DISCUSSION

The KP provides a unique opportunity to study environmental, oceanographic, and biotic changes from the Cretaceous to the Holocene along an extended (18°) latitudinal north-south transect across an elevated Southern Ocean bathymetric feature. This possibility is not available anywhere else in Antarctic waters. Leg 183 has added considerably to the sediment recovery along this feature. However, since all sites were rotary cored during the leg, they are less than ideal for high-resolution paleoceanographic studies, as would be the case if hydraulic piston coring had been used. Nevertheless, the high sedimentation rates often recorded offset this handicap somewhat; therefore, it is useful to discuss some of the biostratigraphic and paleoenvironmental problems encountered. These are discussed in ascending stratigraphic order, beginning with the K/T Boundary.

The K/T boundary at Site 1135 is represented by an unconformity that spans Zones NA1 and NA2 of Wei and Pospichal (1991). On the other hand, a complete boundary appears to have been recovered at Site 1138, although the FO of B. sparsus (Sample 183-1138A-52R-3, 70 cm [490.4 mbsf]) occurs 57 cm above the dramatic color change (Sample 183-1138A-52R-3, 127 cm [490.97 mbsf]) at which the Shipboard Scientific Party placed the boundary. This is not surprising since B. sparsus is usually rare at most sites at which it is encountered (Table T4).

The initial shipboard (whole-core "pass through") paleomagnetic stratigraphy, however, does not support this biostratigraphic placement of the K/T boundary at interval 183-1138A-52R-3, 127-128 cm (490.97 mbsf). Instead, the interval from 489 to 492 mbsf is interpreted as a paleomagnetic normal (Fig. F8) rather than the expected reversed interval (Shipboard Scientific Party, 2000d).

In contrast, however, our stable isotopic analysis reveals a 0.5 decrease in ¹³C across the interval at which the K/T boundary was placed (Fig. F8). This decrease of ¹³C is widely known to occur at the boundary (Pospichal, 1996a; Zachos et al., 1992; Stott and Kennett, 1990) and is similar to the amount measured at other Southern Ocean K/T sites. For example, at Site 750, a ¹³C decrease of ~0.9 (whole-rock sample) across the boundary was reported by Zachos et al. (1992). Site 690 on Maud Rise shows a decrease in ¹³C across the boundary in both benthic and planktonic foraminifers (Stott and Kennett, 1990). Site 527 (Walvis Ridge) also shows a decrease in ¹³C of ~1 across this boundary (Pospichal 1996a).

Some low-latitude sites, such as El Kef, Tunisia, show a greater shift in the carbon isotope values. Keller and Lindinger (1989) reported an isotopic shift of ~2 at the boundary. A 2 negative shift in the carbon isotopes is also reported at Agost, Spain (Smit, 1990) (Fig. F13) (see Pospichal, 1996a, for a summary of these sites). ODP Leg 171B sites (Blake Nose), on the other hand, show a decrease of only 1 across the boundary (Norris et al., 2001).

The overall pattern of a negative carbon isotope shift at the boundary indicates a global productivity collapse (Shackleton and Hall, 1984). The differing magnitudes of the carbon isotope shift at widely separated sites from around the world would seem to indicate different environmental responses to the crisis in different regions (Pospichal, 1996b). It has been suggested that the magnitude of the extinction may have varied significantly from the low to high latitudes (Keller, 1988), but this seems not to have been the case. Pospichal (1996b) and Pospichal and Wise (1990a) clearly demonstrated that the extinction event at the K/T boundary was just as severe in the high latitudes as it was in the low latitudes. The different values for the carbon isotopic shift, therefore, could result from taxonomic variations among the assemblages (mainly calcareous nannofossils) that contributed skeletal material to the fine-fraction carbonate used in isotopic analyses.

In addition, there is the problem of reworked Cretaceous taxa skewing the isotopic signal. This problem is especially acute at sites with heavy bioturbation. There, the more positive ¹³C Cretaceous taxa mix with the more negative ¹³C Danian taxa, resulting in carbon isotope values for the fine fraction somewhere in between the two end-member values. This mixing of Cretaceous and Danian taxa could also explain why ¹³C values reach their most negative values above the K/T boundary at approximately the Zone NA2/NA3 boundary at Sites 527, 738, 690, and 1138. In contrast, at El Kef and Agost, Spain, sites with little reworking of Cretaceous taxa, the ¹³C negative shift is greatest at the K/T boundary (Pospichal, 1996a).

It could also be argued that the location of the most negative ¹³C values indicate the location of the K/T boundary. This assumes that the original carbon isotope event occurred at this point, but that it has been "smeared out" by subsequent bioturbation.

Berger and Heath (1968) proposed a vertical mixing model that suggests that the signal of an event would not be moved up or down as a result of bioturbation and reworking. If this was the case at Site 1138, we would expect that the isotope excursion would have to coincide with the sedimentation change that occurred, indicated by the color change.

According to the model, the darkest part of the core would represent the sediment change, and the lightening of the sediment color away from that level (both above and below) would be due to vertical mixing. In other words, the darkest sediment color in the sequence would coincide with the most negative point in the ¹³C excursion. This is clearly not the case, however, as we find that the more negative values of the isotope excursion are present well above this dark zone, almost at the point where sediment is returning to the background color. The offset between the darkest color and the minimum value in the isotope excursion is 0.65 m. The pattern is best explained by assuming an instantaneous event at the K/T boundary (490.97 mbsf) that produced both a maximum color change plus a maximum change in the ¹³C values. The latter does not coincide with the former, however, because reworking of Cretaceous sediment has diluted the isotope signal.

In short, we believe our placement of the K/T boundary is supported by the isotopic and nannofossil data. Not only is B. sparsus found in several samples within Zone NA1, but there is a greater relative number of M. inversus within that interval in contrast to the underlying Maastrichtian, as is characteristic of the earliest Danian in other boundary sections (e.g., Southern Ocean Site 690; see Pospichal and Wise, 1990a).

We cannot resolve at present, however, the discrepancy between the nannofossil and isotopic data on the one hand and the paleomagnetic results on the other. One can formulate several multiple working hypotheses, such as

1. The preliminary shipboard paleomagnetic measurements for Section 183-1138A-52R-3 provided a false reading.
2. Paleomagnetic Chron C29r is missing because of a hiatus; hence, the interval present where we place the boundary is actually Chron C30 or even C31n, as implied by Petrizzo (2001).
3. B. sparsus may actually range into the Maastrichtian, and therefore cannot be used to distinguish Cenozoic from Cretaceous strata.
4. The dark sediment color in Section 183-1138A-52R-3 may actually be unrelated to the K/T boundary event but rather related to the influx of clay minerals from tectonic activity and subaerial basement exposure at this time to the south at ODP Site 747 (Shipboard Scientific Party, 1989).

The first and most logical test to make relative to these hypotheses would be to confirm or deny the shipboard paleomagnetic data for Section 183-1138A-52R-3. This would require shore-based measurements on discrete samples from that section that have been subjected to more intense demagnetization than is done onboard ship. Such samples were not taken during the cruise because of routine curatorial restrictions placed on any sections suspected to have a complete K/T boundary. We note that shore-based measurements have considerably altered the shipboard magnetic reversal stratigraphy originally suggested for the subjacent core, 183-1138A-53R (Antretter et al., this volume). Until such measurements are made on Section 183-1138A-52R-3, however, we can only speculate about the various hypotheses given above.

In the lower Paleocene, the high-latitude Antarctic zonation scheme developed by Wei and Pospichal (1991) was employed to achieve better biostratigraphic resolution in the study than that offered by either the Okada and Bukry or Martini zonations. Nevertheless, some zones are not present, notably NA4, NA5, and NA6 at Site 1135 and NA5 at Site 1138. Zoning the Paleocene proved difficult because of the absence or rarity of zonal index species. In particular, D. nobilis, D. mohleri, and H. kleinpellii are rare or absent; therefore, zonal boundaries using these datums are marked by uncertainties.

The lower Paleocene samples from Sites 1135 and 1138 do not show a Hornibrookina acme at the Zone NA2/NA3 boundary as reported by Pospichal and Wise (1990a) from Site 690 (Maud Rise) (see also Pospichal, 1996a). The abundance and distribution of the genus Hornibrookina, which evolved in the early Paleocene, can potentially be used to indicate a change in surface water conditions between Site 1135 at 59°S and Site 690 at 65°S at ~64.5 Ma.

Hornibrookina edwardsii constitutes 45% of the assemblage at Site 690 (Pospichal and Wise, 1990a), indicating ideal living conditions for that taxon. The ocean conditions were similar at Site 738 (62°S) at the southern tip of the KP, where Hornibrookina is "abundant" (Wei and Pospichal, 1991). These ocean conditions change, however, going toward the lower latitudes at Site 1135 (59°S), where Hornibrookina does not make up a significant portion of the assemblage. The highest abundance reached is "common" in one sample. Conditions continued to deteriorate for Hornibrookina at Site 1138 (53°S), where it is only "rare" in the assemblage. Even farther north at Site 752, Broken Ridge (then still part of the KP) (Fig. F2), Pospichal (1991) noted only one specimen of Hornibrookina just below Zone NA4. The distribution and abundance of Hornibrookina indicate oceanographic changes between Sites 738 and Site 1135 (Fig. F14). We interpret the data to indicate a water mass boundary or oceanographic front between those sites. The difference in the water masses could be due to temperature or possibly the amount of available nutrients. Further work needs to be done to better define these relationships.

The Paleocene section is marked by unconformities and/or condensed sections (Figs. F9, F10, F11). These condensed sections could reflect in part the moderate core recovery at Sites 1135 and 1138 (average recovery = 62.6% and 58.5%, respectively), which is similar to previous ODP legs to the KP. During Legs 119 and 120, condensed Paleocene sections were also recovered at Sites 738, 747, 748, and 750 (Wei and Thierstein, 1991; Aubry, 1992) and the core recovery was also poor (average = <50%) (Barron, Larsen, et al., 1989). It is likely, however, that several marker species used to define zonal boundaries in the Paleocene are not present or are sporadic in the section. This has hampered Paleocene biostratigraphic resolution on the KP. Nevertheless, Sites 1135 and 1138 significantly increase the amount of biostratigraphic data available for the southern Indian Ocean through the Paleocene.

In the upper Paleocene Zone CP8 at Site 1135, the nannofossil data show an abundance switch from cool water-tolerant Chiasmolithus species to warm water-loving discoasters. This could indicate the presence of the late Paleocene Thermal Maximum in Core 183-1135A-25R. This possibility is currently being investigated using stable isotopic analysis, and those results will be reported elsewhere.

The Eocene section is marked by variable core recovery. Some cores have >100% core recovery, whereas adjacent cores recovered <5%. This is due to chert stringers blocking the core barrel opening. Still, recovery was improved over previous ODP legs to the KP. During Legs 119 and 120, recovery averaged 55% and 40%, respectively, in the Eocene section (Barron, Larsen, et al., 1989; Schlich, Wise, et al., 1989), whereas during Leg 183 recovery averaged ~60% (Shipboard Scientific Party, 2000a) through thicker sections that contained less chert.

Sites 1135, 1136, and 1138 all have an unconformity at the Paleocene/Eocene boundary. The first unambiguously Eocene nannofossil recorded at each site was T. orthostylus, whose FO indicates the boundary between Subzones CP9a and CP9b (53.6 Ma). Overgrowth of Tribrachiatus made identification at the species level difficult, but not impossible. No T. bramlettei was observed below the FO of T. orthostylus, thus indicating the unconformity at all three sites.

Generally, the lower Eocene to lower middle Eocene sections of Holes 1135A, 1136A, and 1138A have abundant and diverse assemblages of discoasters that include D. barbadiensis, D. lodoensis, D. kuepperi, and D. sublodoensis. The sphenolith group also reached its maximum diversity and abundance in the lower Eocene with S. moriformis, S. primus, S. radians, and S. editus. Nevertheless, for this report it was necessary to combine Zones CP10 and CP11 at all sites because of the sporadic nature of T. crassus (the FO of which marks the base of Zone CP11) and to combine Zones CP12 and CP13 at Site 1135 because o f the sporadic presence of N. fulgens (whose FO marks the boundary between Zones CP12 and CP13).

There is an apparent change in the middle Eocene, where discoasters become few or absent and discoaster diversity becomes very low. In the middle Eocene, sphenoliths are represented by only one species, S. moriformis. At the same time that discoasters and sphenoliths were declining in overall abundance, Chiasmolithus and Reticulofenestra began to dominate the assemblage. This was especially true for the small reticulofenestrids during the early middle Eocene. This trend continued into the late middle Eocene, with the development of R. onusta, R. samodurovii, and R. umbilica. These changes in the nannofossil assemblage reflect the cooling trend in the middle Eocene reported by many other authors (Shackleton and Kennett, 1975; Shackleton and Hall, 1984; Oberhänsli et al., 1984; Pospichal and Wise, 1990b; Barrera and Huber, 1991; Aubry, 1992; Wei et al., 1992; Wise et al., 1992).

One of the biostratigraphic problems encountered in the middle Eocene was the Subzone CP13c/CP14a boundary. The FO of R. umbilica is the datum used to define this boundary (Okada and Bukry, 1980). The Okada and Bukry zonation scheme was originally defined using low-latitude material. As drilling in higher latitudes progressed, it was noted that the Okada and Bukry zonal scheme, particularly the subzones, did not always readily apply. For instance, Applegate and Wise (1987) reported for offshore New Jersey that R. umbilica and C. gigas are present in the same samples and that their co-occurrence was not a result of contamination. Subzones CP13b and CP13c, therefore, could not be recognized in that region.

The present paper also reports R. umbilica and C. gigas in the same samples at Site 1138, but not at Sites 1136 and 1135 to the south. This could be due to the rarity of C. gigas at those sites. The co-occurrence of R. umbilica and C. gigas indicates that in higher latitudinal sections the FO of R. umbilica may be older than previously reported or that the LO of C. gigas is present later in the lower latitudes. For this reason the Zone CP13/CP14 boundary was approximated by the FCO of R. umbilica.

One other biostratigraphic question considered here is the extinction of C. solitus. In the lower latitudes, the LO of C. solitus is reported to occur below the FO of R. bisecta (Berggren et al., 1995). The LO of C. solitus at Site 1138, however, clearly occurs above the FO of R. bisecta. The same relationship was noted at Site 512 by Wise (1983) on the Falkland Plateau where R. bisecta was small, indicating the first evolutionary occurrence within a long section dominated by C. solitus. As that site was piston cored, there was no probability of contamination by caving of C. solitus.

Once again we have the problem of a marker species' range apparently being extended in the high latitudes. The alternative is that R. bisecta originated in the low/mid latitudes and subsequently migrated into the high latitudes after the appearance of C. solitus. This is considered less likely as both are cool-water taxa; hence, it is feasible that C. solitus might have persisted in the high latitudes after its disappearance at low latitudes.

The Leg 183 Shipboard Scientific Party placed the LO of C. solitus at 37.9 Ma for the high latitudes (see Wei and Wise, 1990) (Fig. F3), which is used in this paper. Berggren et al. (1995) assigned an age of 40.4 Ma to the LO C. solitus at low latitudes. This discrepancy in the age of the LO of C. solitus needs to be noted for the development of any chronology of events using the FO of C. solitus as a datum.

The upper Eocene was recovered on the KP only in Hole 1138A (Table T3). The overall assemblage reflects the trends of the late middle Eocene in that Reticulofenestra continues to dominate the assemblage along with cool-water Chiasmolithus species and C. pelagicus. The warm-water discoaster group declined in abundance to simple five- and six-rayed forms, now strongly overgrown. Sphenoliths continued to be represented by few to common S. moriformis, sometimes heavily overgrown.