Strontium and oxygen-isotope study of the Dufek intrusion
We found all plagioclase grains to be compositionally zoned, whether in anorthosite or adjacent gabbro. A maximum compositional range of grains within a single sample was found to be 5 to 10 mole percent anorthite; lesser ranges may in some cases have resulted from missing innermost parts of grains in thin section cuts. Although the range of anorthite content of individual grains varies from grain to grain, the difference in average composition of all grains in a sample is always less than 2 mole percent anorthite. Compositional averages and ranges for each analyzed sample are shown in figure 2. Single grains commonly exhibit compositional ranges about the same as the total range in the rock. Our results confirm Abel's et al. (1979) finding that plagioclases of the Dufek Massif section show an overall upward decrease in anorthite content, but they are presently inadequate to define closely stratigraphic trends in compositions within the anorthositic layers being studied. Plagioclase compositions of anorthite-81 and anorthite-78 occur respectively in lowest and highest parts of the Walker Anorthosite; and compositions are in the ranges anorthite-70--67 and anorthite-63-60, respectively, in the lower anorthosite member and the Spear Anorthostie member (figure 2). The generally large compositional range within a sample is a chief difficulty in discernment of trends within a layer. Moreover, in the Spear member, plagioclase of two samples from the same height have compositions with significantly different averages and only slightly overlapping ranges (figure 2). The significance of the smallscale reversals apparent in trends of average plagioclase compositions from gabbro into overlying anorthosite is therefore presently uncertain. Plagioclase zoning patterns are complex. They are characterized by an oscillatory zonation that is commonly nonsymmetrical (zone width unequal around grain core). The zoning patterns of different grains within each of the two Walker Anorthosite samples are similar. In contrast, however, there are significant grain-to-grain zoning pattern differences within individual samples of other anorthosite and gabbro, which suggest that the different grains of a rock did not crystallize in equilibrium. Plagioclases of anorthosites and adjacent gabbros show equally complex crystallization histories. Although the thin, widespread layers of anorthositic cumulate and of gabbro-containing noncumulus anorthositic inclu sions are volumetrically minor, we believe they can provide
Strontium and oxygen-isotope study of the Dufek intrusion A.B. FORD, R.W. KISTLER, and L.D. WHITE U.S. Geological Survey Menlo Park, California 94025
The mostly gabbroic Dufek intrusion (82°30'S 50°W) is one of the world's largest layered igneous complexes. It is more than 1986 REVIEW
important insights into the solidification history of the intrusion. In our continuing study of them, we have begun to analyze plagioclases of much more closely spaced (1 meter or less) samples from across four cyclic units of anorthosite and gabbro of the Stephens Anorthosite member (figure 1). The study will include cumulus and post-cumulus pyroxenes and will incorporate isotopic data (Ford, Kistler, and White, Antarctic Journal, this issue) that can provide insight into some magmatic processes not possible from major- and minor-element compositions alone. Similar to the conclusions drawn from the study of isotopic compositions of rocks and minerals, our study of plagioclases indicates that the crystallization history of the intrusion was complex. This work was supported in part by National Science Foundation grant DPP 80-20753 to the U.S. Geological Survey. References Abel, K.D., G.R. Himmelberg, and A.B. Ford. 1979. Petrologic studies of the Dufek intrusion: Plagioclase variation. Antarctic Journal of the U.S., 14(5), 6-8. Czamanske, G.K., and D.L. Schiedle. 1985. Characteristics of the banded-series anorthosites. In G.K. Czamanske and M.L. Zientek (Eds.), The Stillwater Complex, Montana: Geology and guide. (Montana Bureau of Mines and Geology Special Publication, Vol. 92.) Ford, A.B. 1976. Stratigraphy of the layered gabbroic Dufek intrusion, Antarctica. U.S. Geological Survey Bulletin, 1405(D), D1—D36. Ford, A.B., and G.R. Himmelberg. In press. Geology and crystallization of the Dufek intrusion. In R.J. Tingey (Ed.), Geology of Antarctica. Oxford: Oxford University Press. Ford, A.B., R.W. Kistler, and L.D. White. 1986. Strontium and oxygenisotope study of the Dufek intrusion. Antarctic Journal of the U.S., 21(5). Hess, H. H. 1960. Stillwater igneous complex, Montana—A quantitative mineralogical study. Geological Society of America Memoir, 80, 230. Himmelberg, G. and A.B. Ford. 1976. Pyroxenes of the Dufek intrusion, Antarctica. Journal of Petrology, 17(2), 219-243. Himmelberg, G.R., and A.B. Ford. 1977. Iron-titanium oxides of the Dufek intrusion, Antarctica. American Mineralogist, 62, 623-633. Irvine, TN., D.W. Keith, and S.C. Todd. 1983. The J-M platinumpalladium reef of the Stillwater Complex, Montana: II. Origin by double-diffusive convective magma mixing and implications for the Bushveld Complex. Economic Geology, 78(7), 1287-1334. Wager, L.R., and G.M. Brown. 1968. Layered igneous rocks. Edinburgh: Oliver and Boyd.
50,000 square kilometers in area (Behrendt et al. 1981) and possibly 8-9 kilometers thick at its southern end (Ford 1976). It is about the same age as the dominant phase of Jurassic Ferrar Group tholeitic igneous activity that occurred throughout the Transantarctic Mountains (Ford and Kistler 1980). The stratigraphy and rock types are described by Ford (1976). We have begun a systematic study of the isotopic compositions of oxygen, strontium, and neodymium in rocks and minerals of the intrusion to investigate many questions about its crystallization history raised by previous studies. Work on other layered intrusions (Pankhurst 1969; Taylor and Forester 1979; Gray and Goode 1981; Palacz and Tait 1985; DePaolo 1985) shows that isotopic compositions are more sensitive indicators of many magmatic processes than are major- and minor-element compositions of rocks and minerals. 63
A body of such enormous size, with a volume of possibly 200,000 cubic kilometers or more, very likely formed from multiple and possibly interconnected intrusive centers (Ford and Himmelberg in press), as in South Africa's Bushveld Complex (Hunter 1978). An interpretation of the crystallization record of this mostly ice-covered body must encompass not only the processes that may have operated within a single crystallization chamber, such as fractionation, assimilation, and mixing of internal magma-current flows of various origins (Irvine 1980; Irvine, Keith, and Todd 1983) but also such external processes as additions of "primitive" magma or of variably differentiated magma shifted between interconnected crystallization chambers, material loss by volcanism or ancillary intrustion (Ferrar Group tholeiites?), and meteoric and hydrothermal fluid exchanges with country rocks. General stratigraphically upward increases in the ratios of iron to iron plus magnesium in cumulus pyroxenes (Himmelberg and Ford 1976) and normative albite/(albite + anorthite) in cumulus plagioclase (Abel, Himmelberg, and Ford 1979; Haensel, Himmelberg, and Ford, Antarctic Journal, this issue) resemble trends in other layered intrusions that are interpreted in terms of fractional crystallization (Wager and Brown 1968) or crystallization from a nonhomogeneous magma column (Irvine et al. 1983). The trends extend upward through the 1.8-kilometer-thick lower part of the body exposed in the Dufek Massif and up to about 500 meters height in the upper part exposed in the Forrestal Range, where strong reversals occur in the compositional trends of all cumulus minerals, including ilmeno-magnetite (Himmelberg and Ford 1977). The reversals are the strongest evidence yet found for a large-scale replenishment of magma into the chamber, but the magma's source is
unknown. We are at present in a reconnaissance stage of isotopic studies to investigate relative roles of magma additions and mixing, fractional crystallization, and other mechanisms. The isotopic compositions of oxygen (in per mil of isotopic oxygen-18) in 77 rock samples show that values are much more variable and mostly lower in the Forrestal Range than Dufek Massif (figure 1), in which values are comparable to those of Montana Stillwater Complex (Dingwell and Dunn 1982) and norite of Ontario's Sudbury intrusion (Ding and Schwarcz 1984). Variations are not simply related to stratigraphic height and apparent relative age of crystallization, however, because noncumulus granophyre of a 300-meter-thick layer at the top of the Forrestal Range section has isotopic oxygen-18 values higher than those of many underlying cumulates of the section (figure 1). Three analyzed Dufek Massif plagioclases have isotopic oxygen-18 values in a narrow range (6.2-6.9) similar to "normal" plagioclases of the lower and middle zones of Greenland's Skaergaard intrusion (Taylor and Forester 1979). As in rocks, the isotopic oxygen-18 compositions of plagioclases are lower and more variable (3.2-6.0, n = 12) in Forrestal Range than Dufek Massif cumulates. Similarly low values in higher parts of the Skaergaard intrusion show subsolidus isotopic oxygen-18 depletion probably due to activity of a meteoric-hydrothermal system (Taylor and Forester 1979). Rubidium and strontium analyses of 36 rock and 16 plagioclase samples made in preparation for studies of initial isotopic ratios of 87 strontium/" strontium (87Sr/85Sr,) show that (1) the rubidium/strontium ratio of rocks generally increases with stratigraphic height, from an average of about 0.02 in the Dufek Massif to 0.12 in the Forrestal Range section, in which ratios are more variable (figure 2); and (2) strontium content of pla-
10 C') w —J aC') LL
0 Li
Z I8 (PER MIL) Figure 1. Oxygen isotopic compositions of rocks of the Dufek intrusion. Rocks are cumulates of gabbroic, leucogabbroic, anorthositic, and pyroxenitic composition; except (D)thin dikes of aplite and felsite that intrude the cumulates or other rocks, (C) noncumulus gabbro and diorite at the contact with country rocks near Mount Lechner, (G) the Lexington Granophyre or a 300-meter-thick layer at the top of the intrusion (Ford 1976), and (I) inclusions in gabbro of podiform masses of anorthosite and leucogabbro. ("s 180" denotes "isotopic oxygen.")
64
ANTARCTIC JOURNAL
Rubidium and strontium contents and strontium isotopic composition of minerals of the Dufek intrusion Sample Mineral Rubidiuma Strontium' 87Sr/86 Sr' 875r/86Src
65F23713 Pyroxene 2.94 6.5 0.71083±5 0.70763 65F23713 Plagioclase 5.42 296.0 0.70931±4 0.70918 65F38A Pyroxene 0.54 9.0 0.70841 ±5 0.70799 65F941 Pyroxene 4.80 11.5 0.71130±4 0.70834
(I) LL 10 -J 0
° Rubidium and strontium contents measured by isotope dilution. Given in parts per million. b Measured ratio. Ratios normalized to 16Sr/88Sr = 0.1194. NBS87SrCO3.
87 Sr/ 86Sr = 0.71023±3. Initial ratio, corrected to 172 million years (Ford and Kistler 1980).
C') Li 0
X co Z
Es 0 .05 .10 .15 .20
RUBIDIUM /STRONTIUM Figure 2. Rubidium/strontium ratios in cumulates of the Dufek intrusion.
gioclases also generally increases upward in the stratigraphic sequence, from 216 parts per million in basal anorthosite of the Dufek Massif section and 262 and 309 parts per million in gabbros higher in the section, to greater values in the Forrestal Range where plagioclases from below the mineral-trend reversal at 500-meters height average about 315 parts per million strontium (n = 11) and those above contain much larger amounts (436 parts per million, 491 parts per million). In other igneous complexes an increase in the rubidium/strontium ratio of rocks has been attributed to fractional crystallization (McCarthy and Cawthorn 1980). The greatest increase in the rubidium/ strontium ratio of rocks (and of strontium content of plagioclase) occurs going upward across the mineral-trend reversal, which strongly suggests reflection of differences between added and residual magmas in this part of the body. A preliminary study of initial strontium isotopic compositions of plagioclase and pyroxene from three samples used in our potassium-argon age determinations (Ford and Kistler 1980) suggests that isotopic relations are complex: 87 Sr/ 86 Sr1 differs significantly not only between the same mineral type in different rocks, but between different mineral types in the same rock sample (table). A finding of common occurrence of such isotopically nonequilibrium values would have important consequences for any interpretation of the intrusion's crystallization history. Our continuing study will incorporate work on isotopic compositions of neodymium and hydrogen. The goal of the isotopic 1986
REVIEW
studies is to attempt to decipher the crystallization record of the body and determine relative roles of a variety of possible processes, particularly those of multiple magma emplacement and crystal or liquid fractionation. Additions and mixing of magmas may lead to development of immiscible sulfide liquids and to concentration of platinum and palladium ore zones in layered igneous complexes (Irvine et al. 1983) and, accordingly, a study of isotopic tracers will contribute to understanding the generally minor-known occurrences of disseminated sulfide minerals (Drinkwater, Ford, and Czamanske 1985) and platinum-group metals (Ford et al. 1983) of the intrusion. This work was supported in part by National Science Foundation grant D1'P 80-20753. References Abel, K.D., G.R. Himmelberg, and A.B. Ford. 1979. Petrologic studies of the Dufek intrusion: Plagioclase variation. Antarctic Journal of the U.S., 14(5), 6-8. Behrendt, J.C., D.J. Drewery, E. Jankowski, and M.S. Grim. 1981. Aeromagnetic and radio echo ice-sounding measurements over the Dufek intrusion, Antarctica. Journal of Geophysical Research, 86, 3014-3020. DePaolo, D.J. 1985. Isotopic studies of processes in mafic magma chambers: I. The Kiglapait intrusion, Labrador. Journal of Petrology, 26(4), 925-951. Ding, T.P., and H.P. Schwarcz. 1984. Oxygen isotopic and chemical compositions of rocks of the Sudbury Basin, Ontario. Canadian Journal of Earth Science, 21, 305-318. Dingwell, D., and T. Dunn. 1982. Oxygen isotopic study of the Precambrian Stillwater intrusion, Montana. Geological Association of Canada-Mineralogical Association of Canada Program with Abstracts, 7, 45. Drinkwater, J.L., A.B. Ford., and G.K. Czamanske. 1985. Study of sulfide mineral distribution in the Dufek intrusion. Antarctic Journal of the U.S., 20(5), 50-51. Ford, A.B. 1976. Stratigraphy of the layered gabbroic Dufek intrusion, Antarctica. U.S. Geological Survey Bulletin, 1405(D), D1-D36. Ford, A.B., and G.R. Himmelberg. In press. Geology and crystallization of the Dufek intrusion. In R.J. Tingey (Ed.), Geology of Antarctica. Oxford: Oxford University Press. Ford, A.B., and R. W. Kistler. 1980. K-Ar age, composition, and origin of Mesozoic mafic rocks related to the Ferrar Group in the Pensacola Mountains, Antarctica. New Zealand Journal of Geology and Geophysics, 23, 371-390. Ford, A.B., R.E. Mays, J. Haffty, and B.P. Fabbi. 1983. Reconnaissance of minor metal abundances and possible resources of the Dufek intrusion, Pensacola Mountains. In R.L. Oliver, J.R. James, and J.B. Jago (Eds.), Antarctic Earth science. Canberra: Australian Academy of Science. 65
Gray, G.M., and A.D.T. Goode. 1981. Strontium isotopic resolution of magma dynamics in layered intrusion. Nature, 294(5837), 155-158. Haensel, J.M., Jr., G.R. Himmelberg, and A.B. Ford. 1986. Plagioclase compositional variations in ariorthosites of the lower part of the Dufek intrusion. Antarctic Journal of the U.S., 21(5). Himmelberg, G.R., and A.B. Ford. 1976. Pyroxenes of the Dufek intrusion, Antarctica. Journal of Petrology, 17(2), 219-243. Himmelberg, G.R., and A.B. Ford. 1977. Iron-titanium oxides of the Dufek intrusion Antarctica. American Mineralogist, 62, 623-633. Hunter, D. 1978. The Bushveld Complex and its remarkable rocks. American Scientist, 66, 551-559. Irvine, T.N. 1980. Magmatic density currents and cumulus processes. American Journal of Science, 280A, 1-58. Irvine, T.N., D.W. Keith, and S.G. Todd. 1983. The J-M platinumpalladium reef of the Stillwater Complex, Montana: II. Origin by double-diffusive convective magma mixing and implications for the
Bushveld Complex. Economic Geology, 78(7), 1287-1334. McCarthy, T.S., and R.G. Cawthorn. 1980. Changes in initial "Sr/'Sr ratio during protracted fractionation in igneous complexes. Journal of Petrology, 21(2), 245-264. Palacz, Z. A., and S. Tait. 1985. Isotopic and geochemical investigation of unit 10 from the Eastern Layered Series of the Rhum intrusion, northwest Scotland. Geological Magazine, 122(5), 485-490. Pankhurst, R.J. 1969. Strontium isotope studies related to petrogenesis in the Caledonian basic igneous province of NE Scotland. Journal of Petrology, 10(1), 115-143. Taylor, H.P., Jr., and R.W. Forester. 1979. An oxygen and hydrogen isotope study of the Skaergaard intrusion and its country rocks: A description of a 55-my. old fossil hydrothermal system. Journal of Petrology, 20(3), 355-419. Wager, L.R., and G.M. Brown. 1968. Layered igneous rocks. Edinburgh: Oliver and Boyd.
Apatites of the Dufek intrusion, a preliminary study
succession about 400-500 meters below the (erosional) top of the intrusion. Our more detailed study shows that major amounts (more than 1.5 volume percent) of cumulus apatite first occur about 450 meters below the top (figure 1) but that minor (0.1-1.5 volume percent) or trace (less than 0.1 volume percent) amounts occur at lower stratigraphic levels. We have found local minor amounts of euhedral, cumulus-appearing apatite in an anorthositic (plagioclase cumulate) layer of the Spear Anorthosite Member of the Aughenbaugh Gabbro near the top of the Dufek Massif section, which is the lowest occurrence known. However, gabbro above the layer appears to be free of cumulus apatite. In the lower part of the Forrestal Range section, above the Dufek Massif section, trace or minor amounts of cumulus-appearing apatite also locally occur in some plagioclase cumulate layers of the Stephens Anorthosite Member of the Saratoga Gabbro (figure 2) and in overlying gabbro, but stratigraphic occurrences are discontinuous. The onset of crystallization of major amounts of cumulus apatite occurred ("Ap +" on figure 1B) just before depostion of a thick layer of gabbro containing abundant, large inclusions of noncumulus anorthosite and leucogabbro. The layer has the appearance of a "megabreccia" and may have formed during some kind of disruption event in the magma chamber (Ford and Himmelberg in press). The greatest concentrations of cumulus apatite occur in a 65-meter-thick gabbroic cumulate above the inclusion-bearing layer (figure 1A). Overlying that unit, major amounts of noncumulus apatite occur in a 25-meter-thick unit of noncumulus mafic and intermediate rock below the Lexington Granophyre. Concentrations of apatite (noncumulus) decrease systematically upward in the Lexington at four localities where the granophyre was studied. The stratigraphic variation in apatite abundances, where known, is closely related to P205 content of the rocks (figure 1B). The relations shown in figure 1 suggest a record of apatite crystallization in the Dufek instrusion generally similar to that of the Skaergaard instrusion of Greenland, in which cumulus apatite crystallization occurred after 90-98 percent of the magma had crystallized (Brown and Peckett 1977). An electron microprobe analysis of cumulus apatite (table) from near the base of the gabbroic cumulate unit of figure 1A shows it to be fluorapatite with a composition similar to those of some Skaergaard fluorapatites (Brown and Peckett 1977; Nash
J.L. DRINKWATER, A.B. FORD, and G.K. CZAMANSKE U.S. Geological Survey Menlo Park, California
As part of a continuing study of the mineralogy of the layered gabbroic Dufek intrusion (82°30'S 50°W), we have begun an investigation of the occurrence, distribution, and chemistry of apatite, the only previously unstudied cumulus mineral in the intrusion. Apatite is generally a minor mineral in layered mafic intrusions such as the Dufek and mostly crystallized at a late stage of differentiation from magma enriched in phosphorus (Wager and Brown 1968). The stratigraphy and rock types of the Dufek intrusion are described by Ford (1976) and the terminology we generally follow is discussed by Irvine (1982). In the Dufek intrusion, apatite has three principal modes of occurrence: (1) cumulus crystals with euhedral, prismatic shape, lying more or less parallel to lamination and layering; (2) postcumulus crystals of subhedral to anhedral shape, occurring in interstices between cumulus silicate and oxide grains; and (3) noncumulus crystals of commonly acicular shape, occurring in granophyre and other noncumulus rock. Postcumulus apatite varies up to about 2 millimeters by 4 millimeters in size and is commonly much larger than cumulus apatite, which generally ranges between 1-3 millimeters in length and 0.05-0.2 millimeters in breadth. Apparent length/breadth ratios are generally less than 10 for postcumulus grains but 12-34 for cumulus grains. Our search of more than 1,000 thin sections representing all exposed rock types and stratigraphic parts of the intrusion shows that apatite of any type is rare in the lower 1.8 kilometers of stratigraphic thickness exposed in the Dufek Massif, but is common in the upper 1.7 kilometers of thickness exposed in the Forrestal Range (figure 1). Based on cursory notations during previous studies of other minerals (Himmelberg and Ford 1976, 1977; Abel, Himmelberg, and Ford 1979), we concluded that apatite first appeared as a cumulus phase in the stratigraphic 66
ANTARCTIC JOURNAL
Strontium and oxygen-isotope study of the Dufek intrusion A.B. FORD, R.W. KISTLER, and L.D. WHITE U.S. Geological Survey Menlo Park, California 94025
The mostly gabbroic Dufek intrusion (82°30'S 50°W) is one of the world's largest layered igneous complexes. It is more than 1986 REVIEW
important insights into the solidification history of the intrusion. In our continuing study of them, we have begun to analyze plagioclases of much more closely spaced (1 meter or less) samples from across four cyclic units of anorthosite and gabbro of the Stephens Anorthosite member (figure 1). The study will include cumulus and post-cumulus pyroxenes and will incorporate isotopic data (Ford, Kistler, and White, Antarctic Journal, this issue) that can provide insight into some magmatic processes not possible from major- and minor-element compositions alone. Similar to the conclusions drawn from the study of isotopic compositions of rocks and minerals, our study of plagioclases indicates that the crystallization history of the intrusion was complex. This work was supported in part by National Science Foundation grant DPP 80-20753 to the U.S. Geological Survey. References Abel, K.D., G.R. Himmelberg, and A.B. Ford. 1979. Petrologic studies of the Dufek intrusion: Plagioclase variation. Antarctic Journal of the U.S., 14(5), 6-8. Czamanske, G.K., and D.L. Schiedle. 1985. Characteristics of the banded-series anorthosites. In G.K. Czamanske and M.L. Zientek (Eds.), The Stillwater Complex, Montana: Geology and guide. (Montana Bureau of Mines and Geology Special Publication, Vol. 92.) Ford, A.B. 1976. Stratigraphy of the layered gabbroic Dufek intrusion, Antarctica. U.S. Geological Survey Bulletin, 1405(D), D1—D36. Ford, A.B., and G.R. Himmelberg. In press. Geology and crystallization of the Dufek intrusion. In R.J. Tingey (Ed.), Geology of Antarctica. Oxford: Oxford University Press. Ford, A.B., R.W. Kistler, and L.D. White. 1986. Strontium and oxygenisotope study of the Dufek intrusion. Antarctic Journal of the U.S., 21(5). Hess, H. H. 1960. Stillwater igneous complex, Montana—A quantitative mineralogical study. Geological Society of America Memoir, 80, 230. Himmelberg, G. and A.B. Ford. 1976. Pyroxenes of the Dufek intrusion, Antarctica. Journal of Petrology, 17(2), 219-243. Himmelberg, G.R., and A.B. Ford. 1977. Iron-titanium oxides of the Dufek intrusion, Antarctica. American Mineralogist, 62, 623-633. Irvine, TN., D.W. Keith, and S.C. Todd. 1983. The J-M platinumpalladium reef of the Stillwater Complex, Montana: II. Origin by double-diffusive convective magma mixing and implications for the Bushveld Complex. Economic Geology, 78(7), 1287-1334. Wager, L.R., and G.M. Brown. 1968. Layered igneous rocks. Edinburgh: Oliver and Boyd.
50,000 square kilometers in area (Behrendt et al. 1981) and possibly 8-9 kilometers thick at its southern end (Ford 1976). It is about the same age as the dominant phase of Jurassic Ferrar Group tholeitic igneous activity that occurred throughout the Transantarctic Mountains (Ford and Kistler 1980). The stratigraphy and rock types are described by Ford (1976). We have begun a systematic study of the isotopic compositions of oxygen, strontium, and neodymium in rocks and minerals of the intrusion to investigate many questions about its crystallization history raised by previous studies. Work on other layered intrusions (Pankhurst 1969; Taylor and Forester 1979; Gray and Goode 1981; Palacz and Tait 1985; DePaolo 1985) shows that isotopic compositions are more sensitive indicators of many magmatic processes than are major- and minor-element compositions of rocks and minerals. 63
A body of such enormous size, with a volume of possibly 200,000 cubic kilometers or more, very likely formed from multiple and possibly interconnected intrusive centers (Ford and Himmelberg in press), as in South Africa's Bushveld Complex (Hunter 1978). An interpretation of the crystallization record of this mostly ice-covered body must encompass not only the processes that may have operated within a single crystallization chamber, such as fractionation, assimilation, and mixing of internal magma-current flows of various origins (Irvine 1980; Irvine, Keith, and Todd 1983) but also such external processes as additions of "primitive" magma or of variably differentiated magma shifted between interconnected crystallization chambers, material loss by volcanism or ancillary intrustion (Ferrar Group tholeiites?), and meteoric and hydrothermal fluid exchanges with country rocks. General stratigraphically upward increases in the ratios of iron to iron plus magnesium in cumulus pyroxenes (Himmelberg and Ford 1976) and normative albite/(albite + anorthite) in cumulus plagioclase (Abel, Himmelberg, and Ford 1979; Haensel, Himmelberg, and Ford, Antarctic Journal, this issue) resemble trends in other layered intrusions that are interpreted in terms of fractional crystallization (Wager and Brown 1968) or crystallization from a nonhomogeneous magma column (Irvine et al. 1983). The trends extend upward through the 1.8-kilometer-thick lower part of the body exposed in the Dufek Massif and up to about 500 meters height in the upper part exposed in the Forrestal Range, where strong reversals occur in the compositional trends of all cumulus minerals, including ilmeno-magnetite (Himmelberg and Ford 1977). The reversals are the strongest evidence yet found for a large-scale replenishment of magma into the chamber, but the magma's source is
unknown. We are at present in a reconnaissance stage of isotopic studies to investigate relative roles of magma additions and mixing, fractional crystallization, and other mechanisms. The isotopic compositions of oxygen (in per mil of isotopic oxygen-18) in 77 rock samples show that values are much more variable and mostly lower in the Forrestal Range than Dufek Massif (figure 1), in which values are comparable to those of Montana Stillwater Complex (Dingwell and Dunn 1982) and norite of Ontario's Sudbury intrusion (Ding and Schwarcz 1984). Variations are not simply related to stratigraphic height and apparent relative age of crystallization, however, because noncumulus granophyre of a 300-meter-thick layer at the top of the Forrestal Range section has isotopic oxygen-18 values higher than those of many underlying cumulates of the section (figure 1). Three analyzed Dufek Massif plagioclases have isotopic oxygen-18 values in a narrow range (6.2-6.9) similar to "normal" plagioclases of the lower and middle zones of Greenland's Skaergaard intrusion (Taylor and Forester 1979). As in rocks, the isotopic oxygen-18 compositions of plagioclases are lower and more variable (3.2-6.0, n = 12) in Forrestal Range than Dufek Massif cumulates. Similarly low values in higher parts of the Skaergaard intrusion show subsolidus isotopic oxygen-18 depletion probably due to activity of a meteoric-hydrothermal system (Taylor and Forester 1979). Rubidium and strontium analyses of 36 rock and 16 plagioclase samples made in preparation for studies of initial isotopic ratios of 87 strontium/" strontium (87Sr/85Sr,) show that (1) the rubidium/strontium ratio of rocks generally increases with stratigraphic height, from an average of about 0.02 in the Dufek Massif to 0.12 in the Forrestal Range section, in which ratios are more variable (figure 2); and (2) strontium content of pla-
10 C') w —J aC') LL
0 Li
Z I8 (PER MIL) Figure 1. Oxygen isotopic compositions of rocks of the Dufek intrusion. Rocks are cumulates of gabbroic, leucogabbroic, anorthositic, and pyroxenitic composition; except (D)thin dikes of aplite and felsite that intrude the cumulates or other rocks, (C) noncumulus gabbro and diorite at the contact with country rocks near Mount Lechner, (G) the Lexington Granophyre or a 300-meter-thick layer at the top of the intrusion (Ford 1976), and (I) inclusions in gabbro of podiform masses of anorthosite and leucogabbro. ("s 180" denotes "isotopic oxygen.")
64
ANTARCTIC JOURNAL
Rubidium and strontium contents and strontium isotopic composition of minerals of the Dufek intrusion Sample Mineral Rubidiuma Strontium' 87Sr/86 Sr' 875r/86Src
65F23713 Pyroxene 2.94 6.5 0.71083±5 0.70763 65F23713 Plagioclase 5.42 296.0 0.70931±4 0.70918 65F38A Pyroxene 0.54 9.0 0.70841 ±5 0.70799 65F941 Pyroxene 4.80 11.5 0.71130±4 0.70834
(I) LL 10 -J 0
° Rubidium and strontium contents measured by isotope dilution. Given in parts per million. b Measured ratio. Ratios normalized to 16Sr/88Sr = 0.1194. NBS87SrCO3.
87 Sr/ 86Sr = 0.71023±3. Initial ratio, corrected to 172 million years (Ford and Kistler 1980).
C') Li 0
X co Z
Es 0 .05 .10 .15 .20
RUBIDIUM /STRONTIUM Figure 2. Rubidium/strontium ratios in cumulates of the Dufek intrusion.
gioclases also generally increases upward in the stratigraphic sequence, from 216 parts per million in basal anorthosite of the Dufek Massif section and 262 and 309 parts per million in gabbros higher in the section, to greater values in the Forrestal Range where plagioclases from below the mineral-trend reversal at 500-meters height average about 315 parts per million strontium (n = 11) and those above contain much larger amounts (436 parts per million, 491 parts per million). In other igneous complexes an increase in the rubidium/strontium ratio of rocks has been attributed to fractional crystallization (McCarthy and Cawthorn 1980). The greatest increase in the rubidium/ strontium ratio of rocks (and of strontium content of plagioclase) occurs going upward across the mineral-trend reversal, which strongly suggests reflection of differences between added and residual magmas in this part of the body. A preliminary study of initial strontium isotopic compositions of plagioclase and pyroxene from three samples used in our potassium-argon age determinations (Ford and Kistler 1980) suggests that isotopic relations are complex: 87 Sr/ 86 Sr1 differs significantly not only between the same mineral type in different rocks, but between different mineral types in the same rock sample (table). A finding of common occurrence of such isotopically nonequilibrium values would have important consequences for any interpretation of the intrusion's crystallization history. Our continuing study will incorporate work on isotopic compositions of neodymium and hydrogen. The goal of the isotopic 1986
REVIEW
studies is to attempt to decipher the crystallization record of the body and determine relative roles of a variety of possible processes, particularly those of multiple magma emplacement and crystal or liquid fractionation. Additions and mixing of magmas may lead to development of immiscible sulfide liquids and to concentration of platinum and palladium ore zones in layered igneous complexes (Irvine et al. 1983) and, accordingly, a study of isotopic tracers will contribute to understanding the generally minor-known occurrences of disseminated sulfide minerals (Drinkwater, Ford, and Czamanske 1985) and platinum-group metals (Ford et al. 1983) of the intrusion. This work was supported in part by National Science Foundation grant D1'P 80-20753. References Abel, K.D., G.R. Himmelberg, and A.B. Ford. 1979. Petrologic studies of the Dufek intrusion: Plagioclase variation. Antarctic Journal of the U.S., 14(5), 6-8. Behrendt, J.C., D.J. Drewery, E. Jankowski, and M.S. Grim. 1981. Aeromagnetic and radio echo ice-sounding measurements over the Dufek intrusion, Antarctica. Journal of Geophysical Research, 86, 3014-3020. DePaolo, D.J. 1985. Isotopic studies of processes in mafic magma chambers: I. The Kiglapait intrusion, Labrador. Journal of Petrology, 26(4), 925-951. Ding, T.P., and H.P. Schwarcz. 1984. Oxygen isotopic and chemical compositions of rocks of the Sudbury Basin, Ontario. Canadian Journal of Earth Science, 21, 305-318. Dingwell, D., and T. Dunn. 1982. Oxygen isotopic study of the Precambrian Stillwater intrusion, Montana. Geological Association of Canada-Mineralogical Association of Canada Program with Abstracts, 7, 45. Drinkwater, J.L., A.B. Ford., and G.K. Czamanske. 1985. Study of sulfide mineral distribution in the Dufek intrusion. Antarctic Journal of the U.S., 20(5), 50-51. Ford, A.B. 1976. Stratigraphy of the layered gabbroic Dufek intrusion, Antarctica. U.S. Geological Survey Bulletin, 1405(D), D1-D36. Ford, A.B., and G.R. Himmelberg. In press. Geology and crystallization of the Dufek intrusion. In R.J. Tingey (Ed.), Geology of Antarctica. Oxford: Oxford University Press. Ford, A.B., and R. W. Kistler. 1980. K-Ar age, composition, and origin of Mesozoic mafic rocks related to the Ferrar Group in the Pensacola Mountains, Antarctica. New Zealand Journal of Geology and Geophysics, 23, 371-390. Ford, A.B., R.E. Mays, J. Haffty, and B.P. Fabbi. 1983. Reconnaissance of minor metal abundances and possible resources of the Dufek intrusion, Pensacola Mountains. In R.L. Oliver, J.R. James, and J.B. Jago (Eds.), Antarctic Earth science. Canberra: Australian Academy of Science. 65
Gray, G.M., and A.D.T. Goode. 1981. Strontium isotopic resolution of magma dynamics in layered intrusion. Nature, 294(5837), 155-158. Haensel, J.M., Jr., G.R. Himmelberg, and A.B. Ford. 1986. Plagioclase compositional variations in ariorthosites of the lower part of the Dufek intrusion. Antarctic Journal of the U.S., 21(5). Himmelberg, G.R., and A.B. Ford. 1976. Pyroxenes of the Dufek intrusion, Antarctica. Journal of Petrology, 17(2), 219-243. Himmelberg, G.R., and A.B. Ford. 1977. Iron-titanium oxides of the Dufek intrusion Antarctica. American Mineralogist, 62, 623-633. Hunter, D. 1978. The Bushveld Complex and its remarkable rocks. American Scientist, 66, 551-559. Irvine, T.N. 1980. Magmatic density currents and cumulus processes. American Journal of Science, 280A, 1-58. Irvine, T.N., D.W. Keith, and S.G. Todd. 1983. The J-M platinumpalladium reef of the Stillwater Complex, Montana: II. Origin by double-diffusive convective magma mixing and implications for the
Bushveld Complex. Economic Geology, 78(7), 1287-1334. McCarthy, T.S., and R.G. Cawthorn. 1980. Changes in initial "Sr/'Sr ratio during protracted fractionation in igneous complexes. Journal of Petrology, 21(2), 245-264. Palacz, Z. A., and S. Tait. 1985. Isotopic and geochemical investigation of unit 10 from the Eastern Layered Series of the Rhum intrusion, northwest Scotland. Geological Magazine, 122(5), 485-490. Pankhurst, R.J. 1969. Strontium isotope studies related to petrogenesis in the Caledonian basic igneous province of NE Scotland. Journal of Petrology, 10(1), 115-143. Taylor, H.P., Jr., and R.W. Forester. 1979. An oxygen and hydrogen isotope study of the Skaergaard intrusion and its country rocks: A description of a 55-my. old fossil hydrothermal system. Journal of Petrology, 20(3), 355-419. Wager, L.R., and G.M. Brown. 1968. Layered igneous rocks. Edinburgh: Oliver and Boyd.
Apatites of the Dufek intrusion, a preliminary study
succession about 400-500 meters below the (erosional) top of the intrusion. Our more detailed study shows that major amounts (more than 1.5 volume percent) of cumulus apatite first occur about 450 meters below the top (figure 1) but that minor (0.1-1.5 volume percent) or trace (less than 0.1 volume percent) amounts occur at lower stratigraphic levels. We have found local minor amounts of euhedral, cumulus-appearing apatite in an anorthositic (plagioclase cumulate) layer of the Spear Anorthosite Member of the Aughenbaugh Gabbro near the top of the Dufek Massif section, which is the lowest occurrence known. However, gabbro above the layer appears to be free of cumulus apatite. In the lower part of the Forrestal Range section, above the Dufek Massif section, trace or minor amounts of cumulus-appearing apatite also locally occur in some plagioclase cumulate layers of the Stephens Anorthosite Member of the Saratoga Gabbro (figure 2) and in overlying gabbro, but stratigraphic occurrences are discontinuous. The onset of crystallization of major amounts of cumulus apatite occurred ("Ap +" on figure 1B) just before depostion of a thick layer of gabbro containing abundant, large inclusions of noncumulus anorthosite and leucogabbro. The layer has the appearance of a "megabreccia" and may have formed during some kind of disruption event in the magma chamber (Ford and Himmelberg in press). The greatest concentrations of cumulus apatite occur in a 65-meter-thick gabbroic cumulate above the inclusion-bearing layer (figure 1A). Overlying that unit, major amounts of noncumulus apatite occur in a 25-meter-thick unit of noncumulus mafic and intermediate rock below the Lexington Granophyre. Concentrations of apatite (noncumulus) decrease systematically upward in the Lexington at four localities where the granophyre was studied. The stratigraphic variation in apatite abundances, where known, is closely related to P205 content of the rocks (figure 1B). The relations shown in figure 1 suggest a record of apatite crystallization in the Dufek instrusion generally similar to that of the Skaergaard instrusion of Greenland, in which cumulus apatite crystallization occurred after 90-98 percent of the magma had crystallized (Brown and Peckett 1977). An electron microprobe analysis of cumulus apatite (table) from near the base of the gabbroic cumulate unit of figure 1A shows it to be fluorapatite with a composition similar to those of some Skaergaard fluorapatites (Brown and Peckett 1977; Nash
J.L. DRINKWATER, A.B. FORD, and G.K. CZAMANSKE U.S. Geological Survey Menlo Park, California
As part of a continuing study of the mineralogy of the layered gabbroic Dufek intrusion (82°30'S 50°W), we have begun an investigation of the occurrence, distribution, and chemistry of apatite, the only previously unstudied cumulus mineral in the intrusion. Apatite is generally a minor mineral in layered mafic intrusions such as the Dufek and mostly crystallized at a late stage of differentiation from magma enriched in phosphorus (Wager and Brown 1968). The stratigraphy and rock types of the Dufek intrusion are described by Ford (1976) and the terminology we generally follow is discussed by Irvine (1982). In the Dufek intrusion, apatite has three principal modes of occurrence: (1) cumulus crystals with euhedral, prismatic shape, lying more or less parallel to lamination and layering; (2) postcumulus crystals of subhedral to anhedral shape, occurring in interstices between cumulus silicate and oxide grains; and (3) noncumulus crystals of commonly acicular shape, occurring in granophyre and other noncumulus rock. Postcumulus apatite varies up to about 2 millimeters by 4 millimeters in size and is commonly much larger than cumulus apatite, which generally ranges between 1-3 millimeters in length and 0.05-0.2 millimeters in breadth. Apparent length/breadth ratios are generally less than 10 for postcumulus grains but 12-34 for cumulus grains. Our search of more than 1,000 thin sections representing all exposed rock types and stratigraphic parts of the intrusion shows that apatite of any type is rare in the lower 1.8 kilometers of stratigraphic thickness exposed in the Dufek Massif, but is common in the upper 1.7 kilometers of thickness exposed in the Forrestal Range (figure 1). Based on cursory notations during previous studies of other minerals (Himmelberg and Ford 1976, 1977; Abel, Himmelberg, and Ford 1979), we concluded that apatite first appeared as a cumulus phase in the stratigraphic 66
ANTARCTIC JOURNAL
