mg ratio in the low TiOZ-PZOS basalts and diabase dikes is consistent with variation in olivine fractionation. If the dominant fractionating phase was clinopyroxene, in spite of the lack of petrographic data, the ternary trend would also be the same. The progressive increase in (La + Ce) with decreasing A1203 may be interpreted to reflect plagioclase crystal fractiona- tion. Furthermore, the low TiOz-PZO5 basalts and diabase dikes if they were derived from the same primary magma. However, the lack of Eu depletion in REE patterns may limit plagioclase crystal fractionation less than 20 percent (Philpotts and Schnetzler, 1968; Onion and Gronvold, 1973). Extensive plagioclase fractiona- tion will generate a residual liquid that has Eu depletion since plagioclase preferentially accept Eu2+ in its structure (Philpotts and Schnetzler, 1968; Schnetzler and Philpotts, 1970). 38 A1203 X IO I! I Ll L ‘l ‘1 ll ‘1 I.A+CE ' MG RATIO x 2 AL203 X IO M II H M I! III H H LA-I-CE MG RATIO x 2 ’ Figure 13.--A1§O - (La + Ce) -Mg ratio ternary plots of (a) basalts, (b iabase dikes. ([3) high TiO -P O rocks; (0) . 2 2 5 low T102-P20S rocks. 39 The effect of monomineralic fractional crystallization on REE distributions can be depicted on a plot of the La/Sm ratio (Figure 14), which reflects the relative fractionation of the REE's vs. a total REE abundance indicator, in this case Yb. The crystal/ liquid partition coefficient for these REE's in the various ndnerals (Table 7) are used in the Rahleigh equation to calculate the concentration of the REE various melt fractions as fellows: _ (D-l) cl/co - F where C1 = elemental concentration in the liquid Co = original elemental concentration in the system F = fraction of residual liquid D = partition coefficient TABLE 7.--Partiti0n Coefficients fer Fractional Crystallization Calculations. 015M” 0px(1) Cpx 1(2) Q)x(3) Cara) p1(4) La .0021 .0021 .084 .02 .001 .09 Sm .0098 .0147 .736 .14 .2 .04 Yb .0202 .1443 1.01 .20 4.0 .02 References: (l) Frey et. a1. (1978). (2) Onuma et. a1. (1968). (3) Gruttzek et. a1. (1974). (4) Blanchard et. al. (1976). lA/SM 0.5 Figure 40 (UNOPYIOXENE ‘ I .21) .3 ‘I ClINOPYROXENE- 2 ~05 OIIHOPYROXENE 4: ouvewi‘ n” .I. ann ' .3 PlAGIOClASE b YB l4.--Variation diagram of La/Sm, an index of relative REE abundance, vs. Yb, an index of overall REE enrichment. Line of liquid descent are shown for monanineralic fractional crystallization with mineral names marked. Index nunbers refer to fraction of liquid remaining. Clinopyroxene-l represents high LREE partition coeffi- cient relative to HREE and clinopyroxene-Z represents similar LREE and HREE. 41 The plotted lines represent lines of liquid descent fer olivine, orthopyroxene, clinopyroxene, plagioclase, and garnet. For simplicity, the calculations have been normalized to an initial ratio of l at 0% fractional crystallization of each depicted crystalline phase since the (La/Sm) and Yb ratio cannot be ascertained in the primary liquid. Polymineralic fractionation trends can be approximated from infermation in Figure 14 by adding vectors for each mineralic trend scaled in proportion to their relative fractionation abundances in the polymineralic assemblage. Although fractional crystallization modelling strongly depends on the choice of partition coefficients, it is apparent that the selected coefficients have little effect on relative fractionation of the REE (La/Sm), except for clinopyroxene, because although the absolute magnitude of different mineral partition coefficients may vary, the relative REE abundances fer any given mineral are very similar. TWo sets of clinopyroxene partition coefficients were chosen from the literature so as both to maximize and minimize their relative fractionation. It is apparent from clinopyroxene curves that fractional crystallization models are strongly dependent on the partition coefficients. It is important to note that, with the exception of clinopyroxene-1, because of the low slopes of the liquid descent line wide ranges of Yb will yield only small variations in the La/Sm. Stated otherwise, in a system dominated by olivine 42 and/or plagioclase crystal fractionation the relative fractionation proportions of the REE's remaining in the liquid remains essentially the same. Fractional crystallization of assemblage including clinopyroxene-1 would substantially change La/Sm ratio. Garnet, on the other hand, rapidly depletes the HREE in the residual liquid without substantially changing the La/Sm ratio. Similar observations were reported by Shilling (1971, 1975); Kay et. a1. (1970); Shilling and Bonatti (1975). PARTIAL MELTING CALCULATIONS .A batch partial melting model was chosen for melting calculations in this research since it seems to be more realistic than fractional melting models (e.g. Leeman, 1977; Hartogen and Gubels, 1975) which involves continuous segregation of melt from the residual. In the batch melting model, magma accumulates in a single batch and remains in chemical equilibrium with the residual minerals until the batch of magma is extracted. It is assumed that melting occurs at some invariant (i.e., eutectic or peritectic) composition that can be expressed in terms of the proportion (pi's) in which each phase melts; fer melting interval in which none of the original phases is gone, the value of pi remains constant. Relative concentrations of trace elements in hypothetical derived melts were computed using the batch.melting equation of Shaw (1970) and Hertogen and Gubels (1976) and using a wide range of mantle compositions from the literature (e.g. Leeman, 1976; Carter, 1970; Shilling, 1975; Pankhurst, 1977). The postulated modal abundance for selected models are given in Table 8. The melting modes used in the calculations (Table 9) from various investigators approximate the experimentally determined values of Kushiro (1968, 1969) and Davis and Schairer (1965). The distribu- tion coefficients that were selected are given in Table 10. 43 44 TABLE 8.--Origina1 modes used in melting calculations. ASSEMBLAGE ORIGINAL PROPORTION REFERENCES Olv:Opx:Cpx:Gar .55:.20:.15:.10 Leeman (1976 Olv:Opx:Cpx:Gar .55:.25:.lS:.5 Pankhurst (1977 Olv:Opx:Cpx:Sp .51:.21:.23:.S Carter (1970) Olv:Pox:Cpx:Pl ,.55:.20:.15:.10 Leeman (1976) Olv:Opx:Cpx .60:.25:.15 Shilling (1975) Note: 01v - olivine, 0px - orthopyroxene, Cpx - Clinopyroxene, Gar - garnet, Sp - spinel and P1 - plagioclase. TABLE 9.--Me1ting modes used in melting calculations. ASSEMBLAGE MELTING PROPORTION REFERENCES Olv:Opx:Cpx:Gar .10:.20:.40:.30 Pankhurst (1977) Olv:0px:Cpx:Sp .10:.20:.60:.10 Leeman (1976) Olv:Oox:Cpx:P1 .10:.20:.40:.30 Leeman (1976) Olv:Opx:Cpx .25:.20:.55 Shilling (1975) TABLE 10.--Partition coefficients for melting calculations. Olivine Orthopy- Clinopy- Garnet Spinel Plagio- roxene roxene clase La .0005 .0005 .02 .001 .03 .109 Ce .0008 .0009 .04 .0033 .032 .084 Sm .0019 .0028 .14 .0823 .052 .065 Eu .0019 p .0036 .16 .1333 .055 .066 Yb .0040 .0286 .20 4.0 .17 .040 Lu .0048 .0380 .19 7.0 .081 .031 Data from Frey et.al. (1978) except spinel from Kay and Cast (1973) and plagioclase from Drake and Neil (1975) . 4S Calculated results are plotted on chondrite REE plot (Figure 15) for comparison with the Observed data. 46 Figure 15.--REE abundance patterns as a fUnction of degree of fusion of some hypothetical assemblages. (a) Chondrite- normalized REE pattern of average samples of (l) the high TiOz-PzOs basalts, (2) the high TiOz-PZOS diabase dikes, (3) the low TiOz-PZOS basalts, and (4) the low TiOz-PZOS diabase dikes; b through f, vertical scale represents REE enrichment in derived liquids relative to initial concentrations. Assuming chondrite relative abundances for the initial mantle source, these patterns may be directly compared with chondrite-normalized REE plots. Proportions of phases in the model source rock (x = mode) and in the asswned invariant minimum (P = melt) are given as weight percentage for each model. Symbols for phases in mantle models are OLV - olivine, OPX = Orthopyroxene, CPX = clinopyroxene, GAR = garnet, SP = Spinel and PL = plagioclase. The percent melts are marked on the plotted lines. 47 mmmzaz 8:9; Nb .5 am am ow mm mw mw vw mm «m .m ow mm on em mm «“952 3:21 NH. 2.lo am am hmmw mm vw nHmNHwHLoonImmOprHm mm 2.. m» DU 2m mu ¢J 0. H X X X :8 X .X . v I. I’ll... . m I «Id/I . m no. In OIO . o f 8118 v I 3 8 an .. a. an 3 x u. f .5 E0 30 . l a a 0 3.. m» 8 0. ON 0. x a a. ON mm x «<0 xsv Ito >40 Du :m mu :4 XIXIIIIIIXIIX no» . H .8 I no. u... u. 0 I OI nusoundu 0311 1mm»: amoroz UHZOHG mumEDz u_£oh¢ NH. : or mm mm hm mm mm vo mm Nm Hm ow mm on hm mm .u.. r pr mm mm hm mm mm on mm mm _m 00 mm mm on mm H H H H H H H H H H I H H H r V r Ir H H H H H H H H H H IF H II o 0 34 my Du cw mu m; o 3 me am cm mu c4 0 v H xix xIxIlIIIlex . xix xlxlllxlx . I8 3 _ la . . I H H u _ .TIIT . H u q a v w I sl "Op rylo @116 0. .‘IO Ole am am H E n n H v I. on 3 an o. a 2 3 8 o. a n. o. 2 2 3 x . H 2 a a x a; EU 5 ’6 no ICU K8 2.0 u v a v OI 0' aunts: u—cohc mum—.52 “Etc—d NF 2. or an no em mm mm .0 mm mm —o co mm 0M 5M mm NH. Hh oh me 0% 6% mm me va nHm NW Wm ow mHm on hm mm o o a» 3m .5 mu :4 o 3.. a» 3m cm mu :4 o H v v XIXIIIXIX I“ . ' . m N . m 1. . ... v 1. o . + 0.“ I, v .m . X. n u ,.r m In m W 3 H 3 v 3 [DI 3 v P v on 9 on o. a a. . H O- a. ON an I v H .40 K‘U X8 >6 . a H H r m. o. DISCUSSION Because of the nearly identical chemical characteristics of the reversely polarized diabase dikes in Marquette and Baraga counties and the younger normal polarized Portage Lake Lavas from the Winona Quadrangle they are assumed to have resulted from similar genetic processes even though these suites are temporally and spatially separated. This discussion addresses the question of genesis with specific references to the chemical subgrouped suites and unless specifically stated, ignores the possible restrictions imposed by the spatial-temporal differences. The discussion whiCh fellows begins with mantle melting and examines the primary magma through differentiation and fractional crystal- lization modelling. Several mantle compositions reported in Table 7 have olivine (51 to 60%) and orthopyroxene (20-25%) as major consti- tuents in proportions whiCh vary from 2.2 to 2.75:1. The remaining minerals (up to 25%) are various percentages of some or all of clinopyroxene, spinel, garnet and plagioclase. Garnet, spinel, plagioclase and clinopyroxene peridotite were used in melting calculations to generate l to 30 percent liquid to demonstrate the effect of mineralogy on the liquid REE distribu- tion patterns (Figure 15). It is apparent from these patterns 48 49 that l to 5 percent nentle melting, especially of spinel and clinopyroxene peridotite, could generate liquids which have distribution patterns enriched in the LREE similar to those in the two suites if fractional crystallization is ignored. The signi- ficant difference in the abundance of the REE is attributed to the strong partitioning of the HREE to garnet whiCh leaves the liquid strongly depleted in Yb and Lu relative LREE in the l to 5 percent melting range. .Although small amounts of partial melting of spinel and clinopyroxene peridotite have similar REE distributions to the observed data (Figures 15d, f), the basalts or the diabases could not have formed.by crystallization of these melts fer the fellowing reasons. First, the Mg ratio of the rocks is not indicative of a.primary tholeiite. Second, if major element chemistry is considered, the primary melt would be strongly alkalic in nature since the large ion lithophile elements (LILE) are partitioned toward the liquid. The Observed tholeiite nature of the rocks is in direct contradiction to LILE enrichment. Finally, the dis- tinct groupings based on TiOZ-PZOS concentrations, each with dis- tinctive systematic REE variations (Figures 9 and 11), negates the model of simple mantle melting of a single, homogeneous source to generate both groups. Several experimental studies (Green and Ringwood, 1967; Green, 1970; Kushiro, 1972) show that primary tholeiite magmas are generated upon 20 to 30 percent partial melting of the mantle. 50 Following the above experimental work, from virtually all composi- tions tested with an assumed chondrite REE abundance, the signi- ficant fact which emerged is that 20 to 30 percent melting generated "flat" abundance patterns with the REE in the liquid enriched 3 to 5 times the amount in the starting solid. Shmilar observations are reported by Shilling (1975). From this fact alone it is clear that a partial melting model within the experimentally determined limits of 20 to 30 percent cannot explain the distribution pattern of the REE in the basalts and.the diabases. However, it is possible that such "flat" patterned liquids could be modified by other processes, such as fractional crystallization, to give the Observed REE and major element characteristics of the basalt and diabase suites. If the presence of the olivine and.p1agioc1ase pheno- crysts is assumed to represent crystal fractionation, it was shown above (Figure 14) that the REE's do not appreciably fractionate (e.g. La/Sm ratio remains relatively constant), whereas Yb abundance (a measure of total REE abundance) increases with increasing fractionation of these minerals. Stated other- ‘wise, the fractionation of these minerals will produce "flat," but enriched, abundance patterns similar to the "flat" primary melt produced at 20-30 percent mantle melting. Therefore, it is clear that without fractionation of another phase, crystal frac- tionation of olivine and/or plagioclase from such a melt cannot generate the abundance patterns observed fer the rocks. 51 A.model that evokes mantle melting to create a single primary melt modified by fractional crystallization is fUrther complicated by the two distinct and.1inear trends in the (La/Sm) vs. La and.(La/Sm) vs. Yb plots (Figures 9 and 11) fer the TiOz-PZOS subgrOUping. .A.single linear trend.in these plots is indicative of liquids derived by varying degrees of partial melting of a mantle rock with a specific initial REE content. One is left to conclude that two mantle sources with differing depths and.REE abundances are required to produce the two TiOZ-PZOS sub- groupings. Finally, if a single primary melt gave rise to the two subgroupings, the two linear trends could not be accounted for by simple olivine and plagioclase fractionation since these nfinerals do not effectively change the La/Sm ratio. Chazen and Vogel (1974) noted that the TiOZ/PZOS ratio in mantle melts will depend on the interrelationships of titanium bearing mineral stabilities and apatite stability at a.particular pressure. They further stated (p. 309) that this ratio will be largely independent of reasonableramounts of partial melting since limited melting affects the absolute amount but not their ratios. They argued that the ratio is essentially independent of frac- tionation of primary silicate phases. Thus, the remaining liquid, after some fractionation, will have essentially the same ratio as the initial ratio. It is apparent on the plots of (TiOZ/PZOS) vs. TiO2 (Figure 16) the low TiOZ-PZO5 subgroups have higher TiOZ/PZOS ratios. 52 .mmHHe mmmamHe cam mpHmmmn Ho mHoHa NoHH .mH moNa\NoHH--.OH mesmHa No H H ouum oeuu om”: omuo oo.pu @ Ha 4%? “8% w 0% + .7 .l 4.0 I 0 0 + 0 Z / 1.12. + m 0 8:8 magma mowing: 30.. + 9 8:8 3:38 83-8: :2: 4.. 2. Sega 82-8: :3 e m 0 23:2: 82-3: :2: a o 53» Following the arguments of Chazen and Vogel , the fact that there are two distinct ratio clusters suggests that the subgroups represent the liquids derived from mantle regions of different TiOz-PZO5 abundance, and/or that the liquids were derived by partial melting at different depths. Although no petrographic data were found to confirm clinopyroxene fractionation, this process could adequately explain the observed REE abundances, assuming a "flat" abundance pattern for the primary unfractionated melt. Clinopyroxene frationation can change the relative REE abundance (La/Sm) without a significant variation in Yb (Figure 14) if the high fractionated values of partition coefficients are used in the model. The model must include fractionation of the observed phenocryst phases, olivine and plagioclase, whereby sunning the appropriate proportions of fractionation of these three phases would generate the observed patterns. The amount of clinopyroxene fraction must be unusually high, however, to generate such pattern, especially the low TiOz- P205 basalts and diabase dikes in (La/Sm) vs. La and (La/Sm) vs. Yb plots (Figures 9 and 11) . This can be explained by a higher percentage of clinopyroxene (less fractionation) in the low TiOz-PZO5 basalts and diabase dikes. The amount of clinopyroxene fractiona— tion required to generate the observed REE patterns fran a 20-30% mantle melt is greater than 90%. This excessive amount could be significantly decreased to a reasonable value only if the mantle source was REE enriched. 54 This model ignores the previously stated criterion implicit in the TiOZ/PZO5 ratio, which implies two magma sources generated at different depths. Furthermore, the model needs a very special event of crystallization processes to generate two liquids (the high and low TiOz-PZOS) at the same place in different stratigraphic positions fran the same primary magma. Thus, two magma sources generated at different depths are still needed even in the crystal fractionation of assemblage including clinopyroxene. Morris (1977) and Morris et. al. (1978) reported that a lower Keweenawan peridotite (Yellow Dog Plains peridotite, Marquette county) has an enriched REE pattern similar to the diabase dikes but lower in REE concentration (La/Sm = 1.39, Yb = 3.25). They noted that the REE distribution in peridotite is not easily explained by fractional accumulation from a basaltic parent. Morris et. a1. (1978, p. 279) stated that mass balance calcula- tions indicate there is insufficient LREE's in the diabase to serve as the peridotite parent or to be derived from the peridotite. They also concluded that the high REE and alkali abundance in peridotite could result from melting of a heterogeneous REE enriched portion of the upper mantle, rich in volatiles. This observation supports the mantle heterogeneity as a source of these magmas. In summary, if clinopyroxene fractionation is discounted, the basalts and diabase dikes could have originated fran a primary magma by olivine and subordinate plagioclase fractionation only if 55 the primary liquid had an abundance pattern which is parallel to, but lower than those of the fractionated (Observed) rocks. Thus, the mantle source must have been undepleted and.enriched relative to chondrite values. The inescapable conclusion, no matter whether clinopyroxene fractionation played a role or not, is that there must have been two source liquids to account fer the variation in REE and TiOZ/PZO5 in the subgroups. If clinopyroxene fractionation played a role in the final REE abundance patterns, then the source mantle in this model could have chondrite abundances; and the fractionated rocks derived from "flat" REE abundance pri- mary liquid generated at different depths. C(NCLUSIONS Geochemical and petrographic data have been used in models of basalt petrogenesis for two Keweenawan rock suites. TWo ‘models, mantle melting and fractional crystallization, have been integrated to explain the major and REE abundances. A sunmary of the model characteristics fellows: l) The accepted simple model of 20 to 30 percent melting of mantle material to generate tholeiite is inadequate because the RE abundance of such the primary liquid would have REE values much lower than those observed, and.should have Mg ratios which are approximately 70. Thus, fractionation of such a liquid to obtain values lower than 70 is implied. 2) A model in which the observed phenocryst phases, olivine and plagioclase, are fractionated from a primary melt with chondrite REE abundance would not give the observed LREE enriched patterns. However, an integrated model of melting a chondrite REE abundance peridotitic mantle plus clinopyroxene fractionation could result in the observed REE abundance patterns. This model is not favored because of the large amounts of clinopyroxene fractionation required, and the absence of rocks containing clinopyroxene phenocrysts. 56 3) The linear trend in the La/Sm vs. La and Yb plots are indicative of a series of liquids (rocks) which were derived by varying degrees of partial melting of a common source. Since two linear trends are observed.in each suite, a petrogenetic model must incorporate two sources of differing REE abundance, or melting of a homogeneous source at different depths. On the basis of above it is concluded that the Keweenawan rock suites could have originated fran a primary magma by olivine and subordinate plagioclase fractionation only if the primary liquid had an abundance pattern whiCh is a subdued replica of the derivative rocks. Thus, the mantle source must have been undepleted and enriched relative to chondrite values. The strong segregation of the SUbgTOUpS'Wlth distinctive REE linear variation plots and TiOZ-PZOS contents can be explained if the primary melts which gave rise to the subgroups were derived at different depths. The similarity the REE pattern of the high TiOz-PZO5 lavas to the high TiOz-PZO5 diabases and the corresponding similarity of the low subgroups to one another suggests that similar magmatic processes were operative during middle and lower Keweenawan times. That is, magma generation during the Keweenaw rift interval appears to have fellowed similar paths. This being the case, it may be possible to relate the lower Keweenawan lavas of the South Range Traps and the Powder Mill Group to the diabase dikes on.the basis of their REE abundance patterns. APPENDIX A ANALYTICAL METHODS APPENDIX A ANALYTICALIMETHODS Analytical methods used in this study include X-ray fluorescence, neutron activation and titrametric analyses. X-ray Fluorescence The equipment used in X-ray fluorescence analysis includes: 1. General Electric XRF x-ray generator and dector panel. 2. Flow-proportion counter with p-lO gas (90 percent argon - 10 percent methane) Helium path used when needed. 3. Analyzing crystals, LIF, ADP, PET. 4. High power Mo and Cu target tubes. 5 Other associated electronic equipment. X-ray fluorescence analysis was used for whole rock analysis for the following elements: Si, Al, Fe, Mg, Ca, Na, K, Ti, P, Mn. The sample preparation method used was a fusion technique using the proportions l:l4:3 for samplezfluxzbinder as used by Fabbi (1972). The phosphorous method described by Fabbi (1971) was also used. Neutron Activation Analysis Neutron activation analysis was used to determine whole rock compositions for the rare earth elements, La, Ce, Sm, Eu, Yb, and Lu and for Na. The equipment used in this analysis includes: 58 59 l. A Triga Mark I nuclear reactor. 2. .A GeLi detector--a lithium.drifted germanium crystal held at cryogenic temperatures by liquid.nitrogen. 3. Multichannel analyzer, amplication equipment and other associated electronics. Powdered (1.00000 gram) rock samples, crushed to pass through a 200 mesh seive, were placed in polyvinal vials and sealed for irradiation and analysis. Titrametric Analysis .A titrametric method described by weis (1974) modified after Schafer (1966) was used for the determination of ferrous iron in the analyzed.samples was modified after a procedure used in the Chemical Laboratories of the Geology Division of the Research Council of Alberta (weis, 1974). For detailed procedures of these methods, see weis (1974). 60 TABLE ll.--Comparative standards used in analyses.* X-ray Fluorescence and Neutron Activation W-l BCR-l G-Z AGV -1 GSP-l PCC-l MRG-l DTS-l SY-Z SY-3 Titrametric III-1 *Abby's (1978) concentrations were used. APPENDIX B WHOLE ROCK COMPOSITIONS 61 .HHHQHV mHoz touw< OH.HOH HH. mo. mm.H mo.~ HN. OH.N oo.m mo.m om.v mm.o m~.HH «3.04 we HH.HoH 4H. mo. o¢.H ma.~ HN. Hv.~ mo.oH H~.m HH.¢ Hm.o oo.oH H~.oe no am.mm . Ha. Ho. HH.H m:.n Ho.H m~.~ Ho.a 0:.H me.o Hm.m HH.oH HH.¢¢ ¢~H Hm.mm an. Ho. NH.H Ne.. :H. :m.~ Ho.m HH.: Hm.H mH.v Ho.HH NH.HH H: om.OOH HH. mo. ov.H HH.H an. om.~ em.OH GH.H HH.m Hm.o Hm.oH ca.m. OH mo.mm «H. mo. Hm.H Ho.~ on. m~.~. HH.oH HH.H HH.H am.” .~.mH OH.H¢ H He.ma HH. Ho. mo.H on.. on. oe.~ ~¢.oH mH.H am.o Hm.m H~.oH cc... «N mm.ma mH. mo. ~0.H Ho.m Nm. Hm.~ HH.OH «H.H NH.» Hm.~ Ho.oH Hm.ve H: ~m.mm mH. Ho. Hm.H HH.H HH. HH.~ mm.m am.o 02.. ”H.H m..OH ~H.¢. oN Hm.ma 4H. mo. mm.H H~.H HH.H HH.H Ha.» Hm.» mH.H om.: HH.9H HH.¢H OH mm.mm HH. mo. H..H NH.H om. ~m.~ HH.OH om.o .mH.o HH.H ~H.0H «a.me NH 3238 mo~a - No; 33 HH.OOH HN. HN. mm.~ ~m.n HN. mm.~ Hm.» Hm.m OH.m m~.m HH.HH :H.m¢ can HH.H¢H mH. mH. Hc.~ oe.~ HH. Hm.~ 0:.o Hm.m HH.¢ Ho.cH on.mH mo.ov H HO.O¢H mN. ”H. CH.N HH.. Hm. Hv.~ Hm.» ow.o om.m oo.o wm.mH No.m. can cm.DOH ON. HA. H~.N HH.H HH.H co.» NH.H oo.o mm.m Hm.m mm..H Ho.mv Hm oH.om «N. ”H. HH.N OH.e «H. HH.~ me.: 0H.H oc.HH H~.. oH.mH mH.H4 Hm Ho.HOH NN. HH. H~.~ OH.~ NH. ce.~ mm.m ~H.H om.a mm.¢ oc.oH 4:.cm uHm om.mm HN. NN. mN.N co.~ -.H OH.~ mo.oH cH.¢ Ho.oH HH.H -.HH Ho.m. HHH muHammm mama - NoHH_:mH= m N N N N H .oz HaooH on: o a oHH. Ho: 0 a o a: one on: and awn» Ho~H<_ Hon oHneam 3H3“: 2: mo 0.8.452 3668 3.8: H8“ 32: #3: mafia: N32... 62 9N.CON NN. Ne. Nm. NN.N NN. NN.N NN.NN NN.N NN.N om.N NN.oN Nm.mv 035 NN.ooN NN. No. mm. NN.N NN. oN.N NN.NN NN.N mN.m NN. ON.mN mo.NN Nah No.NON «N. No. 0N.N Nm.N NN. ON.N No.ON cm.m oo.o om.m om.mN No.N N35 moxNo ommnmNa NON; - NONE 3o; NN.OON NN. ON. NN.N NN.N NN. om.N ON.m CN.m NN.N mv.v oN.NN Nm.om . N24 NN.oa 0N. NN. mo.N NN.N NN.N Nm.N No.N No.v NN.NN NN.N NN.NN No.om . N24 Na.ooN NN. oN. Nm.N No.N NN. NN.N mo.N ow.o NN.NN NN.N NN.NN NN.mN Nah NN.NoN NN. NN. NN.N mm.N co. No.n Nm.m om.c NN.NN NN.N NN.NN me.mv N25 NN.NON NN. NN. NN.N NN.N Na. ON.N NN.N No.m mo.NN Nv.m No.NN NN.NN on No.0oN NN. mm. NN.N NN.N NN. cm.m NN.N NN.N NN.NN NN.N NN.NN mm.ov NNN NN.mm NN. NN. Nm.N Nm.N NN. No.m ON.N NN.m ov.NN NN.N ON.NN NN.Nm N25 NN.OON mN. NN. Nm.N oo.N NN. NN.N No.m Ne.o ON.NN NN.N NN.NN NN.NN CNN NN.OON NN. mm. om.N No.N NN. NN.N NN.N Nc.o mo.ON am.m NN.NN Nm.NN N25 NN.OON NN. NN. mm.N am.m NN.N Nm.N Nm.c cm.m NN.oN NN.N NN.NN Nm.Nm . NxaN NN.ma NN. om. NN.N ON.N oo.N NN.N NN.N Na.m mo.NN om.N Nm.NN NN.Nm N NNNN NN.NON NN. NN. NN.N NN.N NN. NN.N oN.N No.N NN.oN mm.. oo.NN Nm.mv . N2 moNNa emanmNo moNN - NONE ngz m N N N N N N m N N 02 Nmuoe as: o a ON» No: C x 9 N2 emu Agw‘ com o on o N< on «Naeam movie 33mg 9:. mo momfig £65me No.34 xuom 30:3 63 ABSOLUTE RARE EARTH ELIE-8325'? CONCIINTPATIQV (PPM) Sample La Ce Sm Eu Yb Lu No. High P20S - T102 Basalts 113 27.48 51.21 7.80 2.16 4.49 .81 91C 24.08 50.51 7.71 2.06 4.50 .79 33 19.24 44.15 6.70 2.02 3.42 .55 87 17.58 34.50 6.28 2.02 3.74 .57 390 17.00 36.33 6.31 1.74 3.33 .68 2 16.64 31.95 5.70 1.73 3.50 .45 39C 16.09 37.96 5.87 1.93 3.62 .71 Low T102 - P205 Basalts 12 18.05 21.86 4.35 1.47 2.66 .40 '16 16.25 22.08 4.28 1.35 2.62 .42 26 14.17 28.62 4.30 1.31 2.70 .56 42 13.81 27.42 4.45 1.49 2.76 .57 28A. 12.90 19.89 4.15 1.35 2.64 .45 4 11.11 28.42 4.13 1.44 2.48 .33 70 10.99 25.24 3.94 1.34 2.13 .43 43 10.70 2 .41 4.15 1.31 2.29 .51 120 10.95 25.42 4.71 1.37 2.61 .36 63 10.66 21.29 4.13 1.37 2.66 .48 64 9.45 24.40 3.79 1.31 2.02 .33 64 ABSOUUTE RARE EARTH ELEMENT CONCENTRATIONS (PPM) Sample La Ce Sm Eu Yb No. High T102 - P 0 Diabase Dikes 2 5 WL* 25.40 48.00 7.13 2.78 3.58 .69 ZPK1* 24.82 44.00 7.28 2.25 3.48 .68 ZPK2* 24.80 48.00 7.74 2.41 3.34 .58 JWZ 21.43 40.77 7.77 2.24 4.01 .71 P10 22.21 47.95 8.19 2.44 4.24 .65 JW1 21.69 47.78 7.52 2.41 3.12 .59 P12 20.67 43.52 8.39 2.66 3.56 .65 PG 20.35 41.66 7.31 2.23 3.45 .54 JW4 15.51 32.72 5.67 1.77 4.24 .70 JW7 13.98 27.51 5.26 1.82 2.76 .62 LM2* 13.50 40.00 6.30 2.13 3.54 .67 HIP“ 9.97 28.00 8.39 1.74 2.40 .56 Low T102 - P205 Diabase Dikes .fiflB 11.89 20.51 3.19 1.43 1.12 .35 JW3 6.26 13.21 2.28 .83 1.61 .44 JW6 2.40 6.05 1.19 .44 .90 .15 *Afterlerris (1977). REFERENCES REFERENCES Abby, S., 1978. Calibration standard. X-ray Spectrometry, 2, 99-121. Anderson, A.T. , and L.P. Greenland, 1969. Phosphorus fractionation diagram as a quantitative indicator of crystallization differ— entiation of basaltic liquids. Geochim. Cosmochim. Acta, 33, 493-505. Annells, R.N., 1973. Proterozoic flood basalts of Eastern Lake Superior: The Keweenawan volcanic rocks of the Mamainse Point area, Ontario. Geol. Survey Canada Special Paper, 72, 51p. , 1974. Keweenawan volcanic rocks of Michipicoten island, Iake Superior, Ontario: an eruptive centre of Proterozoic age. Geol. Survey Canada Bull., 218, 141p. Blanchard, D.P., J.M. Rhodes, M.A. Dungan, K.V. Rodgers, C.H. Davidson, J.C. Brannon, J.W. Jacobs and E.K. Gibson, 1976. The chemistry and petrology of basalts from Leg 36 of the Deep Sea Drilling Project. J. Geophys. Res., 81, 4231-4226. Books, K.G., 1968. Magnetization of lowermost Keweenawan lava flows in the Lake Superior area. U.S. Geol. Survey, Prof. Paper 600-D, 248-254. , 1972. Paleomagnetism of some Lake Superior Keweenawan rocks. U.S. Geol. Survey, Prof. Paper 760, 42p. Bryan, W.B., G. Thompson, F.A. Frey and J.S. Dickey, 1976. Inferred geologic settings and differentiation in basalts from the Deep-Sea Drilling Project. J. Geophys. Res, 81, 4285-4304. Burke, K., and J .F. Dewey, 1973. Plume generated triple junctions: Key indicators in applying plate tectonics to old rocks. Jour. Geology, 81, 406-433. Cann, J.R., 1970. Rb, Sr, Y, Zr, Nb in some ocean floor basaltic rocks. Earth Planet, Sci. Lett., 10, 7-11. Carter, J. , 1970. Mineralogy and chemistry of the earth's upper mantle based on the partial fusion - partial crystallization model. Geol. Soc. America Bull., 81, 2021—2034. 65 66 Case, J .E. and J .E. Gair, 1965. Aeromagnetic map of part of Marquette, Dickinson, Baraga, Alger, and Schoolcraft counties, Michigan and its geologic interpretation. U.S. Geol. Survey Gelphys. Inv. Map Gp-467. Chase, G.G., and T.H. Gilmer, 1973. Precambrian plate tectonics: the Mid-continent Gravity High. Earth Planet. Sci. Lett. , 21, 70-78. Chazen, 8.1., and T.A. Vogel, 1974. Distribution of Ti and P in oceanic basalts as a test of origin. Contr. Mineral. Petrol., 43, 307-316. Cornwall, H.R., 1951. Differentiation in lavas of the Keweenawan series and the origin of the copper deposits of Michigan. Geol. Soc. America Bull., 62, 159-202. Drake, M.J., and D.E. Weill, 1975. The partition of Sr, Ba, Ca, Y, 5112", Eu3+, and other REE between plagioclase feldspar and magmatic silicate liquid: an experimental study. Geochim. Cosmochim. Acta, 39, 689-712. ' Engel, A.E., C.G. Engel, and R.G. Havens, 1965. Chemical character- istics of oceanic basalts and the upper mantle. Geol. Soc. Amer. Bull., 76, 719-734. Fabbi, B.F., 1971. Rapid x-ray fluorescence determination of Phosphorus in geologic samples. Appl. Spec-rosc., 25, 41-43. , 1972. A refined fusion x-ray fluorescence technique and determination of major and minor elements in silicate stan- dards. Amer. Mineral., 57, 237-245. Fleet, A.J., P. Henderson, and D.R.C. Kempe, 1976. Rare earth element and related chemistry of some drilled Southern Indian Ocean basalts and volcanogenic sediments. J. GeOphys. Res., 81, 4257-4268. Floyd, P.A., J .A. Winchester, 1977. Identification and discrimination of altered and metamorphosed volcanic rocks using immobile elements. Chem. Geol., 21, 291-306. Frey, P.A., M.A. Haskin, J. Poetz and L.A. Haskin, 1968. Rare earth abundances in some basic rocks. J. Geophys. Res., 73, 6085- 6098. , W. B. Bryan, and G. Thompson, 1974. Atlantic ocean floor: geochemistry and petrology of basalts from Leg 2 and 3 or the Deep Sea Drilling Project. J. Geophys. Res. , 79, 5507-5527. 67 , D.H. Green, and S.D. Roy, 1978. Integrated models of basalt petrogenesis: a study of quartz tholeiites to olivine melilitites from South Eastern Australia utilizing geochemical and experimental petrological data. Jour. Petrology, 19, 463- 513. Gast, P.W., 1968.1 Trace element fractionation and the origin of tholeiitic and alkaline magma types. Geochim. Cosmochim. Acta, 32, 1057-1086. Gordon, G.B., R. Rondle, G.G. Goles, J.B. Corliss, H.H. Beeson, and 8.8. Oxley, 1968. Instrumental activation analysis of standard rocks with high resolution x-ray detectors. Geochim. Cosmochim. Acta, 32, 369-396. Green, D.H., and A.E. Ringwood, 1967. The genesis of basalts magmas, Contr. Mineral. Petrol., 15, 103-190. , 1970. The origin or basaltic and nephelinitic magmas. Trans. Leichester. Lit. Soc., 64, 28-54. , A.D. Edgar, P. Beasley, E.Kiss, and N.G. Ware, 1974. Upper mantle source for some hawaiites , mugearites and ben- morites. Contr. Mineral. Petrol., 48, 33-43. Green, J .C., 1972. General geology Northern Minnesota, in Sims, P.K., and Morey, G.B. (eds.), Geology of Minnesota: a centermial volume, Minn. Geol. Survey, 299-303. , 1977. Keweenawan plateau volcanism in the Lake . Superior region. Geol. Association Canada Special Paper 16, 408-422. Grutzeck, M. ,, S. Kridelbaugh, and D. Weill, 1974. The distribution of Sr and REE between diopside and silicate liquid. Geophys. Res. Lett., 1, 273-275. ‘ Hart, R. 1970. Chemical exchange between sea water and deep ocean basalts. Earth Planet. Sci. Lett., 9, 269-279. Haskin, L.A., F.A. Frey, R.A. Schmitt, and R.H. Smith, 1966. Meteoric, solar and terrestrial rare earth distributions. Phys. (hem. Earth, 7, 167-321. , M.A., Haskin, R.A. Frey, and T.R. Wilderman, 1968. Relative and absolute terrestrial abundances, Pergamon Press, New York, 889-912. 68 Hellman, P.L., R.E. Smith, and P. Anderson, 1977. Rare earth element investigation of the Chiefden outcrop, N .S.W. , Australia. Contri. Mineral. Petrol., 65, 155-164. Herrman, A.G., M.J. Potts, and D. Knake, 1974. Geochemistry of the rare earth elements in spilites from oceanic and continental crust. Contr. Mineral. Petrol., 44, 1-16. Hertogen, J. , and R. Gubels, 1976. Calculation of trace element fractionation during partial melting. Geochim. Cosmochim. Acta, 40, 313-322. Hinze, W.J., R.F. Roy, and D.M. Davidson, 1972. The origin of late Precambrian rift (abs). Geol. Soc. America Abstracts with Programs, 4, 723. Hubbard, J .A. , 1975. Lower Keweenawan volcanic rocks of Michigan and Wisconsin. U.S. Geol. Survey, Jour. Res., 3. 529-541. Huber, R.N. , 1973. The Portage Lake Volcanics (Middle Keweenawan) on Isle Royale, Michigan. U.S. Geol. Survey Prof. Paper 754C, 32 p. Humphris, S.E., M.A. Morrison, and R.N. Thompson, 1978. Influence of rock crystallization history upon subsequent lanthanide mobility during hydrothermal alteration of basalts. Chem. Geol. 23, 125-237. Irvine, T.N., and W.R.A. Baraga, 1971. A guide to the chemical classification of the conmon volcanic rocks. Canadian Jour. Earth Sci., 8, 523-548. Irving, R.D., 1883. The copper-bearing rocks of Lake Superior. U.S. Geol. Survey Mon. 5, 464p. Jolly, W.T., and R.E. Smith, 1972. Degradation (hydration) - aggrada- tion (dehydration) and low rank metamorphism of mafic volcanic sequence. Rep. 24th Int. Geol. Cong., Section 2, 11-18. , and R.E. Smith, 1974. Degradation and metamorphic differentiation of the Keweenawan tholeiite lavas of Northern Michigan, U.S.A. Jour. Petrology, 13, 273-309. Kay, R.N. and P.W. Gast, 1973. The rare earth content and origin of alkali-rich basalts. Jour. Geol., 81, 653-682. Kay, R.N., H.J. Hubbard, and P.W. Gast, 1970. Chemical characteris- tics and origin of oceanic ridge volcanic rocks. J. Geophys. Res., 75, 1585-1613. 69 King, E.R., and I. Zietz, 1971. Aeromagnetic study of the Mid- continent Gravity High of the Central U.S. Geol. Soc. America Bull., 82, 2187-2207. Kushiro, I., 1968. Compositions of magma formed by partial zone melting of the earth's upper mantle. J. GeOphys. Res. , 73, 619-634. , 1969. The system forsterite-diopside-silica with and without water at high pressures. Am. Jour. Sci., 11, 405-430. , 1972. Partial melting of synthetic and natural periodotites at high pressures. Carnegie Inst. Washington Yearbk., 71, 357-362. Lane, A.G., 1911. The Keweenawan series of Michigan. Michigan Leeman, Geol. and Biol. Survey, pub. 6, Geol. Sur. 4, 2 vols., 983 p. W.P. , 1976. Petrogenesis of Mckinney (Snake River) olivine tholeiite in light of rare earth element and Cr/Ni distribu- tions. Geol. Soc. American Bull., 87, 1582-1586. , 1977. Comparison of Rb/Sr, U/Po, and rare earth Morris , Hraracteristics of subcontinental and suboceanic mantle regions in Dick, J .3. (ed.), Magma genesis, Proceeding of the Amer. Geophys. Union Chapman Conference on partial melting in the earth's upper mantle. State of Oregon, Dept. of Geology and Mineral Industries, 96, 149-168. W.J. , 1977. Geochemistry and origin of the Yellow Dog Plains peridotite, Marquette county, Northern Michigan. M.S. Thesis, Michigan State University, 75 p. , J .T. Wilband, and T.A. Vogel, 1978. Genetic rela- Nathan , tionships of a Keweenawan peridotite and diabase dikes in Marquette county, Michigan. Geol. Soc. American Abstracts with Programs, 10, 279. S. and J .S. Fruchter, 1974. Geochemical and paleomagnetic stratigraphy of the Picture Gorge and Yakima basalts (Colunbia River Group) in central Oregon. Geol. Soc. Amer. Bull., 85, 63-76. Onuna, N., H. Higuch, H. Wakita, and H. Nagasawa, 1968. Trace ele- ment partition between two pyroxenes and the host lava. Earth Planet. Sci. Lett., 5, 47-51. Pankhurst, R.I. , 1977. Open system crystal fractionation and incom- patible element variation in basalts. Nature, 268, 36-38. 70 Petro, W.L., T.A. Vogel and J.T. Wilband, 1979. Major-element chemistry of plutonic rock suites from compressional and extensional plate boundaries. Chem. Geol. , 25, No. 4. (in press). Philpotts, J .A., and C.C. Schnetzler, 1968. Europium anomalies and the genesis of basalts. Chem. Geol., 3, 5-13. Roeder, P.L., and R.F. Emslie, 1970. Olivine-liquid equilibrium. Contr. Mineral. Petrol., 29, 275-289. . Schafer, H.N.S., 1966. The determination of iron (11) oxide in silicate and refractory material - part I , a review. The Analyst, 91, 755-762. Schnetzler, C.C., and J .A. Philpotts, 1970. Partition coefficients of RE between igneous matrix material and rock forming mineral phenocrysts--II. Geochim. Cosmochim. Acta, 34, 331-340. Shaw, D.M. , 1970. Trace element fractionation during anatexis. Geochim. Cosmochim. Acta, 34, 331-340. Shilling, J .G., 1971. Sea-floor evolutions: rare earth evidence. Phil. Trans. Roy. Soc. London, Ser. 4, 268, 663-706. , 1973a. Iceland mantle plume: geochemical study of Reykjanes ridges. Nature, 242, 565-571. , 1973b. Iceland mantle plume, Nature, 246, 141-143. , 1975. Rare earth variations across 'normal segments' of the Reykjanes ridge, 60°-53° N, Mid-Atlantic ridge, 29°S, and East Pacific rise, 2°-19°S, and evidence on the composi- tion of the underlying low velocity layer. J. Geophys. Res. , 80, 1459-1473. , and E. Bonatti, 1975. East Pacific ridge (2°S-19°S) versus Nazca intraplate volcanism: rare earth evidence . Earth Planet. Sci. Lett., 25, 93-102. Stoiber, R.E., and E.S. Davidson, 1959. Amygdule mineral zoning in the Portage Lake Lava Series, Michigan capper district. Econ. Geol., 54, 1440-1460. Sukheswala, R.N. and A. Poldervaart, 1958. Deccan basalts of the Bonbay area, India. Geol. Soc. Amer. Bull., 69, 1475-1494. Sm, 8.8., and R.W. Nesbitt, 1978. Petrogenesis of Archean ultrabasic and basic volcanics, evidence from rare earth elements. Contr. Mineral. Petrol., 65, 301-325. 71 Thompson, R.N., 1975. Primary basalts and magma genesis. Contr. Mineral. Petrol., 52, 213-232. Weiblen, P.W., and G.B. Morey, 1975. The Duluth complex - a petrologic and tectonic sumnary: Proc. 485h An. Mtg. , Minn. Sec. Am. Inst. Mining Metall. Petroleum Engineers, 72-75. Weis, L.A., 1974. A petrochemical investigation of the Winona Quadrangle basalts, Northern Michigan. M.S. Thesis, Univ. of Toledo, 82p. White, W.S., 1972. Keweenawan flood basalts and continental rifting (abs.). Geol. Soc. America Abstracts with Programs, 4, 732-734. Wood, D.A., L.L. Gibson, and R.N. Thompson, 1976. Elemental mobility during zeolite facies metamorphism of ter- tiary basalts of Eastern Iceland. Contr. Mineral. Petrol., 55, 241-254. Wood, W.W. , 1962. Distribution and stratigraphic position of late Precambrian diabase dikes in part of Northern Michigan. M.S. Thesis, Michigan State University, 57p. MICHIGAN STATE UNIV. LIBRARIES ll 1‘I“mlWWIINIIHWIWIHlIHIIIIHIH‘IUIWI 31293100626492