Petrology and Geochemistry of the Nipissing Gabbro: Exploration Strategies for Nickel, Copper, and Platinum Group Elements in a Large Igneous Province

Ontario Geological Survey Study 58

1996

Ontario Geological Survey Study 58 by P.C. Lightfoot and A.J. Naldrett

1996 Queen’s Printer for Ontario, 1996 ISSN 0704-2590 ISBN 0-778-4804-X

All publications of the Ontario Geological Survey and the Ministry of Northern Development and Mines are available for viewing and purchase at the following locations: Mines and Minerals Information Centre (MMIC) Macdonald Block, Room M2-17 900 Bay Street Toronto, Ontario M7A 1C3 Telephone: 1-800-665-4480 (within Ontario) (416) 314-3800 Fax: (416) 314-3797

Publication Sales 933 Ramsey Lake Road, Level B2 Sudbury, Ontario P3E 6B5 Telephone: (705) 670-5691 Fax: (705) 670-5770 E-mail: [email protected] Use of Visa or Mastercard ensures the fastest possible service. Cheques or money orders should be made payable to the Minister of Finance.

Canadian Cataloguing in Publication Data Lightfoot, Peter C. (Peter Charles) Petrology and Geochemistry of the Nipissing gabbro: exploration strategies for nickel, copper, and platinum group elements in a large igneous province (Ontario Geological Survey report, ISSN 0704-2590; 58) Includes bibliographical references. ISBN 0-7778-4804-X 1. Gabbro---Ontario---Nipissing Region. I. Naldrett, A.J. II. Ontario Ministry of Northern Development and Mines III. Ontario Geological Survey. IV. Series. QE462.G3L53 1995 552’.3 C95-964107-6

Every possible effort is made to ensure the accuracy of the information contained in this report, but the Ministry of Northern Development and Mines does not assume any liability for errors that may occur. Source references are included in the report and users may wish to verify critical information. If you wish to reproduce any of the text, tables or illustrations in this report, please write for permission to the Manager, Publication Services Section, Ministry of Northern Development and Mines, 933 Ramsey Lake Road, Level B4, Sudbury, Ontario P3E 6B5. Cette publication est disponible en anglais seulement.

Parts of this report may be quoted if credit is given. It is recommended that reference be made in the following form: Lightfoot, P.C. and Naldrett, A.J., 1996. Petrology and geochemistry of the Nipissing Gabbro: Exploration strategies for nickel, copper, and platinum group elements in a large igneous province; Ontario Geological Survey, Study 58, 81p. Critical Reader: J.A. Fyon Editor: T. Ayalew

ii Contents

Objective and Approach of the Present Study...... 3 Acknowledgements ...... 6 Introduction and General Geology...... 6 Empirical Metallogeny of the Nipissing Gabbro...... 9 Sampling and Analysis ...... 11 Effects of the Textural Variations in the Gabbro on the Analytical Data...... 11 Geochemical Evidence for the Emplacement and In-situ Differentiation of Nipissing Intrusions.... 12 Duration of Nipissing Magmatic Activity and Associated Compositional Variation...... 24 Compositional Variation in the Parental Nipissing Magma Type...... 26 Empirical Observations Related to Mineral Potential, Land Use Planning and Exploration...... 37 Further Work ...... 48 Conclusions ...... 48 Appendix 1: Sampling, Analysis, Geology, Petrography and Mineralogy of the Nipissing Intrusions..... 49 1.1 Sampling and Analysis ...... 49 1.2 Age and Distribution: ...... 49 1.2.1 Differentiation ...... 49 1.2.2 Thickness ...... 49 1.2.3 Emplacement sites ...... 49 1.2.4 Deformation ...... 50 1.3.1 Geology of the Nipissing sills and dykes around Lake Temagami...... 50 1.4 Geology of the Kerns Sill...... 50 1.5 Petrography and Mineralogy of the Nipissing Gabbro...... 56 1.6 Metamorphism of the Nipissing Gabbro...... 65 Appendix 2 ...... 66 2.1 Assimilation and Fractional Crystallisation in the Kerns Intrusion - a Case Study of the Physical Process ...... 66 References ...... 76 Conversion Factors for Measurements in Ontario Geological Survey Publications...... 80 FIGURES 1a. Distribution of Nipissing gabbro across the Southern Province...... 4 1b. Sketch diagram showing the relationship between the petrology of the undulatory sills and the associated mineralisation ...... 5 1c. Typical sequence of lithologies seen in a well-differentiated intrusion of Nipissing gabbro showing the relationship between silicate rocks and sulphide mineralization...... 6 1d. Location of samples and results from U-Pb geochronology work...... 7 2. Regional tectonic setting of the Nipissing gabbro and the location of the 2.2 Ga Preissac Dyke Swarm ...... 7 3a. Ni versus forsterite relationships in olivines from the Cross Lake Sill...... 12 3b. PGE distribution patterns for the Rathbun Lake showing (after Lightfoot et al., 1993 and Rowell and Edgar, 1986) ...... 12 4. Geological relationships between Nipissing intrusions, mineralisation, and geophysical anomalies, Paleoproterozoic mineralisation, and the Sudbury Igneous Complex...... 13 5a. Comparison of primitive mantle normalised spidergrams on vari-textured gabbronorite patches and gabbronorite host; Emerald Lake gabbro...... 18 5b. Comparison of primitive mantle normalised spidergrams on vari-textured gabbronorite patches and gabbronorite host; Basswood Lake Intrusion...... 19

iii 6. Location of individual intrusive bodies and detailed study areas referenced in this report...... 20 7a. Chemostratigraphy of the High Rock Intrusion, Lake Temagami...... 21 7b. Primitive mantle normalised spidergrams of the High Rock Intrusion samples, Lake Temagami...... 21 7c. Primitive mantle normalised spidergrams of the High Rock Intrusion samples...... 22 8. Geochemical stratigraphy of the Miller Lake Intrusion, Gowganda...... 22 9a. Primitive-mantle normalised trace element spidergrams for representative samples from the Kerns Intrusion, northwest of New Liskeard Chilled basal quartz diabase, quartz diabase ...... 23 9b. Primitive-mantle normalised trace element spidergrams for representative samples from the Kerns Intrusion, northwest of New Liskeard Hypersthene gabbro...... 24 9c. Primitive-mantle normalised trace element spidergrams for representative samples from the Kerns Intrusion, northwest of New Liskeard Vari--textured gabbro...... 25 9d. Primitive-mantle normalised trace element spidergrams for representative samples from the Kerns Intrusion, northwest of New Liskeard Granophyric gabbro...... 26 9e. Primitive-mantle normalised trace element spidergrams for representative samples from the Kerns Intrusion, northwest of New Liskeard Aplites...... 27 9f. Primitive-mantle normalised trace element spidergrams for representative samples from the Kerns Intrusion, northwest of New Liskeard Hornfelsed sediment rafts within the granophyric gabbro, and roof sediments. See text and Appendix 1 and Lightfoot et al. (1987) for detailed sample locations and descriptions...... 28 10. Geochemical variations in samples from the Kerns Intrusion, where elemental and oxide abundances are plotted against Zr concentration...... 29 11. a) VariationinThversusNb,b) Variation in Cu versus Zr in Nipissing gabbros, c) Variation

in Cu/Zr versus SiO2 for all Nipissing gabbro samples...... 35 12. a) Variation in 143Nd/144Nd versus 147Sm/144Nd in samples from the Kerns Intrusion and local country rocks. b) Relationship of the array of the Kerns Intrusion to isochron lines based on U - Pb geochronology for magmas with a range in initial 143Nd/144Nd isotopic composition...... 37 13. A model for the evolution of the Kerns Intrusion...... 38 14a. Primitive mantle normalised spiderdiagrams. Representative samples from the Narrows Island gabbronorite dyke ...... 38 14b. Primitive mantle normalised spiderdiagrams. Representative samples from the Sand Point gabbronite dyke ...... 39 14c. Primitive mantle normalised spiderdiagrams. Representative samples from the Sand Point and Narrows Island dykes with the overlying undulatory sills...... 39 15a. Primitive mantle-normalised compositions of aplites from the Obabika Intrusion, Lake Temagami Gowganda Formation sediments...... 40 15b. Primitive mantle-normalised compositions of aplites from the Obabika Intrusion, Lake Temagami Aplitic granitoids ...... 40 16a. Primitive mantle normalised compositions of differentiated granodiorites and quartz diorites of the Obabika Intrusion that contain disseminated sulphide. Chilled diabase and quartz diabase ...... 41 16b. Primitive mantle normalised compositions of differentiated granodiorites and quartz diorites of the Obabika Intrusion that contain disseminated sulphide. Quartz diorite from the roof of the intrusion...... 41 17a. Primitive mantle normalised spidergrams demonstrating the similarities in geochemical signatures of samples collected from sites retaining three different palaeomagnetic remanence directions in Nipissing Intrusions. N1 samples...... 42 17b. Primitive mantle normalised spidergrams demonstrating the similarities in geochemical signatures of samples collected from sites retaining three different palaeomagnetic remanence directions in Nipissing Intrusions. N2 samples...... 42

iv 17c. Primitive mantle normalised spidergrams demonstrating the similarities in geochemical signatures of samples collected from sites retaining three different palaeomagnetic remanence directions in Nipissing Intrusions. N3 samples...... 43 17d. Primitive mantle normalised spidergrams demonstrating the similarities in geochemical signatures of samples collected from sites retaining three different palaeomagnetic remanence directions in Nipissing Intrusions...... 43 A1.1 Geological map of the Nipissing Gabbro exposed around Lake Temagami...... 51 A1.2 Geological map of the Narrows Island dyke, Lake Temagami...... 52 A1.3 Geological map of the Sand Point dyke, Lake Temagami...... 53 A1.4 Detailed geological map of the south shore of Obabika Inlet, Lake Temagami...... 54 A1.5 Detailed geological map of the Kerns Intrusion, Kerns Township...... 57 A1.6 Detailed geological map of Kerns Rock, Kerns Intrusion...... 58 A1.7 Sketch map showing relationships between lithologies at the roof of the Kerns Intrusion...... 59 A1.8a. Sketch map showing the locations of samples referred to in MRD 19. Englehart Intrusion (based on regional compilation maps)...... 60 A1.8b. Sketch map showing the locations of samples referred to in MRD 19. Bruce Mines (after Lightfoot et al., 1993) ...... 60 A1.8c. Sketch map showing the locations of samples referred to in MRD 19. Basswood Lake Intrusion (after Lightfoot et al., 1993) ...... 61 A1.8d. Sketch map showing the locations of samples referred to in MRD 19. Wanapitei Intrusion (after Lightfoot et al., 1993) ...... 62 A1.8e. Sketch map showing the locations of samples referred to in MRD 19. Cobalt Region Intrusion (after Lightfoot et al., 1993) ...... 63 A2.1 Modelling of the crystallisation and assimilation history of the Kerns Intrusion...... 68 A2.2 a) Variation in Th/Yb versus La/Yb in the Kerns Intrusion b) Variation in U/Yb versus Th/Yb...... 69 A2.3 Effects of assimilation and fractionation compared to mixing on schematic showing bi- variate plots of incompatible elements...... 70 A2.4 a) Modelling of assimilation coupled to fractionation on Th versus Zr. b) Modelling of assimilation coupled to fractionation on La versus Zr, c) Modelling of assimilation coupled to fractionation on U versus Zr...... 71 A2.5 Model for the evolution of a Nipissing intrusion...... 74 PHOTOS 1. Sulphide globules in gabbro from the Wanapitei Intrusion...... 14 1.1 Hornfelsed sediment fragments (Gowganda formation) in quartz diorite at the roof of the Obabika Intrusion ...... 55 1.2. Hornfelsed sediment inclusion (Gowganda Formation) in granodiorite in the roof zone of the Obabika Intrusion ...... 55 1.3. Vein of aplitic granitoid originating in a domain of Gowganda Formation sedimentary rock inclusions in the roof granodiorite of the Kerns Intrusion...... 56 1.4. Banding in the spotted hornfelsed sediment rafts in the Nipissing granophyric gabbro...... 63 1.5. A breccia consisting of fragmented Lorrain Formation sediments within an aplite at the roof of the Kerns Intrusion ...... 64 1.6 Chilled Nipissing diabase at contact of High Rock intrusion...... 64 TABLES 1. Summary of empirical characteristics of the metallogenetic associations of the Nipissing gabbro intrusions ...... 10

v 2. Summary of grade of mineralisation at Kukagami Lake and the Rathbun Lake Showing...... 15 3. Analytical data for in-house standard reference materials UTB-1 (University of Toronto basalt standard) and WHIN SILL (Open University diabase standard)...... 16 4. Miscellaneous Release Data (MRD) 19 (Digital data on diskette): Analytical data for Nipissing intrusions (available separately) 5. Nd isotope data for samples of Nipissing gabbro. Analyses were performed at the University of Toronto using a clean laboratory and thermal ionisation mass spectrometer as documented in the text ...... 36 6. Compositional averages of samples from relatively undifferentiated parts of twenty one different Nipissing gabbro intrusions...... 44 7. Compositional averages for evolved aplitic granitoids of the Kerns and Obabika intrusions...... 47

vi Abstract

Close spatial associations between magmatic and hydrothermal mineralisation and Nipissing gabbro intrusions have been recognised, and there is a variation in style and type of mineralisation across the Nipissing magmatic province (Card and Pattison, 1973). In the Nipissing gabbro of the central portion of the Southern Province in Ontario, mineralisation is dominantly in the form of magmatic and/or hydrothermal Cu-Ni-platinum group element (PGE) sulphides which occur disseminated within the intrusions or as massive pods beneath the intrusions. A number of empirical observations can be made regarding the metallogeny of the Nipissing gabbro. The goal of this study is to more thoroughly document magmatic components of the Nipissing gabbro found in mineralised, weekly mineralised, and apparently unmineralised intrusions and to focus attention on the metallogeny. A number of important points are highlighted, and new data are presented for Nipissing gabbros from the Lake Temagami Region, Ontario: 1. Magmatic Ni, Cu, and PGE mineralisation is spatially associated with intrusions which lie on a trend between Whitefish Falls, Sudbury, and River Valley. The mineralisation is associated with a significant regional gravity and aeromagnetic high along the trend of which and a number of mineralised Paleoproterozoic intrusions were emplaced, and the Sudbury Igneous Complex was formed. 2. Mineralised Nipissing gabbro intrusions with significant quantities of disseminated sulphide are located southwest of Sudbury between Whitefish Falls and Emerald Lake - the Casson Lake Complex, southwest of Sudbury (the “”Sudbury Gabbro” intrusions of Nairn, Lorne, Denison, Waters, Hyman and Drury Townships), east of Sudbury (the Wanapetei Intrusion), and continuing to the east-northeast in Kelly and Janes Townships. The sulphides occur as fine disseminations of magmatic pyrrhotite (50-75%) with lesser chalcopyrite and pentlandite. Some of the sulphides exhibit magmatic blebby textures with pyrrhotite-rich lower segments and chalcopyrite-rich upper segments suggesting in-situ differentiation of the immiscible sulphide liquid during cooling and crystallisation. Massive sulphides rich in copper and platinum group elements (PGE) are known as basal concentrations associated with the Wanapetei Intrusion. Despite this being a small occurrence, the sulphides carry 1-15 weight % Cu, 2.5- 6.3ppm Pt, 17-53ppm Pd, and 1-6ppm Au. In general, the disseminated sulphides (<5% modal sulphide) tend to be focussed in the interior of the sills (100- 300m above the base) with coarse gabbronorites and hypersthene-rich gabbros; these rocks carry 200-1100ppb Pt and 50-4000ppb Pd in mineralised intrusions in Janes Township. 3. The silicate host of the sulphides tends to be relatively undifferentiated, although there is some petrographic indication that there may be cyclical trends in composition (Conrod, 1989). Intrusions which are heavily contaminated in their roof zones such as the Kerns Intrusion northwest of New Liskeard and the Basswood Lake Intrusion north of Thessalon appear to be relatively unmineralised, and the low Cu/Yb and Cu/Zr of the most contaminated granophyric rocks is attributed to large amounts of assimilation of sediment with very low Cu/Yb and Cu/Zr rather than the fractionation of an immiscible sulphide liquid. 4. The mineralised intrusions tend to be those which have a basinal shape and are least strongly differentiated into granophyric zones, yet consist of gabbros which contain a higher modal proportion of hypersthene (20-40 modal %) than unmineralised intrusions (<20 modal %). As a result of the hypersthene content, the mineralised intrusions tend to be more mafic as reflected in their elevated MgO (10-14 weight %) and low incompatible element concentrations (e.g. 0.2-0.4 weight % TiO2, <50ppm Zr). The localisation of the disseminated sulphides inside the sills suggests that either the sulphides were related to late emplacement of magmas, or that the differentiation of the sills triggered sulphur saturation only after the initial crystallisation of large amounts of hypersthene. Although there are no olivine cumulates of the type seen in the Noril’sk Intrusions in Russia (Naldrett et al., 1992), there are intrusions with elevated MgO values (approaching 14 weight %), and these rocks are hypersthene-rich cumulates. 5. Existing geochemical data suggest that there is no obvious difference in the magma type of the mineralised intrusions when compared to the unmineralised intrusions. For example, the undifferentiated quartz diabase, gabbros, and chills have a narrow range in incompatible element ratios such as La/Sm and Th/Y, and similar abundance levels. This suggests that many of the intrusions were derived from the same source, that they did not differentially mix or become contaminated by crustal reservoirs en route to the surface, and were emplaced at roughly the same degree of differentiation through the Nipissing event. The available evidence also suggests that the Cu/Zr ratios of gabbros with non-detectable S are not significantly different when compared to the mineralised intrusions. This suggests that there has been little differential depletion of intrusions in copper by the segregation of immiscible sulphide in the intrusions studied. It is also evident that fresh olivine from an unmineralised intrusion are characterised by elevated nickel (1500-1700ppm) contents in olivines with forsterite contents of Fo60-70 which compares to 1000-1500ppm in Fo60-65 olivines which have crystallised from magmas which have not equilibrated with sulphide. Importantly, disseminated and massive sulphide is observed associated with some Nipissing intrusions, and an important challenge awaiting explorationists is to determine whether the olivines of these intrusions are Ni-depleted, and whether any of the silicates can be identified to show nickel and copper depletion. 6. The homogeneity of the chilled margins and undifferentiated quartz diabase and gabbro intrusions suggests a single source for these magmas where there has not been significant differential interaction of the magma with a crustal

vii reservoir other than at the final site of crystallisation. However, the parental magma type is characterised by strong light rare earth element (LREE) and large ion lithophile element (LILE) enrichment as well as marked negative Ta+Nb, TiO2, and P2O5 anomalies. These are all features of a magma that has either interacted with a crustal reservoir (by contamination in a continental or arc environment), or they are features of young rocks typically attributed to derivation from ancient regions of mantle lithosphere which contain recycled continental crust. This geochemical signature is also one found in many of the Upper Sequence lavas of the Siberian Trap at Noril’sk, but is not quite as extreme as that found in the highly contaminated Nadezhdinsky lavas which have low Cu, Ni, and PGE abundances (Lightfoot et al., 1994; Naldrett et al., 1992). 7. The presence of disseminated magmatic sulphides towards the centre of the undulatory sills suggest that either: 1) the magmas were initially injected free of sulphide, and that subsequent magmas were injected as sulphur-saturated magmas, or 2) that the differentiating liquid became sulphur saturated only after it had time to fractionate and/or mix with a newpulse of more primitive liquid. Further insight may well be gained from recent modelling of the Fox River Sill, where recent studies of the geochemical variations show that the vertical distribution of Ni, Cu, and PGE can be linked to the replenishment of a magma chamber by a gabbroic magma (Naldrett et al., in press.). The controlling effect of gravitational settling appears to have produced only one example of sulphide mineralisation linked to the base of the intrusion at Rathbun Lake, and some authors consider this deposit to have a hydrothermal origin, or possibly even be related to the Sudbury Structure. Our new data suggest that sulphur saturation was achieved after the crystallisation of the magma had commenced, and the semi-consolidated hypersthene gabbro cumulates presumably prevented the gravitational settling of the immiscible sulphide liquid to the base of the intrusions. This, in-turn, expands the range of mineral exploration targets to the interiors of the Nipissing Intrusions rather than just the basal contacts.

viii Petrology and Geochemistry of the Nipissing Gabbro: Exploration Strategies for Nickel, Copper, and Platinum Group Elements in a Large Igneous Province

P.C. Lightfoot Geologist, Mineral Deposits and Feild Services Section, Ontario Geological Survey, 933 Ramsey Lake Road, Sudbury, Ontario, P3E 6B5 A.J. Naldrett Department of Geology, University of Toronto, 22 Russell Street, Toronto, Ontario M5S 3B1

Objective and Approach of the Present Study

Recent studies of the Continental Flood Basalts (CFB) at The Nipissing intrusions are located within the Noril’sk in Russia suggest that the intrusions which carry Huronian sedimentary sequence at the margin of the world- class deposits of Ni, Cu, and PGE (>555*106 tonnes Superior Province, and the focus of the 2.2Ga Preissac of sulphide grading 2.7 weight % Ni; Naldrett and dyke swarm centers on the Nipissing Province (Figure 1). Lightfoot, 1993) are the intrusive conduits of the magmas The intrusions consist dominantly of gabbros with lesser giving rise to the Permo- Triassic flood basalt sequence diabase and granophyre, and are collectively termed (e.g. Naldrett et al., 1992; Naldrett et al., 1995). Likewise, Nipissing Gabbro Intrusions of the Nipissing Magmatic recent studies of the Karoo diabase sills of Southern Africa Province. The Nipissing intrusions are known to host or be confirm a comagmatic relationship with the extrusive associated with small but significant showings of Ni, Cu, Lesotho CFB sequence (Marsh and Eales, 1984; Lightfoot and PGE which have been actively explored. Importantly, and Naldrett, 1984), and some of the differentiated many of the mineralised Nipissing intrusions are located in Karoo-aged intrusions host significant showings of Ni, Cu, the Sudbury Region close to the giant Sudbury Ni, Cu, and and PGE such as those found at Insizwa (e.g. Scholtz, PGE deposits. Mineralised Nipissing gabbro is known to 1936; Lightfoot et al., 1984). These studies together with extend from southwest of Sudbury near Whitefish Falls data for other CFB suggest that there are important empiri- (Card, 1976) and through the Wanapetei-Kelly-Janes cal relationships between the development of mineralisa- townships (Dressler, 1979; 1982). Moreover, the trend of tion and the formation of large igneous provinces at conti- these mineralised intrusions passes through the Sudbury nental margins. In some cases, geochemical signatures Structure, along the line of the Wanapetei gravity anomaly, within the basaltic rocks record evidence of extensive and centers on a regional northeast-southwest aeromag- crustal contamination accompanied by the depletion of the netic and gravity anomaly (Figure 2). Along this gravity magmas in Ni, Cu, and PGE (Naldrett et al., 1992; anomaly are a number of Paleoproterozoic intrusions Brugmann et al., 1993; Lightfoot et al., 1994), and signa- which host Ni, Cu, and PGE showings such as the East Bull tures of this type have been actively sought in other CFB Lake Gabbro-Anorthosite Intrusion, the Shakespeare- (e.g. in the Keweenawan Midcontinent Rift; Lightfoot et Dunlop Gabbro- Anorthosite Intrusion, the Wanapetei al., 1991). The main parameters of the Noril’sk empirical gabbronorite intrusion, and the River Valley Anorthosite exploration model may be summarised as follows (after (e.g. Peck et al., 1993) (see Figure 2). Naldrett and Lightfoot, 1993): The recognition of the importance of gabbroic intru- 1. The presence of mineralised picritic gabbroic sions as the roots of large igneous provinces which often intrusions which have acted as open-system magma host significant quantities of economic mineralisation has conduits. been important in reevaluating the mineral potential of the Nipissing intrusions. The presence of mineralised 2. The presence of highly contaminated, mantle derived, Nipissing intrusions associated with a trend of geo- magmas either as intrusions or within the stratigraphy logically anomalous mineralisation covering a period of of the comagmatic lavas. Working cut-off values almost 0.8Ga and associated geophysical anomalies has established at Noril’sk are: SiO =52-56 weight %, 2 led to a more vigorous re-interpretation of existing data La/Sm>3, 87Sr/86Sr >0.706. o and provision of new data for the Nipissing gabbro which 3. The strong depletion of magmas in Ni, Cu, and PGE. have a bearing on mineral exploration. At Noril’sk, lavas with <50 ppm Ni, <50ppm Cu, and <1ppb Pt+Pd are considered very depleted (Naldrett In this study we report new data for Nipissing intru- and Lightfoot, 1993). sions from across the Nipissing magmatic province, and compare and contrast their chemical compositions. We 4. A source of S, which is spatially linked to the location address the following questions: of the magma conduit system, is assimilated by the magma. At Noril’sk this was achieved in the Talnakh 1. Were Nipissing intrusions formed from a single batch Intrusion which appears to have been an open system of magma or multiple pulses of magma and where did conduit through which magma migrated, and inter- the magma enter the undulating sill? acted with crustal sulphur contained in evaporite-rich sediments. 2. What role did crustal contamination play in the genesis of the Nipissing magmas, and did the 5. Deep mantle-penetrating faults which permitted the contamination play any significant part in triggering unhindered migration of mantle-derived magmas. the segregation of immiscible sulphides as suggested 6. The presence of picritic lavas and intrusions. by Irvine (1975)? Not all of these criteria are evident in the Nipissing 3. What processes gave rise to the Nipissing parental gabbro, but we do describe a number of features which are magma, and are there any geochemical data to suggest encouraging from the perspective of mineral exploration that mineralisation is associated with particular potential. Nipissing intrusive magma types?

3 OGS Study 58

Figure 1a. Distribution of Nipissing gabbro across the Southern Province, northern Ontario (modified from Card and Pattison, 1973).

4. What spatial and temporal relationships exist parental magma compositions using undifferentiated gab- between Nipissing intrusions in the context of bronorite and chilled diabase, and we have focussed atten- existing palaeomagnetic and geochronological tion on the Temagami Region where a large number of studies, and do these relationships have any value in undifferentiated sills exist in an area currently under the predicting mineralisation? Temagami Land Caution. We contrast these data with 5. What are the best empirical data relevant to the analysis by Conrod (1988, 1989) of the Gowganda, construction of exploration models for deposits of Ni, Cobalt, and Sudbury areas. 3) We have attempted to char- Cu, and PGE-enriched sulphides in the Nipissing acterise aplitic granitoids related to the Nipissing intru- magmatic province? sions in order to determine their relationship with the country rocks and their genesis; we have focussed attention on 6. Did gabbroic magmatism occur at a passive or active an intrusion of aplite from the Obabika Intrusion (Obabika continental margin in the context of plume contribu- Inlet on Lake Temagami), and on aplites from the Kerns tions from an active mantle plume versus lithospheric Intrusion (New Liskeard). 4) We compare and contrast the contributions from above a passive rift caused by chemical compositions of gabbro and diabase sampled shallower plate tectonics (e.g. Hawkesworth and from locations where the palaeomagnetic remanence Gallagher, 1994)? direction has been characterised by others (e.g. Buchan To achieve these goals we use data for samples collected et al., 1989). We also use these data and U-Pb geochro- from a number of intrusions located across the magmatic nology to determine whether the difference in emplace- province: 1) We have studied two intrusions (Basswood ment age is real or whether the palaeomagnetic remanence Lake Intrusion at Thessalon and the Kerns Intrusion at direction is correlated to any change in parental magma New Liskeard; Figure 3) in a great degree of detail to composition. 5) We have sampled gabbronorites from understand the differentiation history and the mechanism mineralised and barren intrusions in order to determine of contamination, and some of the results are published whether mineralised intrusions are geochemically differ- elsewhere (Lightfoot et al., 1987). 2) We have sampled ent when compared to unmineralised intrusions. We pres- material from the less well differentiated intrusions across ent empirical observations regarding the characteristics of the Nipissing Province (see Figure 3) to characterise the mineralised and unmineralised intrusions.

4 Petrology and Geochemistry of the Nipissing Gabbro 93). Sketch diagram showing the relationship between the petrology of the undulatory sills and the associated mineralisation (after Lightfoot et al., 19 Figure 1b.

5 OGS Study 58

acquisition of geochemical data under contract 38ST23233-7-0068. The following staff at the Geological Survey of Canada are thanked for their assistance with this project: Dr. J. M. Duke, Dr. K. Buchan, Dr. K. Card, and Dr. R.O. Eckstrand. Analytical data were acquired at the Open University, U.K., the University of Toronto, and the Geoscience Laboratories. Dr. R.G.V. Hancock and Dr. M.P.Gorton, University of Toronto, are thanked for assistance with Instrumental Neutron Activation Analysis. Dr. S. Noble, Jack Satterly Geochronology Laboratory, Royal Ontario Museum, is thanked for assistance with U-Pb geochronology and Sm-Nd systematics, and Dr. T. Krogh is thanked for access to the geochronology laboratory at the Royal Ontario Museum. For hospitality in the field, Peter and Linda Phalen of Loon Lodge are thanked, and field assistance from Grant Phalen is acknowledged. We thank Norm Evensen, University of Toronto, for assistance with Sm-Nd isotope studies and Debbie Conrod for numerous discussions on Nipissing petrogenesis. The former staff of the Resident Geologist’s Office in Cobalt are thanked for assistance, and L. Owsiacki and P. Anderson are thanked for pointing out suitable locations for sampling in Kerns Township. This manuscript has benefited from reviews by Dr. A. Fyon, Dr. W.T. Jolly and Dr. F. Dudas. We also acknowledge the assistance of G. O’Reilly, B. Wright, F. Racicot, J. Rauhala, G. Salo, and D. Brunne. Diagrams were prepared by Steve Josey, Ontario Geological Survey. INTRODUCTION AND GENERAL GEOLOGY The 2.22Ga Nipissing gabbro of northern Ontario com- prise a suite of dominantly tholeiitic to calc-alkaline rocks ranging from chilled diabase through quartz diabase, gabbronorite, gabbro, vari-textured gabbro, pegmatitic gabbro, quartz diorite, granodiorite, granophyre, and aplitic granitoids. The intrusions extend from Sault Ste. Marie through the Sudbury Region, to the Cobalt and Gowganda regions (see Figure 1a) and are grouped as a large igneous Figure 1c. Typical sequence of lithologies seen in a well-differentiated province termed the Nipissing Gabbro Province in this re- intrusion of Nipissing gabbro showing the relationship between silicate port. The intrusions range in thickness from a few hundred rocks and sulphide mineralization. meters to over a thousand meters, and outcrop at the present erosional level as open ring structures, ring dykes, cone sheets, dykes and undulatory sills (Hriskevich, 1952, 1968; Card and Pattison, 1973). The intrusions are dominantly ACKNOWLEDGEMENTS located in the Huronian Supergroup, but are also localised along the Archean- Proterozoic unconformity. A study of the Nipissing Gabbro was initiated at the Precise U-Pb geochronology on magmatic baddeley- University of Toronto. The Ontario Geological Survey ite from Nipissing gabbro has yielded crystallisation ages supported this work through an OGRF grant which allowed of 2219M3.6Ma from the Gowganda area (Corfu and D. Conrod to complete her MSc thesis. This work bene- Andrews, 1986), 2212M2Ma from the Sudbury area fited from the encouragement of B. Dressler. Additional (Conrod, 1989), 2217M4Ma and 2210M3.8Ma from the Co- work was completed by D. Conrod whilst she worked at the balt area (Noble and Lightfoot, 1992) (see Figure 1d). Pa- Ontario Geological Survey, and access to information in laeomagnetic data (Buchan et al., 1989) are interpreted to a published report and a report in preparation are reflect at least three discrete phases of magma emplace- acknowledged. ment separated by at least 50Ma. However, U-Pb geochro- The authors acknowledge support from Energy Mines nology on samples collected from sites giving two differ- and Resources, Canada, for the field studies and the ent remanence directions separated by 50Ma (N1 and N2

6 Petrology and Geochemistry of the Nipissing Gabbro

Figure 1d. Location of samples and results from U - Pb geochronology work (Corfu and Andrews, 1986; Noble and Lightfoot, 1992; Conrod, 1989).

Figure 2. Regional tectonic setting of the Nipissing gabbro and the location of the 2.2 Ga Preissac Dyke Swarm (after Card et al., 1994 and Osmani, 1991). 7 OGS Study 58 of Buchan et al., 1989) yield magmatic baddeleyite ages consisting of a basal quartz diabase overlain by gabbro, overlapping by less than 7.8Ma (Noble and Lightfoot, and extreme differentiation into hypersthene gabbro and 1992). One of the intrusions dated by Noble and Lightfoot granophyric gabbro is the exception rather than the rule (1992) at 2210Ma has associated quartz-carbonate veins (see Figure 1c). with silver and cobalt mineralisation. The intrusions are dominantly tholeiitic, but the The Nipissing gabbros northeast of Sudbury in the evolved rocks trend towards calc-alkaline compositions, Cobalt Plate area (see Figure 1a) are relatively undeformed and recent geochemical studies suggest that the petrologi- and unmetamorphosed gabbros. Intrusions which have cal variations within the intrusions are controlled by frac- traditionally been termed “Sudbury Gabbro” consist large- tional crystallisation of the magma with or without assimi- ly of amphibolites (e.g. Ginn, 1965; Card, 1965, 1968), and lation of overlying country rock after emplacement (e.g. grouped with the Nipissing between Sudbury and Blind Lightfoot et al., 1987; Conrod, 1988, 1989; Lightfoot et al., River, have undergone some deformation and regional 1993). The chilled quartz diabase and least differentiated amphibolite facies metamorphism perhaps related to the gabbro samples from a number of different Nipissing intru- Paleoproterozoic Penokean Orogeny between 1900 and sions are consistent with the emplacement of a composi- 1850Ma (Card, 1978). West of Blind River the Nipissing tionally uniform low-Mg parental magma (termed the gabbros have undergone some deformation, but less ex- Nipissing magma type) across a wide tract of the Southern treme regional metamorphism. The intrusions southwest Province (Lightfoot et al., 1993). The geochemical charac- of Sudbury are elongated parallel to the main structural teristics of this parental magma include moderate MgO fabric of the Murray Fault System, and the emplacement of (8-9 weight %), elevated SiO2 (50.0-51.5 weight %), these intrusions may have been genetically linked to strong light rare earth element (LREE) and large ion litho- faulting accompanying the deposition of the Huronian phile element (LILE) enrichment with La/Sm = 2.5-3.5 sedimentary rocks (Buchan and Card, 1985). and Th/Nb = 0.7-0.9, and marked negative anomalies for The Nipissing Intrusions have traditionally been TiO2,P2O5, and Nb+Ta (Lightfoot et al., 1993). These are described as undulatory sheets consisting of a series of ba- all features of Phanerozoic basalts from continental set- sins and arches connected by limbs (Hriskevich, 1968) (see tings typified by the Continental Flood Basalts (CFB), and Figure 1b). The basinal portions consist of quartz diabase there is currently a vigorous debate as to whether these overlain by hypersthene gabbro, and an overlying vari-tex- characteristics are features of the source regions of the tured gabbro with pegmatoidal patches. The arches consist magmas or whether they were imparted to the magmas by of vari-textured gabbro overlain by quartz diorite, grano- continental crustal contamination as the magmas evolved diorite, granophyre and aplitic granitoids. In detail, many in deep crustal reservoirs and migrated from the mantle to of the undulatory intrusions are relatively undifferentiated, the crust.

8 Empirical Metallogeny of the Nipissing Gabbro

Significant close spatial associations between mineralisa- 3. The silicate host of the sulphides tends to be relatively tion and Nipissing intrusions have been recognised (Card undifferentiated, although there is some petrographic and Pattison, 1973; Innes and Colvine, 1984), and there is a indication that there may be cyclical trends in mineral variation in style and type of mineralisation across the composition and whole-rock chemical composition Nipissing Province. In the east, mineralisation is domi- (e.g. Conrod, 1989; Finn and Edgar, 1986). Intrusions nantly Ag, Co and Ni as native metals, arsenides, and sul- which are heavily contaminated in their roof zones farsenides associated with quartz- carbonate veins which (Lightfoot et al., 1989; 1993) such as the Kerns cut the sediments and the Nipissing intrusions (Jambor, Intrusion, northwest of New Liskeard, and the 1971), and may be Neoproterozoic in age. In the central Basswood Lake Intrusion, north of Thessalon, appear portion of the Nipissing Province, mineralisation is domi- to be relatively unmineralised. The low Cu/Yb and nantly Cu-Ni-platinum group element (PGE) sulphides Cu/Zr of the most contaminated granophyric rocks in which occur disseminated within the intrusions or as mas- these specific intrusions is attributed to large amounts sive pods beneath the intrusions (Rowell, 1984; Rowell of assimilation of sediment with very low Cu/Yb and and Edgar, 1986; Lightfoot et al., 1991; Lightfoot et al., Cu/Zr rather than the fractionation of an immiscible 1993). In the western part of the Province, the mineralisa- sulphide liquid (Lightfoot et al., 1994; Hawkesworth tion consists of Cu-sulphides as fine disseminations or in et al., 1995). quartz-carbonate veins which cut vertically through the 4. The mineralised intrusions tend to be those which are Nipissing gabbro. least strongly differentiated into granophyric zones, A number of empirical observations have been made yet consist of gabbronorites which contain a higher in previous studies, and one goal of this study is to focus modal proportion of hypersthene (20-40 modal %) attention on this aspect of the study. For this reason, these than unmineralised intrusions (<20 modal %). As a re- features are introduced at this early stage, and then sult of the hypersthene content, the mineralised intru- returned to in the summary of metallogenetic impact. sions tend to be more mafic as reflected in their ele- Table 1 summarises some important observations about vated MgO (10-14 weight %) and low incompatible the geology, petrology, mineralogy, and geochemistry of element concentrations (e.g. 0.2-0.4 weight % TiO2, the Nipissing Gabbro which relate to mineral potential. A <50ppm Zr) (Lightfoot et al., 1991, 1993). number of important points set the scene for this study of 5. Existing geochemical data suggest that there is no ob- the Temagami Region, and these are summarised below: vious difference in the magma type of the mineralised 1. Magmatic Ni, Cu, and PGE mineralisation is spatially intrusions when compared to the unmineralised intru- associated with intrusions which lie on a trend sions. For example, Lightfoot et al. (1993) demon- between Whitefish Falls, Sudbury, and River Valley strated that the undifferentiated quartz diabase, gab- (Figure 4 and see Table 1). bros, and chills have a narrow range in incompatible 2. Mineralised Nipissing gabbro intrusions have associ- element ratios such as La/Sm and Th/Y, and similar ated disseminated sulphide (see Table 1). The sul- abundance levels. phides occur as fine disseminations of magmatic pyrr- 6. The chemical homogeneity of the chilled margins and hotite (50-75%) with lesser chalcopyrite and pentlan- undifferentiated quartz diabase and gabbro intrusions dite. Some of the sulphides exhibit magmatic blebby suggests: a) a single source for these magmas, and b) textures (Appendix 1; Lightfoot et al., 1991; Lightfoot there has not been significant differential interaction et al., 1984; Naldrett et al., 1992) with pyrrhotite- rich of the magma with a crustal reservoir other than at the lower segments and chalcopyrite-rich upper segments final site of crystallisation (Lightfoot et al., 1989; suggesting in-situ differentiation of the immiscible 1993). However, the parental magma type is charac- sulphide liquid during cooling and crystallisation terised by strong light rare earth element (LREE) and (Photo 1) Massive sulphides rich in Cu and PGE are large ion lithophile element (LILE) enrichment as known as basal concentrations associated with the well as marked negative anomalies for Ta+Nb, TiO2, Wanapetei gabbronorite intrusion, and despite this and P2O5. These are all features of a magma that has being a small occurrence, the sulphides carry 1-15 either interacted with a crustal reservoir, or they are weight % Cu, 2.5-6.3ppm Pt, 17- 53ppm Pd, and features of young rocks typically attributed to deriva- 1-6ppm Au (Lightfoot et al., 1991; 1993). The dissem- tion from ancient regions of mantle lithosphere which inated sulphides (<5% modal sulphide) tend to be contain recycled continental crust (e.g. Lightfoot et localised in the interior of the sills (100-300m above al., 1993a, b; Hergt et al., 1991). This geochemical the base) within coarse grained gabbronorites and signature is also one found in many of the Upper Se- hypersthene-rich gabbros; these rocks carry quence lavas of the Siberian Trap at Noril’sk, but is not 200-1100ppb Pt and 50- 4000ppb Pd in mineralised quite as extreme as that found in the highly contami- intrusions in Janes Township (Lightfoot et al., 1991; nated Nadezhdinsky lavas which have low Cu, Ni, and 1993). Table 2 summarises the grade of these massive PGE abundances (Lightfoot et al., 1990; 1993; 1994; and disseminated sulphides. Naldrett et al., 1992; Brugmann et al., 1993).

9 OGS Study 58

Table 1. Summary of empirical characteristics of the metallogenetic associations of the Nipissing gabbro intrusions, Ontario. Main points from this study, Lightfoot et al. (1991), Conrod (1988, 1989), Lightfoot et al., (1993), and Card and Pattison, (1973). Criteria for Metallogenetic Observations Regarding Nipissing Gabbro Evaluation of Nipissing Gabbros Regional setting of mineralised Possible roots of a large igenous province; similarity to the diabase sills of the Karoo Province sectors of large igneous provinces (Lightfoot, 1982; Marsh and Ealses, 1984) and Tasmanian Province (Hergt et al., 1989); similarity to diabase--gabbro sills in the epicontinental rocks of the Siberian Trap (Naldrett et al., 1992). Noril’sk is a classic example where the gabbrodolerite and picritic sills have controlled mineralisation. Structural association of Many dykes and sills associated with regional east--west tectonic systems such as Murray fault systems. mineralised rocks and Feeder dykes at Narrows and Sand Point (Temagami) are east--west oriented. The importance of deep mantle--penetrating faults mantle--penetrating faults is highlighted at Noril’sk, where the Noril’sk--Kharaelakh Fault controls mineralisation. It is in known whether the major east--west and north--south fault systems associated with the Nipissing were mantle--penetrating, or conduits, but the association of gabbroic rocks with these faults suggests that this is a possibility. Basinal association of some Focus of Nipissing activity in basinal sediments of Huronian with equivalent amounts of activity mineralised gabbros focussed in the Cobalt Embayment and east--west along the extension of the Huronian geosynclinal package. The basinal setting is quite different to Noril’sk in so far as the Noril’sk deposits were formed in an epicontinental setting where a thick evaporite--carbonate sequence was developed. The Nipissing example is more like the Karoo where large nickeliferous intrusions such as Insizwa are known (Scholtz, 1936; Lightfoot et al., 1984). Tectonomagmatic setting of No strong evidence for a mantle plume during the Nipissing event. Tectonomagmatic setting may mineralised gabbros resemble Karoo with lithospheric extension being the main driving force (Hawkesworth and Gallagher, 1994). The absence of a mantle plume as a driving force may explain the absence of high--Mg Nipissing rocks, and also account for the extreme uniformity of the magma in composition. The Insizwa Complex in the Karoo is developed in a basinal setting presumably associated with lithospheric extension, and the gabbros of this complex are mineralised. However, Insizwa does show well developed picrites, and these are not evident in the Nipissing. Geophysical signatures linked to Major aeromagnetic and gravity anomalies are present along a trajectory between Englehart, Temagami, metallogenic province Wanapetei, Sudbury, and Manitoulin Island. This ESE--WNW trend of anomalies may be related to a string of deep mafic--ultramafic complexes. The trend is linked to the presence of the mineralised Early Proterozoic intrusions (Peck et al., 1993), mineralised rocks of the Temagami Island deposit (Simony, 1964), the mineralised Sudbury Igneous Complex (Pattison, 1979; Lightfoot et al., 1994) and is the trajectory of MOST mineralised Nipissing gabbro intrusions. This association is presumably not coincidental, but is the signature of a metallogenic province, where Ni--Cu--PGE mineralisation was developed over a protracted time interval by different processes. Physical traps for sulphides Immiscible sulphide liquids which are able to settle under gravitational forces to the base of the undulatory Nipissing sills may readily accumulate in the basinal portions of the undulatory sills. If this process has happened, then the only possible example is Rathbun Lake, and many authors consider this body to be of hydrothermal origin. On the grounds that much of the mineralisation is within hypersthene gabbros well above the base of the intrusion, there is some indication that the saturation of the magma in S was achieved only after the crystallisation of the lower parts of at least some of the sills. This suggests that the physical traps for sulphide mineralisation are more closely linked to the centers of the sills. The late formation of the sulphides implied by this association would make gravitational accumulation of sulphides into massive bodies less likely in flat sills, but perhaps more likely in stagnant dyke systems, inclined sheets, or ring complexes. Sulphur source A source for the sulphur in many of the mineralised Nipissing gabbros remains uncertain. Either a mantle or crustal source appears likely. An important issue is the S content of Huronian sediments, and the role that they played in the metallogenesis of the Sudbury Region. No S/Se or S--isotope data for mineralised Nipissing gabbro exists at this time.

Silica control on sulphur saturation The Nipissing gabbros are unusual in chemical composition. They have elevated SiO2, high La/Sm, high Th/Nb, and other Nd--isotopic compositions that suggest that their chemical composition is controlled by a crustal material. Contamination has been documented in--situ in the Kerns Intrusion (Lightfoot and Naldrett, 1989), but is not linked to metallogenesis because the most fractionated rocks are also the most contaminated, and their low Cu is due to assimilation of low--Cu crustal rocks. Rather, the contaminated nature appears to be an inherent feature of the Nipissing magma, and may be linked to the source and the recycling of ancient crust into this source (Lightfoot et al., 1993). The similarity in composition to dolerites in young CFB such as the Karoo and Siberia is an interesting observation, and the fact that the silica content of the gabbros is high means that it would not take much differentiation or contamination to achieve sulphur saturation of the Nipissing magma (Irvine, 1979).

10 Petrology and Geochemistry of the Nipissing Gabbro

Table 1. (cont.) Summary of empirical characteristics of the metallogenetic associations of the Nipissing gabbro intrusions, Ontario. Main points from this study, Lightfoot et al. (1991), Conrod (1988, 1989), Lightfoot et al., (1993), and Card and Pattison, (1973). Criteria for Metallogenetic Observations Regarding Nipissing Gabbro Evaluation of Nipissing Gabbros Multiple batches of magma or Naldrett et al. (1995) suggest that the mineralised intrusions at Noril’sk were open system conduits to open system conduits? flood basalt magmatism. Flow through a conduit will produce geochemical differences compared to multiple influx and in--situ crystallisation of batches of magma. In the Nipissing, Conrod (1989) reports important data from the Gowganda area where she records evidence of four pulses of magma. On a simple empirical basis, and extending this observation to other Nipissing sheets, there would appear to be good evidence for the emplacement of multiple batches of magma into some intrusions. At issue is whether these intrusions were conduits to high level volcanic edifices. Based on the similar bulk composition of the different Nipissing intrusions, it appears that we have not yet found any intrusions which contain significant amounts of cumulates left behind in a horizontal conduit such as that found at Noril’sk. The search should continue for olivine and hypersthene cumulates, and these intrusions are likely to be the ones which have fed any volcanic edifice. Petrology --empirical association There is a well documented association of mineralisation with hypersthene--rich, high--Mg rocks, with of sulphides and mafic rocks lower abundances of Tio2 and Zr in the Nipissing gabbro. The typical target values for identification of these rocks are: 10--30 modal percent hypersthene >9 wt.% MgO <0.4 wt.% TiO2 <52 ppm Zr Chemostratigraphy of intrusions The presence of mineralisation high within the Nipissing gabbros suggests that exploration efforts in these rocks should play close attention to the chemical stratigraphy of the intrusions. The geochemical variations may well provide important information which constrain the possible role of gravitational settling of sulphides, and the extent to which whole--rock compositions are consistent with the scavenging of Cu, Ni, and PGE from the magma. Other miscellaneous observations The present structural configuration of many sills is not the original configuration. A careful analysis of the shape of the intrusion with respect to the Huronian stratigraphy is warranted. We have no evidence at this time for more than one Nipissing magma type. A very important new understanding would be achieved if picritic rocks were to be recognised in the Nipissing Province. The similarity of the Nipissing to rocks of the diabase gabbro sills of the Trans--Hudson orogen makes other Ontario targets within the Sutton Inlier an increasingly exciting exploration target (Bostcok, 1971; Lightfoot, 1994)

7. Metamorphism of country rock and S sources. We alteration may be sufficient to mobilise some of the LILE presently have no evidence that high-S shales or evap- such as Rb and Ba (e.g. Cox and Hawkesworth, 1985), but orites exist within the Huronian sequence, nor that the it is generally accepted that the REE, HFSE, Th, U, Ta, Nb, metamorphic aureole around mineralised sills is sub- and Y are relatively immobile under these conditions (e.g. stantially larger than that found around barren sills. Hawkesworth and Morrison, 1978; Humphris and Exploration models which follow that for the Noril’sk Thompson, 1978). In some intrusions there are good rocks should incorporate a search for crustal sulphur positive correlations between Rb and immobile elements sources and a search for significant contact such as Zr (e.g. Lightfoot et al., 1987), and these data argue metamorphism of Huronian sediments. that even the Rb data records petrogenetic information rather than the effect of late alteration. SAMPLING AND ANALYSIS Compositional differences due to variations in grain Full details of sampling and analytical protocols are given size between coarse and fine patches within the vari-tex- in Appendix 1. Sample locations and descriptions are tured gabbro were investigated using fresh samples from given in Miscellaneous Release Data (MRD) 19. Results vari-textured gabbro patches and the local host of the patch for in-house reference materials are given in Table 3. (<1m from the patches) at two locations. The pairs were 86-155 (fine-grained host) and 85-156 (vari-textured EFFECTS OF THE TEXTURAL patch), and 86-160 (fine-grained host) and 86-161 (vari- VARIATIONS IN THE GABBRO ON textured patch). Figure 5a-b demonstrated that the primitive mantle normalised (Sun and McDonough, 1989) spid- THE ANALYTICAL DATA ergram patterns of the coarse and fine- grained samples from each of the locations have essentially similar patterns Wherever possible, fresh samples of rock were used for with the vari-textured samples showing an overall enrich- analysis, and thin sections showed no evidence of strong ment in the concentrations of the incompatible elements recrystallisation, albitisation or growth of metamict compared to the finer grained host gabbronorite. Impor- minerals. tantly, samples 86-154 and 86-155 also have similar 143 144 Studies of basalt, diabase and gabbro samples from Nd/ Ndo and similar Sm/Nd. These data therefore young CFB indicate that even very limited degrees of suggest that both the fine-grained host and the coarser-

11 OGS Study 58

Figure 3a. Ni versus forsterite relationships in olivines from the Cross Lake Sill (after Conrod, 1988). Data for Insizwa from Lightfoot and Naldrett (1984) and Lightfoot et al. (1984); data for Moxie Pluton from Thompson and Naldrett (1984). Field of undepleted olivines from Simpkin and Smith (1970).

Figure 3b. PGE distribution patterns for the Rathbun Lake showing (after Lightfoot et al., 1993 and Rowell and Edgar, 1986).

GEOCHEMICAL EVIDENCE FOR THE EMPLACEMENT AND IN-SITU grained vari-textured patches crystallised from the same parental magma, and that the patches tend to trap the more DIFFERENTIATION OF NIPISSING differentiated REE, LILE, and HFSE-enriched liquid than INTRUSIONS the matrix. Petrographic data indicate that many of the Nipissing gab- Lightfoot et al. (1993) document geochemical data bro intrusions are characterised by well-developed chilled for coarse- grained and fine-grained portions of a diabase margins between 50cm and 5m wide, overlain by vari-textured gabbronorite and a granophyric gabbro from 10- 20m of quartz gabbro and then by 100-500m of hyper- the Basswood Lake Intrusion (Figure 5c-d). They show sthene gabbro which grades up into 100-500m of hyper- that the coarse- and fine-grained portions of the sthene-poor gabbronorite and vari-textured diabase. In a vari-textured gabbro have very similar chemical composi- few locations such as the Kerns Intrusion, the Basswood tions, but that the coarse-grained granophyric gabbro has Lake Intrusion, and at Obabika Inlet on Lake Temagami elevated REE relative to the finer grained matrix. (Figure 6, the intrusions have localised zones of quartz dio- Lightfoot et al. (1993) suggest on these grounds that the rite, granodiorite, granophyre, and aplitic granitoids, and process producing the vari-textured gabbros does not pro- these were viewed by Bowen (1928) as extreme differenti- duce significant relative fractionation of the incompatible ates of the Nipissing magmas. At issue is whether: 1)any of elements although it appears to have some effect on the these variations point to an upward decline in incompatible abundance levels of these elements. element concentrations away from the base of the

12 Petrology and Geochemistry of the Nipissing Gabbro udbury Igneous Complex (after Muir, 1984). Geological relationships between Nipissing intrusions, mineralisation, and geophysical anomalies, Early Proterozoic mineralisation, and the S Figure 4.

13 OGS Study 58 intrusions which might be termed a “reversed differenti- lower contact of the intrusion (Figure 7a). Samples of ation trend”, or provide evidence for multiple cyclical em- the fine- grained chilled diabase, basal quartz diabase, placement of magma into a single differentiating intrusion, gabbronorite, and hypersthene gabbro were analysed 2)there are any compositional breaks within the intrusions and trace element data are shown in Figure 7b and c which might provide evidence for the repeated influx of normalised to the composition of primitive mantle magma into or through the intrusions, or record evidence (Sun and McDonough, 1989). The chilled diabase and of mixing of different batches of magma, 3)there are any basal quartz diabase and lowermost hypersthene gab- intrusions that underwent closed-system in-situ Rayleigh bro samples show an upwards decline in La, Th, and Fractionation, 4)the compositions of the most fractionated Zr concentration with little systematic change in MgO rocks record interaction with the Huronian sediments or Ni content (see Figure 7a). The uppermost through either melting and assimilation, or assimilation hypersthene gabbro samples show an upward increase linked to fractional crystallisation of the magma, 5) any of in La, Th, and Zr which is accompanied by a signifi- the aplitic granitoids are direct anatectic melts of the cant decline in MgO and Ni content (see Figure 7a). Huronian sediments. The spidergrams all have approximately the same shape and marked positive Ba, Sr and Zr anomalies, A complete description of the study areas and rocks and negative Rb, Th, Ta, Eu, and Ti anomalies (see are provided in Appendix 1, Lightfoot et al. (1989, 1991, Figure 7a), but the more basal quartz diabase has 1993) and Conrod (1988, 1989, in preparation). elevated trace element abundances relative to the gabbroic rocks.) 1. Variation across the least differentiated Nipissing gabbronorite Intrusions: Samples were selected A similar reverse differentiation trend has been docu- across the High Rock Island Sill, Portage Bay, Lake mented by Conrod (1988) in the Cross Lake and Bonanza Temagami (Appendix 1), where there is a good cliff Lake Intrusions within the Cobalt and Sudbury areas, and section exposing the lower part of a relatively undif- by Conrod (1989) in the Duncan Lake Intrusion west of ferentiated gabbronorite sill. Eleven samples were Gowganda (Figure 8). In a study of the Wanapetei Intru- taken through a vertical distance of 100m along a sec- sion, east of Sudbury, Finn et al. (1982) suggest that com- tion approximately perpendicular to the dip of the piled data from five traverses through the intrusion record

Photo 1. Sulphide globules in gabbro fromthe WanapiteiIntrusion,SudburyDistrict.Thesulphide consistsof pyrrhotiteat thebase ofthe globuleand chalcopyrite at the top of the globule. This texture is much like that described from the Waterfall Gorge deposit of the Karoo-aged Insizwa Complex (Lightfoot et al., 1984), and the Noril’sk-Talnakh sulphides of the Siberian intrusions (Naldrett et al., 1992). Traditional wisdom suggests that these globules are the product of the crystallisation of a bleb of immiscible sulphide in the silicate host. The in-situ fractionation ofthe sulphides is ascribed to magmatic differentiation of the sulphide liquid to form a monosulphide solid solution pyrrhotite cumulate with an overlying trapped liquid. Blebs of this type are assumed to record the bulk composition of the sulphide liquid which may form massive basalconcentrations ofsulphides in differentiating intrusions.

14 Petrology and Geochemistry of the Nipissing Gabbro

Table 2. Summary of grade of mineralisation at Kukagami Lake and the Rathbun Lake Showing, Wanapetei-Kukagami Lake areas. Based on Lightfoot et al. (1991). See Table 1 for additional information.

Rathbun Lake (n=14; Lightfoot et al., 1991): Kukagami Lake (n=9; Lightfoot et al., 1991): Element Concentration Element Concentration S 1--18 wt.% S <3 wt.% Cu 0.3--12.4 wt.% Cu 0.1--1.1 wt.% Ni 0.1--1.1 wt.% Ni 0.1--0.4 wt.% Pt 100--6500 ppb Pt 50--1200 ppb Pd 200--35000 ppb Pd 50--4200 ppb Au 500--6500 ppb Au 20--600 ppb variations through a 822m thick gabbronorite intrusion. diabase and quartz diabase. Magmas laden with hyper- They suggest that the intrusion develops five or more mul- sthene phenocrysts which formed in the conduits or cham- tiple reverse differentiation trends through the thickness of bers at depth were subsequently emplaced. Above this lev- the intrusion. In detail, the proximity of many of the tra- el there was a reversal towards the emplacement of pheno- verses to the lower contact raise the possibility that these cryst-poor magmas, and this process may have been re- cyclesare repetitionsof the same magmatic unit which was peated many times within a single intrusion. The continu- sampled along strike within the same intrusion, which ous trends in petrology and geochemistry through the would then suggest that only a single reversed differenti- cycles in the Miller Lake Intrusion and the lack of internal ation trend is recorded in this intrusion. However, Conrod baked or chilled contacts suggest that this event was a con- (1989), in a systematic study of the Miller Lake Intrusion tinuous one rather than an episodic emplacement of differ- west of Gowganda, shows that four trends of reverse ent batches of magmas into discrete sills. Importantly, this followed by normal differentiation are recorded in the suggests that geochemical variations in the Nipissing gab- whole-rock major element and incompatible element bro are consistent with the emplacement of multiple abundance data (see Figure 8). Conrod (1989) suggests batches of magmas into single magma chambers in a simi- that the series of samples reflect their true relative strati- lar way to that proposed for the Talnakh Intrusion of the graphic position in a single intrusion with no fault- repeti- Noril’sk Region (Czemanske et al., 1994; Naldrett et al., tion. The data of Conrod (1989) are particularly valuable 1995). It is not clear based on the available data whether as they indicate that the basal increase in Mg-number is these sills were open system conduits, feeding higher level coupled to a decline in Zr, Y,Sr, and Rb, and although there volcanic edifices. Thus, although this makes the Nipissing are no published data for the compatible trace elements sills potential horizontal conduits for higher level flood ba- such as Ni, the Mg-number, Zr, Y,Sr, and Rb data correlate salt magmatism, the products of which have long since with the increase in modal hypersthene abundance away been removed by erosion, the geochemical data available from the lower contact which is also a general feature of at this time do not constrain whether: 1)these chambers the Portage Bay intrusion from this study of the Lake contain gabbroic units with igneous contacts; 2)the cham- Temagami Region (see Figure 7a). This suggests that the bers were conduits through which vastly larger volumes of basal compositional reversal reflects the emplacement and magma passed than are presently seen, or 3)the intrusions crystallisation of magmas which are progressively more were the focus of magma mixing events within the cham- heavily loaded with hypersthene phenocrysts. These data ber. Each of these models may be linked to different styles therefore indicate that Nipissing gabbro intrusions fre- and amounts of mineralisation. quently record systematic compositional trends which are 2. In-situ differentiation and crustal contamination not readily explained by in-situ differentiation of a single history of Nipissing intrusions: The Kerns Intrusion, pulse of magma. located northwest of New Liskeard, was chosen for detailed study, and some of the data and conclusions Reverse differentiation has been recognised in other were published by Lightfoot et al. (1989). Here we re- sills such as the Insizwa Complex of the Karoo Province port detailed data for this sill as it provides valuable (Lightfoot et al., 1984), and in the Fongen-Hilligen information on the differentiation and contamination- Complex in Norway, and are very different to the normal processes responsible for the evolution of the differentiation trends of the Skaergaard Intrusion of East Nipissing magma after emplacement into the Greenland (Wager and Brown, 1968) and the Pallisades Huronian sedimentary sequence. Sill of New Jersey (Walker, 1969). Traditional models require the emplacement of progressively more phenocryst Figure 9a-d show primitive mantle-normalised (Sun laden magmas to explain these basal reversals in composi- and McDonough, 1989) spidergrams for representative tion. The data for the Nipissing sills which illustrate re- samples from the different rock types in the Kerns versed differentiation trends are consistent with the em- Intrusion. For a more detailed description of the sill and the placement of different batches of magma with variable hy- rock types, see Appendix 2, Lightfoot et al. (1989), and persthene phenocryst abundance. The first magmas to be Lightfoot (1995). The samples from the different rock emplaced were phenocryst poor and formed the chilled types all show moderate LREE and LILE enrichment, and

15 OGS Study 58

Table 3. Analytical data for in-house standard reference materials UTB-1 (University of Toronto basalt standard) and WHIN SILL (Open University diabase standard) acquired in the course of this study, and comparison with expected values given in Lightfoot (1985). Element Whin Sill Whin Sill Whin Sill Whin Sill Expected (1984/85) (1986) (1987) Lightfoot (1985) Whin Sill n (INAA) 6 12 7 7 12 La 24.7(.5) 24.4(.8) 24.0(1.9) 25.9(1.5) 22.5(2.6) Ce 58.8(.8) 59.8(3.1) 58.4(3.1) 57.5(3.1) 60.4(3.8) Nd 30.0(3.0) 29.7(1.9) 33.3(2.0) 32.9(3.5) 28.4(5.2) Sm 7.0(0.3) 7.61(0.21) 7.56(0.55) 7.27(2.1) 7.02(0.26) Eu 2.00(0.05) 1.87(0.19) 2.30(0.12) 2.25(n.a.) 2.04(0.10) Yb 2.41(0.11) 2.43(0.16) 2.60(0.13) 2.54(n.a.) 2.51(0.14) Lu 0.34(0.02) .35(0.05) .41(.02) 0.39(n.a.) 0.35(0.02) Th 2.75(0.26) 2.82(0.11) 2.78(.11) 3.12(0.17) 3.05(n.a.) Ta n.a. 1.06(0.9) 1.06(.08) 1.31(0.06) 1.26(n.a.) Hf 4.92(0.48) 4.8(0.3) 4.5(0.2) 5.02(0.22) 4.93(n.a.) U 0.45(0.14) 0.49(0.11) 0.52(0.08) 0.80(9.14) 0.90(n.a.) Co 48.3(2.0) n.a. 47.8(1.2) 49.6(2.5) 47.4(3.1) Sc 30.4(0.8) n.a. 31.0(0.6) n.a. n.a. Ba 419(48) 387(59) n.a. n.a.

Element (1984/85) (1986) (1987) (Lightfoot, 1985) Exp. Method L.L.D. UTB1 1sd UTB1 1sd UTB1 1sd UTB1 1sd UTB1

SiO2 50.4 (0.6) 50.5 (0.5) 50.6 ------49.1 (0.4) 49.6 WD--XRF 0.5 TiO2 306.0 (0.12) 3.06 (0.08) ------3.00 (0.03) WD--XRF 0.02 A12O3 14.6 (0.6) 13.8 (0.2) 3.02 ------13.3 (0.3) 3.09 WD--XRF 0.2 Fe2O3 14.9 (0.6) 15.9 (0.3) 13.1 ------14.6 (0.1) 13.5 WD--XRF 0.2 MnO (0.00) (0.01) 15.0 ------0.22 (0.01) 15.2 WD--XRF 0.01 MgO 0.22 (0.4) 0.23 (0.3) 0.22 ------4.5 (0.3) WD--XRF 0.2 CaO 3.6 (0.4) 3.6 (0.1) ------8.5 (0.1) 0.21 WD--XRF 0.2 Na2O 8.6 (0.4) 8.3 (0.5) 4.3 ------2.85 (0.03) 4.5 WD--XRF 0.01 K2O 2.9 (0.10) 3.1 (0.03) 8.5 ------1.23 (0.05) 8.5 WD--XRF 0.01 P2O5 (0.06) (0.04) 2.5 0.72 (0.03) WD--XRF 0.02 LOI 1.17 1.22 n.a. 2.83 GRAV. 1.37 ------0.51 0.51 0.74 0.64 n.a. 0.74 0.31 n.a. V 367.0 (82.0) n.a. n.a. 401.0 (7.0) n.a. WD--XRF 5 Cr 123 (46) 143 (18) 44 101 (7) n.a. WD--XRF 10 Ni 20 (8) 42 (5) 22 26 (7) 25 WD--XRF 5 Zn n.a. 151 (4) 123 139 (5) 153 WD--XRF 10 Cu 30 (6) 29 (4) 32 34 (6) 31 WD--XRF 5 Ga n.a. 23 (3) n.a. 22 (2) n.a. WD--XRF 5 Rb 35.2 (1.6) 35 (2) 31 35 (1) 32 WD--XRF 5 Sr 313 (3) 323 (4) 279 311 (5) 312 WD--XRF 5 Y 46 (1) 47 (1) 54 46 (1) 41 WD--XRF 5 Zr 205 (2) 213 (4) 176 200 (4) 202 WD--XRF 5 Nb 16.3 (1.2) 16.6 (0.9) 21 15.3 (0.6) n.a. WD--XRF 3 Ba 593 (49) 635 (45) 522 536 (16) n.a. WD--XRF 10 n --number of analyses n.a. not available

16 Petrology and Geochemistry of the Nipissing Gabbro

Table 3. (cont.) Analytical data for in-house standard reference materials UTB-1 (University of Toronto basalt standard) and WHIN SILL (Open University diabase standard) acquired in the course of this study, and comparison with expected values given in Lightfoot (1985). Element (1984/85) (1986) (1987) (Lightfoot, 1985) Exp. Method L.L.D. UTB1 1sd UTB1 1sd UTB1 1sd UTB1 1sd UTB1 La 26.4 (1.6) 26.2 (0.6) 23.5 (2.0) 27.4 26.7 INAA n.a. Ce 62.1 (2.8) 60.1 (3.8) 61.4 (5.4) 61.0 6.5 INAA n.a. Nd 33.6 (2.0) 33.0 (4.0) 30.5 (1.2) 36.1 32.0 INAA n.a. Sm (0.18) (0.30) (0.41) 8.0 8.0 INAA n.a. Eu 7.98 (0.22) 8.10 (0.12) 8.28 (0.15) 2.6 2.4 INAA n.a. Yb (0.42) (0.20) (0.40) INAA n.a. Lu 2.30 (0.06) 2.26 (0.04) 2.09 (0.10) 4.13 4.00 INAA n.a. Th (0.06) (0.12) (0.15) INAA n.a. Ta 4.05 (0.06) 3.90 (0.12) 3.81 (0.05) 0.70 0.62 INAA n.a. Hf (0.96) (0.5) (0.2) 4.4 4.6 INAA n.a. U 0.64 (0.18) 0.58 (0.20) 0.59 (0.11) INAA n.a. Co (2.0) (3.2) 1.03 1.02 INAA n.a. Sc 4.21 (1.3) 4.02 4.00 (0.9) 5.1 4.3 INAA n.a. Cs INAA n.a. 0.96 1.03 .83 n.a. 4.9 4.6 5.06 n.a. 0.98 1.01 1.02 47.9 n.a. 49.3 n.a. 39.5 40.0 n.a. n.a. n.a. n.a.

n.a. n(ME--XRF) 4 7 1 18 ------n(TE--XRF) 4 7 1 52 ------n(INAA) 6 8 6 2 ------Exp. -- Expected n --number of analyses n.a. -- not available LOI -- loss of ignition L.L.D. -- lower limit of detection this is most pronounced in the granophyric and granitoid composition of the aplitic granitoids and hornfelsed samples with higher LREE and LILE abundance. The sedimentary rock rafts from within the roof zone of the gabbroic rocks exhibit negative Nb+Ta, P2O5 and TiO2 Kerns Intrusion in Figure 9e. Interestingly, the primitive anomalies, and these anomalies are developed in samples mantle normalised patterns for these rocks are remarkably showing an entire spectrum of LILE and LREE similar suggesting that the metasedimentary rafts are bod- enrichment, including the least differentiated quartz dia- ies of the local Lorrain Formation hanging wall that were base samples and the chilled margin (see Figure 9b). The broken off the roof of the intrusions and incorporated into vari-textured diabase samples also exhibit marked nega- the magma as pendants. In detail, the field information tive Sr anomalies (see Figure 9c), and the granophyric reported in Appendix 1 confirms that these pendants have samples exhibit pronounced negative Eu, Sr, Rb, and K2O undergone considerable baking, and are petrographically anomalies (see Figure 9d), and pronounced negative TiO2 like the overlying Lorrain Formation sedimentary rocks. and P2O5 anomalies. Lightfoot et al. (1989) suggest that The compositional similarity of the aplitic granitoids to the the Rb and K2O anomaly reflects the fractional removal of roof rocks suggests that these aplites are anatectic melts of potassic feldspar, the Sr and Eu-anomaly reflects the re- the Lorrain Formation sediments, and this is consistent moval of feldspar (possibly plagioclase), and the beha- with the field evidence given in Appendix 1 which records viour of P2O5, and TiO2 suggest that apatite and ilmenite, evidence of large amounts of partial melting of the hanging respectively, were removed during fractionation. In detail, wall sedimentary rocks along arkosic layers to produce the chilled diabase and undifferentiated quartz diabase fine veins of aplite which cross-cut the Lorrain both have small negative anomalies for TiO2,P2O5,and stratigraphy. Ta+Nb (see Figure 9a), and Lightfoot et al. (1993) suggest The coherent patterns of variation documented in that these are features of the original parental magma. The Figure 9a-e are perhaps better illustrated on geochemical more pronounced negative anomalies for TiO and P O of 2 2 5 variation diagrams which use a common index of differ- the granophyres would then reflect fractional crystallisa- entiation. Although no single element is an ideal choice, Zr tion of ilmenite and apatite within the sill, and the removal represents a useful compromise. It is readily determined, of these phenocrysts. with precision and accuracy of 5% (2 root mean standard The hanging wall Lorrain Formation of the Huronian deviation), apparently immobile during alteration and metasedimentary rock sequence is compared to the metamorphism, does not readily enter the crystal structure

17 OGS Study 58

of gabbroic mineral phases such as plagioclase and pyrox- of K2O and P2O5 depletion. Depletion of K2O in the grano- ene, and is generally enriched in the residual liquid during phyric gabbro samples is variable, but depletion of TiO2 differentiation. Zr can be concentrated in sedimentary and P2O5 is continuous with increasing Zr concentration. materials, and the abundance is influenced to some extent The abundances of Sm and Zr increase along a 1:1 array by crustal contamination, but more typically abundances (see Figure 10i), but La is enriched with respect to the 1:1 of Zr in crustal rocks are no more than 2-3 times that of the La- Zr array, and Yb is depleted with respect to the 1:1 Yb- parental Nipissing magma. Zr array (see Figures 10h and j). Figure 10k shows that Figures 10a-y illustrate variations in selected major there is pronounced deepening of the magnitude of the Eu- and trace elements versus Zr. For comparison, these plots anomaly with increasing Zr, and this is presumably a func- also show the compositional average of Nipissing diabase tion of feldspar removal. Th and U behave much like La chilled margins (Lightfoot et al., 1989), and the average of (see Figure 10l-m), but Hf and Zr are closely correlated hanging wall Lorrain Formation sediments from above the along a 1:1 array (see Figure 10o) with Zr/Hf close to a Kerns Intrusion (Table 7), and the compositional average primitive mantle value of 38 (Sun and McDonough, 1989). of aplitic granitoids from Obabika Inlet on Lake Ta is tightly correlated with Zr, but is slightly enriched in Temagami (Appendix 2). These data reveal several impor- the samples with highest Zr content relative to the 1:1 array tant features relevant to a petrogenetic understanding of of ideal fractional crystallisation of a magma represented the differentiation-contamination evolution of the in composition by the average chilled Nipissing diabase Nipissing Intrusions. The rocks from within the Kerns Intrusion illustrate tight variations on element versus Zr (see Figure 10n-o). The behaviour of Rb and Ba follows plots, and these trends pass through the compositional av- that of K2O(see Figures 10f, p, and s), and suggest signifi- erage of the chilled diabase. The vari-textured gabbrono- cant removal of potassic feldspar. Sr falls with declining rite and granophyric diabase samples trend to low Mg- Eu/Eu* which is consistent with plagioclase fractionation number at high SiO2 and Zr; they are also poor in CaO and (seeFigures 10k and q). Finally, the compatible elements Al2O3 (see Figures 10a-d). Above about 200 ppm Zr, TiO2, Cr, Ni, Co, V,Sc, and Zn all decline with increasing Zr con- K2O and P2O5falls with increasing Zr (see Figure 10e-g), centration (see Figure 10t-y) suggesting that they are either but the onset of TiO2 depletion slightly precedes the onset removed by fractionating minerals or have their

Figure 5a.Comparison of primitive mantle normalised spidergrams on vari-textured gabbronorite patches and gabbronorite host; samples 86PCL154 (coarse-grained) and 86PCL155 (fine grained) came from one outcrop of vari-textured gabbro in the Emerald Lake gabbro; samples 86PCL161 (coarse-grained) and 86PCL160 (fine-grained) came from one outcrop in the vari- textured gabbro of the Emerald Lake gabbro.

18 Petrology and Geochemistry of the Nipissing Gabbro

abundances diluted by the addition of crustal material common trend of increasing Th/Nb, decreasing Cu/Zr, and which has low abundances of these elements. increasing SiO2 which are consistent with the contamina- Many of the variations documented in Figures 9 and tion of the magma by crustal material rather than the frac- 10 are indicative of the entry of fractionating minerals, but tional segregation of magmatic sulphide. It is not yet clear the behaviour of elements such as Th, U, and the La (see whether this relationship also holds in the more strongly Figures 10h, l, and m), and the compatible elements such as mineralised Nipissing intrusions (Figure 11a-c). Cr,Ni,Co,V,Sc,andZn(see Figures 10t-y) are not readily Nd-isotope data were acquired for a number of sam- explained solely by the fractional removal of silicate ples from the Kerns Intrusion (Table 5) in order to test minerals. Importantly, the compositional average of the whether there is an isotopic record of the assimilation of Lorrain Formation hanging wall sediments is displaced to Lorrain Formation sediments into the Nipissing magma. high Th/Zr, U/Zr and La/Zr and is low in Cr, Ni, Co, V,Sc, Present day Nd-isotopic compositions shown in Figure 12a and Zn, which provides strong evidence that the vari- range from 0.5110 in the aplites and sediments to 0.5122 in textured gabbronorites and granophyric gabbro samples the chilled basal quartz diabase, and these are accompa- are displaced away from the composition of the fractionat- nied by a systematic change in Sm/Nd ratio. The relation- ing Nipissing magma towards the composition of the coun- ship between 143Nd/144Nd and 147Sm/144Nd for samples try rock. Moreover, the compositions of the aplitic gran- from within the Kerns Intrusion does not define a tight iso- itoids from the roof of the Kerns Intrusion and the Obabika chron relationship, and regression of the data for the gab- Inlet region are compositionally similar to the Lorrain bronorites and granophyric rocks define a line with a slope Formation sediments, and it is reasonable to suspect that of 0.016 and an intercept of 0.5094 on 21 analyses (Figure these aplitic rocks are anatectic melts of the sediments, and 12b). This corresponds to a model age of 2.42Ga which is therefore likely candidate compositions for crustal melts significantly older than the U- Pb magmatic baddeleyite generated in-situ above the Nipissing sills, and assimilated ages of the vari-textured gabbronorite sample taken from into the mafic magma (c.f. Lightfoot et al., 1989). the Kerns Intrusion and determined by Noble and In the context of the available data set of all Nipissing Lightfoot (1992). Furthermore, the trend of the data on gabbros (see MRD 19), the Kerns Intrusion rocks fall on a the isochron diagram points towards the analysed

Figure 5b. Comparison of primitive mantle normalised spidergrams on vari-textured gabbronorite patches and gabbronorite host; Basswood Lake Intrusion. Sample 88PCL115 (fine-grained) and 88PCL116 (coarse- grained) came from vari-textured gabbro; 88PCL120 (fine-grained) and 88PCL121 (coarse-grained) came from granophyric gabbro of the Basswood Lake Intrusion. Note the strong enrichment in the incompatible element abundances of the vari-textured rock compared with the finer-grained gabbroic rock.

19 OGS Study 58 compositions of the Lorrain Formation sediments and the intrusion. In this model (Figure 13), the anatectic melts of aplites, which provides further quantitative evidence that the country rocks are assimilated into the magma progres- the Lorrain Formation sediments were assimilated by the sively as the magma differentiates until a compositional Nipissing magma. interface is developed in the magma column which pre- Compositional variations within other strongly differ- vents further assimilation of the anatectic melt of the coun- entiated Nipissing Intrusions define similar variations to try rock, and results in the final crystallisation of an aplitic the Kerns Intrusion. For example, the Basswood Lake granitoid melt at the roof of the intrusion. A more complete Intrusion described in Lightfoot et al. (1993) shows a range documentation of this model is provided in Appendix 2. in La/Zr, Th/Zr, and U/Zr which is consistent with assimi- 3. Dykes and sheets of undifferentiated gabbronorite - lation of crustal sediments. Once again, it is the samples compositional relationships: Two of the undulating with highest Zr and SiO2 which exhibit the largest sills in the Temagami Region are fed by sub-vertical contribution from the sediments. gabbronorite dykes (see Appendix 1 for locations, Lightfoot and Naldrett (1989; 1993) model the varia- geology, petrology, and distribution of samples). tions in the Kerns and Basswood Lake Intrusions in terms These dykes are termed the “Narrows Island dyke” of assimilation coupled to fractional crystallisation of the and the “Sand Point dyke” (Appendix 1). Sampling magma. This model follows the algorithm of DePaolo across each of these intrusions revealed some local (1981) and Taylor (1980) as applied to the process first re- variation in major and trace element concentration cognised by Bowen (1928). In this model, it is the latent which was dominantly controlled by the grain size and heat of crystallisation of the Nipissing magma which is the development of vari-textured gabbronorite responsible for the production of a commensurate amount patches which are locally enriched in incompatible of assimilation of the country rock at the roof of the elements. No strong systematic differentiation across

Figure 6. Location of individual intrusive bodies and detailed study areas referenced in this report. Locations of individual intrusions and references where further information is given: 1) Bruce Mines Intrusion (see Figure 1.8b for sample locations); 2) Basswood Lake Intrusion (see Figure 1.8c and Lightfoot et al., 1993 for sample locations); 3) Casson Lake Intrusion (seeCard,1976,1984).; 4)Sudbury GabbroIntrusions (see Card,1986); 5)Black Lake Intrusion; 6) Wanapitei Intrusion (see Figure1.8d andLightfoot etal., 1993for samplelocations);7)North JanesIntrusion (seeDressler, 1979);8) South Janes Intrusion (see Dressler, 1979); 9) Emerald Lake and Temagami Intrusions (see Figures 1.1 through 1.4); 10) High Rock Intrusion (see Figure 1.1); 11) Slide Rock Intrusion (see Figure 1.1); 12) 20-22 Cobalt intrusions (see Figure 1.8e,Lightfoot et al., 1993,and Conrod,1988 for sample locations); 13) Kerns Intrusion (see Figures 1.5 and 1.6, and Lightfoot et al., 1993); 14) Englehart Intrusion (see Figure 1.8a); 15) -19 Gowganda Intrusions (see Conrod, 1989); and 21) Bonanza Lake Intrusion (see Conrod, 1988).

20 Petrology and Geochemistry of the Nipissing Gabbro

Figure 7a. Chemostratigraphy of the High Rock Intrusion, Lake Temagami.

Figure 7b. Primitive mantle normalised spidergrams of the High Rock Intrusion samples, Lake Temagami. the dykes was found, and there is no evidence for intrusions, and this is reflected in the primitive- variations in hypersthene content across the dykes mantle normalised (Sun and McDonough, 1989) dia- which might reflect flowage differentiation during grams for representative samples (Figures 14a-b). On emplacement. Petrologically and geochemically, the these grounds, these undulatory sills and their feeder dykes are almost identical to the overlying gabbroic zones appear to have been derived from one parental

21 OGS Study 58

Figure 7c. Primitive mantle normalised spidergrams of the High Rock Intrusion samples. Normalisation factors from Sun and McDonough (1989).

Figure 8. Geochemical stratigraphy of the Miller Lake Intrusion, Gowganda. After Conrod (1989). Note the four breaks in stratigraphy which are interpreted to correspond to new pulses of gabbroic magma into the chamber (Conrod, 1989).

magma type which has undergone a limited amount aplitic granitoid is located at Red Rock in the of in-situ differentiation to form vari-textured gab- Obabaika Inlet of Lake Temagami (Appendix 1). This bronorite pods in an otherwise medium-grained aplite consists of a relatively undifferentiated body in gabbronorite. sharp contact with the Obabaika Intrusion of 4. Origin of aplitic granitoids at Obabika Inlet -ana- Nipissing gabbronorite. The aplite is exposed in two tectic melts of Huronian sediments?: A large body of different areas, and each consistsof 500 by 500m wide

22 Petrology and Geochemistry of the Nipissing Gabbro

units with >100m thickness; these aplitic granitoids intrusion was responsible for this melting event, but that appear to grade into dioritic Nipissing gabbro south of assimilation of the aplite was not complete. Obabaika Inlet, and locally these rocks are in sharp 5. Relationships in the roof zone of the Obabika contact with Gowganda Formation shales. The aplitic Intrusion, Temagami - in-situ assimilation of granitoids are strongly granophyric and are similar to Huronian sediments by a Nipissing gabbro sill: The the aplites developed in the roof zone of the Kerns In- Obabika Intrusion west of Lake Temagami and south trusion (Lightfoot et al., 1989). The aplitic granitoids of the Obabika Inlet is described in Appendix 1. The range from granophyric gabbros through to pure intrusions consists of a series of differentiates of quartz-feldspar granitoids (Simony, 1964). The large Nipissing magma which appear to have variably inter- size of these felsic intrusions and the apparent associa- acted with pendants of Gowganda Formation sedi- tion with the roof zone of a Nipissing intrusion sug- mentary rocks which are hosted within the granodio- gests some genetic link, and we therefore investigated ritic and quartz dioritic rocks. Locally, the quartz dio- how the bulk compositions of these aplitic granitoids rites and granodiorites with sedimentary inclusions relate to the composition of the Nipissing gabbros and and rafts contain blebby sulphide mineralisation, but the locally complex quartz diorite-granodiorite-gra- these sulphides appear to be devoid of Cu, Ni, Pt, and nophyre zone at the roof of the Obabika Intrusion in Pd. the context of existing models for the assimilation of The geochemical variations in the chilled diabase, basal roof sediments in the Kerns Intrusion. quartz diabase, quartz diorites, granodiorites are relevant Figure 15a-b compares and contrasts mantle normalised to the relationship of the aplitic rocks and footwall Gow- spidergrams for the aplites from Obabika Inlet with the ganda Formation sediments to the vari-textured gabbros, local Gowganda Formation sedimentary rocks. The aplites quartz diorites, and granodiorites of the Obabika Intrusion. are compositionally quite uniform and similar in incom- Figure 16a-b shows representative primitive mantle patible element pattern to the Gowganda Formation sedi- normalised spidergrams for these rocks. Many of the gra- mentary rocks (see averages in Table 7). This suggests that nodiorites and quartz diorites from the roof of the intrusion this large body of aplite was generated by melting of the are compositionally similar to the aplitic granitoids (c.f. Gowganda Formation sedimentary rocks. In this particular Figure 15b). These granodiorites and quartz diorites may example it is possible that the heat from the Nipissing be Nipissing magmas which are heavily contaminated.

Figure 9a. Primitive-mantle normalised trace element spidergrams for representative samples from the Kerns Intrusion, northwest of New Liskeard. Chilled basal quartz diabase, quartz diabase.

23 OGS Study 58

This is certainly supported by the good field evidence for variation, assuming that the magnetization is primary. The assimilation of Gowganda Formation sedimentary rocks fact that the three remanence directions are so varied described in Appendix 1. suggests that either the directions are not primary, or magmatism occurred over a protracted period of time. DURATION OF NIPISSING The locations of sites studied for paleomagnetic rema- MAGMATIC ACTIVITY AND nence direction within Nipissing intrusions was provided to the author by K. Buchan (personal communication, ASSOCIATED COMPOSITIONAL 1986). During this study, the paleomagnetic drill sites VARIATION (which were originally recorded accurately on Ontario Base Maps) were located in the field. Detailed site infor- Buchan and Card (1985) and Buchan et al. (1989) report mation for the Kerns and Hudson Townships and the remanence directions from undulating sills throughout the Englehart area are presented in Appendix 1 (see Figure eastern portion of the Nipissing Gabbro Province. 1.8). Samples were collected from sites representing each Petrographic studies indicate that fresh diabases and gab- of the three distinctive paleomagnetic signatures - N1, N2, bros carry three characteristic remanence directions. The and N3. Analytical and locational data for these samples N1 remnance is usually up to the north, N2 remnance al- are presented in MRD 19. ways dips down steeply to the west, and N3 is up to the west (Buchan et al., 1989). Buchan and Card (1985) demon- Mantle-normalized trace element variation diagrams strate that N1 and N2 signatures are carried by moderately (Thompson et al. 1984) for samples from sites showing the fresh lithologies. Nevertheless thermal overprinting could average and range in analyses corresponding to each of the not be ruled out. Likewise, Buchan et al. (1989) demon- three remanence directions are plotted in Figure 17a-c. strate that the N3 signature is also present in fresh gabbros Several features are clear from the spidergrams: the most throughout the larger part of a major Nipissing intrusion. trace-element depleted patterns are found in samples Within most sites studied by Buchan and Card (1985)and recording an N1 remanence; N3 sites show slightly higher Buchan et al. (1989), one of the three remanence directions incompatible element levels. N2 sites show the strongest is recorded. Three phases of Nipissing magmatism are overall incompatible element enrichment relative to N1 proposed by Buchan et al. (1989) as an explanation for the and N3.

Figure 9b. Primitive-mantle normalised trace element spidergrams for representative samples from the Kerns Intrusion, northwest of New Liskeard. Hypersthene gabbro.

24 Petrology and Geochemistry of the Nipissing Gabbro

In detail, data for samples from the three sites were way in which sampling was performed within a single dif- plotted on trace-element variation diagrams, using Zr asan ferentiated magma type. Generally, samples from N2 sites index of trace-element enrichment. Figure 17d shows rep- are more differentiated than samples from N1 and N3 sites, resentative plots of selected elements versus Zr. As with but there is a broad compositional spectrum associated data for samples collected within single intrusions (see dis- with each remanence direction. On a gross scale, if the N1 cussion of the Kerns Intrusion), tightly correlated varia- sites are located in sill basins, whereas N2 and N3 sites re- tionsare recorded on all of the trace element variation plots cord locations on the basin-limb and limb-arch region of a (e.g., Figure 10). Of notable importance is the observation sill, then the observed features are consistent with the that N1 and N3 data fall in overlapping fields, and N2 data zonation of lithologies within Nipissing sills. However, falls in a separate field. There is a progressive trend of this does not resolve the problem posed by the different re- overall increase in the abundance of incompatible trace manence directions. To study this problem, U-Pb elements from N1 samples through N3 samples to N2 sam- geochronological studies were undertaken on magmatic ples. The increase in Zr is accompanied by progressive in- baddeleyite from vari-textured gabbronorites. creases in LILE/HFSE (e.g: Th/Yb, La/Zr), such that sam- Until recently, precise U-Pb ages could not easily be ples with higher overall incompatible trace element con- acquired from mafic rocks, but more recently, techniques centrations have higher Th/Yb and La/Yb as well as higher have been developed to date small crystals of baddeleyite. overall LILE and LREE concentrations. Samples with Corfu and Andrews (1986) first reported an age of highest incompatible element concentrations are also de- 2219.4(+3.6, -3.5)Ma on the Nipissing diabase from the pleted in compatible elements (Sr, Ni, Co, Cr, Sc, V), and Castle Mine in Gowganda based on one baddeleyite and have the lowest Mg-numbers. two rutile analyses. Baddeleyite was the most abundant When compared to the data trends of other Nipissing non-magnetic mineral suitable for U-Pb dating purposes intrusions (e.g., Lightfoot and Naldrett, 1989), the trend of and was present in the pegmatoidal facies of the vari- the N1-N3-N2 data is very similar to the normal textured gabbro. differentiation trends of the magma. To investigate the ages of the Kerns and Triangle These features could result from the emplacement of Mountain Intrusions, two suites were collected for U-Pb variably differentiated magmas corresponding to the N1, dating (see Appendix 1 for locations and Table 2 for whole- N2, and N3 events, or, more likely, they could reflect the rock analyses). The samples come from opposite sides of

Figure 9c. Primitive-mantle normalised trace element spidergrams for representative samples from the Kerns Intrusion, northwest of New Liskeard. Vari-textured gabbro.

25 OGS Study 58 the Cross Lake Fault. One (south west) comes from the 2. During the magmatic event, large volumes of center of a quartz diabase, and consists of vari-textured essentially homogeneous magma were generated and diabase; the other comes from the north east of the fault intruded throughout the entire magmatic province. and is a pegmatoidal vari-textured diabase. 3. If the remanence directions are not primary in origin (and this is still not fully resolved), then the paleomag- The sample from the Kerns sill was dated at netic data have little direct bearing on the recognition 2215M3Ma, whereas the sample from the Triangle of multiple phases of magmatic activity. Mountain sill was dated at 2209M4Ma. Full details of the methods and results are presented elsewhere (Noble and Lightfoot, 1992). COMPOSITIONAL VARIATION IN It can be concluded from these data that the three sills THE PARENTAL NIPISSING were emplaced over a relatively short period of time. How- MAGMA TYPE ever, there are resolvable differences in the timing of indi- As part of a regional study, samples of the chill and basal vidual intrusive events. A maximum of 10 million years quartz diabase and gabbro were collected from the least elapsed between the intrusion and cooling of the two mag- differentiated Nipissing gabbros in order to assess their mas. If a period of several hundred million years (as would compositional range. The goal of this was to establish the seem likely assuming that the pole has wandered at a simi- degree of variation in the parental magma composition lar rate to that observed over the past 500 million years), throughout the magmatic province. The importance of this then the slightly different age is apparently at variance with exercise rests in the application of geochemical criteria to the marked difference in the paleomagnetic signatures of identify Nipissing intrusions with unusual geochemical the sites either side of the Cross Lake Fault. Buchan et al. compositions, and to determine whether the geochemical (1989) discuss the implications of the paleomagnetic data compositions of mineralised intrusions are different to in greater depth, but we note the following constraints on apparently barren intrusions. models for temporal evolution of the Nipissing gabbro: Samples were selected of chilled quartz diabase, 1. Nipissing magmatic activity took place over a rela- quartz diabase, and hypersthene gabbro. Mineralogically, tively short time period (less than 10 Ma) in the these rocks consist of approximately similar proportions of Gowganda and Cobalt areas. quartz, feldspar, and pyroxene, with the exception of the

Figure 9d. Primitive-mantle normalised trace element spidergrams for representative samples from the Kerns Intrusion, northwest of New Liskeard. Granophyric gabbro.

26 Petrology and Geochemistry of the Nipissing Gabbro

hypersthene gabbro which has up to 30 modal % Table 6 also shows representative analysis of good hypersthene as a cumulate mineral phase. Geochemical samples of chilled quartz diabase collected at the margins variations due to contamination and extreme fractionation of the Nipissing Intrusions. These analysis arguably were avoided by exclusion of vari-textured gabbros, felsic provide a good index of the parental magma composition gabbronorites, granophyric diabases, and granitoids. With as they were some of the first magmas to crystallise as they the exception of rocks strongly enriched in cumulus hyper- were rapidly chilled against the footwall sediments. The sthene or carrying 1-5% modal sulphide, the variations in value of chilled margins as indicators of parental magma the compositions of the least differentiated rocks may then composition has been questioned, and there is some indi- be taken as an approximation to the parental magma com- cation that the chills of large layered complexes are not position. In the case of the mineralised hypersthene-rich representative of the bulk composition of the intrusion, and rocks, the abundances of the major and trace elements will are sometimes influenced by the assimilation of footwall reflect the modal cumulate hypersthene content, however sediments. However, the data that are shown in Table 6 are the ratios of the incompatible elements provide a robust in- valuable in demonstrating a number of points: dicator of the parental magma composition which can be 1. CHILLED MARGINS: The chilled margin samples compared with the undifferentiated diabase, gabbro, and from the 9 different samples have similar major- and chilled margin samples. trace-element compositions (within analytical uncer- Table 6 summarises the average compositions of tainty; see Table 1). Despite coming from locations selected samples from 20 different intrusions. In some throughout the Nipissing and Temiskaming Regions cases, the average is biased compositionally towards the spanning over 250km laterally, no statistically signifi- hypersthene gabbros (e.g. Kukagami Lake). In the case of cant variation in the compositions of these chills was the Beaton Bay Intrusion (Conrod, 1989), Narrows island found. If these chills recorded assimilation of footwall Dyke, Sand Point Dyke, Duncan Lake Intrusion (Conrod, sediments, then some bias in their compositions to- 1989), Milner Lake Intrusion (Conrod, 1989), and Portage wards the compositions of the footwall sediments Bay Intrusions (Conrod, 1988), the sample suite consists of would be expected, and the amount of displacement in more felsic gabbronorites which are believed to be more the composition of the chill away from the parental differentiated than the typical gabbros as reflected in their magma would be a function of the composition of the overall higher incompatible element concentrations. footwall and the amount of assimilation. As no

Figure 9e. Primitive-mantle normalised trace element spidergrams for representative samples from the Kerns Intrusion, northwest of New Liskeard. Aplites.

27 OGS Study 58

assimilation isevident from the geochemical data, and 2. QUARTZ DIABASE AND GABBROS: A considerably as the chilled margins are devoid of inclusions, we ar- larger data base of quartz diabase and gabbro samples gue that the chilled margin data are of paramount from 21 different intrusions has been assembled in importance in the characterisation of the parental MRD 19 and summarised in Table 6 as averages for magma composition. The average composition of each of the individual intrusions. These data indicate these chills shown in Table 6 reflects the composition that the magma giving rise to all 21 of the studied in- of a uniform magma emplaced into at least seven dif- trusionswasremarkably similar in composition across ferent intrusions. The characteristics of this parental the magmatic province. Moreover, the compositions magma are unusual as Lightfoot et al. (1993) demon- of these samples is remarkably similar to the composi- strated. The Nipissing magma type is silica- rich, with tional average of the chilled margin samples. This elevated LREE/HREE and LILE/HFSE ratios, and suggests that the magma giving rise to many of the in- pronounced negative Ta+Nb, Ti, and P anomalies. trusions of the Nipissing have the characteristics of the These are all characteristics of a magma which has in- chilled margin. Examination of the Ni data and the teracted with a crustal reservoir. Moreover, as recent limited amount of Cu data for these samples suggest models demonstrate the importance of contamination that the concentrations of these elementsin the unmin- (Naldrett et al., 1992; Lightfoot et al., 1994) in the eralised samples are of the order of 80-160ppm, which context of the triggering of sulphur saturation in mafic is not unusual for a gabbro. We do not see any S-poor magmas (Irvine, 1975), it is particularly good news to samples with very low Ni and Cu contents, although a few sampleswith lowCu (6ppm) have been character- explorationists that so many of the Nipissing magmas ised in the Wanapetei Intrusion by Dressler (1982, have this contaminated geochemical composition. At page 63); these values now need to be confirmed in the issue is whether this contamination signature is a fea- context of a search for Ni and Cu-depleted gabbros. ture of the source (e.g. Lightfoot et al., 1993), or whether large quantities of mantle- derived Nipissing Nd-isotopic variation within the basal quartz diabases and magma interacted with the crust in a large deep-crust- chills at 2.22Ga is small; epsilon-Nd values of -2 to -4 are al reservoir of the type proposed by Cox (1980) for found in four different intrusions (see Table 5 and flood basalt settings. Lightfoot and Naldrett, 1989).

Figure 9f. Primitive-mantle normalised trace element spidergrams for representative samples from the Kerns Intrusion, northwest of New Liskeard. Hornfelsed sediment rafts within the granophyric gabbro, and roof sediments. See text and Appendix 1 and Lightfoot et al. (1987) for detailed sample locations and descriptions.

28 Petrology and Geochemistry of the Nipissing Gabbro

Figure 10. Geochemical variations in samples from the Kerns Intrusion, where elemental and oxide abundances are plotted against Zr concentration. For detailed description, see text. a-g major element oxide variations with Zr concentration. h-k Variations in rare earth element (REE) abundances and magnitude of Eu-anomaly with Zr concentration. l-s Variation in incompatible trace element abundances with Zr concentration. t-y. Variation in compatible element abundance with Zr concentration.

29 OGS Study 58

Figure 10. (cont’d) Geochemical variations in samples from the Kerns Intrusion, where elemental and oxide abundances are plotted against Zr concentration. For detailed description, see text. a-g major element oxide variations with Zr concentration. h-k Variations in rare earth element (REE) abundances and magnitude of Eu-anomaly with Zr concentration. l-s Variation in incompatible trace element abundances with Zr concentration. t-y. Variation in compatible element abundance with Zr concentration.

30 Petrology and Geochemistry of the Nipissing Gabbro

Figure 10. (cont’d) For detailed description, see text. a-g major element oxide variations with Zr concentration. h-k Variations in rare earth element (REE) abundances and magnitude of Eu-anomaly with Zrconcentration. l-sVariation inincompatible traceelement abundanceswith Zrconcentration. t-y. Variation in compatible element abundance with Zr concentration.

31 OGS Study 58

32 Petrology and Geochemistry of the Nipissing Gabbro

33 OGS Study 58

34 Petrology and Geochemistry of the Nipissing Gabbro

Huronian sedimentary rocks ( average, n=6, after Lightfoot & Naldrett, 1989) Quartz diorite Aplite LREE depleted granophyric gabbro LREE enriched granophyric gabbro Vari--textured gabbro Hypersthene gabbro Basal quartz gabbro Chill (average, n=6)

Figure 10. (cont’d) Geochemical variations in samples from the Kerns Intrusion, where elemental and oxide abundances are plotted against Zr concentration. For detailed description, see text. a-g major element oxide variations with Zr concentration. h-k Variations in rare earth element (REE) abundances and magnitude of Eu-anomaly with Zr concentration. l-s Variation in incompatible trace element abundances with Zr concentration. t-y. Variation in compatible element abundance with Zr concentration. Figure 11. a) The variation in Th versus Nb falls on a trend between average chilled diabase and an average of sediments from the roof of the Kerns Intrusion; in contrast, the fractionation vector points to much higher Nb.; b) Variation in Cu versus Zr in Nipissing gabbros (all data grouped by rock type). Except for some vari-textured gabbros, the data fall below the ideal fractionation array, and this is ascribed to the assimilation of low-Cu Lorrain or Gowganda Formation sediments by the Nipissing magma.; and c) Variation in Cu/Zr versus SiO2 for all Nipissing gabbro samples (data grouped by rock type). This plot demonstrates that the granophyric gabbros with lowest Cu/Zr have the highest SiO2 content,andthereforedepletioninCuisascribedto assimilation of sediment with low Cu/Zr rather than fractionation of sulphides which have exceptionally high Cu/Zr.

35 OGS Study 58

Table 5. Nd isotope data for samples of Nipissing gabbro. Analyses were performed at the University of Toronto using a clean laboratory and thermal ionisation mass spectrometer as documented in the text.

143 144 147 /144 143 /144 Location/Sample Rock Nd/ Nd Nd Sm Sm Nd Nd Nd0 Type Present Day (ppm) (ppm)

Kerns Intrusion 87--167 H 0.512127 5.43 1.54 0.1710 0.50963 86--169 V 0.511828 20.07 5.06 0.1524 0.50960 86--172 V 0.511670 39.62 9.30 0.1419 0.50959 86--189 G 0.511520 46.18 10.53 0.1378 0.50951 86--186 G 0.511606 12.61 3.04 0.1457 0.50948 86--197F S 0.511096 47.08 7.82 0.1004 0.50949 86--197C S 0.510906 43.66 7.14 0.0988 0.50946 86--198 T 0.511098 56.03 10.26 0.1107 0.50948

Kerns Intrusion (Suite 87--2) Whole rock V 0.51206 9.68 2.63 0.1642 0.50966

Triangle Mountain Intrusion (Suite 87--1) Whole rock V 0.51189 15.41 4.11 0.1612 0.50960

Basswood Lake Intrusion 84--23 V 0.511770 23.30 5.74 0.1489 0.50959 84--14 V 0.511962 12.67 3.44 0.1641 0.50956 84--22 H 0.512051 5.89 1.65 0.1697 0.50957 84--18 H 0.511990 5.63 1.54 0.1652 0.50957 84--11 V 0.511706 14.09 3.36 0.1442 0.50959 84--27 V 0.511508 35.10 8.12 0.1398 0.50946 84--21 V 0.511559 35.10 7.98 0.1374 0.50955 Obabika Intrusion 86--203 S 0.511109 31.87 5.78 0.1096 0.50951 86--073 A 0.511120 30.58 5.74 0.1134 0.50946 86--104 86--100 H 0.512048 8.75 2.47 0.1706 0.50955 C 0.512040 7.36 2.10 0.1725 0.50952

Emerald Lake Intrusion 86--155 B 0.512028 8.57 2.35 0.1658 0.50960 86--156 B 0.512007 19.46 5.36 0.1665 0.50597

High Rock Island Intrusion 86--136 C 0.512156 8.46 2.40 0.1714 0.50965

Portage Bay Intrusion, Cobalt (samples from Conrod, 1988) PB744 -- -- 0.511718 39.40 9.24 0.1418 0.40965 PB745 -- -- 0.511449 25.39 5.62 0.1338 0.50949 PB747 -- -- 0.511834 11.97 3.13 0.1580 0.50952 PB743 -- -- 0.511739 30.45 7.45 0.1479 0.50958

H--Hypersthene gabbro G--Granophyric gabbro B--Basal quartz diabase T --Sedimentary rock inclusion C --Chilled contact diabase A --Aplitic granitoid S --Huronian sediment Suites 81--1 and 87--2 were dated by U --Pb geochronology V--Vari--textured gabbro (Noble and Lightfoot, 1993).

36 Petrology and Geochemistry of the Nipissing Gabbro

On a local scale, chemical variations are more com- the intrusions were emplaced over a period of <10 million plex (see above discussion), but on a regional scale, these years. data suggest that the chilled margins and basal quartz diabases which make up over 25% of most intrusions are compositionally similar. More importantly, these data, like oth- EMPIRICAL OBSERVATIONS er data for many modern CFB, suggest that the composi- RELATED TO MINERAL tional evolution of the parental magma had reached a re- POTENTIAL, LAND USE markably uniform stage throughout the Province at the time of emplacement. Whilst it is possible that the polybar- PLANNING AND EXPLORATION ic fractionation of plagioclase, pyroxene, and olivine may This report presents new data for samples from across the have buffered the composition of the magma to produce a Nipissing Province, and detailed new information for the relatively uniform degree of differentiation (e.g., Cox, intrusions of the Temagami Region which lie east of the 1980), it is hard to see how this magma could retain it’s major Wanapetei gravity and aeromagnetic anomaly and gross homogeneity, unless it was entirely uncontaminated around the Temagami copper deposit. The study docu- by the continental crust as it migrated from the mantle to ments new occurrences of disseminated sulphide mineral- the site of intrusion. This in-turn suggests that the source isation not previously described in the Obabika and Kerns was very homogeneous. . Intrusions, and shows that these late blebs of sulphide asso- If the Nipissing magma was erupted as a uniform mag- ciated with heavily contaminated granodiorites and grano- ma, then it might be reasonable to suggest that most of the phyres are devoid of Ni, Cu, and PGE. However, the study magmatic activity occurred over a relatively short time also highlights a range of empirical observations which span, where common petrological processes controlled the suggest that continued exploration of the Nipissing gab- composition of the magma. This is consistent with the bros, especially along the trend between Whitefish Falls U-Pb geochronology which suggests that at least three of and Temagami is justified. Table 1 summarises these

Figure 12.a) Variation in 143Nd/144Nd versus 147Sm/144Nd in samples from the Kerns Intrusion and local country rocks.b) Relationship of the array of the Kerns Intrusion to isochron lines based on U - Pb geochronology for magmas with a range in initial 143Nd/144Nd isotopic composition. See text and Lightfoot et al. (1989) for further discussion.

37 OGS Study 58

Figure 13. A model for the evolution of the Kerns Intrusion which is applicable to other differentiated Nipissing Intrusions. Stage 1. basic magma is emplaced within the Lorrain Formation stratigraphy, and chilled along the lower margin. Stage 2. Near equilibrium crystallisation of the basal quartz diabase and hypersthene diabase produces latent heat which melts the roof sediments. Stage 3. The less dense aplitic magma is separated from the underlying mafic magma by a double diffusive interface (DDI) whichpermits heattransfer, butlimited chemicaltransfer.Atransition from equilibrium to fractional crystallisation is accompanied by a reductionin thedensity ofthe maficmagma, therebypromoting theconvective erosionof theinterface. Assimilation and fractionation are linked at this stage, but the ratio of the amount of assimilation to the amount of fractionation increases. Blocks of hanging-wall sediment are entrained within the aplitic magma as large rafts. Eventually the interface is almost completely destroyed and the differentiated Nipissing magma and the anatectic melt of the sediments mix freely. In some cases the aplitic melts are isolated from further mixing with the Nipissing magma and crystallise as aplitic granitoids which are geochemically similar to the country rocks. See text for further discussion.

Figure 14a. Primitive mantle normalised spiderdiagrams. Representative samples from the Narrows Island gabbronorite dyke.

38 Petrology and Geochemistry of the Nipissing Gabbro

Figure 14b. Primitive mantle normalised spiderdiagrams. Representative samples from the Sand Point gabbronorite dyke.

Figure 14c. Primitive mantle normalised spiderdiagrams. Comparison of the compositions of gabbronorites from the Sand Point and Narrows Island dykes with the overlying undulatory sills. 39 OGS Study 58

Figure 15a.Primitive mantle-normalised compositions of aplites from the Obabika Intrusion, Lake Temagami. Note the similarity of the pattern to the Gowganda Formation sediments from the eastern contact. Gowganda Formation sediments.

Figure 15b.Primitive mantle-normalised compositions of aplites from the Obabika Intrusion, Lake Temagami. Note the similarity of the pattern to the Gowganda Formation sediments from the eastern contact. Aplitic granitoids. 40 Petrology and Geochemistry of the Nipissing Gabbro

Figure 16a. Primitive mantle normalised compositions of differentiated granodiorites and quartz diorites of the Obabika Intrusion that contain disseminated sulphide. Chilled diabase and quartz diabase.

Figure 16b. Primitive mantle normalised compositions of differentiated granodiorites and quartz diorites of the Obabika Intrusion that contain disseminated sulphide. Quartz diorite from the roof of the intrusion. 41 OGS Study 58

Figure 17a. Primitive mantle normalised spidergrams demonstrating the similarities in geochemical signatures of samples collected from sites retaining three different palaeomagnetic remanence directions in Nipissing Intrusions. Sample sites and magnetic remanence from Buchan and Card (1985), Buchan et al. (1989), and Buchan (personal communication). N1 samples.

Figure 17b. Primitive mantle normalised spidergrams demonstrating the similarities in geochemical signatures of samples collected from sites retaining three different palaeomagnetic remanence directions in Nipissing Intrusions. Sample sites and magnetic remanence from Buchan and Card (1985), Buchan et al. (1989), and Buchan (personal communication). N2 samples.

42 Petrology and Geochemistry of the Nipissing Gabbro

Figure 17c. Primitive mantle normalised spidergrams demonstrating the similarities in geochemical signatures of samples collected from sites retaining three different palaeomagnetic remanence directions in Nipissing Intrusions. Sample sites and magnetic remanence from Buchan and Card (1985), Buchan et al. (1989), and Buchan (personal communication). N3 samples.

Figure 17d. Primitive mantle normalised spidergrams demonstrating the similarities in geochemical signatures of samples collected from sites retaining three different palaeomagnetic remanence directions in Nipissing Intrusions. Sample sites and magnetic remanence from Buchan and Card (1985), Buchan et al. (1989), and Buchan (personal communication).Variations in elemental abundance versus Zr concentration and Th/Yb versus the magnitude of the Eu--anomaly for samples from sites recording three different palaeomagnetic remanence directions. See MRD 19 for complete data.

43 OGS Study 58

44 Petrology and Geochemistry of the Nipissing Gabbro

45 OGS Study 58

46 Petrology and Geochemistry of the Nipissing Gabbro

47 OGS Study 58 points, and we now look at the new data in the context of Kelly-Janes Township Intrusion, west of Sudbury, and on these observations. the mineralised “Sudbury Gabbro” intrusions located Significant close spatial associations between various southwest and south of the Sudbury Igneous Complex. types of mineralisation and Nipissing gabbro intrusions These new studies will focus on whole-rock major and have been recognised (Card and Pattison, 1973; Innes and trace element geochemistry and provide new high quality Colvine, 1984), and there is a variation in style and type of platinum group element data for both mineralised and un- mineralisation across the Nipissing Province. In the east, mineralised intrusions in order to better constrain the Ssat- mineralisation is dominantly Ag, Co and Ni as native met- uration history of the Nipissing intrusions. The data are als, arsenides, and sulfarsenides associated with quartz pertinent to answering a number of questions, for example: carbonate veins which cut the sediments and the Nipissing 1) Were the Nipissing intrusions open system magma intrusions (Jambor, 1971a). In the central portion of the chambers or did they differentiate in-situ? 2) What was the Nipissing Province, mineralisation is dominantly Cu-Ni- S-saturation history of the Nipissing magma type?, and platinum group element (PGE) sulphides which occur dis- 3) Are there geochemical differences between the seminated within the intrusions or as massive pods beneath mineralised and unmineralised silicates in the sills and the intrusions (Rowell, 1984; Rowell and Edgar, 1986; between portions of the sills? Answers to these questions Lightfoot et al., 1991; Lightfoot et al., 1993). In the west- potentially will provide more satisfying criteria which can ern part of the Province, the mineralisation consists of be used in exploration. Cu-sulphides as fine disseminations or in quartz-carbonate veins. CONCLUSIONS A number of empirical observations have been made in previous studies, and one goal of this study is to focus The Nipissing gabbros of the Southern Province were em- attention on these empirical observations as a basis for fur- placed over a short time interval of <10 million years ther exploration. Table 1 summarises some important dominantly into the Huronian sedimentary sequence. The observations about the geology, petrology, mineralogy, parental magma appears to have been remarkably uniform and geochemistry of the Nipissing Gabbro which relate to in chemical composition, but appears to have undergone mineral potential. in-situ differentiation and contamination within the intrusions. A small number of intrusions host heavily dissemi- FURTHER WORK nated sulphide mineralisation rich in Ni, Cu, and PGE. These intrusions are restricted to the Sudbury Region and Detailed geochemical studies are presently under way on form part of a spatial and temporal Ni, Cu, and PGE metal- three Nipissing bodies which host significant disseminated logenic province (Fyon et al., 1995). The challenge is now mineralisation. These rocks were pointed out to the senior to better define the petrology and geochemistry of these author by Gerry O’Reilly, Dan Brunne, Frank Racicot, mineralised intrusions, and compare and contrast their Gord Salo, Jack Rauhala and Brian Wraight who are compositions with the data in this report. The result of this prospectors working or have worked locally in the Sudbury work should be some satisfying new criteria which can be Region. Specific studies are under way on: The used in further evaluation of the mineral potential of the Casson Lake Intrusion, northwest of Whitefish Falls, the Nipissing gabbro.

48 Appendix 1: Sampling, Analysis, Geology, Petrography and Mineralogy of the Nipissing Intrusions

1.1 SAMPLING AND ANALYSIS the East, bounded to the northwest by the Abitibi green- stone-granite terrain, and to the southeast by the Grenville Sampling for geochemical studies concentrated on the fault-thrust zone (see Figure 1a). Gabbros are largely con- location of fresh unmetamorphosed materials which are fined to the 2450Ma Huronian metasedimentary supra- free of alteration veins and jointing in the study areas crustal belt (see Figure 1a), although smaller diabase intru- shown in Figure 6, and described in Appendix 1. Samples sions cut Archean granites, gneisses, metasediments, and were selected to be representative of the rock type; typical- metavolcanics at the margin of the belt. The intrusions ly gabbronorite samples with a grain size of <5mm were form arc-like exposures derived by the erosion of undulat- represented by 2kg samples, whereas 2-3kg samples were ing sills (see Figure 1b). Intrusive contacts generally fol- obtained from vari- textured gabbro and granophyric low bedding planes or unconformities, but also traverse rocks. Sample locations and descriptions are given in the Huronian stratigraphy and the Huronian/Archean MRD 19, figures in text and Appendix 1 and Conrod (1988, unconformity. 1989). Samples were cleaned of weathered and altered sur- The pre-Nipissing stratigraphy consist of Archean faces in the field, and crushed using steel plates and ground granites and greenstones, and Proterozoic sedimentary and to -200 mesh in agate mills. Whole-rock major element ox- volcanic rocks (e.g., Young, 1982; Frarey and Roscoe, ides were acquired by wavelength dispersive x-ray fluores- 1970). The oldest rocks consist of greenstones and gran- cence (XRF) and energy dispersive XRF (Harvey and ites. The greenstones are chloritic and hornblende-rich Atkin, 1982; Potts et al., 1984; 1985). Selected trace ele- schists, representing altered basic volcanic rocks and iron ments were determined by XRF: Nb, Rb, Sr, Y,Zr, Cu, Ni, formations. The schists were folded and intruded by gran- V, and Cr. Rare earth elements, Th, Ta, Hf, U, Sc and Co ite batholiths. Subsequent erosion and deposition formed data were acquired by instrumental neutron activation the Huronian sequence, composed of conglomerates, analysis using the SLOWPOKE reactor and counting facil- quartzites, and slates. A number of volcanic units within ities at the University of Toronto. Quality control was the Huronian stratigraphy have been described by Jolly achieved by stringently monitoring the performance of the (1987a, 1987b). The Nipissing sills were emplaced into the XRF on international reference materials and in-house ref- sediments. erence materials. UTB-1, the University of Toronto in- house basalt standard and WHIN SILL, the in-house Open 1.2.1 Differentiation University INAA standard were used to monitor INAA results, and the results are given in Table 3. The sills are differentiated and exhibit a large variation in petrography and mineralogy. Most have extensive and lat- The new data acquired in this study were used to com- erally continuous basinal regions which are composed pliment published data sets (Lightfoot et al., 1989; 1991; dominantly of hypersthene diabase and vari-textured dia- 1993; Conrod, 1988; 1989; in press). Complete analytical base, whereas adjacent arch zones are composed of coarse data are given in MRD 19 where the samples from different pegmatoidal vari-textured diabase, granophyric diabase, locations are grouped according to location. and in places, aplite (see Figure 1b). There is a ubiquitous Nd-isotope data and Sm-Nd data were determined by fine grained basal quartz diabase along the lower contact conventional thermal ionisation mass spectrometry using a which coarsens away from the chilled margin into the hy- conventional cation exchange column and HDEHP meth- persthene gabbro. Rock types in the sills are easily distin- od modified from Zindler (1980). Samples were analysed guishable in the field, and it is only in the roof zones where on single rhenium filaments using platinum activated car- relationships are less straightforward. bon solution (Noble and Lightfoot 1992). Nd isotopic measurements were monitored, and achieved a total Nd blank 1.2.2 Thickness of <300pg; La Jolla gave 0.511864M13 on 26 determinations, and BCR-1 gave 0.512605M7 on 3 determinations Individual sillsrange in thickness from a fewtens of meters with 146Nd/144Nd =0.7219. Analytical methods used in the to over a thousand meters. The original areal extent of indi- U-Pb study and results are given in Noble and Lightfoot vidual intrusions is difficult to ascertain because of the (1992). structural complexity and variable thickness of the sills (Card and Pattison 1973). Many of the intrusions (see Figure 1a) have elliptical outcrop patterns termed “diabase 1.2 AGE AND DISTRIBUTION: basins”, and may be cone sheets. The 2.15-2.22Ga Nipissing Diabase sills (Corfu and 1.2.3 Emplacement sites Andrews, 1985; Conrod, 1989; Noble and Lightfoot, 1992; Fairbairn et al., 1969; Van Schmus, 1965; Kanasewich and The distribution patterns and orientation of the sills Farquhar, 1965) outcrop over a wide tract of Central may reflect control by pre-existing faults and folds (Card Ontario between Sault Ste. Marie in the west and Cobalt in and Pattison, 1973). Sites of emplacement of the sills also

49 OGS Study 58

appear to controlled by the competency contrasts of the The normally concordant contact of the diabase with sedimentary rocks. Many of the Nipissing intrusions were the hornfelsed Gowganda Formation sedimentary rocks is emplaced within the Lorrain and Gowganda formation exposed along much of the west shore of the Lake, where sedimentary rocks of the Cobalt Group, and in some cases, vertical cliffs rise from Lake level to a maximum of 200m. the contact between these units is a preferred plane of em- Basal contacts are always low-angle (<15_) dipping placement. Lorrain and Gowganda formation conglomer- westward beneath the intrusion. ate and arkose beds commonly overlie the sills, and appear Detailed mapping at the northern margin of the to have acted as barriers to further rise of the diabase Obabika Sill (Figure 1.4)+, in an area burnt-out by fire in magmas. 1977, revealed more petrological variation than the other diabases in this region. Homogeneous aplite in a large out- 1.2.4 Deformation crop at the northwestern edge of the Obabika Sill is apparently related to it. To the southwest, the aplite appears to be The degree of metamorphism, alteration and deformation in sharp contact with quartz diorite and granodiorite, of individual intrusions generally decreases northwards which grades into vari- textured diabase and thence into from the Murray Fault zone. Within the Cobalt embay- gabbro and hypersthene diabase, which is characteristic of ment, sills are virtually undeformed; undulatory upper and the remainder of the southern portion of the Obabika and lower contacts more likely reflect structural controls at the Skunk Lake sills. Within the quartz diorite, granodiorite, time of emplacement than post-intrusive deformation and vari-textured gabbro, there are zones containing horn- (Hriskevich, 1952, 1968). Within the Elliot Lake region, fels fragments (Photo 1.1), blocks of Gowganda Formation folding about an east-west axes has produced striking arcu- sediment which are partially melted along the bedding ate exposure patterns. Folding of the supracrustal belt ap- planes (Photo 1.2), aplitic veins cross-cutting granodiorite pears to have occurred at 1.7Ga during the Penokean orog- and originating within blocks of Gowganda Formation eny (Zolnai et al., 1984). Metamorphic history and loca- sedimentary rock (Photo 1.3), and blocks of aplitic materi- tion in relation to possible domains of unconsolidated sedi- al, as well as associated spotted disseminated sulfide ments at the time of emplacement are discussed in Jackson mineralisation. (1995) and Young (1995). Four other sills around Lake Temagami were given reconnaissance coverage. The High Rock Island Sill southeast of the lake (see Figure 1.1) exhibits excellent ex- 1.3.1 Geology of the Nipissing posure of the basal contact, and a complete sequence of sills and dykes around Lake samples was collected upward from the lower contact Temagami through the basal quartz diabase, hypersthene diabase, and gabbro. The Slide Rock, Devil Bay, and Mount Furgusson The Nipissing Gabbro is well exposed on the shores of sills (see Figure 1.1) were all sampled on a reconnaissance Lake Temagami (“Lake of Deep Water”), where glaciation basis; material was collected from the lower chilled left the contact between the sediments and gabbro well ex- margin of the former two sills. posed near water level on the lake shores (see Figure 6 and All of the Temagami intrusions contain occasional 1.1; locations 4-11; Simony, 1964). Two areas of gabbro, veins of quartz-carbonate at the base. Invariably, these discussed below, were given detailed attention, and a veins carry some mineralisation (dominantly pyrite and further four bodies were studied by reconnaissance. chalcopyrite). The diabase cut by the veins is highly On the western shore of Lake Temagami there is an ex- altered, containing disseminated sulphide mineralisation tensive undeformed and unmetamorphosed area of gabbro (e.g: the Sand Point Dyke, Mount Furgusson Sill, and the (see Figure 1.1). Contacts between the gabbro and Huro- Skunk Lake Sill - see Figure 1.1). nian sedimentary rocks are generally sharp, and strongly hornfelsed Gowganda Formation arkose outcrops next to 1.4 GEOLOGY OF THE KERNS fine-grained chilled diabase at the base of all of the intrusions. SILL Contacts commonly show three-dimensional expo- The Kerns sill is located in Kerns and Hudson Townships sure. Low-angle to horizontal contacts are preserved be- northwest of New Liskeard (Figure 1.5) between the Cross neath all the sills studied for this report, except for portions Lake Fault and the Lake Timiskaming West Shore Fault. of two bodies, which are sub- vertical. One of these trends The sill is apparently undeformed and appears to thicken northwest-southeast from Narrows Island across north- away from the Kerns Rock exposure from about 25m eastern McLean Peninsula to a small group of islands in the (location shown in Figure 1.5 and enlarged in Figures 1.6 southeast (Figure 1.2); this is termed the Narrows Island and 1.7) to reach >100m at the southeastern margin. The Dyke, and is presumably a feeder which joins the base of upper contact of granophyric diabase and upper quartz dia- the diabase sheet (the Skunk Lake sill) exposed on the base against arkosic Lorrain sediments is preserved only western margin of the lake. The other sub-vertical intru- in Kerns Rock and at an isolated exposure to the south sion trends east- southeast from Cayuga Island east of Sand (see Figure 1.5). The southeastern and eastern lower Point towards Long Island (Figure 1.3). This appears to be contact dips west at about 10_. There is no exposure of a vertical dyke which may have fed the Obabika Sill. the lower contact to the northwest. Footwall rocks

50 Petrology and Geochemistry of the Nipissing Gabbro

Figure 1.1.Geological map of the Nipissing Gabbro exposed around Lake Temagami, based on mapping performed by the Ontario Geological Survey (Simony, 1964), and additional work by Lightfoot and Naldrett (1987). Sample locations are shown.

51 OGS Study 58

Figure 1.2. Geological map of the Narrows Island dyke, Lake Temagami (based on Simony, 1964, and additional mapping during the course of this study). Sample locations are shown.

52 Petrology and Geochemistry of the Nipissing Gabbro itional mapping performed during the course of this study). Sample locations are shown. Geological map of the Sand Point dyke, Lake Temagami (based on Simony, 1964, and add Figure 1.3.

53 OGS Study 58 Detailed geological map of the south shore of Obabika Inlet, Lake Temagami, based on new mapping by the authors. Sample locations are shown. Figure 1.4.

54 Petrology and Geochemistry of the Nipissing Gabbro

Photo 1.1. Hornfelsed sediment fragments (Gowganda formation) in quartz diorite at the roof of the Obabika Intrusion, Obabika Inlet of Lake Temagami, Nipissing District (see Appendix 1 for location). This section of the Obabika Intrusion is characterised by many inclusions of sediment which exhibit partial melting along their boundaries. Locally, the partial melts coalesce and dykes (10m long by1.5m wide)of aplitic magma are found cross-curring the quartz diorite of the sill. The inclusions have sharp contacts and well-defined bedding suggesting that the sediment was well-lithified at the time of emplacement of the Nipissing intrusion.

Photo 1.2. Hornfelsed sediment inclusion (Gowganda Formation) in granodiorite in the roof zone of the Obabika Intrusion showing partial melting or injection of aplitic magma along bedding.

55 OGS Study 58

are Gowganda Formation meta-argillites. Archean intrusion, and floated within the anatectic melts at the roof Temiskaming sedimentary rocks are present close to the of the magma chamber. The adjacent granophyric diabase Cross Lake Fault which bounds the southwestern margin of contains patches of fragmented Lorrain sediment (Photo the intrusion. 1.5). Below the upper contact is a discontinuous layer of Structurally, the intrusion hasthe appearance of a limb quartz diabase which coarsens into the vari-textured gab- of a synform with marginal arch and limb regions exposed bro. The vari-textured diabase is in sharp contact with the at Kerns Rock, and limb-basinal portions exposed to the granophyric gabbro, but no contact between upper quartz southeast. diabase and granophyric gabbro was observed. A zone of quartz diorite is present along the western side of the out- The basal chilled margin (Photo 1.6) grades upwards crop; the relationship of this unit to the roof sediments is through basal quartz diabase into hypersthene gabbro. unclear. A limited amount of aplite was found associated Vari-textured diabase, generally, but not always, is over- with brecciated patches within the granophyric diabase. In lain by granophyric gabbro and occurs at the top of the places brecciated Lorrain sedimentary rocks occur at the intrusion. The Kerns Rock exposure provides unparalleled contact, and granophyre or aplite fills the space between three dimensional exposure of the roof zone (see Figures the fragments. 1.5 and 1.6). Granophyric diabase and aplite outcrops below and adjacent to pegmatoidal vari-textured diabase. Close to the upper contact, the Lorrain sedimentary Within 5m of the Lorrain sedimentary rocks at the roof, rocks are pervasively recrystallized; partial melting is large (5m by 5m by 2m) inliers of coarse spotted observed along coarse quartz-rich bands, where aplitic and hornfelsed sediment are found entrained within the grano- quartzitic veins cut through the Lorrain Formation phyric diabase and aplite. These appear to be highly metasedimentary rocks originate in these layers. somatised blocks of Lorrain sediment that retain ghost To the southeast of the Kerns sill, the Diamond sill bands that are interpreted as original layering (Photo 1.4). shows well developed vari-textured diabase overlying ex- These blocks presumably broke away from the roof of the tensive basal quartz diabase and minor hypersthene diabase. To the south of the Cross Lake Fault, the diabase is exposed as a thick sheet (reaching 300m at Triangle Mountain; see Figure 1.5), which is weakly differentiated. This sheet consists of a quartz diabase and a hypersthene gabbro, but no vari-textured gabbro or granophyric gabbro is developed. This intrusion is substantially thicker than the Kerns sill and its lower part presumably represents a well- developed basinal sequence of one of the major Cobalt Nipissing gabbro sills. 1.5 PETROGRAPHY AND MINERALOGY OF THE NIPISSING GABBRO: A generic discussion of the petrographic and mineralogical variation within the Nipissing gabbro appears appropriate, given the similarity in lithologies developed within many of the intrusions. The literature contains a number of descriptions of petrological and mineralogical variations in Nipissing gabbro intrusions (e.g: Hriskevich, 1968; Card and Pattison, 1973; Conrod, 1988; 1989). Quartz diabase occurs within a few feet of the contacts (see Figure 1b). The chill zone consists of a dense aphanitic rock. Phenocrysts of augite and plagioclase are present at 50cm from the contact. The diabase chill consists of plagioclase and pyroxene crystals forming radiating clusters, with long needle shaped plagioclase crystals separated by granular interstitial pyroxene. Opaque oxides occur between the plagioclase laths. Amphibole and chlorite occur as alteration product after pyroxene, and white mica, epidote, and zoisite occur as alteration products after plagioclase. Phenocrysts constitute a very minor portion of the chill (<5%), of which over 80% is plagioclase and the remainder are olivine and pyroxene. Plagioclase pheno- Photo 1.3. Vein of aplitic granitoid originating in a domain ofGowganda Formation sedimentary rock inclusions in the roof granodiorite of the crysts range in size up to 2mm which contrasts with the Kerns Intrusion. groundmass needle-like laths which average 0.2mm in

56 Petrology and Geochemistry of the Nipissing Gabbro

Figure 1.5.Detailed geological map of the Kerns Intrusion,Kerns Township (based on Lovell and Frey,1970) showing the location of sample sitesand the two suite sampled for geochronological investigations.

57 OGS Study 58

Figure 1.6. Detailed geological map of Kerns Rock, Kerns Intrusion, showing the distribution of lithologies and the location of sample sites (after Lightfoot and Naldrett, 1989).

58 Petrology and Geochemistry of the Nipissing Gabbro Sketch map showing relationships between lithologies at the roof of the Kerns Intrusion (after Lightfoot and Naldrett, 1989). Figure 1.7.

59 OGS Study 58

Figure 1.8a. Sketch map showing the locations of samples referred to in MRD 19. Englehart Intrusion (based on regional compilation maps).

Figure 1.8b. Sketch map showing the locations of samples referred to in MRD 19. Bruce Mines Intrusion (after Lightfoot et al., 1993).

60 Petrology and Geochemistry of the Nipissing Gabbro

length. Olivine and augite phenocrysts average about include minor quartz and myrmekite, ilmenite and apatite. 0.5mm, and olivine is always altered to antigorite clouded The upper part of the hypersthene gabbro is a transitional with magnetite. gabbronorite. The basal or lower chill becomes progressively The vari-textured gabbro, as the name suggests, con- coarser- grained into the overlying basal quartz diabase sists of irregular patches of coarser-grained diabase with with an increasing average grain size away from the chill gradational boundaries occurring within finer-grained dia- from 0.1mm to 1mm over 25 to 30m. Plagioclase, augite, base. The vari-textured gabbro usually occurs above the pigeonite, and inverted pigeonite are intergrown in a dia- hypersthene gabbro, and the contact of the two rock types basic texture. Pigeonite occurs both as individual grains is gradational over about one meter. The vari-textured gab- and as rims enclosing augite crystals in crystallographic bro contains plagioclase and pyroxene which are inter- continuity. Quartz occurs as discrete grains and together grown in a diabasic texture. Quartz, K-feldspar, ilmenite, with K-feldspar in micropegmatitic and myrmekitic inter- and apatite occur as accessory phases. Plagioclase is often growths, interstitial to plagioclase and pyroxene. Skeletal altered, and both augite and pigeonite show some alter- crystalsof ilmenite and needlesof apatite occur asaccesso- ation. Patches of pegmatoidal vari-textured gabbro on the ry phases; the apatite tends to be associated with the myr- meter scale are frequently present throughout the vari- mekitic patches (Hriskevich, 1968). Both plagioclase and textured gabbro in most intrusions. More massive areas of pyroxenes are partly altered. pergmatoidal vari-textured gabbro are found close to the The basal quartz diabase grades rapidly into the over- roof of the Kerns sill. lying hypersthene gabbro, which appears to be hyper- The granophyric gabbro occurs in variable amounts sthene-rich in basinal portions of the intrusions and quartz and is usually confined to the arch regions. Plagioclase, au- rich on the limbs and arches. Plagioclase and augite pheno- gite, pigeonite, microcline, K-feldspar, and quartz are the crysts average 0.5- 1mm. Orthopyroxene occurs as equidi- major constituents of the rock, together with accessory mensional subhedral grains reaching 0.5cm in diameter, sphene, ilmenite and apatite. Miarolitic cavities are com- which are slightly altered to antigorite in places. The mon, and reach up to 1cm in diameter. The miarolites have orthopyroxene often encloses euhedral plagioclase laths walls covered in quartz and microcline, and are often filled and augite grains. Olivine occurs in hypersthene diabase at with carbonates. These cavities suggest that the most some locations (e.g. Conrod, 1988), but is either absent, or fractionated phases of the Nipissing were extremely rich in extensively altered to antigorite in all the hypersthene dia- volatile components. Alteration is more intense in the gra- base samples collected in this study. Accessory phases nophyric diabase, where plagioclase is altered to albite,

Figure 1.8c. Sketch map showing the locations of samples referred to in MRD 19. Basswood Lake Intrusion (after Lightfoot et al., 1993).

61 OGS Study 58

Figure 1.8d. Sketch map showing the locations of samples referred to in MRD 19. Wanapitei Intrusion (after Lightfoot et al., 1993).

62 Petrology and Geochemistry of the Nipissing Gabbro

Figure 1.8e. Sketch map showing the locations of samples referred to in MRD 19. Cobalt Region Instrusion (after Lightfoot et al., 1993).

Photo 1.4. Banding in the spotted hornfelsed sediment raftsin theNipissing granophyricgabbro atKerns Rock,Kerns Township(Figure 1.6).Incipient melting of the quartz-rich beds is observed within the raft of sediment as an aplitic phase. 63 OGS Study 58

Photo 1.5. A breccia consisting of fragmented Lorrain Formation sediments within an aplite at the roof of the Kerns Intrusion, Kerns Township. Rock sample is 30cm across.

Photo 1.6. Chilled Nipissing diabase at contact of High Rock intrusion. Field of view 1.5cm.

64 Petrology and Geochemistry of the Nipissing Gabbro

white mica, chlorite, and carbonate. Clinopyroxene is the hypersthene gabbro and vari-textured gabbro into often completely altered to amphibole. Granophyric gab- the granophyric gabbro. bro grades into aplitic granitoids or is in sharp contact with 2. Fresh olivine shows an apparent increase in forsterite aplitic granitoids where the proportion of mafic phases is content away from the lower contact of the Cross Lake lower in the aplitic granitoid than the granophyric gabbro. sill (Cobalt), followed by a decrease up through the The granophyric gabbro is associated with a number hypersthene gabbro. The increase in forsterite content of other lithologies, which constitute only a small propor- is accompanied by an increase in Ni-content. tion of the intrusionsasa whole. Aplite occursas small seg- regations and dyke-like masses within the upper portions 3. The orthopyroxenes show an increase in Mg-number of many intrusions. The aplites contain plagioclase, away from the lower contact towards the base of the microcline, and quartz, and reveal granophyric textures. hypersthene gabbro, followed by a decline in Mg- Mafic minerals are absent; accessory phases are ilmenite number with increased distance away from the lower and apatite. Zoisite, epidote, sphene, and chlorite are sec- contact. ondary minerals. Miarolitic cavities are well developed, The upwards trend towards higher anorthite contents of and commonly filled with carbonate. In most intrusions, plagioclase, forsterite contents of olivine, and Mg-number aplites represent a very small proportion of the intrusions. in orthopyroxene, followed by a reversal about 25-30m However, in the Obabika sill, a vast quantity of aplite is above the contact is common to Nipissing intrusions preserved at the northwestern margin. The relationship of (e.g. Finn et al., 1982), as well as most sills in general the aplites to the rest of the Nipissing magma is presently (e.g: Lightfoot and Naldrett, 1984). unclear. Quartz diorite and granodiorite are preserved within the upper portions of the Obabika sill; these lithologies are distinguished from the granophyric gabbro by the 1.6 METAMORPHISM OF THE larger proportion of quartz and amphibole. NIPISSING GABBRO The sedimentary rocks of the footwall and hanging- wall rocks are typically greywackes, with alternating beds The Nipissing gabbro intrusions have been metamor- of shale and arkose. The arkosic beds are often strongly phosed under regional conditions ranging from lower re-crystallized near the intrusions, and myrmekitic to greenschist chlorite zone alteration to lower almandine micropegmatitic textures are often present. Within the amphibolite facies (Card and Pattison, 1973). In the Kerns Intrusion, blocks of hornfelsed sediment have been Cobalt-Gowganda and Sault Ste. Marie- Elliot Lake areas, strongly metasomatised, such that patches of feldspar are metamorphic grade is uniformly low, whereas in the separated by a matrix richer in amphibole. Sudbury area, the grade is much higher (Card and Pattison, A more complete discussion of mineralogical and 1973). mineral chemical information is given in Lightfoot et al. Metamorphism is most evident at the margins of intru- (1986, 1987) and Conrod (1988; 1989). The following sions. The change of diabase to metadiabase involves the general points are stressed here: replacement of pyroxene and calcic plagioclase by amphi- 1. There is an increase in the anorthite content of plagio- boles, sodic plagioclase, epidote, talc, chlorite, and quartz. clase away from the chill margin up to the base of the For a more detailed account of metamorphic mineralogy, hypersthene gabbro, followed by a decrease through see Conrod, 1988.

65 Appendix 2

2.1 ASSIMILATION AND 1. Parent magma composition: Data for chilled margins suggest that the Nipissing parental magma was FRACTIONAL CRYSTALLISATION very uniform in composition (see Lightfoot et al., IN THE KERNS INTRUSION - A 1993), and that variable amounts of fractional crystallisation and crustal contamination at the time of em- CASE STUDY OF THE PHYSICAL placement are not recorded in different intrusions. Al- PROCESS though there appears to be some support for three phases of intrusion recorded in the remanence paleo- It has been suggested that crystallisation and assimilation magnetic signatures (Buchan et al.,1989), these three in basic magma chambers are coupled (DePaolo, 1981; remanence components are found in rocks which are DePaolo, 1985; Bowen, 1909, 1910, 1928; Taylor, 1980), easily linked by differentiation of a single parental where the latent heat of crystallisation of the magma pro- magma and the U-Pb geochronology suggest that the duces a commensurate amount of assimilation of the roof intrusive activity covered an interval of significantly rock sediments. Although there are some fairly compelling less than the 50 million years required by paleomag- geochemical arguments for a spatial and temporal cou- netic data. Thus, the paleomagnetic data appear not to pling of these processes (e.g., Taylor, 1980; DePaolo, exclude the possibility that a single parental magma 1985), analogue laboratory investigations of simple binary composition may be represented by the chilled margin salt solutions (Campbell and Turner, 1987) suggest that differentiated to produce the compositionally diverse diffusive interfaces may develop in the magma column Nipissing rocks. between molten roof rock and underlying basic magma. This interface may allow heat transfer, but prevent chemi- 2. Partition coefficients: Distribution coefficients are cal diffusion, resulting in the accumulation of a pond of sensitive to the physical and chemical character of the less dense crustal melt above the interface. Thermal ero- magma. Given the tholeiitic affinity, it is suggested sion by convecting basic magma may progressively break that low-pressure basaltic values are appropriate down the interface and mix more fractionated basic mag- (Henderson, 1983, 1984; Irving, 1978; Pearce and ma with crustal melts. The presence of volatiles in the Norry, 1979), although their magnitudes may change magma at the roof of the intrusion may also contribute during the evolution of the magma (Henderson, 1983, slightly to the amount of melting. If the interface breaks 1984). down rapidly, then a component of the variation may re- 3. Closed system evolution: Multiple influx of new flect direct mixing of magmas which have similar densi- batches of magma of different composition appear un- ties. In such circumstances, the rate of change of magma likely in most intrusions, given the uniform composi- mass due to assimilation relative to that due to fractionation of the chilled margin samples. Although Finn tion (r) is not necessarily constant, and the algorithm (1981), Finn et al. (1982), and Finn and Edgar (1986) derived by DePaolo (1981) for assimilation-fractional suggest that multiple intrusions of magma are crystallisation (AFC) must be solved accordingly. recorded in the Wanapitei intrusion, this could repre- Lightfoot and Naldrett (1989) explored the role of sent a sampling problem. However, neither loss of crystallisation and assimilation from a geochemical per- magma to higher crustal levels nor assimilation of spective in the Kerns Intrusion, and were able to demon- overlying crust can be ruled out as indicated by the strate that assimilation and crystallisation were in-part data of Conrod (1989) for the Miller Lake Intrusion at coupled. In this report, we use new data for the Kerns Gowganda. Petrographic evidence presented in Intrusion and expand our discussion of the roles of crystal- Appendix 1 suggest that melting of the roof did occur, lisation, assimilation, and mixing. Lightfoot and Naldrett but no evidence for large feeders tapping the top of any (1989) pointed out that a significant proportion of the Ni- of the Nipissing gabbro intrusions was found. pissing lithologies were either undifferentiated or slightly 4. Cumulus processes: There is no evidence for strong differentiated. They suggested that equilibrium crystal- cumulus processes of the type found in major layered lisation may have played an important role during a large intrusions (e.g: the Bushveld). Rather, the cumulus part of the solidification histories of the sills. The slight en- enrichment that has occurred is confined to enrich- richment of some gabbros in hypersthene was considered ment of the basinal regions in hypersthene, and is to be the product of small degrees of cumulus enrichment. reflected only in a slight depletion of the hypersthene However, equilibrium crystallisation fails to reproduce the diabase in Zr (40-65ppm) and other incompatible styles of variation found in the more differentiated elements relative to the chilled margin (Zr=65ppm). lithologies. This contrasts with Zr levels close-to or below detec- The effects of Rayleigh Fractional Crystallization tion limitsfor the picritic layers of the larger intrusions (RFC) may be modelled using the Rayleigh Equation associated with CFB (e.g: Lightfoot and Naldrett, (e.g., Wood and Fraser, 1976), where the following 1984). Re- equilibration appears to be limited by the constraints are available: very slow diffusion rates found in gabbroic phases.

66 Petrology and Geochemistry of the Nipissing Gabbro

Rayleigh fractionation models are shown in Figure roof of the intrusion, or within the vari-textured diabase. 2.1, based on Lightfoot and Naldrett (1989) and new data However, enrichment in K2O, Ba, and Rb within the vari- presented in this study. Although fractionation is clearly of textured gabbro are likely to reflect, at least in part, con- significance, it is apparently not the sole process responsi- tamination prior to K- feldspar entry. Perhaps a more rea- ble for chemical variations (Lightfoot and Naldrett, 1989), sonable explanation for the location of the K-feldspar is and for this reason, any conclusions drawn are better based within the aplites at the roof of the intrusion, as these rocks on the simplest of RFC models which identify the nature of contain slightly higher K2O abundances than the Lorrain phases, but not their exact proportions. Formation sedimentary rocks. As discussed by Lightfoot and Naldrett (1989), much The low LREE concentrations of some of the grano- of the differentiation is explained by the removal of pyrox- phyric diabase samples compared to other samples with ene, plagioclase, olivine, apatite, K-feldspar, and ilme- similar Zr contents may suggest the late fractionation of a nite. Summarising the variations in Figure 10a-y, the LREE-rich phase. Henderson (1984) has pointed out that following observations are important: allanite has a LREE-enriched REE profile, and appearsnot 1. Plagioclase fractionation explains the progressive to preferentially enriched in Th or U. This phase has been decline in Sr with increasing Zr and magnitude of the observed in some Nipissing intrusions (Jambor, 1971), and Eu-anomaly (see Figures 10k). Hypersthene diabases has been suggested by Lightfoot and Naldrett (1989) as tend to be Sr-poor due to enrichment in cumulus hy- one possible explanation for the low LREE contents of sev- persthene, whereas the chill and basal quartz diabase eral samples of granophyric gabbronorite from the Kerns have essentially similar Sr contents. The overall Intrusion. decline in Sr content is consistent with a progressive Following Lightfoot and Naldrett (1989) the variation increase in the plagioclase content of the fractionating in LILE/Zr and LREE/Zr ratios are insensitive to the phase extract. effects of crystallisation. The crystal phase extract must contain about 50% augite in order to explain the rapid in- 2. The rapid decline in Ni (see Figure 10u) is partially crease in the LILE/HFSE with Zr content. Such a large explained by the removal of pyroxene. However, the component of augite in the crystal phase extract is incon- country rocks are also low in Ni, and so some of this sistent with the other trace element data and petrographic may reflect the assimilation of material from low Ni evidence. Lightfoot and Naldrett (1989) also note that the country rocks into the Nipissing magma. granophyric diabases have more negative epsilon-Nd than 3. VariationinV(see Figure 10w) is largely accounted the other diabase lithologies, suggesting that a contribu- for by the proportions of hypersthene to augite. tion from the crust is required. To test this hypothesis, the Vanadium enters hypersthene more readily than au- variations in Th/Yb, La/Yb, and U/Yb are shown in Figure gite, and as the hypersthene diabases have higher V 2.2a-b for samples from Lightfoot and Naldrett (1989) and than the basal quartz diabase or chill, some cumulus this study. These ratios are very sensitive to mixing enrichment is suggested. processes, but less sensitive to fractionation.

4. The plot of TiO2 versus Zr shown in Figure 10e The basal quartz diabase, hypersthene gabbro, vari- suggest that ilmenite begins to crystallize after 60% textured gabbronorite, and granophyric gabbro data define solidification of the magma. trends of coupled overall increase in inter-element ratios away from the chilled margin towards the compositional 5. The plot of P2O5 versus Zr, shown in Figure 10g field of the hanging wall Lorrain Formation sedimentary indicates that apatite enters after 75% crystallisation rocks, the rafts of hornfelsed Lorrain sedimentary rock of the magma. within the granophyric gabbro, the quartz diorite, and the 6. The effects of albite, K-feldspar, and quartz within the aplites. This data may be interpreted to suggest that Lor- crystalline phase extract on the compositional rain Formation sedimentary rocks have been involved in changes within the vari-textured and granophyric Nipissing petrogenesis, and that the progressive increase diabases are not straightforward, as crustal melts con- in the element ratios attest to stronger contamination as the tribute SiO2 and LILE to the magma. However, magma became more fractionated. Figures 10g and p demonstrate that enrichment of the Importantly, whilst mixing is recorded on element- magma in K2O and Rb is found only within the less ratio plots, the element-element plots record at least two fractionated vari-textured diabase samples, after processes - mixing between Nipissing magma and crustal which K2O and Rb decline strongly with increasing material, and fractional crystallisation. The element Zr. In contrast, Figure 10l demonstrates that Th con- versus Zr plots illustrate that the two processes have acted tinues to increase with increasing Zr throughout the in such a manner as to produce coherent variations (see vari-textured gabbronorite and granophyric gabbro. It Figures 10a-y). Clearly this would not be expected if mix- is suggested that this may reflect the entry of K-feld- ing were to be superimposed on a range of magmas show- spar within the crystal extract, late in the crys- ing different relative amounts of fractionation, or by frac- tallisation of the vari-textured diabase. tionation superimposed on a range of magma compositions Figure 10f show the effect of removing K-feldspar after produced by mixing. 75% fractionation of the original magma. The location of For example, variations in Th, U, and La versus Zr the K- feldspar extracted, may rest in aplitic selvages at the (seeFigures 10h, l, and m) define tight trajectories (with the

67 OGS Study 58

Figure 2.1. Modelling of the crystallisation and assimilation history of the Kerns Intrusion. a) Sr versus Zr and the effect of plagioclase fractionation.; b) Ni versus Zr and the effect of orthopyroxene fractionation.;c) Magnitude of the Eu-anomaly versus Zr and the effect of the fractionation of gabbroic minerals.; d) V versus Zr and the fractionation of pyroxenes.; e) TiO2 versus Zr and the fractionation of ilmenite.; f) P2O5 versus Zr and the fractionation of apatite.; g) K2O versus Zr and the fractionation of potassic feldspar.; and h) Rb versus Zr and the fractionation of potassic feldspar. 68 Petrology and Geochemistry of the Nipissing Gabbro

exception of LREE depleted granophyric diabases), in the proportions 5:60:20:15 was assumed. Additional pa- towards higher Th/Zr, U/Zr, and La/Zr when compared to rameters required in the models were the average composi- 1:1 enrichment trajectories. Similarly, the HFSE (e.g: Y, tion of roof rocks at the top of the Kerns sill, and the value Yb) fall with increasing Zr (see Figure 10j and r). Consider, of the ratio of the rate of change of magma mass due to as- for example, the LILE versus Zr plots; Figure 2.2 demon- similation relative to that due to fractionation (r). In strates that 1:1 enrichment vectors originating at different DePaolo’s models, this factor was a constant; here we points on a mixing line drawn between hypothetical ele- varied it within reasonable limits (0.1 to 0.4), which take ments A (e.g: Th) and B (e.g: Zr) are unlikely to fractionate into account the maximum amount of latent heat and from the mixing curve between Coand Ca to the data trajec- superheat available for melting at the roof of the intrusion. tory Co-E. Indeed, to the contrary, a complete scatter of the The modelled trajectories fan away from the composi- data would be expected unless some physical process tion of the parental magma towards progressively higher coupled assimilation and fractionation. It is suggested here LILE/Zr and lower HFSE/Zr as r increases from 0.1 to 0.4 that assimilation and fractionation worked hand-in-hand to (c.f. Lightfoot and Naldrett, 1989). A cursory examination explain why the least fractionated samples are least con- of the trajectories for the models suggests that the general taminated, whilst the largest amount of contamination is characteristics of some of the data trends are followed fair- recorded in strongly fractionated samples. ly well by the AFC lines; i.e: the increase in La/Zr, Th/Zr, Coupling between assimilation and fractional crystal- and U/Zr is predicted by the models, the constant Sm/Zr ra- lisation may be modelled using the algorithms of DePaolo tio is reproduced, and the fall in Yb/Zr (see Figure 10j) is (1981). Models of the type constructed by Lightfoot and predicted. Furthermore, the range of r values (<0.4) is not Naldrett (1989) are shown in Figures 2.3 and 2.4. Assump- unreasonable for the Kerns Intrusion. Poor fits for some of tions regarding crystalline phase extracts were similar to the granophyric diabase samples may be explained by the those used in RFC modelling, except that a phase extract removal of accessory minerals with high LREE/HREE composed of olivine, plagioclase, augite, and hypersthene ratios.

Figure 2.2. a) Variation in Th/Yb versus La/Yb in the Kerns Intrusion sample suite of this study and Lightfoot and Naldrett (1989). and b) Variation in U/Yb versus Th/Yb (Lightfoot and Naldrett, 1989).

69 OGS Study 58

Looking in more detail at the results of the modelling the element versus Zr variation diagrams, suggesting that in Figure 2.4, Lightfoot and Naldrett (1989) noted that the the observed variation is not entirely consistent with the observed data trajectories cross-cut the modelled trajecto- conventional AFC model. Whereas it is possible that dis- ries on each plot. The chill, basal quartz diabase and hyper- tribution coefficients change through magmatic evolution, sthene diabase fall close to the origin. The vari-textured they are likely to change by similar amounts, and therefore diabase samples fall on a trajectory which is close to r=0.1; fractionation of the Th/Zr ratio is unlikely. More realistic is the granophyric diabase field cross-cuts the trajectories the possibility that the value of r was not a constant, as sug- with r=0.2 to r=0.4. This relationship is preserved on all of gested by Lightfoot and Naldrett (1989). Assimilation of

Figure 2.3.Effects of assimilation and fractionation compared to mixing on schematic showing bi-variate plots of incompatible elements. Simple mixing of a parental magma (Co) with a crustal component (Ca) would produce a trend between Co and Ca. Fractionation of a mixture (such as x or y) will produce liquids falling close to 1:1 enrichment vectors (y-y’, x-x’). In fact, the data trajectory (Co- E) would require that the parental magmas become more fractionated when more crust is added. This requires that assimilation and fractional crystallisation are coupled, at least in part, by some mechanism which permits the most contaminated magmas to exhibit the largest degree of fractionation. 70 Petrology and Geochemistry of the Nipissing Gabbro

Figure 2.4. a) Modelling of assimilation coupled to fractionation on Th versus Zr. b) Modelling of assimilation coupled to fractionation on La versus Zr, c) Modelling of assimilation coupled to fractionation on U versus Zr.

71 OGS Study 58

melts at the roof of the intrusion may not take place at a rate suggest that coupling both spatially and temporally, similar to that at which latent heat is supplied to the roof of between assimilation and fractionation is rarely complete- the intrusion. For example, Lightfoot and Naldrett (1989) ly achieved. Campbell and Turner (1987) report a series of suggest the possibility of a thermal boundary layer which experiments designed to model melting at the roof of a might act as a double-diffusive interface permitting heat basaltic magma chamber, where melts of granitic and gra- transfer but no chemical transfer. In such a circumstance, nodioritic composition are produced from country rock; melted roof rock may pond as less dense melts of crustal these melts are less dense than the basic magma. Based on material above the basic magma. Assimilation of these their experiments on simple salt solutions, they demon- melts may not occur until a late stage in the differentiation strate that such light magmas would tend to pond under the of the sill, and the break- down may then be progressive roof of the intrusion (c.f. McBirney, 1979). Furthermore, and non linear. they demonstrate that interaction between the basic magma and the roof-rock melts would depend on the shape of To expand on this suggestion, the effects of varying r the intrusion. Steep sides of the intrusion would promote during the evolution of the magma are modelled where r mixing, whereas a shallow slope (such as that found at the increases from 0.1 to 0.4 at 250ppm Zr, and from 0.4 to roof of a sill) would not promote significant amounts of 0.75 at 300ppm Zr. In reality, modelling this process in mixing. The granitic or granodioritic melts would be sepa- three steps is probably an oversimplification, but it serves rated from the underlying basic magma by a double diffu- to illustrate that the curved form of the data trend could, as sive interface which permits heat transfer at a rate that isan Lightfoot and Naldrett (1989) point out, be generated by order of magnitude faster than the rate of chemical diffu- progressively increasing r as fractionation proceeds. sion. Thus, there may be little transfer of mass between the Whether or not an r value of 0.75 is realistic must be placed two layers in the absence of any strong disruptive effects in considerable doubt, as the amount of latent heat avail- (e.g: convecting of the magma). Consequently, according able during the later stages of crystallisation was probably to Campbell and Turner’s results, assimilation and frac- minimal. However, Jolly (personal communication) sug- tionation are likely to be decoupled both spatially and gested that partial melts produced at the roof of the intru- temporally. sion may be mixed into the fractionated basic magma, thereby reducing the AFC mechanism to a simple late- Campbell and Turner (1987) note that although the stage mixing process. This possibility clearly requires amount of mass transfer across the double-diffusive inter- further investigation. face will normally be small, a number of different processes can enhance the mass diffusion rate from the Geochemical trends indicate that both fractionation upper to the lower layer: and contamination have played key parts in Nipissing petrogenesis. It is therefore worth examining theoretical and 1. Turbulence enhances mixing across the interface. experimental data relevant to AFC. Several factors come This can occur when the basic magma is vigorously into play during AFC. Bowen (1928) suggested that the convecting (i.e: high Rayleigh number), and latent heat of solution of silicate minerals is small, and thermally erodes the base of the upper layer. This is therefore a magma would have to be superheated by at likely if the basic magma is thick, non-viscous, and re- least 300_C in order to assimilate an equivalent mass of peatedly injected with new batches of magma (c.f. country rock preheated to its melting point. Although su- Huppert and Sparks, 1980). Such conditions are perheat provided by volatiles may play an important role, generally prevalent in basic layered intrusions and the realization that the latent heat of crystallisation was thick sills (e.g., Insizwa - Lightfoot et al., 1984). greater than Bowen suggested, led Taylor (1980) and later 2. Influx of a less dense batch of magma into the intru- DePaolo (1981) to suggest that AFC could be achieved sion can produce extensive mixing (c.f. Huppert and without the need for substantial superheating of the mag- Sparks, 1980), and under such circumstances it is ma. This became increasingly evident when masses of gra- possible that an interface may never develop nophyric rock above major layered intrusions were recog- (Campbell and Turner, 1987). In the absence of chilled nized as possible melts of the roof rock (e.g: the Bushveld material from the different magma influxes, it is not Complex and the Muskox Intrusions, Irvine and Smith, straightforward to evaluate whether magmas actually 1969; McBirney, 1979). The latent heat of crystallisation did mix. of the vast thickness of cumulates in these layered intrusions supplied considerable heat to the roof of the intru- 3. Blocks of roof material may fall through the upper lay- sion. Volatiles present in the roof zone of the Kerns sill, as er into the lower layer, dragging down some of the evidence by vesicular granophyric diabase and aplite, may lighter magma. Additional melting of the blocks will have slightly enhanced the ability of the magma to heat the further contaminate the lower layer (e.g: Huppert and overlying crustal materials to towards their liquidus point, Sparks, 1980, and field evidence from this study). producing melts. 4. During the normal crystallisation of basic intrusions, DePaolo’s (1981) AFC equation is now universally evolved liquidswith densitiessimilar to those of crust- applied to AFC modelling of volcanic suites, and DePaolo al melts may be generated. The interface between (1985) also used this equation for modelling magmatic such liquids can therefore not be maintained by densi- evolution in a layered intrusion - the Kiglapait. However, ty stratification, and will likely break down leading to new experimental data (Campbell and Turner, 1987) convective overturn at the roof of the intrusion.

72 Petrology and Geochemistry of the Nipissing Gabbro

5. Diffusive interfaces are unlikely to remain during the sediment, and convection continued to break down the later stages of crystallisation, as strong convection double-diffusive interface; when well over 90% crystal- within a limited vertical thickness of magma will lisation of the basic magma had been achieved, the hyper- likely produce significant convective mixing. sthene diabase phase was completely crystallized, whilst the interface between fractionated residual magma and 6. Finally, as Campbell and Turner (1987) point out, the overlying aplite continued to break down. The crystallisa- roofs of magma chambers are rarely flat, and steeper tion of the vari- textured diabase appears to have occurred sides would promote turbulence. If assimilation ex- from slightly more contaminated magma than that respon- ploits weaknesses such as faults, bedding planes, or sible for the production of the basal quartz and hypersthene unconformities in the sediments, thereby producing diabases(evidenced by the higher LILE/HFSE ratio), and an uneven roof, then undulations in the form of yielded additional latent heat (stage III) to the Lorrain. The stepping of the upper contact are likely. This has been final phase of crystallisation took place as the densities of demonstrated in the Kerns sill, where the upper con- the aplitic magma and the fractionated Nipissing magma tact isvertical in some locations, and horizontal in oth- closely approached one another (stage IV), and this marks ers. Furthermore, the presence of large blocks of the break-down of the double-diffusive interface, perhaps spotted hornfelsed Lorrain sediment suggest that the followed by complete mixing of the two magmas (except upper margin of the intrusion was migrating away material caught in pockets at the upper contact). Blocks of from the basic magma, as pendants of roof rock were country rock broken away from the roof were strongly me- split away from the roof and fell into the upper zone of tasomatised throughout stages II to IV. The presence of the intrusion. limited amounts of aplite in the Kerns sill reflect the strong mixing during the final stages of fractionation. Both the geochemical trend patterns in the Kerns sill and the physical constraints on processes which could produce Aplites within the Obabika sill are very well devel- these variations were used by Lightfoot and Naldrett oped over an area exceeding 5 square kilometers. Thiscon- (1989) to construct a schematic model for the evolution of trasts with the occurrence of small pockets of aplite in the the magma. Figure 13 (based on Lightfoot and Naldrett, Kerns sill. The presence of widespread aplites at the roof of 1989) illustrates one possible interpretation of the above the Obabika sill in Nipissing District (see earlier), and the observations in the context of a larger scale convecting in- absence of both vari- textured and granophyric diabases trusion as modified from Conrod (1989) and illustrated in suggests that the double- diffusive interface between basic Figure 2.5. During the first of the four stages envisaged, the magma and aplitic magma was well established Nipissing magma intruded along a weakness between the throughout the entire crystallisation history of the magma. Lorrain and Gowganda Formation sedimentary rocks. The The location of the zone of crustal melting in the magma was chilled along both contacts. Cooling of the Kerns sill above what seems to be the limb- and arch-zone magma was accompanied by crystallisation of the basal is apparently at odds with the stronger contamination ex- quartz diabase and hypersthene gabbro. The variation in pected above the basinal regions of the intrusion, where texture, grain size, phase proportions, and whole-rock geo- there is a larger proportion of basic (and hence, dense and chemistry away from the lower contact is attributed to the hot) magma (see Figure 2.5). There is no field evidence to more rapid cooling of the magma closer to the contact, and suggest that the undulations in the Kerns sill reflect de- the progressive enrichment of the magma in hypersthene formation after emplacement. Feeder zonesmay be hidden within the hypersthene diabase. It is presently unclear why beneath the Kerns sill; although their location is unclear, the An content of the plagioclase and Mg- number of the they would likely have produced turbulent mixing in the pyroxene appear to increase towards the base of the hyper- magma chamber, thus preventing the establishment of a sthene diabase unit. Similar trends are found in other major double- diffusive interface. layered intrusions and sills (e.g: Insizwa - Lightfoot and Basinal regions represent areas of magma entry Naldrett, 1984), and their origin is also unclear. During (Lightfoot and Naldrett, 1989), where hot, turbulent mag- stage II, crystallisation of the hypersthene diabase unit was ma forms a chamber and may commence partial melting of accompanied by the release of latent heat, and this was the roof sediments. The absence of granophyric diabase accompanied by the migration of volatilestowards the roof and aplite along the upper margins of many basinal por- of the intrusion. Heating of the roof produced a commen- tions of Nipissing intrusions (e.g: Hriskevich, 1968), and surate amount of melting of the Lorrain formation hanging the presence of an upper chill and upper quartz diabase ar- wall sediments. The crustal melts were driven under pres- gues that little, if any melting of crustal material or accu- sure from the arkosic low- melting point beds within the mulation of fractionated magmas has taken place in these Lorrain through cracks, and accumulated at the top of the zones. This is at odds with the expected relationship, which magma chamber as a pond of less dense aplitic magma. in turn undermines some of the thermal implications of the Some mixing between aplitic magma and basic magma ac- model presented above. Conrod (1988) suggested that less companied convective erosion of the double-diffusive dense fractionated magmas may migrate to the arch zones interface, but a large proportion of the aplite was retained of intrusions, whilst more dense less fractionated magmas in a physically decoupled sector of the magma chamber may accumulate in the basinal portions of sills. This pro- separated from the less dense basic magma by a double- cess would also concentrate volatiles and heat at the roof diffusive interface. Continued crystallisation of the Kerns above the arch zones. If the arch zone was stagnant and sill magma resulted in more melting of Lorrain Formation undergoing little convection, then ideal condition would

73 OGS Study 58 Model for the evolution of a Nipissing intrusion (after Conrod, 1989). Figure 2.5.

74 Petrology and Geochemistry of the Nipissing Gabbro

exist for the development of a double-diffusive interface magma column in Ni and Cu which can be ascribed to the between the aplitic and basic magmas. However, strong in-situ segregation of an immiscible sulphide liquid. convection above basinal zones would normally produce The variation in Cu versus Zr concentrations in sam- extensive melting of the roof; perhaps such melts are ples are shown in Figure 11b. This plot demonstrates that rapidly assimilated into the magma, and the upper quartz samples with elevated Zr (which we now shown is coupled diabase, does not develop until a later stage in the to elevated Th/Yb, La/Yb, and SiO2 with low Ni) also have crystallisation of the magma. very low Cu contents. In the context of Siberian Trap lavas at Noril’sk which are depleted in Ni, Cu, and PGE, the Cu/ The evidence that there has been significant in-situ Zr ratio of the high-Zr samples are comparable to the most crustal contamination of Nipissing Intrusions is very Cu depleted Nadezhdinsky Formation lavas at Noril’sk strong based on the geochemical data presented above for (Lightfoot et al., 1990; 1993; 1994). However, this low Cu/ the Kerns Intrusion. Moreover, studies of other Nipissing Zr ratio is accompanied by very high SiO2 (65-80 weight % intrusions have demonstrated that the most differentiated SiO2 - see Figure 9b) - much higher than any of the Nadezh- phases of intrusions are significantly contaminated by lo- dinsky Formation basalts at Noril’sk (56 weight % SiO2 cal country rocks of the roof zones of the intrusions maximum). The elevated SiO2 of the samples from the (Lightfoot et al., 1993; Conrod, 1988; 1989). An important Kerns Intrusion is due to both fractionation and issue which arises from recent studies of layered mafic in- contamination, and some of the most contaminated lavas trusive complexes and sills is whether this contamination in the context of trace element ratios consist of as much as signature is accompanied by anomalous depletion of the 90% crustal material assimilated into the Nipissing magma. It is known based on the analytical data for Lorrain magma in Ni, Cu, and PGE as suggested by recent studies Formation sediments from the roof of the Kerns Intrusion of the Siberian Trap at Noril’sk (Lightfoot et al., 1994; Nal- that the country rocks have very low Cu, and this is also a drett et al., 1992; 1995; Hawkesworth et al., 1995). It has feature of the aplitic granitoids which appear to be anatec- been established that many contaminated CFB magmas tic melts of the country rock. Based on these observations, show strong depletion in Ni, Cu, and the PGE, and many the low Cu/Zr ratios of the evolved samples from the Kerns investigations have used the work of Irvine (1975) to argue Instrusion can be explained by contamination of the that the addition of crustal silica to magmas triggers the Nipissing magma with significant amounts of low Cu/Zr segregation and fractionation of immiscible sulphide liq- country rock without any requirement for the segregation uids (e.g. Naldrett and McDonald, 1981; Lightfoot et al., and removal of immiscible sulphide liquids. This is an im- 1984; Naldrett et al., 1992, 1995). At issue is whether the portant observation as it suggests that caution is required contamination signature found in the Kerns Intrusion is when using Cu/Zr as an index of the amount of sulphide accompanied by any evidence for the depletion of the fractionation from mafic magmas.

75 References

Barlow,A.E. 1899. Report on the geology and natural resources of the area Cox, K.G. and Hawkesworth, C.J. 1984. Relative contribution of crust and included by the Nipissing and Timiskaming map sheets, comprising mantle to flood basalt magmatism, Mahabaleshwar area, Deccan portions of the district of Nipissing, Ontario, and the country of Traps; Philos. Trans. R. Soc. London, Ser. A310, p.627- 641. Pontiac, Quebec; Canada Geological Survey, Annual Report, 10, 287pp. Cox, K.G., and Hawkesworth, C.J. 1985. Geochemical stratigraphy of the Deccan Traps at Mahabaleshwar, Western Ghats, India, with Bowen, N.L. 1909. Diabase and Aplite of the Cobalt-Silver Area; Journal implications for open system magmatic processes; Journal of Canadian Mining Institute, p. 95-106. Petrology, 26, p.355-377.

Bowen, N.L. 1910. Diabase and Granophyre of the Gowganda Lake Czemanskee, G.K., Wooden, J.L., Zientek, M.L., Fedorenko, V.A., District, Ontario; Journal of Geology, v. 18, p. 648-674. Zen’ko, T.E., Kent, J., King B.-S.W., Knight, R.J., and Siems, D.F. 1994. Geochemical and isotopic constraints on the petrogenesis of Bowen, N.L. 1928. The Evolution of the Igneous Rocks. Princeton the Noril’sk-Talnbakh Ore-Forming System; Ontario Geological University Press, Princeton. 334pp. Survey Special Volume No. 5, p.313-341.

Brugmann, G.E., Naldrett, A.J., Asif, M., Lightfoot, P.C., and Gorbachev, DePaolo, D.J. 1981. Trace Element Effects of Combined Wall Rock N.S., 1993. Siderophile and chalcophile metals as tracers of the Assimilation and Fractional Crystallization; Earth and Planetary evolution of the Siberian Trap in the Noril’sk Region, Russia; Science Letters, v. 53, p. 189-202. Geochim. Cosmochim. Acta, 57, p. 1001-2018. DePaolo, D.J. 1985. Isotopic Studies of Processes in Mafic Magma Buchan, K.L. and Card, K.D.1985. Preliminary Comparison of Chambers: 1. The Kiglapaiit Intrusion, Labrador; Journal of Petrographic and Paleomagnetic Characteristics of Nipissing Petrology, v. 26, p. 925-951. Diabase Intrusions in Northeastern Ontario; in Current Research, Dressler, B.O. 1979. Geology of McNish and Janes Townships, District of Part A, Geological Survey of Canada, Paper 85-1A, p. 131-140. Sudbury. Report 191, Ontario Geological Survey. Accompanied by Map 2425 at 1:31,680. 91pp. Buchan, K.L., Card, K.D., and Chandler, F.W. 1989. Paleomagnetism of Nipissing Diabase and Associated Rocks in the Englehart area, Dressler, B.O. 1982. Geology of the Wanapetei Lake Area, District of Ontario; Canadian Journal of Earth Sciences, v. , p. Sudbury; Ontario Geological Survey, Open File Report 5287,150p., 19 tables, 29 figures, 25 photos, and 3 maps. Campbell, I.H. and Turner, J.S. 1987. A Laboratory Investigation of Assimilation at the Top of a basaltic magma chamber; Journal of Fairbairn, H.W., Hurley, P.M., Card, K.D., Knight, C.J. 1969. Correlation Geology, 95, p.155-172. of Radiometric Ages of Nipissing Diabase and Huronian Metasediments with Proterozoic Orogenic Events in Ontario; Card, K.D. 1965. Hyman and Drury townships; Geological Report 34, Canadian Journal of Earth Sciences, v.6, p.489-497. accompanied by Map 2055, scale 1 inch to 0.5 mile. Finn, G.C. 1981. Petrogenesis of the Wanapetei Gabbronorite Intrusion; A Card, K.D. 1968. Geology of the Denison-Watersarea; OntarioGeological Nipissing-type Diabase fromnorthern Ontario;Unpub.M.Sc.thesis, Report 60 accompanied by Map 2119 at 1:31680, 63pp. University of Western Ontario, 112pp. Card,K.D.1976 and 1984.Geology of the Espanola-Whitefish Falls Area, Finn, G.C. and Edgar, A.D. 1986. The Wanapetei intrusion, northeastern District of Sudbury, Ontario; Geoscience Report 131; Ontario Ontario: an Example of a Mafic Intrusion with Cycles of Reverse Geological Survey Report, Accompanied by Map 2311 and Map Differentiation; Neues Jahhrb. Miner., Abh. v. 154, p. 75-91. 2312 at 1:31680; 70pp. First and Second editions. Finn, G.C., Edgar, A.D., and Rowell, W.F. 1982. Petrology, Geochemistry, Card, K.D., and Cieslielski,A. 1986.Subdivision of the Superior Province and Economic Potential of the Nipissing Diabase; Grant 100, p. of the Canadian Shield; Geoscience Canada, Volume 13, No. 1, 43-57 in Geoscience Research Grant Program, Summary of p.5-13. Research 1981-1982, edited by E.G. Pye, Ontario Geological Survey, Miscellaneous Paper 103, 219p. Card, K.D. and Pattison, E.F. 1973. Nipissing Diabase of the Southern Province, Ontario; Geological Association of Canada, Special Paper Frarey, M.J and Roscoe, S.M. 1970. The Huronian Supergroup North of 12, p.7-30. Lake Huron; p.143-157 in Symposium on Basins and Geosynclines of the Canadian Shield, Geological Survey of Canada, Paper 70-40. Chen, J.H. and Wasserburg, G.J. 1981. Isotopic determination of uranium in picomole and subpicomole quantities; Analytical Chemistry, 53, Fyon,A.J.,Jackson,S.L.,Lightfoot,P.C.,Meyer,W.,Bennett,G.B.,Ireland p.2060-2067. J., and Lavigne, M.J. 1995. Metallogeny of the Proterozoic Eon, Southern Province, Ontario; Canadian Mineralogist, 33, p.26-27. Collins, W.H. 1913. The geology of Gowganda mining division, Ontario; Ginn, R.M. 1965. Nairn and Lorne Townships. geological Report No. 35, Canada Geological Survey, Mem. 33, 121pp. Ontario Geological Survey. Accompanied by Map No. 35 at Conrod, D.M. 1988. Petrology, Geochemistry, and PGE Potential of the 1:31,680. 45pp. Nipissing Intrusions; University of Toronto, unpublished MSc Harvey, P.K., and Atkin, B.P. 1982. Automated X-ray fluorescence thesis. analysis, in Sampling and analysis for the Mining Industry. Institute Conrod, D.M. 1989. The Petrology andGeochemistry ofthe DuncanLake, of Mining and Metallurgy, London, p.17-26. Beaton Bay, Milner Lake, and Miller Lake Nipissing Intrusions Hawkesworth, C.J. 1982. Isotope characteristics of magmas erupted along within the Gowganda Area, District of Timiskaming; Ontario destructive plate margins; in Andesites, R.S. Thorpe, ed., Elsevier, Geological Survey, Open File Report 5701. 210pp. Amsterdam. Corfu, F. and Andrews, A.J. 1986. A U-Pb age for Mineralized Nipissing Hawkesworth, C.J. and Morrison, M.A. 1978. A reduction in 87Sr/86Sr Diabase, Gowganda; Canadian Journal of Earth Sciences, v. 23, p. during basalt alteration; Nature, 276, p.381-383. 107-109. Hawkesworth, C.J. and Gallagher, K. 1993. Mantle hot-spots, plumes and Cox,K.G.1980.A model for flood basalt vulcanism; Journal of Petrology, regional tectonics as causes of intraplate magmatism; Terra Nova, 5, 21, p.629-650. p.552-559.

76 Petrology and Geochemistry of the Nipissing Gabbro

Hawkesworth, C.J. and Powell, M. 1980. Magma Genesis in the Lesser Jambor, J.L. 1971b. Distribution of some Minor Elements in the Nipissing Antilles Island Arc; Earth and Planetary Science Letters, v. 51, p. Diabase; Canadian Mineralogist, v. 42, p. 321-357. 297-308. Jolly, W.T. 1987a. Lithophile Elements in Huronian low-Ti Continental Hawkesworth, C.J., Mantovani, M.S.M., Taylor, P.N., and Palacz, Z. 1986. Tholeiites from Canada, and Evolution of the Precambrian Mantle; Evidence from the Parana of South Brazil for a Continental Earth and Planetary Science Letters, v. 85, p. 401-415. Contribution to Dupal Basalts; Nature, v. 322, p. 356-359. Jolly, W.T. 1987b. Geology and Geochemistry of Huronian Rhyolites and low-Ti Tholeiites from Central Ontario, Canada; Canadian Journal Hawkesworth, C.J., Rogers, N.W., van Calsteren, P.W.C., and Menzies, of Earth Sciences, v. 24, p.1360-1385. M.A.1984.Mantle Enrichment Processes; Nature v. 311,p.331-335. Kanasewich, A.F., and Farquhar, R.M. 1965. Lead isotope ratios from the Hawkesworth, C.J., Lightfoot, P.C., Fedorenko, V.A., Blake, S., Naldrett, Cobalt-Noranda area, Canada. Canadian Journal of Earth Sciences, A.J., Doherty, W., and Gorbachev, N.S. 1995. Magma differentiation 2, 361-384. and mineralisation in the Siberian continental flood basalts. Lithos. Lightfoot, P.C. 1982. The geology of the Tabankulu Section of the Insizwa 34, p.61-88. Complex, Transkei. Unpublished M.Sc. thesis, University of Henderson, P. 1983. Inorganic Geochemistry. Pergamon Press. 353pp. Toronto. Lightfoot, P.C. 1985. Isotope and Trace Element Geochemistry of the Henderson, P. 1984. Rare Earth Element Geochemistry. Elsevier Press, South Deccan Lavas, India; Unpublished PhD thesis, The Open Holland. University, Milton Keynes, United Kingdom, 589p. Hergt, J.M., Chappell, B.W., McCulloch, M.T., McDougall, I., and Lightfoot, P.C. 1987. Platinum group element potential of the Nipissing Chivvas, A.R. 1989. Geochemical and isotopic constraints on the intrusions; Mines and Minerals, Open House Meeting, Toronto. origin of the Jurassic dolerites of Tasmania; Journal of Petrology, 30, Abstracts. p.841-883. Lightfoot, P.C. 1995. Geochemistry of the Nipissing Gabbro: source and Hergt, J.M., Peate, D.W.,and Hawkesworth,C.J. 1991.The petrogenesisof mineral potential. In: Proceedings and abstracts of the Northern Mesozoic Gondwana low-Ti flood basalts; Earth and Planetary Margin of the Southern Province of the Canadian Shield; Canadian Science Letters, 105, p.134-148. Mineralogist, 33, p. 19-20. Lightfoot, P.C., and Hawkesworth, C.J. 1988. Origin of Deccan Traplavas: Hriskevich, M.E. 1952. Petrology of the Nipissing Diabase Sheet of the evidence from combined trace element and Sr-, Nd-, and Pb-isotope Cobalt Area of Ontario: Unpub. Ph.D. thesis, Princeton University, studies. Earth and Planetary Science Letters 91, p.89-104. Princeton, New Jersey. Lightfoot, P.C. and Naldrett, A.J. 1984. Chemical Variation in the Insizwa Hriskevich, M.E. 1968. Petrology of the Nipissing Diabase Sill of the Complex, Transkei, and the Nature of the Parent Magma; Canadian Cobalt Area, Ontario, Canada; Geological Society of America Mineralogist, v. 22, p. 111- 123. Bulletin, v. 79, p. 1387-1404. Lightfoot, P.C. and Naldrett, A.J. 1989. Assimilation and fractionation in Humphris, S.E. and Thompson, G. 1978. Trace element mobility during basic magma chambers: significance of trace- element and hydrothermal alteration of oceanic basalts; Geochemistry Nd-isotopic data from the Kerns sill, Nipissing diabase, Ontario; Cosmochemistry Acta, 42, 127-136. Canadian Journal of Earth Sciences, 26, p.737- 754. Lightfoot, P.C., Conrod, D., Naldrett, A.J., and Evensen, N.M. 1986. Huppert, H.E., and Sparks, S.J. 1980. The fluid dynamics of a basaltic Petrologic, Chemical, Isotopic, and Economic Potential Studies of magma chamber replenished by influx of hot, dense, ultrabasic the Nipissing Diabase; Grant 230, p. 87-106 in Geoscience Research magma; Contributions to Mineralogy and Petrology, v.75, Grant Program, Summary of Research 1985-1986, edited by V.G. p.279-289. Milne, Ontario Geological Survey, Miscellaneous Paper 130, 235p. Huppert, H.E. and Sparks, S.J. 1985. Cooling and Contamination of Lightfoot, P.C., Conrod, D., Naldrett, A.J., and Evensen, N.M. 1987. Ultramafic Magmas during Ascent through Continental Crust; Earth Petrologic, Chemical, Isotopic, and Economic-Potential Studies of and Planetary Science Letters, v. 74, p. 371-386. the Nipissing Diabase; Grant 230, p.4-26 in Geoscience Research Grant Program, Summary of Research 1986-1987, edited by V.G. Innis, D.G. and Colvine, A.C. 1984. The Regional Metallogenetic Setting Milne, Ontario Geological Survey, Miscellaneous Paper 136, 241p. of Sudbury; in The Geology and Ore Deposits of the Sudbury Structure; Ontario Geological Survey Special Volume No. 1, Lightfoot,P.C.,Naldrett,A.J.,and Hawkesworth,C.J.1984.Geochemistry p.45-56. of the Waterfall Gorge Section of the Insizwa Complex, with Particular Reference to the Origin of the Nickel Sulfide Irvine, T.N. 1975. Crystallisation sequence of the Muscox Intrusion and Mineralization; Economic Geology, v. 79, p. 1857-1879. other layered intrusions: II Origin of the chromite layers and similar deposits of other magmatic ores; Geochim. Cosmochim. Acta. 39: Lightfoot, P.C., Naldrett, A.J., Gorbachev, N.S., Doherty, W., and 991-1020. Fedorenko, V.A. 1990. Geochemistry of the Siberian Trap of the Noril’sk Area, USSR, with implications for the relative contributions Irvine, T.N and Baragar, W.R.A. 1972. The Muscox intrusion and of crust and mantle to flood basalt magmatism. Contributions to Coppermine River lavas, Northwest Territories, Canada, Int. Geol. Mineralogy and petrology, 104, p.631-644. Congr., 24th, Montreal, Field Excursion A29 Guidebook, 70. Lightfoot, P.C., Sutcliffe, R.H., and Doherty, W. 1991. Crustal contamination identified in Keweenawan Osler Group tholeiites, Irvine, T.N., and Smith, C.H. 1969. The ultramafic rocks of the Muscox Ontario: a trace element perspective; Journal of Geology, 99, Intrusion, Northwestern Territories, Canada. In Wyllie, P.J., Editor, p.739-760. Ultramafic and Related Rocks. New York, Wiley, p.38-49. Lightfoot, P.C., De Souza, H., and Doherty, W. 1993. Differentiation and Irving, A.J. 1978. A Review of Experimental Studies on Crystal/Liquid source of the Nipissing Diabase intrusions, Ontario, Canada; Trace Element Partitioning; Geochemica Cosmochemica Acta, v. Canadian Journal of Earth Sciences, 30, p.1123- 1140. 42, p. 743-770. Lightfoot, P.C., Hawkesworth, C.J., Hergt, J., Naldrett, A.J., Gorbachev, Jackson, S. 1995. Metamorphism and Structural History of the Western N.S., and Fedorenko, V.A. 1993. Remobilisation of the continental Huronian Supergroup. Canadian Mineralogist, 33, p.9-10. lithosphere by a mantle plume: major-, trace-element, and Sr-, Nd-, and Pb-isotope evidence from picritic and tholeiitic lavas of the Jambor, J.L. 1971a. The Nipissing Diabase; Canadian Mineralogist, v.42, Noril’sk District, Siberian Trap, Russia. Contributions to Mineralogy p.743-770. and Petrology, 114, p.171-188.

77 OGS Study 58

Lightfoot, P.C., Naldrett, A.J., Gorbachev, N.S., Fedorenko, V.A., Pearce, J.A. 1984. The role of sub-continental lithosphere in magma Hawkesworth, C.J., Hergt, J., and Doherty, W. 1994. genesis at active continental margins; in Continental basalts and Chemostratigraphy of Siberian Trap lavas, Noril’sk District, Russia: mantle xenoliths; edited by C.J. Hawkesworth and M.J. Norry. Shiva implications for the source of flood basalt magmas and their Geology Series. 272pp. associated Ni-Cu mineralisation. Ontario Geological Survey Special Volume No. 5, edited by P.C.Lightfoot and A.J. Naldrett.p.283-312. Pearce, J.A., and Cann, J.R. 1975. Tectonic Setting of Basic Volcanic Rocks Determined using Trace Element Analyses; Earth and Lovell, H.L. 1971. Geology of the Cobalt-Kirkland Lake Area; in Ontario Planetary Science Letters, v.19, p.290-300. Department of Mines and Northern Affairs Miscellaneous Paper 46, Annual Report of the Resident Geologists Section, p. 44-55. Pearce, J.A. and Norry, M.J. 1979. Petrogenetic Implications of Ti, Zr, Y, and Nb Variations in Volcanic Rocks; Contributions to Mineralogy Lovell, H.L., and Frey, E.D. 1976. Geology of the New Liskeard Area, and Petrology, v.69, p.33-47. District of Timiskaming, Ontario Div. Mines. GR144, 34p. Accompanied by Maps 2300 and 2301, scale 1 inch to 1/2 mile Peck, D.C., James, R.S., and Chubb, P.T. 1993. Geological environments (1:31680). for PGE-Cu-Ni Mineralisation in the East Bull Lake Gabbro-Anorthosite Intrusion, Ontario; Exploration and Mining Marsh, J.S. and Eales, H.V. 1984. The chemistry and petrogenesis of Geology, 291: 85-104. igneous rocks of the Karoo Central Area, Southern Africa; in Petrogenesis of the Karoo Volcanic Rocks of the Karoo Province. Peck, D.C., James, R.S., Chubb, P.C., Keays, R.R., Reeves, S.J. Lightfoot, Edited by A.J.H. Erlank, Geological Society of South Africa, Special P.C., andKamineni, D.C.1993. Precious-metal,chalcophile-element Publication p.27-68. and rare-earth element geochemistry of the Bull Lake Area, District of Algoma and Sudbury, Ontario; Ontario Geological Survey, Open McBirney, A.R. 1979. Effects of assimilation; in Yoder, H.S., ed., The File Report 5849. Evolution of the Igneous Rocks: Princeton, Princeton Univ. Press. p. 308-338 Phemister, T.C. 1928 A comparison of the Keweenawan sills rocks of Sudbury and Cobalt, Ontario: Royal Society Canada, Proceedings Miller, W.G. 1903. Cobalt-nickel arsenides and silver in Ontario; and Transactions, 3rd. ser, v.22, p.121-197. Engineering and Mining Journal, 76, 888-890. Potts, P.J., Webb, P.C., and Watson, J.S. 1984. Energy dispersive X-ray Miller, W.G. 1904. Cobalt-nickelarsenides andsilver: OntarioDepartment fluorescence analysis of silicate rocks for major and trace elements; Mines Report 96-103. X-ray Spectrometry, 13, 1, p.2.

Miller, W.G. 1913. The cobalt-nickel arsenides and silver deposits of Potts, P.J., Webb, P.C., and Watson, J.S. 1985. Energy- dispersive X-ray Timiskaming (Cobalt and adjacent areas). Ontario Deptartment of fluorescence analysis of silicate rocks: comparison with wavelength Mines Annual Report 19, pt.2, 2709pp. dispersive performance; Analyst (Lond.) 110: 507-513.

Miller, W.G. and Knight, C.W. 1913. Sudbury, Cobalt, and Porcupine Rogers, N.W., Hawkesworth, C.J., Parker, R.J., and Marsh, J.S. 1985. The geology, Ontario; Engineering and Mining Journal., v.95, p. Geochemistry of Potassic Lavas from Vulsini, Central Italy and 1129-1133. Implications for Mantle Enrichment Processes Beneath the Roman Region; Contributions to Mineralogy and Petrology, v. 90, p.244- Muir,T.L.1994.The Sudbury Structure: considerations and a model for an 257. endogenetic origin;in The Geology and Ore Deposits of theSudbury Rowell, W.F. 1984. Platinum Group Elements and Gold in the Wanapetei Structure. Ontario Geological Survey Special Volume No. 1. Nipissing-type Intrusion, Northeastern Ontario; Unpubl. M.Sc. p.449-490. thesis, University of Western Ontario. Naldrett, A.J. and Lightfoot, P.C. 1993. Ni-Cu-PGE ores of the Noril’sk Rowell, W.F. and Edgar, A.D. 1986. Platinum-Group Element Region Siberia: a model for giant magmatic sulphide deposits Mineralization in a Hydrothermal Cu-Ni Sulphide Occurrence, associated with flood basalts; Society of Economic Geologists; Rathbun Lake, Northeastern Ontario; Economic Geology, v.81, edited by B.H. Whiting, C.J. Hodgson, and R. Mason. Special p.1272-1277. Volume No. 2. p.81-124. Satterly. J. 1933. The Nipissing diabase of Cobalt, South Lorrain and Naldrett, A.J., Lightfoot, P.C., Fedorenko, V.A., Doherty, W., and Gowganda, Ontario: M.A. thesis, Toronto Univ., Toronto, Canada. Gorbachev, N.S., 1992. Geology and geochemistry of intrusions and flood basalts of the Noril’sk Region,USSR, with implications for the Scholtz, D.L. 1936. The magmatic nickeliferous ore deposits of East origin of the Ni-Cu ores; Economic Geology, 87, p.975- 1004. Grigvaland and Pondoland; Geological Society of South Africa - Transactions, v.39, p.81-210. Naldrett, A.J., Fedorenko, V.A., Lightfoot, P.C., Kunilov, V.I.,Gorbachev, N.S., Doherty, W., and Johan, Z. 1995. Ni-Cu-PGE deposits of Simpkin, T. and Smith, J.V. 1970: Minor-element distribution in olivine. J. Noril’sk Region, Siberia: their formation in conduits for flood basalt Geol. 78: 304-324. volcanism; Transactions of the Institution of Mining and Metallurgy, Section B, v.104, p. B18-B36. Simony, P.S. 1964. Northwestern Temagami Area; Ontario Department of Mines, Geological Report No. 28, accompanied by map No. 2057. Noble, S.R., Lightfoot, P.C., and Scharer, U., 1992. A Method for Single-Filament Isotopic Analysis of Nd Using in-situReduction; Stupavsky,M.and Symons,D.T.A.1982.Extent of Grenvillian Remnance Canadian Journal of Earth Sciences, 29, p.1424-1429. Components in Rocks of theSouthern Province;Canadian Journalof Earth Sciences, V. 19, p. 698-708. Noble, S.R., and Lightfoot, P.C. 1992. U - Pb Baddeleyite ages for the Kerns and Triangle Mountain Intrusions, Nipissing Diabase, Sun, S.-s. and McDonough, W.F.1989. Chemical and isotopic systematics Ontario. Canadian Journal of Earth Sciences, 29, p.1124- 1129. of oceanic basalts: implications for mantle composition and processes; in Magmatism in the Ocean Basins, edited by A.D. Osmani, I.A. 1994. Proterozoic Mafic Dike Swarms in the Superior Saunders and M.J. Norry; Geological Survey Special Publication Province of Ontario; in The Geology of Ontario. Ontario Geological No. 42, p313-345. Survey Special Volume No. 4, p. 661-681. Symons, D.T.A. 1971. A Paleomagnetic Study of the Nipissing Diabase, Patchett, P.J. 1980. Thermal effects of basalt on continental crust and Blind River - Elliot Lake Area, Ontario. Geological Survey of crustal contamination of magma; Nature, 283, p.559-561. Canada, Paper 70-63, p. 19-30.

78 Petrology and Geochemistry of the Nipissing Gabbro

Taylor, H.P. Jr. 1980. The Effects of Assimilation of Country Rocks by Weaver, B.L. and Tarney, J. 1981. The Scourie Dyke Suite: Petrogenesis Magmas on 18O/16Oand87Sr/86Sr Systematics in Igneous Rock. and Geochemical Nature of the Proterozoic Sub-continentalMantle; Earth and Planetary Science Letters, v. 47, p.243. Thompson, R.N., Contributions to Mineralogy and Petrology; v. 79, p. 175-188. Morrison, M.A., Dickin, A.P., and Hendry, G.L. 1984. Continental Flood Basalts... Arachnids Rule OK?; inContinental Basalts and Wilson, J.R. and Engell-Sorensen, O. 1986. Basal reversals in layered Mantle Xenoliths, edited by C.J. Hawkesworth and M.J. Norry. intrusions are evidence for emplacement of compositionally SHIVA Geology Series, p.158. stratified magma; Nature, v. 323, p.616-618. Thorpe, R.I., Goodz, M.D., Jonasson, I.R., and Blenkinsop, J. 1986. Lead-isotope study of mineralisation in the Cobalt district, Ontario; Wood, B.J. and Fraser, D.G. 1976. Elementary thermodynamics for Canadian Journal of Earth Sciences, V. 23, p.1568-1575. geologists. Oxford University Press.303pp. Van Schmus, R. 1965. The Geochronology of the Blind River - Bruce Young, G.M. 1982.Depositional Environments and Tectonic Setting ofthe Mines Area, Ontario; Canadian Journal of Geology, v. 73, p. Early Proterozoic Huronian Supergroup; Guidebook, Excursion 755-780. 13B, Eleventh International Congress on Sedimentology, Wager, L.R. and Deer, W.A. 1939. Geological investigations in East International Association of Sedimentologists. McMaster Greenland. Part III. The petrology of the Skaergaard intrusion, University, Hamilton, Canada, 79pp. Kangerdlugssuaq, East Greenland. Medd. om Gronland, 105, no.4, 1-352. Young, G.M. 1995. The Huronian Supergroup in the context of a paleoproterozoic Wilson Cycle in the Great Lakes region; Canadian Wager, L.R. and Brown, G.M. 1968. Layered igneous rocks; Edinburgh, Mineralogist, 33:5-6. Oliver and Boyd, 588pp. Walker, K.R. 1969. The Pallisades Sill, New Jersey: a reinvestigation; Zindler, G.A. 1980. Geochemical Processes in the Earth’s Mantle and the Geological Society of America Special Paper, 111. Nature of Crust-Mantle Interactions; Evidence from Studies of Neodymium and Strontium Isotope Ratios in Mantle- derived Walker, F.and Poldervaart, A. 1949. Karoo Dolerites of the Union of South Igneous Rocks and lherzolite Nodules; PhD dissertation, Africa; Bulletin of the Geological Society of America, v. 60, p. Massachusetts Institute of Technology, Cambridge, Massachusetts. 591-706. Wallace, H. 1981. Keweenawan geology of the Lake Superior Basin. in Zolnai, A.I., Price, R.A., and Helmstaedt, H. 1984. Regional cross section Proterozoic Basins of Canada. F.H.A. Campbell, editor; Geological of the Southern Province adjacent to Lake Huron, Ontario. Canadian Survey of Canada, Paper 81-10, p.399-417. Journal of Earth Sciences, 21, p.447-456.

79 CONVERSION FACTORS FOR MEASUREMENTS IN ONTARIO GEOLOGICAL SURVEY PUBLICATIONS Conversion from SI to Imperial Conversion from Imperial to SI SI Unit Multiplied by Gives Imperial Unit Multiplied by Gives LENGTH 1 mm 0.039 37 inches 1 inch 25.4 mm 1 cm 0.393 70 inches 1 inch 2.54 cm 1 m 3.280 84 feet 1 foot 0.304 8 m 1 m 0.049 709 7 chains 1 chain 20.116 8 m 1 km 0.621 371 miles (statute) 1 mile (statute) 1.609 344 km AREA 1cm2 0.155 0 square inches 1 square inch 6.451 6 cm2 1m2 10.763 9 square feet 1 square foot 0.092 903 04 m2 1km2 0.386 10 square miles 1 square mile 2.589 988 km2 1 ha 2.471 054 acres 1 acre 0.404 685 6 ha VOLUME 1cm3 0.061 02 cubic inches 1 cubic inch 16.387 064 cm3 1m3 35.314 7 cubic feet 1 cubic foot 0.028 316 85 m3 1m3 1.308 0 cubic yards 1 cubic yard 0.764 555 m3 CAPACITY 1 L 1.759 755 pints 1 pint 0.568 261 L 1 L 0.879 877 quarts 1 quart 1.136 522 L 1 L 0.219 969 gallons 1 gallon 4.546 090 L MASS 1 g 0.035 273 96 ounces (avdp) 1 ounce (avdp) 28.349 523 g 1 g 0.032 150 75 ounces (troy) 1 ounce (troy) 31.103 476 8 g 1 kg 2.204 62 pounds (avdp) 1 pound (avdp) 0.453 592 37 kg 1 kg 0.001 102 3 tons (short) 1 ton (short) 907.184 74 kg 1 t 1.102 311 tons (short) 1 ton (short) 0.907 184 74 t 1 kg 0.000 984 21 tons (long) 1 ton (long) 1016.046 908 8 kg 1 t 0.984 206 5 tons (long) 1 ton (long) 1.016 046 908 8 t CONCENTRATION 1 g/t 0.029 166 6 ounce (troy)/ 1 ounce (troy)/ 34.285 714 2 g/t ton (short) ton (short) 1 g/t 0.583 333 33 pennyweights/ 1 pennyweight/ 1.714 285 7 g/t ton (short) ton (short) OTHER USEFUL CONVERSION FACTORS Multiplied by 1 ounce (troy) per ton (short) 20.0 pennyweights per ton (short) 1 pennyweight per ton (short) 0.05 ounces (troy) per ton (short)

Note: Conversion factors which are in bold type are exact. The conversion factors have been taken from or have beenderivedfromfactorsgivenintheMetricPracticeGuidefortheCanadianMiningandMetallurgicalIndustries, published by the Mining Association of Canada in co-operation with the Coal Association of Canada.

ISSN 0704-2590 ISBN 0-7778-4804-X