OCEAN DRILLING PROGRAM LEG 183 PRELIMINARY REPORT KERGUELEN PLATEAU_BROKEN RIDGE: A LARGE IGNEOUS PROVINCE Dr. Millard Coffin Co-Chief Scientist Institute for Geophysics University of Texas at Austin 4412 Spicewood Springs Road Building 600 Austin, TX 78759-8500 U.S.A. Dr. Frederick Frey Co-Chief Scientist Department of Earth, Atmospheric and Planetary Sciences Massachusetts Institute of Technology Building 54-1226 77 Massachusetts Avenue Cambridge, MA 02139 U.S.A. Dr. Paul Wallace Staff Scientist Ocean Drilling Program Texas A&M University 1000 Discovery Drive College Station, TX 77845-9547 U.S.A. _________________________ Dr. Jack Baldauf Deputy Director of Science Operations ODP/TAMU _________________________ Dr. Paul Wallace Leg Project Manager Science Services ODP/TAMU April 1999 This informal report was prepared from the shipboard files by the scientists who participated in the cruise. The report was assembled under time constraints and is not considered to be a formal publication which incorporates final works or conclusions of the participating scientists. The material contained herein is privileged proprietary information and cannot be used for publication or quotation. Preliminary Report No. 83 First Printing 1999 Distribution Electronic copies of this publication may be obtained from the ODP Publications Homepage on the World Wide Web at http://www-odp.tamu.edu/publications. D I S C L A I M E R This publication was prepared by the Ocean Drilling Program, Texas A&M University, as an account of work performed under the international Ocean Drilling Program, which is managed by Joint Oceanographic Institutions, Inc., under contract with the National Science Foundation. Funding for the program is provided by the following agencies: Australia/Canada/Chinese Taipei/Korea Consortium for the Ocean Drilling Deutsche Forschungsgemeinschaft (Germany) Institut Franáais de Recherche pour l'Exploitation de la Mer (France) Ocean Research Institute of the University of Tokyo (Japan) National Science Foundation (United States) Natural Environment Research Council (United Kingdom) European Science Foundation Consortium for the Ocean Drilling Program (Belgium, Denmark, Finland, Iceland, Italy, The Netherlands, Norway, Spain, Sweden, and Switzerland) Marine High-Technology Bureau of the State Science and Technology Commission of the People's Republic of China Any opinions, findings, and conclusions or recommendations expressed in this publication are those of the author(s) and do not necessarily reflect the views of the National Science Foundation, the participating agencies, Joint Oceanographic Institutions, Inc., Texas A&M University, or Texas A&M Research Foundation. The following scientists were aboard JOIDES Resolution for Leg 183 of the Ocean Drilling Program: Millard F. Coffin Co-Chief Scientist Institute for Geophysics University of Texas at Austin 4412 Spicewood Springs Road Building 600 Austin, TX 78759-8500 U.S.A. Internet: mikec@utig.ig.utexas.edu Work: (512) 471-0429 Fax: (512) 471-8844 Frederick A. Frey Co-Chief Scientist Department of Earth, Atmospheric and Planetary Sciences Massachusetts Institute of Technology Building 54-1226 77 Massachusetts Avenue Cambridge, MA 02139 U.S.A. Internet: fafrey@mit.edu Work: (617) 253-2818 Fax: (617) 253-7102 Paul J. Wallace Staff Scientist Ocean Drilling Program Texas A&M University 1000 Discovery Drive College Station, TX 77845-9547 U.S.A. Internet: Paul_Wallace@odp.tamu.edu Work: (409) 845-0879 Fax: (409) 845-0876 Maria J. Antretter Paleomagnetist Institut fÅr Allgemeine und Angewandte Geophysik Ludwig-Maximilians-UniversitÑt MÅnchen Theresienstrasse 41 MÅnchen 80333 Federal Republic of Germany Internet: maria@geoelek.geophysik.uni- muenchen.de Work: (49) 89-2394-4206 Fax: (49) 89-2394-4205 Nicholas T. Arndt Igneous Petrologist GÇosciences UniversitÇ de Rennes I Avenue de GÇnÇral Leclerc Rennes Cedex 35042 France Internet: arndt@univ-rennes1.fr Work: (33) 2-99-28-67-79 Fax: (33) 2-99-28-67-80 Jane Barling Igneous Petrologist Department of Earth and Environmental Sciences University of Rochester Hutchison Hall 227 Rochester, NY 14627 U.S.A. Internet: barling@earth.rochester.edu Work: (716) 275-2514 Fax: (716) 244-5689 Florian Boehm Sedimentologist GEOMAR Wischhofstr. 1-3 D-24248 Kiel Germany Internet: fboehm@geomar.de Work: 49-431-600-2842 Fax: 49-431-600-2941 Mai Kirstine Borre Physical Properties Specialist Department of Geology and Geotechnical Engineering Technical University of Denmark Building 204 DK-2800 Lyngby Denmark Internet: iggmkb@pop.dtu.dk Work: 45-45 25 21 63 Fax: 45-45 88 59 35 Helen K. Coxall Paleontologist Department of Geology University of Bristol Wills Memorial Building Queens Road Bristol BS8 1RJ United Kingdom Internet: h.k.coxall@bris.ac.uk Work: (44) 117-928-9000 Fax: (44) 117-925-3385 John E. Damuth Sedimentologist Department of Geology University of Texas at Arlington P.O. Box 19049 Arlington, TX 76019-0049 U.S.A. Internet: damuth@uta.edu Work: (817) 272-2976 Fax: (817) 272-2628 Heike Delius Logging Scientist Angewandte Geophysik Rheinisch-WestfÑlischen Technischen Hochschule Aachen Lochnerstrasse 4-20 Aachen 52056 Federal Republic of Germany Internet: heike@sun.geophac.rwth-aachen.de Work: (49) 241-804831 Fax: (49) 241-8888-132 Robert A. Duncan Igneous Petrologist College of Oceanography Oregon State University Oceanography Administration Building 104 Corvallis, OR 97331-5503 U.S.A. Internet: rduncan@oce.orst.edu Work: (541) 737-5206 Fax: (541) 737-2064 Hiroo Inokuchi Paleomagnetist School of Humanity Environment Policy and Technology Himeji Institute of Technology Shinzaikehonmachi 1-1-12 Himeji, Hyogo 670-0092 Japan Internet: inokuchi@hept.himeji-tech.ac.jp Work: (81) 792-92-1515, ext 319 Fax: (81) 792-93-5710 Laszlo Keszthelyi Volcanologist Hawaii Institute of Geophysics and Planetology University of Hawaii at Manoa 2525 Correa Road Honolulu, HI 96822 U.S.A. Internet: lpk@pirl.lpl.arizona.edu Work: (808) 967-8825 Fax: (808) 967-8890 John J. Mahoney Igneous Petrologist School of Ocean and Earth Science and Technology University of Hawaii at Manoa 2525 Correa Road Honolulu, HI 96822 U.S.A. Internet: j.mahoney@soest.hawaii.edu Work: (808) 956-8705 Fax: (808) 956-2538 C. Leah Moore Volcanologist Department of Geology Australian National University CRC LEME ACT 0200 Australia Internet: leah@basins.anu.edu.au Work: (61) 2-6201-5296 Fax: (61) 2-6201-5728 R. Dietmar MÅller JOIDES Resolution Logging Scientist Department of Geology and Geophysics University of Sydney Building F05 Sydney, NSW 2006 Australia Internet: dietmar@es.su.oz.au Work: (61) 2-9351-2003 Fax: (61) 2-9351-0184 Clive R. Neal Igneous Petrologist Department of Civil Engineering and Geological Sciences University of Notre Dame Notre Dame, IN 46556 U.S.A. Internet: neal.1@nd.edu Work: (219) 631-8328 Fax: (219) 631-9236 Kirsten E. Nicolaysen Igneous Petrologist Department of Earth, Atmospheric and Planetary Sciences Massachusetts Institute of Technology Building 54-1218 77 Massachusetts Avenue Cambridge, MA 02139 U. S. A. Internet: knic@mit.edu Work: (617) 253-2869 Fax: (617) 253-7102 Malcolm S. Pringle Igneous Petrologist Isotope Geosciences Unit Scottish Universities Research and Reactor Centre Scottish Enterprise Technology Park East Kilbride, Glasgow G75 0QU United Kingdom Internet: m.pringle@surrc.gla.ac.uk Work: (44) 1355-223332 Fax: (44) 1355-229898 Douglas N. Reusch Sedimentologist Department of Geological Sciences University of Maine 5790 Bryand Global Sciences Center Orono, ME 04469-5790 U.S.A. Internet: doug@iceage.umeqs.maine.edu Work: (207) 581-2186 Fax: (207) 581-2202 Peter J. Saccocia Metamorphic Petrologist Department of Earth Sciences and Geography Bridgewater State College Bridgewater, MA 02325 U.S.A. Internet: psaccocia@bridgew.edu Work: (508) 697-1200, ext 2124 Fax: (508) 697-1785 Damon A.H. Teagle Metamorphic Petrologist Department of Geological Sciences University of Michigan 2534 C.C. Little Building Ann Arbor, MI 48109-1063 U.S.A. Internet: teagle@umich.edu Work: (313) 763-8060 Fax: (313) 763-4690 Veronika WÑhnert Palynologist/Sedimentologist Museum fÅr Naturkunde Institut fÅr PalÑontologie Invalidenstrasse 43 Berlin 10115 Federal Republic of Germany Internet: h1782bvx@rz.hu-berlin.de Work: 0049-30-2093-8855 Fax: 0049-30-2093-8868 Dominique A.M. Weis Igneous Petrologist DÇpartment des Sciences de la Terre et de L'Environnement UniversitÇ Libre de Bruxelles C.P. 160/02 Avenue F.D. Roosevelt 50 Bruxelles 1050 Belgium Internet: dweis@resulb.ulb.ac.be Work: (32) 2-650-3748 Fax: (32) 2-650-2226 Sherwood W. Wise Paleontologist Department of Geology Florida State University Carraway Building Tallahassee, FL 32306-4100 U.S.A. Internet: wise@.gly.fsu.edu Work: (904) 644-6265 Fax: (904) 644-4214 Xixi Zhao Physical Properties Specialist Earth Sciences Department University of California, Santa Cruz Institute of Tectonics Santa Cruz, CA 95064 U.S.A. Internet: xzhao@earthsci.ucsc.edu Work: (408) 459-4847 Fax: (408) 459-3074 SCIENTIFIC REPORT ABSTRACT Most of the Kerguelen Plateau and Broken Ridge formed as a single giant oceanic plateau in Cretaceous time. During Ocean Drilling Program Leg 183, igneous basement rock and sediment cores were obtained from five sites on the Kerguelen Plateau and two on Broken Ridge. Based on the recovery of basalt, other igneous rocks, and interbedded and overlying sediment, we found that From south to north, the age of the uppermost crust forming this very large igneous province (LIP) decreases, possibly in steps (i.e., ~110 Ma in the southern Kerguelen Plateau, ~85 to 95 Ma in the central Kerguelen Plateau, Broken Ridge, and Elan Bank, and Û35 Ma in the northern Kerguelen Plateau); the submarine igneous basement of Elan Bank and the northen Kerguelen Plateau had not been previously sampled. The growth rate of the LIP at five of seven basement sites was sufficient to form a subaerial landmass. This was most spectacularly revealed at central Kerguelen Plateau Site 1138 by wood fragments in a dark brown sediment overlying the subaerially erupted lava flows, a result consistent with the charcoal and wood fragments in sediments overlying igneous rocks at Site 750 in the southern Kerguelen Plateau. The terminal stage of volcanism forming the LIP included explosive eruptions of volatile- rich felsic magmas formed from cooling basaltic magmas that were trapped within the crust when the flux of basaltic magma from the mantle decreased. Previous geochemical studies of basalt from the southern Kerguelen Plateau and eastern Broken Ridge had identified a component derived from continental crust (e.g., Mahoney et al., 1995), but the mechanism for incorporation of a continental component into the oceanic plateau was unconstrained. Possible processes range from recycling of continental material into a deep mantle plume to contamination of mantle-derived basaltic magma by fragments of continental crust isolated in the embryonic Indian Ocean crust during rifting of Gondwana. At Site 1137 on Elan Bank, ~26 m of a braided river conglomerate was intercalated with basaltic flows; the clasts in this conglomerate show the wide range of rock types that were subaerially exposed on Elan Bank. Most notable are clasts of garnet-biotite gneiss, a rock type that is characteristic of continental crust, thereby showing that a continental fragment is present in this oceanic environment. INTRODUCTION Large igneous provinces (LIPs) are a significant type of planetary volcanism found on the Earth, the moon, Venus, and Mars (Coffin and Eldholm, 1994; Head and Coffin, 1997). They represent large volumes of magma emplaced over relatively short time periods, such as expected from decompression of upwelling, relatively hot or wet mantle. This process explains hot spot magmatism at the Earth's surface and is conceptually described by various plume head and tail models applicable to the Earth's sublithospheric mantle. In such models, the plume head leads to oceanic plateaus and continental flood basalts, and the tail leads to volcanic chains known as hot spot tracks. Terrestrial LIPs are dominantly mafic rocks formed during several distinct episodes in Earth's history, perhaps in response to fundamental changes in the processes that control energy and mass transfer from the Earth's interior to its surface. The ocean basins contain several Cretaceous LIPs; the two largest are the Kerguelen Plateau_Broken Ridge in the Indian Ocean (Fig. 1) and the Ontong Java Plateau in the Pacific Ocean. Both are elevated regions of the ocean floor encompassing areas of ~2 x 106 km2 (Coffin and Eldholm, 1994). These giant LIPs are important for several reasons. They provide information about mantle compositions and dynamics that are not revealed by volcanism at spreading ridges. For example, today's plume-associated volcanism (principally, oceanic islands) accounts for only 5% to 10% of the magma and energy expelled from the Earth's mantle, but the giant LIPs may have contributed as much as 50% in Early Cretaceous time (Coffin and Eldholm, 1994), thereby indicating a substantial change in mantle dynamics from Cretaceous to present time (e.g., Stein and Hofmann, 1994). Because magma fluxes represented by oceanic plateaus are not evenly distributed in space and time, their episodicity punctuates the relatively steady-state production of crust at seafloor spreading centers. These intense episodes of igneous activity temporarily increase the flux of magma and heat from the mantle to the crust, hydrosphere, and atmosphere, possibly resulting in global environmental changes, such as excursions in the composition and isotopic characteristics of seawater (e.g., Larson, 1991; Ingram et al., 1994; Jones et al., 1994; Bralower et al., 1997). Finally, because oceanic LIPs apparently resist subduction, they contribute to the growth of continents. The Kerguelen Plateau_Broken Ridge LIP is interpreted to represent voluminous Cretaceous volcanism associated with the arrival of the Kerguelen plume head below young Indian Ocean lithosphere (Fig. 2) (e.g., Duncan and Storey, 1992; M.F. Coffin et al., unpubl. data). Subsequently, rapid northward movement of the Indian plate over the plume stem formed a 5000- km-long, ~82 to 38 Ma, hot spot track, the Ninetyeast Ridge (Duncan, 1991). At ~40 Ma the newly formed Southeast Indian Ridge (SEIR) intersected the plume's position. As the SEIR migrated northeast relative to the plume, hot spot magmatism became confined to the Antarctic plate. From ~40 Ma to the present, the Kerguelen Archipelago, Heard and McDonald Islands, and a northwest-southeast_trending chain of submarine volcanoes between these islands were constructed on the northern and central sectors of the Kerguelen Plateau (Figs. 1, 2, 3, 4). Thus, an ~115-m.y. record of volcanism is attributed to the Kerguelen plume (e.g., Mahoney et al., 1983; Weis et al., 1992; M.F. Coffin et al., unpubl. data). Despite their huge size and distinctive morphology, oceanic plateaus remain among the least understood features in the ocean basins. This drilling leg focused on sampling the Kerguelen Plateau_Broken Ridge LIP to determine (1) the age and composition of the basement volcanic rocks in all major parts of the LIP, (2) the mantle and crustal components that contributed to the magmatism, (3) the mass transfer and chemical fluxes between the volcanic crust and atmosphere- hydrosphere-biosphere system, and (4) the tectonic history of the LIP beginning with the mechanisms of growth and emplacement and continuing with the multiple episodes of postconstructional deformation that created the present complex bathymetry (Figs. 3, 4). STUDY AREA Physical Description The Kerguelen Plateau is a broad topographic high in the southern Indian Ocean surrounded by deep ocean basins_to the northeast by the Australian-Antarctic Basin, to the south by the 3500-m-deep Princess Elizabeth Trough, to the southwest by the Enderby Basin, and to the northwest by the Crozet Basin (Fig. 3). The plateau stretches ~2300 km between 46¯S and 64¯S in a southeast-trending direction toward the Antarctic continental margin. It is between 200 and 600 km wide and stands 2_4 km above the adjacent ocean basins. Variable age oceanic crust abuts the Kerguelen Plateau (Fig. 3). As summarized by Schlich and Wise (1992), the oldest magnetic anomalies range from Chron C11 (30.1 Ma) in the northeast, to Chron C18 (40.1 Ma) off the central part of the eastern plateau (we use the geomagnetic polarity time scale of Cande and Kent, 1995). Farther south, the eastern flank of the Southern Kerguelen Plateau is bounded by the Labuan Basin. Basement of the Labuan Basin has not been sampled by drilling, but its structure resembles that of the main Kerguelen Plateau (Rotstein et al., 1991; Munschy et al., 1992). To the northwest, magnetic anomaly sequences from Chrons C23 to C34 have been identified in the Crozet Basin, but to the southwest no convincing anomalies have been identified in the Enderby Basin, although Mesozoic anomalies have been suggested (Li, 1988; Nogi et al., 1996). An Early Cretaceous age for the Enderby Basin is assumed in most plate reconstructions (e.g., Royer and Coffin, 1992). Beginning with early studies (Schlich, 1975; Houtz et al., 1977), the Kerguelen Plateau province has been divided into distinct domains. Coffin et al. (1986) and Kînnecke et al. (1998) recognize five domains: northern, central, and southern Kerguelen Plateau; Elan Bank; and the Labuan Basin (Figs. 3, 4). The northern Kerguelen Plateau (NKP), ~46¯S to 50¯S, has shallow water depths (<1000 m) and basement elevations 3000_4000 m above adjacent seafloor, with maximum elevations forming the Kerguelen Archipelago. A lack of rocks older than ~40 Ma from the Kerguelen Archipelago (K.E. Nicolaysen et al., unpubl. data), as well as plate reconstructions (Royer and Sandwell, 1989; Royer and Coffin, 1992), suggests that the age of the NKP is Û40 Ma, whereas the central and southern domains of the submarine Kerguelen Plateau appear to be of Cretaceous age (M.F. Coffin et al., unpubl. data). However, the basement of the submarine NKP had never been sampled before Leg 183. The central Kerguelen Plateau (CKP), ~50¯S to 55¯S, is also relatively shallow, contains a major sedimentary basin (Kerguelen-Heard Basin), and includes the volcanically active Heard and McDonald Islands. Broken Ridge and the CKP are conjugate Late Cretaceous provinces (Fig. 1) that were separated by seafloor spreading along the SEIR during Eocene time (Mutter and Cande, 1983). The southern Kerguelen Plateau (SKP) apparently formed in Early Cretaceous time (M.F. Coffin et al., unpubl. data). Relative to the NKP, water depths are greater (1500 to 2500 m), and it is tectonically more complex (Figs. 3, 4). There are several large basement uplifts and evidence for multiple stages of normal faulting, graben formation, and strike-slip faulting (e.g., Coffin et al., 1986; Fritsch et al., 1992; Rotstein et al., 1992; Royer and Coffin, 1992; Angoulvant-Coulon and Schlich, 1994; Kînnecke and Coffin, 1994; Gladczenko et al., 1997). Elan Bank, a salient extending westward from the boundary between the CKP and SKP, has water depths from <1000 to 2000 m. Before Leg 183, basement had not been sampled from Elan Bank and its age was, therefore, unknown. Labuan Basin, which adjoins the CKP and SKP to the east, is a deep (>3500 m), extensively faulted, thickly sedimented (>2 s two-way traveltime in places, or >2000 m, assuming a sediment velocity of 2000 m/s) basin. Dredging of an exposed faulted basement block in the Labuan Basin recovered metamorphic and granitic rock; these rocks have been interpreted as ice-rafted debris (Montigny et al., 1993). Thus, the basin's age and the nature of its crust (i.e., oceanic or continental) remain uncertain. The CKP was contiguous with Broken Ridge when these domains formed during Cretaceous time (Houtz et al., 1977; Coffin et al., unpubl. data; Duncan, 1991). Subsequently, at ~40 Ma, Broken Ridge and the CKP began to separate along the nascent SEIR. Broken Ridge, now ~1800 km north of the Kerguelen Plateau, is a narrow and elongated oceanic plateau (100_200 km by ~1000 km at ~2000 m water depth) that trends west-northwest (Figs. 5, 6). It is markedly asymmetric in cross section, dipping gently (<2¯) toward the north but with a steeply dipping (>10¯) southern face (Fig. 5). This southern flank was uplifted, perhaps more than 2000 m, during the early Tertiary breakup between Broken Ridge and the Kerguelen Plateau (Weissel and Karner, 1989; Peirce et al., 1989). Crustal Structure Ocean Drilling Program (ODP) Legs 119 and 120 drilling results (Barron, Larsen, et al., 1989; Schlich, Wise, et al., 1989), dredging data (Leclaire et al., 1987; Davies et al., 1989; Duncan, 1991; Weis et al., 1998), and multichannel seismic reflection data (Coffin et al., 1990; Schaming and Rotstein, 1990; Schlich et al., 1993) have shown that igneous basement of the Kerguelen Plateau and conjugate Broken Ridge is basaltic. Numerous dipping intrabasement reflections interpreted as flood basalts have been identified in the crust of the CKP and SKP and on Elan Bank (Kînnecke et al., 1997). Wide-angle seismic data from the Kerguelen Archipelago on the NKP show an upper igneous crust 8_9.5 km thick and a lower crust 6_9.5 km thick (Recq and Charvis, 1986; Recq et al., 1990, 1994; Charvis et al., 1995). Wide-angle reflection and refraction experiments employing ocean-bottom seismometers have been undertaken recently on both the CKP and SKP (Charvis et al., 1993, 1995; Operto and Charvis, 1995, 1996; Kînnecke et al., 1998; Charvis and Operto, 1998). The crustal structure beneath the Kerguelen Archipelago differs significantly from that of the CKP. Igneous crust of the CKP is 19 to 21 km thick and is composed of three layers. The upper layer is 1.2 to 2.3 km thick, and velocities range from 3.8 to 4.9 km/s. It could be composed of either lava flows or interlayered volcanic and sedimentary beds. The second layer is 2.3 to 3.3 km thick, and velocities increase downward from 4.7 to 6.7 km/s. In the ~17-km-thick lower crust, velocities increase from 6.6 km/s at ~8.0 km depth (near the top of the layer) to 7.4 km/s at the base of the crust, with no internal discontinuity. On the southern plateau, the ~22-km- thick igneous crust can be divided into three layers: (1) an upper crustal layer ~5.3 km thick with velocities ranging from 3.8 to 6.5 km/s; (2) a lower crustal layer ~11 km thick with velocities of 6.6 to 6.9 km/s; and (3) a 4- to 6-km-thick transition zone at the base of the crust characterized by velocities of 6.7 to 6.9 km/s (Operto and Charvis, 1995, 1996). This low- velocity, seismically reflective transition zone at the crust/mantle interface has not been imaged on the NKP or CKP; it is the basis for the hypothesis that parts of the SKP contain fragments of continental crust (Operto and Charvis, 1995; 1996). Previous Sampling of Igneous Basement: Ages and Geochemical Characteristics In this section we summarize results of previous sampling (Legs 119 and 120 and dredging) of the Kerguelen Plateau_Broken Ridge LIP. Based in large part on ODP- related studies, there is a consensus that the Kerguelen plume was the major source of magma for constructing the Kerguelen Plateau_Broken Ridge LIP. Although sampling and dating of the entire LIP are grossly insufficient, sampling of the SKP at four spatially diverse locations (Sites 738, 749, and 750, and dredge site MD48-05; see Figs. 3, 4) shows that the uppermost igneous crust of SKP formed over a relatively short interval at ~110 Ma (K/Ar data from Leclaire et al., 1987; Whitechurch et al., 1992; and 40Ar/39Ar data from Pringle et al., 1994; Storey et al., 1996; M.F. Coffin et al., unpubl. data). In contrast, basement basalts from Site 747 on the CKP may be much younger, ~85 Ma (M.F. Coffin et al., unpubl. data). This age is similar to the 83_88 Ma age for lavas from Broken Ridge dredge sites 8 and 10 (40Ar/39Ar data from Duncan, 1991), which coincide spatially with the prebreakup position of Site 747 (Figs. 1, 2). Also, piston coring of sediments on the northeast flank of the CKP between the Kerguelen Archipelago and Heard Island (MD35-510 in Fig. 4) recovered cherts and calcareous oozes of probable Santonian age (Frîhlich and Wicquart, 1989). In summary, we have very few high-quality age data for the 2.3 x 106 km2 (equivalent to approximately eight Icelandic plateaus) of the Kerguelen Plateau_Broken Ridge LIP. These sparse data support the hypothesis that large magma volumes erupted over short time intervals, possibly as two pulses during Cretaceous time_the SKP at ~110 Ma; the CKP, Broken Ridge, and perhaps Elan Bank at ~85 Ma (Fig. 7). In contrast, Cenozoic volcanism (~38 Ma to present) has formed the Kerguelen Archipelago (e.g., K.E. Nicolaysen et al., unpubl. data), Heard and McDonald Islands (Clarke et al., 1983; Quilty et al., 1983), and the bathymetric/gravity highs between the Kerguelen Archipelago and Heard Island (Weis et al., 1998). A major goal of Leg 183 was to drill at other sites throughout the plateau to determine if formation of this LIP was truly episodic or if there was a continuous south to north decrease in age of volcanism. Although the southern and central Kerguelen Plateau formed in a young oceanic basin (Royer and Coffin, 1992; Munschy et al., 1994; Coffin et al., unpubl. data), evidence is equivocal as to whether it formed at a spreading center, like Iceland, or off-ridge, like Hawaii (Coffin and Gahagan, 1995). Before Leg 183, several observations had indicated that much of the uppermost basement of the southern and central Kerguelen Plateau erupted in a subaerial environment_specifically, (1) oxidized flow tops and vesicularity of lava flows at Sites 738 and 747; (2) nonmarine, organic-rich sediments containing up to 5-cm pieces of charcoal overlying the basement at Site 750; and (3) claystone overlain by a basalt cobble conglomerate and glauconitic sediment with wood fragments in the lowermost core at Site 748 (Schlich et al., 1987). Coffin (1992) concluded that the drill sites in the SKP had long (>10 to Û50 m.y.) histories of subaerial volcanism and erosion, followed by subsidence caused by cooling. Zeolite mineralogy of the basaltic basement indicates erosion to deeper levels at Site 749 than at Sites 747 and 750 (Sevigny et al., 1992). The islands on the Kerguelen Plateau are dominantly formed of <40- Ma transitional and alkaline lavas (Fig. 8) (Weis et al., 1993, 1998; Barling et al., 1994; Yang et al., 1998; K.E. Nicolaysen et al., unpubl. data). Before Leg 183, the only alkaline basalt recovered from the Kerguelen Plateau, Broken Ridge, and Ninetyeast Ridge was a flow ~200 m above basement at Site 748. Tholeiitic basalt of Cretaceous age has been recovered from four dredge and four drill sites on the central and southern Kerguelen Plateau and three dredge sites on Broken Ridge (Figs. 3, 4, 5, 6, 8); seven drill sites on Ninetyeast Ridge have yielded solely tholeiitic basalt ranging from 38 to 82 Ma (Fig. 1). Although the tholeiitic basalts from several of these sites are geochemically distinct, their incompatible element abundances resemble those of ocean-island tholeiitic basalts, rather than typical mid-ocean-ridge basalts (MORBs) (Kerguelen Plateau and Broken Ridge: Davies et al., 1989; Weis et al., 1989; Storey et al., 1992; Mahoney et al., 1995; Ninetyeast Ridge: Frey et al., 1991; Saunders et al., 1991; Frey and Weis, 1995). We infer that tholeiitic basalt was the dominant magma type produced by the Kerguelen plume from ~110 to 38 Ma during formation of the Kerguelen Plateau, Broken Ridge, and Ninetyeast Ridge. The significance of this result is that tholeiitic basalts are derived from relatively high (>5%) extents of partial melting (Kent and McKenzie, 1994), and the inference is that the Kerguelen plume was a high-flux magma source for a long time (Figs. 2, 7). However, the MgO-rich melts expected from large extents of melting of high-temperature plumes (e.g., Storey et al., 1991) have not been recovered from Cretaceous parts of the Kerguelen Plateau. Picritic (i.e., olivine rich) alkaline lavas of Quaternary age are found on Heard Island (Barling et al., 1994), and transitional picritic lavas (14 to 19 Ma) were recently dredged from one of the bathymetric/gravity highs between the Kerguelen Archipelago and Heard Island (Weis et al., 1998) (see Figs. 3, 4). All of these picrites are olivine-rich cumulates rather than crystallized MgO-rich melts. Most lavas from the Kerguelen Plateau and Broken Ridge have Sr and Nd isotopic ratios that range from the high 87Sr/86Sr-low 143Nd/144Nd end of the field for SEIR MORB to the field proposed for the Kerguelen plume (Fig. 9). In Pb-Pb isotopic plots, Kerguelen Plateau lavas from Sites 747, 749, and 750 define an elongate field subparallel to that for SEIR MORB (Fig. 10); however, like lavas forming the Kerguelen Archipelago, the Kerguelen Plateau lavas are offset from the MORB field to higher 208Pb/204Pb and 207Pb/204Pb at a given 206Pb/204Pb ratio. In addition, submarine Kerguelen Plateau lavas extend to lower 206Pb/204Pb than Kerguelen Archipelago lavas (Fig. 10). These Sr, Nd, and Pb isotopic data have been interpreted as a result of mixing between the Kerguelen plume and entrained depleted (MORB related) asthenosphere (e.g.,Weis et al., 1992). In contrast, basalts from Site 738 on the southernmost SKP and dredge 8 from eastern Broken Ridge (Figs. 1, 2, 3, 4, 5) have atypical geochemical characteristics for oceanic lavas. These lavas have very high 87Sr/86Sr, low 143Nd/144Nd, and very high 208Pb/204Pb and 207Pb/204Pb ratios that accompany relatively low 206Pb/204Pb (Figs. 9, 10). They also have relative depletions in abundances of Nb and Ta, and there is a positive correlation between 87Sr/86Sr and the extent of Nb depletion (Fig. 11). Mahoney et al. (1995) concluded that these isotopic characteristics, coupled with depletions of Nb and Ta, arose from a continental lithosphere component that contributed to these basalts, a hypothesis also proposed by Storey et al. (1989) to account for Ta depletion in basalts dredged from the Kerguelen Plateau. Basalts from Site 738 on the SKP and dredged from eastern Broken Ridge are relatively depleted in Nb (Fig. 12A). Significant relative depletion in Nb is also evident in basalts dredged from the 77¯ graben on the SKP and from Site 747 on the CKP. Trends to anomalously high Th/Nb and La/Nb and a positive correlation between 87Sr/86Sr and La/Nb are also defined by the Bunbury Basalt, southwest Australia, and the Rajmahal Basalt, northeast India (Figs. 11, 12B). These continental basalts, erupted at ~123_130 Ma and 116 Ma, respectively, are contaminated to varying degrees by continental crust (Frey et al., 1996; Kent et al., 1997). The combination of geochemical features in basalts from Site 738 and eastern Broken Ridge (i.e., very high 87Sr/86Sr and low 143Nd/144Nd; high 208Pb/204Pb and 207Pb/204Pb ratios that accompany relatively low 206Pb/204Pb; anomalously high Th/Nb and La/Nb; see Figs. 9, 10, 11, 12) is consistent with continental crust as the continental component. In particular, the low 206Pb/204Pb ratios require aged crust with low U/Pb, such as some types of Archean crust. In detail, the trend for Site 747 lavas (Fig. 12A) differs from that of other Kerguelen Plateau basalts because Site 747 lavas trend to high La/Nb without elevated Th/Nb. This trend is similar to that for North Atlantic MORB from the lower flow units at Hole 917A (Fig. 12B), which are contaminated by the Archean crust of eastern Greenland, specifically, lower crustal granulite-facies gneiss (Fitton et al., 1998a, 1998b). The effects of continental contamination are very evident in these Hole 917A basalts because their parental magmas were MORB-like with much lower incompatible element abundances than are found in most plume-related lavas. Note that the combination of elevated La/Nb with low Th/Nb is unlike recent estimates of average lower crust composition (e.g., LC in Fig. 12; Rudnick and Fountain, 1995) but is typical of Lewisian granulites (Fig. 12; lower crust estimates are from Weaver and Tarney, 1984). Although not as extreme as some basalts from the lower units of Hole 917A, several geochemical characteristics of Site 747 basalts are consistent with crustal contamination_namely, (1) the trend to high La/Nb without high Th/Nb; (2) the offset to low 143Nd/144Nd from the 87Sr/86Sr-143Nd/144Nd trend defined by Kerguelen Archipelago lavas; and (3) the low 206Pb/204Pb ratios, which are lower than those of all other lavas from the Kerguelen Plateau, Ninetyeast Ridge, Kerguelen Archipelago, and Heard Island (Figs. 9, 10, 11, 12). These characteristics are consistent with ancient continental crust as the contaminant; the high La/Nb_low Th/Nb (Fig. 12A) trend suggests a component similar to Lewisian granulites. Archean granulites are on the conjugate Antarctic and Indian margins (e.g., Black et al., 1992), and it is possible that fragments of such crust were incorporated into the embryonic Indian Ocean and subsequently sampled during formation of the Kerguelen Plateau. In addition, the Os and Pb isotopic ratios of peridotite xenoliths in basalts from the Kerguelen Archipelago are interpreted as reflecting Gondwana lithospheric mantle that was incorporated into the Indian Ocean mantle during rifting (Hassler and Shimizu, 1998; Mattielli et al., 1999). The geochemical evidence for a continental component in basalts forming the Kerguelen Plateau and Broken Ridge is consistent with a crustal velocity structure suggesting that the SKP contains a stretched continental fragment (Operto and Charvis, 1995; 1996); this geophysical evidence is at ~58¯S in the vicinity of the basalts dredged from the 77¯ graben and cored at Site 750, whereas Site 738 is much farther south at ~63¯S and Site 747 is to the north at ~55¯S (Fig. 3, 4). These results suggest that continental lithosphere may be widespread in the Kerguelen Plateau. SCIENTIFIC OBJECTIVES Leg 183 objectives focused on four major problems related to the formation and evolution of a giant LIP: 1. Chronology of Kerguelen Plateau_Broken Ridge magmatism: The goal was to quantify magma flux as a function of time. 2. Petrogenesis of basement igneous rocks: The goal was to constrain the mineralogy and composition of the mantle sources that contributed to the magmatism, the melting processes that created the magmas, and the postmelting magmatic evolution; in particular, we sought to evaluate the role of continental lithosphere in the magmatism that formed the different domains of this LIP. 3. Environmental impact: The goal was to understand the postmagmatic processes that affected the igneous crust and evaluate the effects of LIP magmatism on the environment. 4. Tectonic history: The goal was to identify and interpret relationships between LIP development and tectonism. Chronology of Kerguelen Plateau_Broken Ridge Magmatism The most significant question to answer is "how much magma was erupted over what time interval?"_more specifically (1) "what is the age of the uppermost volcanic basement?" (2) "do eruption ages vary systematically with location on the plateau?" (3) "was the growth episodic or continuous?" and (4) "did the plateau grow by lateral accretion (i.e., similar to Iceland) or by vertical accretion and underplating?" Answers to these questions are provisionally provided by dating the oldest sediment above basaltic basement; more definitive results will come from postcruise 40Ar/39Ar dating of the lavas. Other important questions related to magma flux are (1) "did volcanism end abruptly or gradually?" (2) "did volcanism change from tholeiitic /transitional basalt to alkaline basalt as in the Kerguelen Archipelago, or did it remain exclusively tholeiitic like the Ninetyeast Ridge?" and (3) "were evolved (nonbasaltic) magmas erupted?" These questions were answered by drilling several holes with >100-m basement penetration. Answers to all of these questions related to magma flux are required to understand the physical and chemical processes that formed the Kerguelen Plateau_Broken Ridge LIP. Petrogenesis of Basement Igneous Rocks Several lines of evidence support the interpretation that the Kerguelen plume has been a long- term source of magma for major bathymetric features in the eastern Indian Ocean. For example, the systematic south to north age progression on Ninetyeast Ridge is consistent with a hot spot track formed as the Indian plate migrated northward over the Kerguelen plume (Mahoney et al., 1983; Duncan, 1991). Also, isotopic similarities among lavas from the Ninetyeast Ridge, the younger lavas forming the Kerguelen Archipelago and Heard Island, and the older lavas forming the Kerguelen Plateau and Broken Ridge (Figs. 9, 10) are consistent with the Kerguelen plume as an important source component (Weis et al., 1992; Frey and Weis, 1995, 1996). The preservation of a LIP resulting from partial melting of a decompressing plume head and its associated hot spot track derived from the plume stem presents an excellent opportunity to understand the evolution of a long-lived plume. Many studies of oceanic island volcanoes have demonstrated that geochemically distinct sources (e.g., plume, entrained mantle, and overlying lithosphere) contribute to plume-related volcanism. Because isotopic characteristics of plume, asthenosphere, and lithosphere sources are usually quite different, temporal geochemical variations in stratigraphic sequences of lavas can be used to determine the relative roles of different sources in plume-related volcanism. Establishing how the proportions of these sources change with time and location aids our understanding of how plumes "work" (Chen and Frey, 1985; Gautier et al., 1990; White et al., 1993; Peng and Mahoney, 1995). What was the role of depleted asthenosphere in creating the Kerguelen Plateau_Broken Ridge LIP? We use "depleted" to indicate relative depletion in abundances of highly incompatible elements (e.g., depleted asthenosphere is the source of most MORB [see Hofmann, 1997, for additional discussion]). Such asthenosphere can be entrained into an ascending plume head, or, when a plume is located at a spreading ridge axis, there can be mixing between plume-derived and MORB magmas. For example, much of the Ninetyeast Ridge formed when the Kerguelen plume was close to a ridge axis (Royer et al., 1991). Weis and Frey (1991) inferred that the relatively low 87Sr/86Sr and high 143Nd/144Nd ratios of lavas from Deep Sea Drilling Project (DSDP) Site 756 on the Ninetyeast Ridge are a consequence of the plume being close to a spreading ridge axis during formation of the Ninetyeast Ridge. There is also recognition that depleted material may be intrinsic to a plume (e.g., Saunders et al., 1998). Fitton et al. (1998) suggested that a plot of Zr/Y vs. Nb/Y is a geochemical discriminant for distinguishing depleted material intrinsic to the Icelandic plume from North Atlantic MORB. In the case of the Kerguelen plume, however, this discriminant is compromised by the presence of a continental lithosphere component. A continental lithospheric component has been recognized geochemically in lavas from the southern SKP (Site 738 at ~63¯S) and in dredge 8 lavas from eastern Broken Ridge (Mahoney et al., 1995). A less obvious but significant continental lithospheric component is also present in basalts from Site 747 on the CKP and in basalts dredged from the 77¯ graben in the SKP (Figs. 9, 10, 11, 12). In addition, wide-angle seismic data from the Raggatt Basin (58¯S) of the SKP show a reflective zone at the base of the crust that has been interpreted to be stretched continental lithosphere (Operto and Charvis, 1995, 1996). In contrast, there is no compelling geochemical evidence for a continental lithosphere component in lavas from the Ninetyeast Ridge and the Kerguelen Archipelago (Frey et al., 1991; Weis et al., 1993, 1998; Frey and Weis, 1995, 1996; Yang et al., 1998), but such a component is present in the Big Ben basaltic series on Heard Island (Barling et al., 1994) (Figs. 9, 11, 12A) and in mantle xenoliths found in Kerguelen Archipelago lavas (Hassler and Shimizu, 1998; Mattelli et al., 1999). Determination of the spatial and temporal role of continental lithosphere components in the Kerguelen Plateau_Broken Ridge LIP is required to evaluate whether these continental components are a piece of Gondwana lithosphere that was incorporated into the plume. Relatively shallow basement holes (>100 m) in the Kerguelen Plateau and Broken Ridge can be used to define spatial and short-term variability in the geochemical characteristics of the lavas erupted during the waning phase of plateau volcanism. A surprising result of previous drilling (Legs 119 and 120) on the Kerguelen Plateau is that sampling of the uppermost 35 to 50 m of igneous basement at several plateau sites shows that lavas at each site have distinctive geochemical characteristics (e.g., basalts from each site have a distinct combination of Sr and Nd isotopic ratios [Fig. 9] and incompatible element abundance ratios, such as Ti/Zr [Fig. 13]). The latter ratio is useful because it can be precisely determined by the shipboard X-ray fluorescence (XRF) spectrometer. In many continental flood basalts, relatively low Ti/Zr is diagnostic of significant contamination by continental crust (e.g., the Bunbury and Rajmahal Basalts; Frey et al., 1996; Kent et al., 1997). In summary, we sought to determine whether the geochemical heterogeneity of basalts from different domains of the Kerguelen Plateau and Broken Ridge reflect spatial and temporal heterogeneities in a plume or localized differences in mixing proportions of components derived from asthenosphere, plume, and slivers of continental lithosphere. Answering this question requires knowledge of temporal variations in geochemical characteristics at several locations within this LIP. Preliminary data relevant to these questions are provided by shipboard geochemical results. These topics are also the main focus of postcruise geochemical and isotopic studies. Environmental Impact A major goal of Leg 183 was to address the environmental impact of the formation of the Kerguelen Plateau and Broken Ridge. Important goals for this assessment are to (1) define postmagmatic compositional changes resulting from interaction of magmas with the surficial environment, (2) determine the relative roles of submarine and subaerial volcanism in constructing the upper part of the plateau, (3) estimate volatile contents of magmas from compositional studies of phenocrysts and their inclusions, and (4) evaluate the extent of degassing by determining the abundance and distribution of vesicles. The study of altered and metamorphosed basement rocks will be a major source for this information, but overlying sediments will also provide important data (e.g., the presence of terrestrial and terrigenous sedimentary components, as at Site 750, establishes an important role for subaerial volcanism). For subaerial eruptions, the input of volcanic gases into the atmosphere is controlled by the volatile content of the magmas and eruption rate. For submarine eruptions, it is essential to determine if hydrothermal systems developed that were significant in controlling local, regional, and global elemental and isotopic fluxes. Our overall goal was to assess the environmental impact of the Kerguelen/Broken Ridge LIP by estimating fluxes of elements, volatiles, particulates, and heat into the atmosphere- hydrosphere-biosphere system. Tectonic History To understand relationships between tectonism and LIP magmatism, we will study the seismic volcanostratigraphy of the Kerguelen Plateau and Broken Ridge by linking seismic facies analysis with petrophysics, borehole data, and synthetic seismic modeling. We sought to determine stratigraphic and structural relationships both within the various Kerguelen Plateau domains and Broken Ridge and between these features and adjacent oceanic crust. Seismic volcanostratigraphic studies can reveal temporal and spatial patterns of LIP extrusion in a regional tectonic framework, as well as test for synchronous or asynchronous postemplacement tectonism of the Kerguelen Plateau, Broken Ridge, and adjacent ocean basins. Knowledge of the uplift and subsidence histories of the Kerguelen Plateau and Broken Ridge will provide much- needed boundary conditions for models of mantle upwelling, crustal thinning, crustal growth, and postconstructional subsidence and faulting. Observations of physical volcanology, such as (1) flow thicknesses and directions, (2) flow morphologies, (3) relative thickness of marginal breccia zones and massive interiors of flows, (4) vesicle distribution within flows, (5) the presence and nature of interbeds, and (6) evidence for subaerial vs. submarine extrusion, will provide important information on eruption parameters and the distribution of melt conduits. In addition, physical volcanological observations coupled with shipboard measurements of physical properties and downhole logging data will provide ground truth for seismic volcanostratigraphy. We seek to determine the types of eruptive activity that formed the volcanic rocks and sediments of the Kerguelen Plateau and Broken Ridge. This includes distinguishing between subaerial and subaqueous eruptions, for example, by recognizing oxidized crusts on lava flow surfaces and analysis of intercalated soils developed on flow tops. In conjunction with seismic volcanostratigraphic studies, we will attempt to identify surficial and shallow subsurficial sources for the basalts (discrete volcanoes or feeder dikes) and to assess the effects of preexisting bathymetry and topography on flow distribution. DRILLING STRATEGY LIPs are enormous constructions that present considerable challenges for adequate sampling to address the major questions outlined in the preceding sections. Our knowledge of oceanic Cretaceous LIPs is rudimentary, similar, perhaps, to that of mid-ocean ridges before the general acceptance of the plate tectonics paradigm in the late 1960s. Geophysical surveys and a grid of shallow (100 to 200 m) basement drill holes are necessary to address Cretaceous mantle dynamics, the physical and chemical processes involved in construction of these LIPs, and their environmental consequences. Understanding the temporal and compositional history of the Kerguelen Plateau and Broken Ridge requires a multifaceted drilling strategy (Fig. 14), including (1) transects of shallow basement holes across the surface of the LIP; (2) offset drilling in tectonic windows that expose deeper levels of the LIP that are otherwise inaccessible; (3) intermediate (1000_2000 m) and deep (>2000 m) basement holes at carefully chosen locations; and (4) reference holes on older adjacent oceanic crust. Leg 183, complemented by Legs 119 and 120, is part of the fundamental and necessary reconnaissance phase of sampling. To obtain a comprehensive database of eruption ages and lava compositions for the entire LIP, we sampled igneous basement to depths of ~100 to 150 m at as many morphologically and tectonically distinct regions of the Kerguelen Plateau_Broken Ridge LIP as possible during one drilling leg (Table 1; Figs. 3, 4, 5, 6). In addition, the sedimentary section immediately overlying the basement provides estimates of minimum ages for extrusive basement, important information regarding eruption and weathering in a subaerial vs. submarine environment, and evidence for tectonic events in the plateau's history. Neogene to Cretaceous sediments overlying basement also provided significant paleoceanographic information for high southern latitudes. At some sites, tephra horizons of various ages provide information on explosive eruptions at nearby islands (McDonald, Heard, and Kerguelen Archipelago) and perhaps more distant volcanoes. PRINCIPAL RESULTS Sites 1135 and 1136 Sites 1135 and 1136 (water depths of 1567 and 1931 m, respectively) are on the southern Kerguelen Plateau, approximately midway between two ODP Sites (738 and 750) where basaltic basement has previously been recovered. Sites 1135 and 1136 are ~350 km north of Leg 119 Site 738 and 300 km south of Leg 120 Site 750 (Figs. 3, 4). Major objectives of drilling on the southern Kerguelen Plateau were to obtain 150 m of igneous basement to characterize the age, petrography, and compositions of the lavas, the physical characteristics of the lava flows, and the environment of the eruption (subaerial or submarine). A specific goal was to evaluate the areal extent of the continental lithosphere component that has been recognized in the Site 738 lavas using trace element and isotope geochemistry; such a component is not present in the more northerly Site 750 lavas (Figs. 9, 10, 11, 12, 13). Sedimentary objectives at Sites 1135 and 1136 were to determine sequence facies, to define the ages of seismic sequence boundaries, to estimate the duration of possible subaerial and shallow marine environments, and to obtain minimum estimates for basement age. Hole instability forced us to abandon Site 1135 after drilling to ~70 m above acoustic basement (Fig. 15), but we were able to accomplish some of our basement- oriented objectives at Site 1136 (Fig. 16), located ~30 km east of Site 1135. In particular, we penetrated 33.3 m of basaltic basement that included three inflated pahoehoe flows, two of which are characterized by massive, relatively unaltered interiors. These rocks provide excellent samples for radiometric dating and geochemical analyses. The 526-m-thick late Pliocene to Late Cretaceous sedimentary sequence recovered at Site 1135 is almost entirely pelagic calcareous ooze and chalk (Fig. 17). Chert nodules are common from ~140 m below seafloor (mbsf) to the bottom of the hole. We recovered an expanded (238 m thick) middle Eocene to latest Paleocene nannofossil ooze section, an interval not well represented during previous coring on the Kerguelen Plateau or elsewhere in the Southern Ocean at these high latitudes (~60¯S). The study of this section will improve high-latitude biostratigraphic correlations. Furthermore, a Cretaceous/Tertiary boundary section at ~260 mbsf is possibly marked by a bed of light greenish gray calcareous clay with an irregular upper contact and scattered well-rounded clasts of white nannofossil ooze. These features suggest an erosional or mass-wasting event. We have tentatively identified Chron C29n above, and Chron C29r below the boundary, respectively. Near the boundary, velocity, magnetic susceptibility, and natural gamma ray intensity change significantly; in addition, water content, porosity, and carbonate content decrease below the boundary. Sedimentation rates were high in the Paleogene ooze (up to 15 m/m.y.) and Cretaceous chalks (8_10 m/m.y.); the Paleocene section, however, is abbreviated by hiatuses. The 128-m-thick sedimentary sequence recovered at Site 1136 (Fig. 18) includes an expanded upper lower Eocene to lower middle Eocene section of pelagic calcareous ooze and chalk (Unit II) that is not well represented in other drill holes on the Kerguelen Plateau or in any other southern high-latitude sites. Study of these sediments will refine high-latitude middle Eocene biostratigraphic zonations. Underlying these pelagic sediments is calcareous zeolitic volcanic clayey sand (Unit IV), probably deposited in a high-energy neritic (shelf) environment, and a carbonate-bearing zeolitic silty clay (Unit V). Fossil debris in Unit V is common and suggests deposition at shallow paleodepths (upper bathyal to outer neritic) in more tranquil conditions than prevailed during deposition of the overlying clayey sand. The sands and clays overlying basement basalt contain diluted but relatively well-preserved micro- and nannofossil faunas of middle Albian age, thereby providing a minimum age for the section and underlying basalts of ~104.5_106.5 Ma. Lower Albian marine sediments have not been recovered during previous drilling on the Kerguelen Plateau, but sands and clays recovered in Hole 1136A resemble lower Albian sediment drilled on the Falkland Plateau. The sands and clays may correspond to nonmarine, palynomorph- bearing Albian sediment found in silt and claystone cored at Site 750 on the SKP. Albian and Late Cretaceous nannoplankton, foraminifers, and pollen assemblages will provide information on regional paleoceanographic conditions during those times. The epiclastic succession (Units IV and V) and overlying calcareous sediments reflect increasing water depths with time concomitant with a decreasing volcaniclastic component in the sediments. Basaltic volcanic components in the epiclastic sediments at this site are probably derived from erosion of the basaltic plateau. At Site 1136, from 128.1 to 161.4 mbsf, we cored three normally magnetized tholeiitic basalt flows (55% recovery; Fig. 18). We infer that the vesicular tops of the two upper flows were not recovered. Basalt from the uppermost flow (6.2 m recovered from an ~10- m-thick flow) varies downward from moderately altered to a massive interior to a fine- grained, vesicle-rich (~10% clay-filled vesicles) and oxidized base. Horizontal vesicle sheets and the general vertical distribution of vesicles within the massive interior and lower crust of the upper flow imply that it formed as an inflated pahoehoe flow. Basalt from the middle flow (13.3 m recovered from an ~20-m-thick flow) varies downward from a massive interior to a fine- grained, vesicle-rich (10%_15%) base. This flow is also probably an inflated, large-volume pahoehoe flow. We only recovered 53 cm of a vesicular basalt breccia that forms the rubbly flow top of the lower flow. Although we cannot unambiguously determine the eruption environment of these flows, the inference that they are inflated pahoehoe flows and the absence of features indicating submarine volcanism (e.g., pillows and quenched glassy margins) suggest subaerial eruption. All lavas are sparsely to moderately phyric basalts containing phenocrysts of plagioclase with lesser amounts of clinopyroxene olivine. Phenocrysts are found as either isolated grains or as two texturally distinct types of glomerocrysts. Corroded plagioclase cores in one glomerocryst type resemble those in small (~1 cm) microgabbro xenoliths. Vesicle-rich segregations (1_2 cm wide) contain 10%_30% vesicles in a nonporphyritic, fine- to medium-grained matrix rich in glass and titanomagnetite. The basaltic rocks are slightly to completely altered to low temperature secondary phases that partially replace primary minerals, completely replace mesostasis, fill veins, and partially to completely fill vesicles. The most common secondary minerals are clays (Mg-saponite and celadonite), calcite, and zeolites. In general, clay minerals abound at all depths, whereas the abundances of calcite and zeolites exhibit more pronounced downhole variations. The wide variation of K and Rb contents in the lavas analyzed by XRF reflects formation of these secondary phases. The four least altered samples (loss on ignition [LOI] = 0.9% to 2.1%) from the upper and middle flows have 50.0_51.0 wt% SiO2, 6.4_6.7 wt% MgO, and 1.60_1.76 wt% TiO2. Both flows are quartz-normative tholeiitic basalts (Fig. 19) with relatively low MgO and Ni contents and low Mg numbers; they are similar to basement rocks from other parts of the Kerguelen Plateau. In detail, the upper flow has marginally lower Ti, Nb, Zr, Y, and Ce, distinctly lower V, and higher Cr abundances than the middle flow. Primitive-mantle- normalized abundances of highly incompatible trace elements (Ba, Nb, and Ce) are only slightly greater than those of less incompatible elements (Ti and Y). Site 1136 lavas do not have the anomalously low Nb/Ce and Ti/Zr ratios that have been used in conjunction with isotope data to infer a continental lithospheric component in basalt from Site 738 on the southern plateau (Fig. 20). In many geochemical characteristics, Site 1136 lavas are similar to the low Al2O3 group at Site 749. No evidence indicates that these lavas contain a component derived from continental lithosphere. Major results of drilling at Sites 1135 and 1136 on the SKP include the following: 1. Expanded middle Eocene to uppermost Paleocene (Site 1135) and upper lower Eocene to lower middle Eocene (Site 1136) pelagic calcareous sediment sections not previously recovered at any Southern Ocean sites will improve our understanding of high-latitude paleoceanography at critical times of Paleogene cooling as well as aid high-latitude biostratigraphic correlations. 2. Albian and Late Cretaceous nannoplankton, foraminifers, and pollen assemblages will provide information on regional paleoceanographic conditions during those times of high relative sea level and high global temperatures. 3. Paleoenvironments of volcanic rock and overlying sediment range from subaerial (basalt) to neritic (clay and sand) to pelagic (chalk and ooze), documenting subsidence of the Kerguelen Plateau since Early Cretaceous time. 4. The >105-Ma basement basalts at Site 1136 are 10- to 20-m-thick inflated pahoehoe flows similar to continental flood basalts, such as the Columbia River Basalt; there is no geochemical evidence for the continental lithosphere component present in Site 738 basalts. Site 1137 Site 1137 lies on Elan Bank, a large western salient of the main Kerguelen Plateau, at a water depth of 1016 m (Figs. 3, 4). Elan Bank, flanked on three sides by oceanic crust of the Enderby Basin, had not been sampled before our drilling at Site 1137; therefore, the age and geochemistry of its igneous crust, as well as the feature's relationship to the contiguous central and southern Kerguelen Plateau, were completely unknown. Site 1137 lies on the eastern portion of the crest of Elan Bank; we chose the location as representative of the entire Elan Bank on the basis of its relatively simple structural setting, thin sedimentary section, and the presence of intrabasement seismic reflections (Fig. 21). The major objective at Site 1137 was to obtain igneous basement to characterize the age, petrography, and compositions of the lavas, the physical characteristics of the lava flows, and the environment of eruption (subaerial or submarine). We were especially interested in constraining the age of the uppermost igneous basement at Elan Bank for comparison with the proposed ~110 and ~85 Ma volcanic pulses on the southern and central Kerguelen Plateau, respectively. Sedimentary objectives at Site 1137 were to determine sequence facies, to define the ages of seismic sequence boundaries, to estimate the duration of possible subaerial and shallow marine environments, to obtain minimum estimates for basement age, and to determine the paleoceanographic history of this high-latitude site. As discussed below, we largely achieved our goals at Site 1137. We cored basaltic basement and interbedded sediment from 219.5 to 371.2 mbsf. One of the most significant and unexpected results of the leg was the discovery of garnet gneiss clasts in a fluvial conglomerate interbedded with basaltic basement at this site. This provides unequivocal evidence of continental crust in Elan Bank. In addition, the geochemical characteristics of these basalts clearly indicate a continental crustal component. We recognize three sedimentary lithologic units (I_III) in the upper 219.5 mbsf (Fig. 22). They rest unconformably on basaltic basement (Unit IV). Unit I (0_9.5 mbsf) consists of Pleistocene foraminifer-bearing diatom ooze with apparent ice-rafted sand and pebbles. Unit II (9.5_199.5 mbsf) is Miocene to uppermost Eocene white nannofossil ooze with rare chert. Some intervals contain diatoms or foraminifers. Units I and II represent marine pelagic deposition and are characterized by compressional (P-) wave velocities of 1564 to 1785 m/s that show little scatter. Porosity in the upper two units clusters between 50% and 63%. Unit III (199.5_219.5 mbsf) is a 20-m-thick sequence of glauconite-bearing sandy packstone with abundant shell fragments that was probably deposited in a neritic environment. In the core overlying basaltic basement, the packstone contains well-preserved late Campanian (72_76 Ma) foraminifers, calcareous nannofossils, and dinoflagellates. P-wave velocities in Unit III vary considerably from 3120 and 4340 m/s, and porosity, which ranges between 4.7% and 24.9%, averages 11.7%. Natural gamma-ray intensities are relatively high in the glauconitic sand. Basalt and interbedded volcaniclastic sediment comprise the 151.7- m basement sequence in Hole 1137A (Unit IV; 219.5_371.2 mbsf), which we subdivide into basement Units 1_10 (Fig. 22). The 10 units include seven basaltic lava flows, totaling ~90 m in thickness, and three sedimentary units. All basement units are clearly distinguishable in downhole logging data. Relatively good core recovery and high quality logs enable us to constrain true thicknesses of the basement units through core-log integration. All basalts are normally magnetized; in light of biostratigraphic ages in Unit III and thickening of this unit to the east, the basalts likely acquired their magnetization during the long Cretaceous Normal Superchron prior to 83 Ma. Inclinations after thermal demagnetization of basalts from the seven flows range from _55¯ to _72¯, with a mean of _66¯. We calculate a paleolatitude of ~48¯S, which is 8.5¯N of Site 1137. P-wave velocities in basement rocks range from 2648 to 6565 m/s, averaging ~4650 m/s. Velocities within lava flows correlate inversely with the degree of alteration estimated from visual inspection. Six of the seven basaltic flow units erupted subaerially; one may have erupted into shallow, wet sediment. We interpret each basaltic unit as a single lava flow. The flows are 7 to 27 m thick, and we recovered three flow-top breccias, a pahoehoe surface, and two basal contacts where lava apparently baked the underlying units. Flow-top breccias appear to have formed by breaking up of small pahoehoelike fingers that are repeatedly intruded into older breccia. This style of autobrecciation is atypical of pahoehoe, aa, or Hawaiian transitional lava flows, but is common in western United States flood basalts. The three or possibly four pahoehoe flows in Hole 1137A are inflated; in morphology, they resemble large-volume lava flows forming continental flood basalts. Multiple horizontal vesicular zones within the 27-m-thick flow suggest a complicated inflation history of perhaps four separate pulses of lava. Sediments intercalated with the lowermost flow top breccia suggest that the lava intruded wet sediment. With the exception of this lowermost unit, the flows show oxidation zones and morphologies consistent with subaerial emplacement. Basement Units 1_4 are aphyric to moderately plagioclase Ò clinopyroxene Ò olivine-phyric basalt, whereas Units 7, 8, and 10 are moderately to highly plagioclase Ò clinopyroxene-phyric basalt. Like all other Cretaceous basement basalt recovered from the Kerguelen Plateau, Site 1137 basalts are tholeiitic to transititional in composition (Fig. 19); they have 50.4 to 52.7 wt% SiO2 and 4.4 to 7.3 wt% MgO. However, abundances of incompatible minor and trace elements (Fig. 23), as well as the ratio of more incompatible to less incompatible elements (e.g., Zr/Y; see Fig. 24), are higher than in other Cretaceous basement tholeiites recovered from the plateau. Consequently, Site 1137 basalts form a distinct geochemical group that may reflect a relatively lower extent of melting or a source more enriched in incompatible elements. A continental lithospheric component, probably crust, in Site 1137 basalts is indicated by their relatively low Nb/Ce and high Zr/Ti ratios, as well as their trend of Nb/Y vs. Zr/Y (Fig. 25). Such a component is also present in Kerguelen Plateau basalts from Sites 738 and 747, but not in basalts from Sites 749, 750, and 1136 (Figs. 24, 25). Lava flow interiors at Site 1137, as at Site 1136, are relatively unaltered compared to Cretaceous basement lavas previously recovered from the Kerguelen Plateau. For example, LOI for least altered samples is <2.2% and averages 1.2%. The mobility of Rb and K during postmagmatic alteration shows in the poor correlation of Rb and Nb abundances (Fig. 23) and the wide abundance range of these elements compared to other incompatible elements (Fig. 26). Basement rocks vary from slightly to completely altered with low- temperature secondary phases replacing primary minerals and mesostasis and filling veins, fractures, and vesicles. Clay minerals are the dominant secondary minerals in all basement units. More permeable horizons (such as brecciated and/or vesicular flow tops and bases, and zones with high vein and fracture densities) exhibit higher degrees of alteration and more diverse secondary phases that include calcite, zeolite, quartz, and amorphous silica. In downhole logs, highly altered flow tops are characterized by higher potassium content caused by the increased abundance of clay minerals. Alteration of Hole 1137A lavas likely results from both weathering and low-temperature alteration. Unlike postmagmatic submarine alteration of typical oceanic crust, subaerial weathering and low- temperature interaction of basalts and interbedded sediments with groundwater at Site 1137 preceded submarine alteration. Basement Units 5, 6, and 9 consist of volcaniclastic sedimentary rocks. Basement Unit 5 (286.7_291.0 mbsf) is a succession of interbedded volcaniclastic siltstones and sandstones. Many beds are normally graded, and others show parallel laminations. These sediments overlie Basement Unit 6 (291.0_317.2 mbsf), which consists of volcaniclastic conglomerate. Clasts range from well-rounded granules to small boulders. Most intervals are clast supported, but matrix- supported intervals also are present. The depositional environment of Basement Units 5 and 6 appears fluvial, perhaps associated with a braided river. These units represent a significant hiatus of unknown duration between eruptions of basaltic flow Units 1_4 and 7 and 8. Furthermore, fluvial facies of Units 5 and 6 corroborate our interpretation of subaerial lava flow effusion. Diverse clasts within the conglomerate (Unit 6), and to a lesser extent similar lithic clasts within the underlying crystal-vitric tuff (Unit 9), constitute a greater variety of rock types than usually recovered from a single drill hole into igneous basement. In particular, the predominant clast lithologies in Units 6 and 9 include porphyritic trachyte, flow- banded rhyolite, plagioclase phyric basalt, and a variety of small, highly altered, sparsely phyric and aphyric basalts. The most unexpected clasts, however, are rounded cobbles of garnet-biotite-gneiss and granitoid. We also find single grains of garnet and perthitic alkali feldspar, presumably weathered from sources similar to those of the gneiss and granitoid cobbles, in the sand fraction of Unit 5 and as xenocrysts in parts of the Unit 9 crystal-vitric tuff. It is difficult to imagine anything but an originally continental source for such material. Basement Unit 9 (344.0_360.7 mbsf) consists of altered crystal- vitric tuff composed of ~40% coarse (1_2 mm) angular crystals of sanidine and <5% lithic clasts enclosed within a light- to dark-green dense matrix. Cuspate and tricuspate glass shards, now partially to completely altered to clay minerals, and sanidine phenocrysts (<5 mm) are the principal components of the tuff; minor components include amphibole, plagioclase, quartz, and opaques. Broken bubble-wall shards and abundant embayed and broken crystals indicate that the tuff formed in an explosive volcanic eruption. Suspended within the tuff are 1%_2% subangular to rounded, granule- to pebble-sized lithic clasts of varying lithologies, mostly basalt. The coarse grain size of both the enclosed pebbles and the primary sanidine crystals precludes deposition of this material by settling from an ash cloud; transport in a pyroclastic flow is more likely. However, we infer that the material was reworked because of the absence of (1) internal stratification, (2) a basal breccia or fine flow top, or (3) normal grading of lithics and crystals. The even distribution of crystals and pebbles throughout the tuff and the massive internal texture of the deposit provide evidence for mass flow redeposition of these sediments to their present locations. Small fault zones with offsets of Û5 cm are highly altered and contain locally abundant (Û5%) pyrite and possibly chlorite. Native copper surrounds lithic fragments in more intense alteration zones (<1 cm wide). In summary, significant results bearing on the origin and evolution of Elan Bank include the following: 1. Subsidence of the Kerguelen Plateau is recorded by the paleoenvironments of volcanic rocks and sediment that range from subaerial or fluvial (basalt and interbedded sediment) to neritic (packstone) to pelagic (ooze). 2. The volcaniclastic conglomerate contains clasts of trachyte, rhyolite, granitoid, and garnet- biotite gneiss; the garnet-biotite gneiss, in particular, indicates continental crust at this south Indian Ocean location. 3. The sanidine-bearing crystal-vitric tuff, as well as the trachyte and rhyolite clasts in the conglomerate, indicates that highly evolved magmas erupted, in some cases explosively, during the final stages of the volcanism that formed Elan Bank. 4. Most of the seven basement flows erupted in a subaerial environment. These inflated pahoehoe and transitional rubbly flows are typical of continental flood basalts, such as the Columbia River Basalt. 5. Like other Cretaceous igneous basement rocks of the Kerguelen Plateau, the seven basement lava flows are tholeiitic to transitional basalts; however, Site 1137 basalts are more enriched in incompatible elements, perhaps a result of lower extents of partial melting or derivation from a source more enriched in incompatible elements. Also, we infer a continental crust component in Site 1137 basalts from their less- than-primitive mantle ratios of Nb/Ce and Zr/Ti and their Nb/Y vs. Zr/Y trends. Site 1138 Site 1138 lies on the CKP ~150 km north-northwest of Site 747 (Leg 120) and 180 km east- southeast of Heard Island (Figs. 3, 4). In the vicinity of Site 1138, geological structure and seismic stratigraphy are relatively simple, and interpreted igneous basement contains some internal reflections (Fig. 27). Basalts at Site 747 erupted at ~85_88 Ma, as determined from 40Ar/39Ar data and from the biostratigraphy of the overlying sediments. In contrast, Heard Island is dominated by Quaternary volcanism. A major objective at Site 1138 was to determine if the uppermost basaltic crust of the CKP is ~85 Ma at more than one location. Also, geochemical characteristics of Site 747 basalts indicate a continental crust component, possibly Archean granulite, which differs from the continental component in basalt from the SKP at Site 738. During continental breakup, continental lithosphere along the conjugate Antarctic and Indian margins may have been fragmented and incorporated into embryonic Indian Ocean mantle. Subsequently, in localized areas this continental material may have interacted with basaltic magmas forming the Kerguelen Plateau. Therefore, we were especially interested in comparing the petrology and geochemistry of basaltic basement from this second CKP drill site with basalt from the southern, northern, and Elan Bank domains, as well as Heard Island and the Kerguelen Archipelago. Additional basement objectives were to determine the physical characteristics of the lava flows and the environment of the eruption (subaerial or submarine). The sedimentary objectives at Site 1138 were to determine sequence facies, to define the ages of seismic sequence boundaries, to estimate the duration of possible subaerial and shallow marine environments, to obtain minimum estimates for basement age, and to determine the paleoceanographic history of the CKP. At Site 1138 our objectives were achieved by coring ~144 m of volcanic basement and ~698 m of overlying sediment. We recovered Upper Cretaceous through Pleistocene sediment from the upper 698 mbsf of Hole 1138A, whereas the lower 144 m of the hole yielded multiple, ~5-m- thick basalt flows overlain by volcaniclastic and minor sedimentary rocks (Fig. 28). We recognized seven lithologic units in Hole 1138A; Units I_VI are sedimentary rocks resting unconformably on the volcanic basement (Unit VII). The upper 650 m of sediment is biosiliceous and carbonate pelagic ooze, of which the top 110-m section comprises a relatively complete and expanded sequence of Quaternary and Pliocene biosiliceous sediments. The lower ~50 m of the sedimentary section consists of Upper Cretaceous shallow marine and terrestrial sediments. Unit I (0_112.0 mbsf) consists of foraminifer-bearing diatom clay with interbedded foraminifer-bearing diatom ooze in the upper portion. We found a few thin volcanic ash layers in this late Pleistocene to late Miocene unit. Grain density averages 2.38 g/cm3; porosity, 77%; and P-wave velocity, 1568 m/s in Unit I. Unit II (112.0_265.9 mbsf) is composed of foraminifer-bearing nannofossil clay (Subunit IIA) that overlies foraminifer-bearing nannofossil ooze (Subunit IIB). The carbonate/silica ratio of the 153.9-m-thick Miocene Unit II is much higher than that of Unit I. Volcanic material is disseminated in the sediment as well as in rare distinct tephra layers. In Unit II, grain density averages 2.61 g/cm3, porosity 60%, and P-wave velocity 1672 m/s. Unit III (265.9_601.8 mbsf) is late Oligocene to middle Campanian in age. It consists of foraminifer- bearing chalk and contains scattered chert nodules in its lower part. Cyclic color variations (white to greenish gray) are common. The Cretaceous/Tertiary boundary near the base of Subunit IIIA (Core 183-1138A-52R) is possibly complete, but lithologies do not change across it. In Unit III, grain density averages 2.70 g/cm3, porosity averages 48%, and P-wave velocity averages 2310 m/s. Unit IV (601.8_655.6 mbsf), of middle Campanian to Cenomanian(?) age, consists of cyclic alternations of light gray foraminifer-bearing chalk with gray through greenish gray to black intervals of nannofossil claystone. The dark gray to black beds become prominent and increase in clay content in the lower portion. Chert nodules are present in the upper part of the unit. grain density averages 2.67 g/cm3, porosity 35%, and P-wave velocity 2665 m/s. A bed of black, faintly laminated (unburrowed) claystone with high organic carbon content (2.22%) is at the base of Unit IV. Units I through IV represent deep-marine pelagic sedimentation; however, the relatively high clay content of sediments in Unit I and Subunit IIA suggests terrigenous input from overbank flow of turbidity currents moving down a submarine canyon ~45 km west- northwest of Site 1138. The black claystone at the base of Unit IV reflects an oxygen- starved environment that may be the oceanic anoxic event marking the Cenomanian/Turonian boundary. Unit V (655.57_671.88 mbsf) consists predominantly of glauconitic calcareous sandstone of Turonian_Cenomanian age deposited in a neritic environment. Serpulid worm tubes and large bivalve fragments are common. Grain density averages 2.71 g/cm3, porosity averages 42%, and P-wave velocity averages 2719 m/s in Unit V. The gradual transition from neritic oxidized sediment (Unit V) to interbedded black claystone and chalk (Unit IV) to pelagic sediments (Unit III) supports the postulated major transgression causing the Cenomanian_Turonian oceanic anoxic event. This hypothesis will be tested by shore-based studies. Unit VI (671.88_698.23 mbsf) consists of Upper Cretaceous fossil- rich, dark brown silty claystone with interbedded sandstone of fluvial or shallow marine origin. The silty claystone contains many wood fragments, possible sporangias, a seed, and fossil spores and pollen. The sandstone beds contain well-rounded pebbles and sand grains of volcanic material. At the bottom of Unit VI, silty claystone rests upon volcanic basement rocks (Unit VII). Large rounded pebbles of weathered basalt close to the base of Unit VI suggest a regolith formed by weathering of volcanic basement. In Unit VI, grain density averages 2.72 g/cm3, porosity averages 37%, and P-wave velocity averages 2328 m/s, the latter defining a pronounced velocity inversion from overlying Unit V. The seismic sequence containing the deepest marine sediments cored at Site 1138 thickens to the northeast, suggesting that basaltic basement rocks could be significantly older than the minimum age indicated by biostratigraphy. We recognize 22 units within the 144 m of igneous basement (Unit VII) drilled at Site 1138 (Fig. 28). Basement Unit 1 includes rounded cobbles of flow-banded, aphyric to sparsely sanidine-phyric dacite. Unit 2 is a complex succession of volcaniclastic rocks overlying basalt lava flows_Units 3 through 22. The 20-m-thick volcaniclastic succession comprising Unit 2 contains six variably oxidized and altered pumice lithic breccias. We interpret these as unwelded, subaerial pyroclastic flow deposits. The pumice clasts are typically aphyric, and the bulk composition of a pumice-rich sample is trachytic. The volcaniclastic sequence also includes pumice beds, reworked volcaniclastic sediments, and highly altered ash deposits that contain accretionary lapilli. Basement Units 3_22 are ~5-m-thick subaerial basaltic lava flows that range from inflated pahoehoe to classic aa. Several boundaries are oxidized, suggesting subaerial weathering between eruptions. The relatively thin flows at Site 1138 resemble Hawaiian lavas and contrast with the generally thicker flows drilled at Sites 1136 and 1137. Most flows have unique flow top breccias, which are not easily classified. Some breccias contain slabs of pahoehoe mixed with aa clinker; others are a jumble of pahoehoe lobes. The breccias contain varying amounts of sediment; some may be reworked, perhaps in a fluvial environment. Most flows probably erupted on a moderate slope (1¯ to 4¯) under conditions of high shear resulting from a high eruption rate or topographic confinement. Several observations indicate that these are near vent flows; specifically, aa and slab pahoehoe flows rarely travel more than a few tens of kilometers from vents; abundant small vesicles indicate that the lavas did not flow far enough for vesicles, which formed at vents, to coalesce; and clasts in some of the welded basal breccias appear to be spatter, which only forms close to vents. All basalts show normal magnetic inclinations. We calculated a mean inclination of _60.8¯, which corresponds to a paleolatitude of 46.4¯S, assuming a geocentric dipole field. The paleolatitude is thus 7¯N of Site 1138. This southward shift in latitude since Late Cretaceous time is consistent with the 8.5¯ difference we found at Site 1137 on Elan Bank. The basalts have average (range) grain densities of 2.90 g/cm3 (2.44_3.13), porosities of 25% (9_55), and P-wave velocities of 4014 m/s (1884_7491). The massive parts of flow Units 3_22 are slightly to locally highly altered, whereas alteration ranges from high to complete in the brecciated zones. Rubbly flow tops are partly to completely altered to clay minerals, and abundant euhedral zeolites form the matrix, fill veins, and partially fill vesicles and large voids. Lava clasts are commonly completely altered to brown clay minerals. Multiple generations of zeolite (clinoptilolite) exhibit many crystal shapes, predominantly equant and prismatic, but fluffy forms frequently fill fissures. Sediment filling breccia void space is variably indurated, perhaps caused by silicification. Calcium carbonate is absent except from the uppermost basalts directly underlying the volcaniclastic sequence. Most of the basalts are moderately to highly vesicular and aphyric to sparsely plagioclase- phyric tholeiites (Fig. 19, 29). Units 9 and 19 contain clinopyroxene phenocrysts, and Units 5_16 and 19 contain 1%_5% olivine microphenocrysts, now completely replaced by secondary clays. The relatively unaltered (LOI of only 0.5 to 2 wt%) massive parts of these basaltic flows have similar major element compositions (e.g., MgO contents vary only from 4.5 to 7 wt%). However, with increasing depth, Mg/Fe, Ni, and Cr contents decrease, and abundances of most incompatible elements (Sr is an exception) increase by nearly a factor of two (Fig. 30), thereby defining a trend to Fe- and Ti-rich basalt. This systematic downhole trend is consistent with extensive fractionation of the phenocryst phases, plagioclase, olivine, and clinopyroxene. Basalts from the two drill sites on the CKP (Sites 747 and 1138) overlap in a Nb/Y vs. Zr/Y plot (Fig. 31). The major results of drilling Site 1138 on the CKP include 1. Paleoenvironments of the Cenomanian/Turonian boundary event appear to be preserved in the transition from oxidized neritic sediment to black claystone (shale) to pelagic sediment, which may enable testing of the hypothesis that a major transgression caused this oceanic anoxic event. 2. The sedimentary sequence overlying basement contains Upper Cretaceous shallow marine and terrestrial sediments. Turonian (and older?) silty claystone contains well-preserved wood fragments, a seed, spores, and pollen, documenting for the first time that the CKP was subaerial after volcanism ceased. 3. The inferred minimum basement age, Turonian (89_93 Ma), is older than the 85_88 Ma proposed for Site 747, only 150 km to the southeast. 4. Volcanic growth of the Kerguelen Plateau at this site on the CKP terminated with eruptions of highly evolved magma that include dacitic lavas and pyroclastic flow deposits of trachyte. 5. Basaltic basement underlying evolved rocks is represented by 20 thin, Hawaii-like subaerial flows that erupted onto moderate slopes of 1¯ to 4¯. These are tholeiitic basalt flows whose compositions define a systematic downhole trend to FeTi- rich basalt. Such highly evolved basalts have not been previously recovered from the Kerguelen Plateau; incompatible element abundance ratios, such as Nb/Zr and Nb/Y, of Site 1138 basalts, however, overlap with the field defined by basalts from Site 747, the other drill site on the CKP. Site 1139 Site 1139 lies on Skiff Bank (Leclaire Rise), a bathymetric and gravimetric high ~350 km west-southwest of the Kerguelen Archipelago (Figs. 3, 4). Flanked to the south and west by Cretaceous oceanic crust of the Enderby Basin, Skiff Bank appears to be structurally related to, and bathymetrically continuous with, the NKP. At least two major faults, however, offset interpreted igneous basement between Skiff Bank and the large massif containing the Kerguelen Archipelago. Skiff Bank has been proposed to be the current site of the Kerguelen hot spot (Fig. 2), but hundreds of meters of sediment on parts of the elevated feature argue against Skiff Bank originating entirely by recent volcanism. Both Skiff Bank and Elan Bank trend east-west, approximately perpendicular to the trends of fracture zones in the Enderby Basin and thus parallel to the axis of breakup between Antarctica and India. The free-air gravity signatures of the two features are also similar; pronounced negative anomalies flank their southern margins, but not their northern margins (Fig. 4). Many rock types, including both aphyric basalt and plutonic rocks such as alkali granite, were recovered in a single dredge haul from Skiff Bank, quite close to Site 1139. The plutonic rocks were interpreted as ice-rafted debris. Hence, the age and composition of Skiff Bank's igneous crust and its relationship to the contiguous northern Kerguelen Plateau are not established. The NKP is commonly believed to have formed since ~40 Ma, when the SEIR separated Broken Ridge and the CKP (Fig. 2), but submarine igneous basement of the NKP has never been drilled. Site 1139 lies at a depth of 1427 m on Skiff Bank's southwestern terrace, which is >1000 m lower than the crest, located <50 km to the northeast (Fig. 3). We chose this location as representative of the entire Skiff Bank on the basis of its relatively simple structural setting and thin sedimentary section. The top of acoustic basement is flat lying, and the overlying basement is a sediment sequence ~500 m thick (Fig. 32). The fault scarp marking the boundary between Skiff Bank and the Enderby Basin lies ~20 km southwest of Site 1139 and offsets the basement by more than 2700 m. The major objectives at Site 1139 were to obtain igneous basement to characterize the ages, petrography, and compositions of the lavas, the physical characteristics of the lava flows, and the environments of eruption (subaerial or submarine). We were especially interested in testing the hypothesis that the age of the uppermost igneous basement at Skiff Bank is <40 Ma. The sedimentary objectives at Site 1139 were to determine sequence facies, to define the ages of seismic sequence boundaries, to estimate the duration of possible subaerial and shallow marine environments, to obtain minimum estimates for basement age, and to determine the paleoceanographic history of this high latitude site. As discussed below, we largely achieved our goals at Site 1139. We drilled 233 m into igneous basement that is overlain by early Oligocene shallow marine sediments. Sediments were recovered from the upper 461 mbsf of Hole 1139A, whereas extensively altered felsic volcaniclastic rocks and mafic to intermediate composition lava flows were recovered from the lower 233 m of the hole (Fig. 33). We recognize six lithologic units. Units I_V are sediment and sedimentary rock resting on volcanic basement (Unit VI). Unit I (0 to 47.5 mbsf) consists of foraminifer-bearing diatom-bearing nannofossil ooze (Subunit IA) of Quaternary age and foraminifer-bearing nannofossil ooze (Subunit IB). Scattered basaltic sand grains and rare pebbles as well as traces of pumice are present in Subunit IA. Unit II (47.5 to 380.7 mbsf) consists of nannofossil-bearing clay and claystone with interbedded nannofossil-bearing ooze and chalk of early late Miocene to mid-Oligocene age. Trace fossils are very common. Unit II records a substantial influx of terrigenous clay from an adjacent volcanic landmass. In Subunit IB and the upper portion of Unit II (to ~107 mbsf) P-wave velocity averages 1822 m/s, bulk density ranges from 1.5 g/cm3 to 1.7 g/ cm3, grain density ranges between 2.6 and 2.8 g/cm3, and porosity changes from 60% to 74%. Sediments become semilithified by 100_110 mbsf. An unusual nannofossil (Braarudosphaera) bloom in late Oligocene time, reported previously on the SKP, may have been synchronous with other occurrences in the Atlantic and Indian Oceans. Minimum sedimentation rates were ~16 m/m.y. in the Miocene and ~20 m/m.y. in the Oligocene. We observed very rare tephra layers and disseminated volcanic ash, locally concentrated in burrows. Chert nodules appear only at the base of Unit II. We correlate normal and reverse magnetic polarities between ~100 and ~380 mbsf to early Miocene to early Oligocene geomagnetic Chrons C5D to C12 (or C13). From 108.9 mbsf to the base of Unit II at 380.7 mbsf, velocities increase linearly with depth, from 1785 to 4331 m/s. Within this depth interval, three volcanic ash layers have high P-wave velocities. In the same interval, bulk density increases from 1.3 g/cm3 to 2.1 g/cm3 with a mean of 1.7g/cm3, and porosity decreases from a maximum of 75% to 42%. grain density maintains a nearly constant value of ~2.8 g/cm3. Unit III (380.7 to 383.5 mbsf) is foraminifer nannofossil chalk of anomalous brownish to reddish yellow color. The P-wave velocity averages 3616 m/s. Units I_III represent deep-marine pelagic sedimentation. The base of the pelagic section is earliest Oligocene in age. Unit IV (383.5 to 384.9 mbsf) consists of dusky red to greenish pink sandy packstone with rare planktonic foraminifers and bivalve shell fragments (Fig. 33). Index properties change significantly near the boundary between Units III and IV, from 381.4 to 384.4 mbsf. The bulk density in this zone averages 2.0 g/cm3, grain density averages 2.8 g/cm3, and porosity ranges between ~50% and ~31%. Grains are predominantly highly altered volcanic lithic fragments. P-wave velocity averages 3616 m/s. Unit V (384.9 to 461.7 mbsf) consists of interlaminated grainstone and sandstone with some thin interbeds of rudstone and cross- bedded intervals. Bryozoans, bivalves, and echinoids are the major biogenic components. Units IV and V were deposited in a shallow-marine neritic environment in low-energy and very high-energy (near shore) settings, respectively. Well-rounded cobbles at the top of the basement suggest a beach deposit at the base of the sedimentary succession. At Site 1139, we drilled 232.5 m into igneous basement with a 37.4% recovery rate (Fig. 34). We identify 19 basement units; an upper succession of variably welded trachytic to rhyolitic volcanic and volcaniclastic rocks (Units 1_5) underlain by 14 lava flows (Units 6_19). All basement units are highly altered and fractured. The high degree of alteration and poor core recovery in Units 1_5 make it difficult to identify physical volcanic features and interpret modes of emplacement. However, these units can be distinguished in the natural gamma ray logs. Rocks in Units 1_5 have P-wave velocities varying from 2577 to 4770 m/s, with a mean value of 3616 m/s. Their bulk densities average 2.3 g/cm3, grain densities range from 2.6 to 2.9 g/cm3, and porosities average ~25%. The underlying 14 subaerial lava flow units (Units 6_19) have P-wave velocities that are typically >3000 m/s, with a mean of 4416 m/s. Bulk densities vary widely, with a mean of 2.4 g/cm3; grain density approaches a mean of 2.8 g/cm3 and decreases slightly with depth; and porosity varies widely, from 65% to 3%. All basement units have positive magnetic inclinations, corresponding to reversed polarity. Unit 1, which had poor recovery (57 m thick; 5.3 m recovered), contains a variety of felsic volcanic and volcaniclastic rocks (Figs. 34). Unit 1A consists of rounded, massive to flow banded rhyolite cobbles. Unit 1B is a lens of bioclastic sandstone that resembles the grainstone at the base of the sedimentary section. Beneath this, a thin felsic pumice breccia (Unit 1C) overlies a zone of altered, perlitic felsic glass that contains lithic fragments (Unit 1D). We interpret the glassy zone to be the densely welded core of a pyroclastic flow deposit. The base of Unit 1D is a silicified basal breccia with lithic fragments and pumice. Beneath this are highly sheared and altered, clay-rich volcaniclastic sediments (Unit 1E) that we interpret as a fault zone. Within Units 1C through 1E, both clasts and the matrix commonly display cataclastic fabrics, and slickensides are ubiquitous on broken clay-rich surfaces. Unit 2 (10.5 m thick; 1.35 m recovered) consists of dark red (oxidized) rhyolite with ~10% sanidine and minor quartz phenocrysts. Flattening and agglutination textures suggest that this is a welded pyroclastic flow deposit. Unit 3 (9.7 m thick; 4.6 m recovered) is a green, highly altered, crystal-vitric tuff. It contains abundant sanidine phenocrysts, minor quartz, and lithic clasts, in a perlitic glassy matrix that is locally banded. As with Unit 1C, we interpret Unit 3 to be the densely welded core of a pyroclastic flow deposit. Unit 4 (30.1 m thick; 5.9 m recovered) contains massive to brecciated, dark red (oxidized) rhyolite that is similar to Unit 2. Unit 5 (17.4 m thick; 4.2 m recovered) is highly altered, sanidine-phyric trachyte that consists of a massive central zone bounded by a brecciated top and base; this unit is probably a lava flow. Basement Units 6_17 (65.7 m drilled; 41.4 m recovered) consist of aphyric to sparsely plagioclase-phyric volcanic rocks ranging in composition from trachybasalt to trachyandesite (Fig. 35). We subdivide this sequence into individual lava flow units on the basis of brecciated flow tops and increased vesicularity toward the margins. The flow units vary from 1.8 to 19.8 m in thickness, but most are <6 m thick. Unit 10 consists of small pahoehoe lobes, Unit 11 is an aa flow, and the other flow units have brecciated margins of indeterminate character. The breccias are highly altered and sheared, with both matrix and clasts nearly completely altered to clay minerals. Breccia clasts are oxidized and cemented by calcite and siderite(?) as well as clay minerals. The relatively thin massive portions of the flows have many moderately to steeply dipping fractures and pronounced streaks of mesostasis, now altered to green clay minerals. Thin veins, commonly containing carbonate, pervade the basalt units. Rarely, the rock has a pale gray hue, and the groundmass is bleached because of the replacement of igneous minerals by secondary calcite. Basement Units 18 and 19 are highly to completely altered highly sanidine-phyric trachyandesite and trachyte, respectively. Alteration consists of either intense hematitic red staining or white/pink bleaching. The minerals in the bleached rocks include quartz, sanidine, and siderite. The latter mineral is a common phase that cements groundmass, replaces primary phases, and fills veins and vesicles. The veins within the bleached intervals have prominent red alteration halos (hematite) and are filled with hematite, quartz, siderite, and calcite. The alteration of these rocks is probably the result of interaction between felsic igneous rocks with large volumes of hydrous fluids in some form of geothermal (subaerial?) system. Although their compositions were affected by posteruption alteration, the major element compositions of the volcanic and volcaniclastic rocks comprising the basement at Site 1137 clearly form a series from trachybasalt to trachyte and quartz-bearing rhyolite (Fig. 35). These lavas are significantly more alkaline than the dominantly transitional to tholeiitic basement lavas recovered from all other Kerguelen Plateau drill sites. However, with the exception of the rhyolites, the alkaline Skiff Bank lavas are quite similar to alkaline lava series erupted in the Southeast Province of the Kerguelen Archipelago in the early Miocene and again in the Pliocene and Pleistocene. In summary, significant results bearing on the origin and evolution of Skiff Bank (Site 1139) are 1. Subsidence of the NKP is recorded by the paleoenvironments of volcanic rocks and overlying sediments; since Eocene time, environments have changed from subaerial (volcanic and volcaniclastic rocks) to intertidal (beach deposits) to very high-energy, near- shore (grainstone and sandstone) to low-energy, offshore (packstone) to bathyal pelagic (ooze). 2. The oldest sediments overlying igneous basement are earliest Oligocene; this minimum age for basement is consistent with a <40 Ma age for the uppermost igneous basement of Skiff Bank. 3. The 233 m of igneous basement consists of an uppermost 124-m succession of variably welded trachytic to rhyolitic volcanic and volcaniclastic rocks; in addition, the two lowermost lava flow units comprise 33 m of sanidine-phyric trachyandesite and trachyte. As at Elan Bank (Site 1137), highly evolved magmas erupted, in some cases explosively, during the final stages of volcanism that formed Skiff Bank. 4. The volcanic basement at Skiff Bank, with a minimum age of earliest Oligocene, includes an alkaline lava series ranging from trachybasalt to trachyte. Similar alkaline lavas have erupted in the Kerguelen Archipelago in early Miocene and Pliocene/Pleistocene time; therefore Skiff Bank, which crests <500 m below sea level, may have been a somewhat older island analogous to the Kerguelen Archipelago 350 km to the east- northeast. Site 1140 Site 1140 lies on the northernmost Kerguelen Plateau ~270 km north of the Kerguelen Archipelago (Fig. 4). Flanked to the north and east by Eocene and younger oceanic crust of the Australian-Antarctic Basin, and to the west by Cretaceous oceanic crust of the Crozet and Enderby basins, the NKP is believed to have formed since 40 Ma via Kerguelen hot spot magmatism (Royer and Sandwell, 1989; Royer and Coffin, 1992; M.F. Coffin et al., unpubl. data). The boundary between the northern Kerguelen Plateau and the Australia- Antarctic Basin lies ~5 km north of Site 1140 and offsets basement by ~400 m. The Kerguelen Archipelago is part of the NKP; its igneous rocks yield dates from 39 Ma to recent (K.E. Nicolaysen et al., unpubl. data). However, submarine igneous basement of the NKP has never been sampled, so its age and composition, as well as its relationship to the central and southern plateau sectors and to Skiff and Elan banks, are unknown. Site 1140 lies at a depth of 2450 m on the northern flank of the NKP. We chose this location as representative of the NKP on the basis of its relatively simple structural setting and thin sedimentary section (Fig. 36). The top of acoustic basement is flat lying, and the overlying basement is a sediment sequence ~350 m thick. The major objectives at Site 1140 were to obtain igneous basement to characterize the ages, petrography, and compositions of the lavas and the environments of eruption (subaerial or submarine). We were especially interested in (1) testing the hypothesis that at least the uppermost igneous basement of the NKP is <40 Ma and (2) comparing the submarine NKP lavas with the subaerial lavas forming the Kerguelen Archipelago. The sedimentary objectives at Site 1140 were to determine sequence facies, to define the ages of seismic sequence boundaries, to estimate the duration of possible subaerial and shallow-marine environments, to obtain minimum estimates for basement age, and to determine the paleoceanographic history of this moderate latitude site. As discussed below, we largely achieved our goals at Site 1140. We drilled 87 m into pillow basalt flows that are intercalated with thin chalk beds containing late Eocene nannofossils and foraminifers. The sedimentary section above igneous basement consists entirely of pelagic ooze and chalk and appears to rest unconformably on the underlying submarine basalt flows. We recognize only one sedimentary unit (lithologic Unit I) overlying volcanic basement rocks (Fig. 37). Unit I (0 to 234.5 mbsf) predominantly consists of light greenish gray foraminifer- bearing nannofossil ooze and nannofossil chalk. Biostratigraphic data, as well as preliminary interpretation of reversed and normal magnetic Chrons, indicate that lithologic Unit I is middle Miocene to early Oligocene or latest Eocene in age. We divide this unit into two subunits (IA and IB) based on the presence of diatom ooze in the uppermost part of the unit. Subunit IA (0 to 10.0 mbsf) consists of white diatom nannofossil ooze with interbeds of dark brown silty diatom ooze, light brown silty foraminifer-bearing diatom ooze, and yellowish brown nannofossil- bearing diatom ooze. Subunit IB (10.0 to 234.5 mbsf) comprises most of the sedimentary section and is predominantly light greenish gray foraminifer-bearing nannofossil ooze, which contains middle Miocene nannofossil and planktonic foraminifer species of warm-water affinity not found elsewhere on the Kerguelen Plateau. Physical properties in Subunit IA and the upper part of Subunit IB (0_180 mbsf) vary only slightly; bulk density ranges from 1.4 to 1.7 g/cm3, grain densities range between 2.1 and 2.8 g/cm3, and porosity changes from 57% to 76%. P-wave velocities show little scatter, ranging from 1491 to 1852 m/s. As the ooze becomes semilithifed nannofossil chalk downhole (~180_234 mbsf), bulk density gradually increases from 1.5 to 2.0 g/cm3 (mean = 1.7 g/cm3), with porosity decreasing from a maximum value of 74% to 44% (mean = 60%). Grain density is nearly constant at ~2.7 g/cm3 throughout this interval, and velocity increases from 1578 to 2018 m/s. At the base of Subunit IB, just above igneous basement, clear rhombic dolomite crystals are disseminated throughout the sediments. Nannofossils and planktonic foraminifers in the ooze directly overlying igneous basement indicate a minimum basement age of early Oligocene (30.0_34.3 Ma). All physical properties change abruptly at the sediment/basalt boundary. From 235 to 250 mbsf, porosity decreases sharply from a mean of 60% in Lithologic Subunit IB to 6% in basalt flows, and grain density increases from 2.7 to 2.9 g/cm3. P- wave velocity varies from 5484 to 6859 m/s. Drilling at Site 1140 penetrated 87.4 m of basement rocks, which we divide into six units, five submarine basaltic flows (Units 1_3, 5, and 6) and an ~1-m-thick layer of dolomitized nannofossil chalk (Unit 4). Two other thin calcareous-dolomitic sedimentary interbeds are between basalt flows at the Unit 2_3 and Unit 5_6 boundaries. We observe a magnetic reversal at the boundary between basement Units 1 and 2. Unit 1 is normal polarity, and Units 2 through 6 are reversed, indicating a hiatus in volcanism as construction of the northernmost Kerguelen Plateau's igneous crust ended. Downhole logs of density, resistivity, and velocity show high values in the interiors of basalt flows and lower values at flow margins and in the interbedded sediments. At the top of basement Unit 3, a thin bed of well-burrowed, greenish white nannofossil chalk is latest Eocene in age. Basement Unit 4 contains a sedimentary bed with a top and bottom composed of rusty orange dolomite separated by a bed of well-burrowed, very pale brown dolomitic nannofossil chalk. Index properties change sharply at the boundary between Units 3 and 4; bulk density decreases from 2.8 to 2.1 g/cm3, grain density decreases to a mean of 2.8 g/cm3, and porosity increases to a mean of 41%. In Units 5 and 6, bulk density ranges from 2.5 to 3.0 g/cm3, porosity changes from 4% to 24%, grain density varies between 2.9 and 3.1 g/cm3, and velocity ranges from 5099 to 6829 m/s (mean = 6055 m/s). The top interval of basement Unit 6 is rusty brown dolomite resembling that of basement Unit 4. The interbedded sediments indicate bathyal water depths during late Eocene to early Oligocene extrusion of the lava flows. Pelagic deposition in a bathyal environment continued uninterrupted until at least middle Miocene time. Basement Units 1 and 6 each contain a ~5-m-thick massive lobe in addition to ~30 small (50 to 100 cm) basaltic pillows. Only <1-m-diameter pillows were recovered from Units 2 and 3. Unit 5 contains similar pillows and an ~10-m-thick massive lobe. Comparison between cores and logging data indicate that these flow units are 4.4 to 23.4 m thick. The thick massive lobes are probably sheet flows. Thick sheet flows and absence of rubbly talus suggest low to moderate slopes. Although small pillows cannot advance far before freezing, the larger sheets could efficiently transport magma from a distant vent. The flows are cryptocrystalline to fine grained and generally only sparsely vesicular. Vesicles are largely restricted to chilled margins. Vesicularity varies within the units but is consistently low, suggesting the deep water corroborated by bathyal sediments. Pillow margins are fine grained with 1- to 2-cm-wide unaltered glassy rims (Fig. 38). Calcareous sediment or carbonate veins commonly fill sutures between pillows. The fine-grained pillow margins consist of moderately plagioclase Ò olivine Ò clinopyroxene-phyric basalt, whereas pillow interiors range from plagioclase-phyric to aphanitic. Olivine is a minor phenocryst and groundmass phase in Units 1 and 2, (Fig. 39) but it is rare to absent in the lower basaltic units, in which clinopyroxene is a phenocryst phase. Units 4 and 6 are moderately plagioclase-phyric, whereas the others are essentially aphanitic with <1% phenocrysts in the massive portions of the flows. Groundmass phases are calcic (An60_70) plagioclase (20%_40%), augite (25%_40%), olivine (0%_5%), titanomagnetite (1%_2%), and altered glass. Textures range from ophitic, intergranular, or intersertal in pillow interiors to glassy at pillow margins. Alteration of Site 1140 lavas strongly resembles that of young mid- ocean ridge pillows from the uppermost ocean crust (e.g., DSDP/ODP Holes 504B and 896A, located in 5.9-Ma-old crust in the eastern equatorial Pacific Ocean). Glass on the margins of lava pillows is fresh and isotropic in thin section. Glassy margins are crosscut by numerous calcite and dolomite veins that developed concentrically to the pillow rinds. These veins are generally wide (2_3 mm), and the carbonates exhibit dog-tooth, sparry habits. Baked, highly indurated chalk- derived marbles are commonly preserved in the pillow interstices. Rarely, these sediments apparently penetrated the magma, resulting in internal glassy quenched zones in the pillow interiors. Crystalline interiors of the lavas are slightly to moderately altered. The most common feature of the alteration is development of brown to orange oxidation halos concentric to the pillow rinds, as well as along clay and carbonate veins. Orange brown clays and iron oxyhydroxides pseudomorphically replace groundmass mafic minerals. The gray to greenish gray portions of the basalts are generally fresh, except for the replacement of mesostasis by green clays and the partial filling of rare vesicles with green-blue clay and coarse-grained pyrite. Oxidation halos are less common in the more massive, fine-grained interiors of the thicker lava units, except where these rocks are intercalated with ~1-m-thick beds of dolomitized and oxidized chalk. The sedimentary rocks may have acted as channels enabling the access of large volumes of seawater-derived fluids into the basement sequences, resulting in the precipitation of abundant, euhedral, colorless dolomite crystals in the chalk and numerous sparry carbonate veins in the pillow lavas. Compared to other basement basalt recovered from the Kerguelen Plateau during Leg 183, the Site 1140 basalts are distinctive in that they (1) were erupted in a submarine environment, as indicated by their pillowed structure and the intercalcated nannofossil- bearing calcareous sediments and (2) are relatively unaltered, as indicated by fresh glass preserved at pillow margins and the lack of alteration in massive interiors of flow units. The five basement flow units at Site 1140 are tholeiitic basalts that are poorer in alkalis than lavas at other locations on the Kerguelen Plateau, except for Site 750 (Fig. 19). Basalts from Site 1140 range to higher MgO (8.1%) and Ni (100 ppm) contents than basalts from other Leg 183 drill sites. They form two distinctive geochemical groups. Relative to Units 1, 5, and 6, Units 2 and 3 are enriched in highly incompatible elements, such as P, Zr, and Nb, by factors of two to four (Fig. 40). Units 1, 5, and 6 have near chondritic ratios of Nb/Zr and Zr/Y; in this respect they are similar to the ~110-Ma lavas from Site 749 on the SKP. Despite their eruption in late Eocene time, when the SEIR was <50 km away, Site 1140 lavas are not geochemically similar to depleted MORB. They are geochemically similar to other tholeiitic basalts associated with the Kerguelen plume (Fig. 31). Unlike basalts from Elan Bank (Site 1137), basalts from Site 1140 show no evidence for a component derived from continental lithosphere. Major results of drilling Site 1140 on the northern flank of the NKP include 1. Miocene nannofossil and planktonic foraminifer species of warm water affinity that had not been recovered previously by drilling on the Kerguelen Plateau. 2. Five submarine basalt flows with intercalated sediments. The ~1- m-thick dolomotized nannofossil chalk and other thinner chalk interbeds within the submarine basaltic flow units indicate episodic eruptions in a bathyal environment. Nannofossils and foraminifers in these chalks indicate a latest Eocene age, confirming that the uppermost igneous basement of the northernmost Kerguelen Plateau formed at <40 Ma. A magnetic reversal at the boundary between Basement Units 1 (normal) and 2_6 (reversed) corroborates nonvolcanic intervals as the uppermost igneous crust of the NKP formed. 3. Basalt flows, ~4 to 23 m thick, which are dominantly <1 m pillows with a few massive, 5- to 10-m lobes. The ~1-cm-thick quenched pillow margins are fine grained and contain macroscopically unaltered glass that is isotropic in thin section. Basaltic glass has not been recovered at other drill sites on the Kerguelen Plateau; shore-based studies of this glass will provide geochemical data, especially abundances of H2O, CO2, and S, that cannot be obtained from studies of altered crystalline rocks. 4. Five pillow lava flow units consist of tholeiitic basalts that form two distinct geochemical groups; both groups have incompatible element abundance ratios within the range of other tholeiitic basalts associated with the Kerguelen plume. Unlike basalts from Elan Bank (Site 1137), there is no evidence for a continental lithosphere component. Sites 1141 and 1142 Sites 1141 and 1142 are situated near the crest of Broken Ridge ~350 km east of DSDP/ODP Sites 255, 752, 753, 754, and 755 (Figs. 5, 6). Flanked to the south by Eocene and younger oceanic crust of the Australia-Antarctic Basin and to the north by Cretaceous oceanic crust of the Wharton Basin, Broken Ridge appears to have formed during Late Cretaceous time as a result of Kerguelen hot spot magmatism (Duncan, 1991; Duncan and Storey, 1992; M.K. Coffin et al., unpubl. data). Subsequently, Broken Ridge and the Kerguelen Plateau began to separate along the nascent SEIR at ~40 Ma. Igneous basement of Broken Ridge had not been sampled previously by drilling; dredge samples from three locations along the feature's southern, faulted boundary yield dates of ~62, ~83, and 88_89 Ma (Duncan, 1991). Because of the scatter in ages of the dredged rocks and the absence of in situ basement samples from Broken Ridge, knowledge of Broken Ridge's age and composition remains extremely limited. We located Sites 1141 and 1142 on the JOIDES Resolution single-channel seismic Line JR183-101. Sites 1141 and 1142 lie at depths of 1197 m and 1201 m, respectively, ~3_4 km north of the crest of Broken Ridge. We chose this location primarily on the basis of its thin sedimentary section (Fig. 41). The top of acoustic basement has an apparent dip to the north-northeast of 0¯ at Site 1141 and 2.5¯ at Site 1142. An ~100-m-thick sediment sequence overlies igneous basement. Since basement of Broken Ridge had never been drilled, our major objective at Sites 1141 and 1142 was to determine its age and composition. Additional basement objectives were to determine the physical characteristics of the lava flows and the environment of eruption (subaerial or submarine). The sedimentary objectives at Site 1141 were to determine sequence facies, to define the ages of seismic sequence boundaries, to estimate the duration of possible subaerial and shallow marine environments, to obtain minimum estimates for basement age, and to determine the paleoceanographic history of Broken Ridge. At Sites 1141 and 1142, these objectives were achieved by coring 72 and 51 m of volcanic basement, respectively, and ~113 m of overlying sediment at Site 1141 (Figs. 42, 43). The abrupt termination of Hole 1141A led to an unanticpated experiment whereby we compared two basement sections separated by only 800 m. At Site 1141, sediments were recovered from 0 to 103.8 mbsf (Fig. 42). We recognize only one sedimentary unit, lithologic Unit I. The basement volcanic rocks are designated lithologic Unit II. Unit I (0_113.5 mbsf) consists of white foraminifer nannofossil ooze of Pleistocene to early Miocene age. Core 183-1141A-1R consists of nannofossil-bearing foraminifer ooze that is predominantly composed of sand-sized foraminifers and displays slight normal size-grading. Traces of aragonite are present in Core 183-1141A-1R. Temperate calcareous microfaunas and floras characterize the current-winnowed Neogene calcareous ooze recovered at Site 1141. They are joined by subtropical index taxa in the Pliocene_Pleistocene section, a result of northward movement of Broken Ridge into warmer, lower latitude waters. The average sedimentation rate of 6 m/m.y. for the entire carbonate ooze section is the lowest Neogene rate for Leg 183. In Cores 183-1141A-8R and 9R, we obtained reliable remanent magnetization and correlated normal and reversed polarities with middle Miocene Chrons C5 to C5AD. Bulk densities in Unit I vary from 1.6 to 1.8 g/cm3, and porosity ranges from 54% to 65%, with a mean of 62%. P-wave velocities in Unit I show very little scatter, with a mean value of ~1860 m/s. The base of Unit I consists of a layer of sandy foraminifer limestone with abundant sand- to pebble-sized rock fragments and mineral grains. The limestone, late-middle to late Eocene in age (35_38 Ma), postdates rifting and separation of Broken Ridge and the CKP. The pebbles include basalt with ferromanganese crusts. The thickness of this basal layer is uncertain as only two small fragments were recovered. The pelagic sedimentary succession at Site 1141 indicates that Broken Ridge has been at bathyal water depths since at least early Miocene time. Neritic fossils in the basal limestone indicate redeposition from shallow-water areas to a bathyal environment during the Eocene or later. Unit II consists of basalts, which are highly altered in the upper portion of the section. It is subdivided into six basement units. Below the boundary between Units I and II (~114 mbsf), index properties change abruptly. From 115.8 to 116.7 mbsf in the upper part of basement Unit 2, bulk densities increase to a mean value ~2.0 g/cm3, grain densities increase to 2.9 g/cm3, and porosities decrease to 48%. At Site 1142 (Fig. 43) no sediments were recovered from the drilled interval (0_91 mbsf; Core 183-1142A-1W, except for some small fragments of sandy pebbly foraminifer limestone with Oligocene or Eocene nannofossils. The six basement units at Site 1141 represent 71.2 m of basement penetration and consist of five mafic lava flows overlain by a coarse-grained sedimentary deposit of three small fragments of moderately altered, medium-grained, plagioclase-clinopyroxene-olivine gabbro, which could be from a dike, sill, or gravel bed. These lava flows appear to have been erupted subaerially; no evidence suggests interaction with water during emplacement. The top of most volcanic units in Hole 1141A have been highly to completely altered to clay; in some cases this intense alteration affects entire units. Red flow tops and green to gray flow interiors suggest decreasing oxidation with depth. In many intervals, traces of native copper are in the groundmass, and abundant native copper line some fracture surfaces. Vesicles are filled with dark green clay, calcite, zeolite, amorphous silica, and quartz and have well-developed colloform textures. Slickensides are numerous along some fractured surfaces. Perhaps the most noteworthy alteration within Hole 1141A is the spectacular alteration halos associated with quartz veins. In some instances, a single quartz vein extends for >120 cm with multiple, symmetrical alteration halos progressively altering the surrounding wall rock. Common calcite veins are generally <0.5 mm wide and crosscut the quartz-filled veins. Basement Unit 2 (20.0 m thick) is fine-grained, aphyric basalt. The least altered portion of Basement Unit 3 (8.3 m thick) is fine- to medium-grained, sparsely plagioclase-phyric basalt. Basement Unit 4 (19.6 m thick) is aphyric to moderately plagioclase- or plagioclase-olivine-phyric basalt. Most of the massive interior of this unit is moderately altered, but a fresher, denser, and finer-grained zone might represent a dike intruded into the basement Unit 4 flow; however, no contacts were recovered. Basement Unit 5 (8.0 m thick) is sparsely olivine-phyric. Basement Unit 6 (15.3 m thick) ranges from aphyric to moderately olivine-and- plagioclase-phyric basalt. In addition to olivine phenocrysts, groundmass olivine and minor apatite in the lower part of basement Unit 4, and throughout basement Units 5 and 6, suggest that these basalts are alkalic. Thin sections show that carbonate, clay, and iron oxides completely replace the mafic phases and groundmass glass from the top of Unit 1 to the upper part of Unit 6; the bottom part of Unit 6 retains a large proportion of relatively fresh phenocryst and groundmass olivine. All index properties change markedly near the boundary between basement Units 2 and 3 and reach extremes in basement Unit 6, where bulk densities vary from 2.6 to 2.9 g/cm3 with a mean of 2.7 g/cm3, grain density approaches a mean of 2.8 g/cm3, and porosity varies from 16% to 3%. P-wave velocities in basement Unit 6 increase gradually with depth, from 4276 to 6902 m/s. At Site 1142, 50.9 m of basement penetration recovered six basement Units (Fig. 43). They include a diverse range of lithologies, including olivine-phyric basalt lava flows, possible pillow basalts, subaerial deeply weathered (felsic?) lavas, and volcanic sediments. Alteration and weathering of these basement rocks suggest subaerial exposure. The lithologies and alteration intensity within Hole 1142A are heterogenous. Some units are relatively massive and only slightly altered basalt, whereas other volcanic units are variably brecciated by both volcanic and tectonic processes and have been completely altered to clay. Primary igneous textures are still visible in most units and are typically accentuated by the replacement of feldspar by light green clay and mafic minerals by red-brown clay. Basement Unit 1 (1.9 m thick) is a slightly to moderately altered, massive, fine-grained, aphyric to sparsely plagioclase-and-olivine-phyric basalt; the upper portion has prominent oxidation halos, suggesting a period of exposure and weathering, possibly related to Eocene rifting and breakup between Broken Ridge and the CKP. Basement Unit 2 (1.2 m thick) consists of a single section containing 20 moderately to completely altered cobble- sized pieces of genetically unrelated rock types, including volcanic breccia, clinopyroxene-phyric, plagioclase-phyric, olivine- phyric, and aphyric basalt, and feldspar- and feldspar-quartz-phyric felsic volcanic rocks. Some of the pieces have abraded, slightly weathered surfaces, suggesting the unit may be a near-source debris flow or a talus pile. Basement Unit 3 (19.3 m thick) is a completely altered, aphanitic, aphyric to moderately olivine-plagioclase-phyric basaltic breccia with three subunits defined on the basis of textural characteristics. Basement Unit 4 is a 4.3-m-thick, well- indurated, normally graded claystone or mudstone with very highly to completely altered, very coarse sand-sized to small granule-sized lithic clasts and crystals of quartz and altered feldspar in a red clay matrix. We interpret this unit to be a mudflow deposit. Basement Unit 5 (8.8 m thick) is a very highly to completely altered, aphanitic, aphyric volcanic breccia. The only hint of the original rock type is provided by very rare quartz crystals, which suggest an evolved composition. Basement Unit 6 is composed of Ú9.6 m of nonvesicular, massive, and possibly pillowed basalt. It contains several fine-grained clay margins that strongly resemble highly altered glass. However, dark green to black alteration halos adjacent to calcite veins within massive basalt appear similar to features interpreted as highly altered glassy pillow margins. Thus, evidence is equivocal as to whether these are actually pillow basalts. Basalt in Unit 6 is aphyric and fine-grained to aphanitic. Spectacular sinuous and semicircular red-brown oxidation fronts and halos are common and are often crosscut and offset by calcite veins. These oxidation bands, similar to those in weathered pillow basalts, may represent subaerial weathering after the uplift of Broken Ridge that accompanied its breakup with the CKP. All basement rocks in Holes 1141A and 1142A are normally magnetized. Despite a lateral separation of only 800 m, we cannot correlate the basement units at Sites 1141 and 1142. Seismic reflection data collected over the sites during Leg 183, as well as seismic reflection data and drilling results from Sites 752_755 lying ~350 km to the west, indicate that to the north of the bathymetric crest of Broken Ridge, prebreakup sediment and presumably igneous basement dip consistently to the north. This may account for the differences in basement units between Sites 1141 and 1142, in that Site 1142 penetrated a deeper stratigraphic section than Site 1141 did. Alternatively, lateral variability of igneous basement in Broken Ridge could be much greater than that in continental flood basalts. The major results of drilling Sites 1141 and 1142 include 1. Six basement Units at Site 1141 include five mafic flows that appear to have been erupted subaerially. 2. Several of these lava flows, 8 to 20 m thick, contain phenocryst and groundmass plagioclase and olivine. 3. Six basement Units at Site 1142 range from olivine-phyric basaltic flows to subaerial deeply weathered (felsic?) lavas and volcanic sediments. A surprising result is that basement Unit 6 may be a pillow basalt that was subsequently oxidized in a subaerial environment, perhaps after the ~40-Ma uplift of Broken Ridge that accompanied its breakup with the CKP. 4. Differences between igneous basement sections at Sites 1141 and 1142 may result from penetration of different stratigraphic levels or from considerable, unanticipated lateral variability in volcanism. SUMMARY Coring Summary During Leg 183, we drilled eight widely spaced holes in different domains of the Kerguelen Plateau_Broken Ridge LIP (Fig. 1): the SKP (Sites 1135 and 1136), Elan Bank (Site 1137), CKP (Site 1138), Skiff Bank on the NKP (Site 1139), northern NKP (Site 1140) and Broken Ridge (Sites 1141 and 1142). Except at Site 1135, igneous basement rocks were recovered with penetrations ranging from 33 to 233 m (Table 1). Chronology of Kerguelen Plateau_Broken Ridge Magmatism Biostratigraphic studies of sediment directly overlying igneous basement at Leg 183 sites provide minimum ages for the volcanic and volcaniclastic rock. Middle Albian (~104.5_106.5 Ma) shallow-water sands and clays overlie inflated pahoehoe flows at Site 1136, suggesting that the age of the lavas is close to the ~110 Ma age of all other basalt samples recovered to date from the SKP. Site 1138 on the CKP yielded undifferentiated Upper Cretaceous claystone and sandstone on top of igneous basement; these sediments are overlain by Cenomanian_Turonian (~93.5 Ma) sandstone. This age is older than, but close to, the ~85 Ma date for basalt at Site 747 (M.F. Coffin et al., unpubl. data). Drilling at the conjugate Broken Ridge region (Site 1141; see Fig. 2) did not provide useful minimum ages because the oldest sediment is Miocene, postdating Eocene separation of Broken Ridge and the CKP and, therefore, much younger than the formation age of Broken Ridge. The first igneous basement ever recovered from Elan Bank at Site 1137 is overlain by late Campanian (73.5_74.6 Ma) packstone. The lavas and volcaniclastic sediments forming basement are likely to be somewhat older, as the packstone is at the top of a basal sedimentary sequence that thickens markedly to the east of Site 1137. Submarine igneous basement of the NKP was cored for the first time at Site 1139 on Skiff Bank and Site 1140 on the northernmost Kerguelen massif (Fig. 3). On Skiff Bank, chalk at the base of the pelagic sedimentary section is earliest Oligocene (32.8_34.3 Ma) in age. Igneous basement is probably older, as grainstone, packstone, and sandstone lie between it and the overlying pelagic section. Nevertheless, the minimum age is not inconsistent with the oldest rocks from the Kerguelen Archipelago (Giret and Beaux, 1984; K.E. Nicolaysen et al., unpubl. data). At Site 1140, lowermost Oligocene (34.3 Ma) pelagic sediment directly overlies basement, and pelagic sediment of late Eocene age (~35 Ma) is intercalated within basalt flows that form the uppermost basement. Petrogenesis of Basement Igneous Rocks Except for Site 1139 on Skiff Bank (NKP), the dominant lava type at all drill sites from diverse domains of the Kerguelen Plateau_Broken Ridge LIP are basalt with transitional to tholeiitic compositions (Fig. 19). These basalts range in inferred age from Albian to early Oligocene and are from the SKP (Sites 738, 749, 750, and 1136), Elan Bank (Site 1137), CKP (747 and 1138), NKP (1140), and Broken Ridge (Sites 1141 and 1142). The basalts have relatively low MgO and Ni contents, and their compositions are not similar to primary melts of peridotite. An example of the important role for fractional crystallization in controlling the compositions of these lavas are basalts from Site 1138, which show a systematic downhole trend to Fe-Ti-rich basalt reaching TiO2 and total Fe2O3 contents of 3.2% and 19.2%, respectively. In general, evolved tholeiitic basalt compositions are typical of many flood basalts. The likely explanation is that the youngest magmas in a LIP, like Kerguelen Plateau_Broken Ridge, must ascend through relatively thick lithosphere, thereby promoting cooling and partial crystallization of the magma. Subsequent segregation of olivine- and pyroxene-rich cumulates then forms the high- velocity lower crust that is typical of oceanic plateaus such as the Kerguelen Plateau (Charvis et al., 1993, 1995; Operto and Charvis, 1995, 1996; Farnetani et al., 1996; Kînnecke et al., 1998; Charvis and Operto, 1999) and led to the complementary evolved residual melts. In regard to this scenario, it is important to realize that we have only sampled the upper 30_200 m of a ~20-km- thick mafic crust. The basement lavas at Site 1139 (Skiff Bank on the NKP) are an exception to the previous generalization; they form an alkaline lava series ranging from trachybasalt to trachyte and rhyolite; the lowermost flow at Site 1139 is a trachyte (Figs. 34, 35). Similar alkaline lava series have erupted in lower Miocene and Pliocene/Pleistocene time in the Southeast Province of the Kerguelen Archipelago (Weis et al., 1993, 1998; Frey et al., in press). The simplest interpretation is that Skiff Bank, which reaches <500 m water depth, is a submerged island analogous to but slightly older than the Kerguelen Archipelago, also on the NKP but 350 km to the east-northeast (Fig. 3). An important objective of Leg 183 was to evaluate the role of continental crust in constructing the Kerguelen Plateau_Broken Ridge LIP. Previous evidence pointing to a significant role for continental crust in diverse parts of the LIP includes isotopic and trace element abundance data for basalts from the SKP (Site 738), CKP (Site 747), and basalts dredged from the SKP and eastern Broken Ridge (Figs. 11, 12, 13; Storey et al., 1989, 1992; Mahoney et al., 1995) and the seismic structure of the crust in the northern part of the SKP (Operto and Charvis, 1995, 1996). Some mantle xenoliths in the Kerguelen Archipelago lavas also show evidence for a continental lithosphere component (Hassler and Shimizu, 1998; Mattielli et. al., 1999), as does a trachyte from Heard Island (Barling et al., 1994). Drilling at Site 1137, however, recovered the strongest evidence to date for a component of continental crust in the Kerguelen Plateau. A ~26-m-thick fluvial conglomerate (Fig. 44) intercalated with basaltic basement contains clasts of trachyte, rhyolite, granitoid, and garnet-biotite gneiss (Fig. 45); the garnet-biotite gneiss, in particular, indicates that continental crustal rocks were once exposed at Elan Bank. Furthermore, although it is difficult to use shipboard geochemical data to identify continental material in mantle-derived basaltic rocks, our study of Site 1137 cores builds a compelling case. The basement basalts at Site 1137 are geochemically distinctive; they have atypically high Zr/Y and Zr/Ti (Figs. 24) and a slight relative depletion in Nb abundance (Figs. 24, 25)_both characteristics are consistent with a component derived from continental crust. A Nb/Y vs. Zr/Y plot has been used to distinguish between lavas derived from the Icelandic plume and North Atlantic MORB (Fitton et al., 1997, 1998). In this plot (Fig. 31), it is important to realize that different plumes are geochemically distinct; in particular, lavas of varying age, from ~82 Ma to Pliocene_Pleistocene, associated with the Kerguelen plume define a Nb/Y-Zr/Y trajectory along the lower boundary for the Icelandic plume (Fig. 31). We conclude that the Kerguelen plume plots on this line in the lower left portion of the figure. Basement basalts from several locations on the Kerguelen Plateau (and Broken Ridge?) also lie along this trend (namely, Sites 747, 1138, 1138, and 1140). Two other locations are in the Icelandic field (Sites 749 and 750), and two others lie in the MORB field (Sites 738 and 1137); however, basalts from these latter two sites do not have MORB geochemical characteristics (Figs. 9, 25). We infer that basalts from Sites 738 and 1137 are in the MORB field because they are plume- derived basalts that have been contaminated by continental crust (Fig. 25). The effects of crustal contamination are obvious in the Sr, Nd, and Pb isotopic characteristics of Site 738 basalts (Figs. 9, 10). Isotopic data are not yet available for Site 1137 basalts, but the clasts of garnet-biotite- gneiss in a conglomerate intercalated with these basalts (Fig. 45) unambiguously show that continental crust is present in the oceanic environment of Elan Bank. Continental material, whether derived from continental crust or subcrustal continental lithosphere, is occasionally incorporated into oceanic lithosphere. At one end of the spectrum are microcontinents such as Seychelles and Jan Mayen, which maintain normal continental crustal thicknesses when isolated in ocean basins by jumps of seafloor spreading centers, with or without involvement of mantle plumes. At the other end of the spectrum are subtle geochemical signatures in otherwise typi