At the request of shore-based participants, we sampled interstitial water at Site 1098 with unusually high resolution, squeezing a total of 30 whole-round core samples, one per core section, from Hole 1098C (Table T29). The resulting profiles of interstitial water chemistry (Fig. F24) illustrate a sharply stratified diagenetic environment, influenced at least partly by variations in sediment lithology. Chloride concentrations remain constant through the upper 20 mbsf and decrease slightly by 2% from 20 to 40 mbsf, then remain constant at greater depths. Any stronger trends seen in profiles of other dissolved constituents should reflect chemical reaction processes.

Sulfate concentrations decrease linearly from a seawater value (28 mM) to zero in the interval from 20 to 30 mbsf (Fig. F24), and this gradient delineates the sulfate reduction zone at Site 1098. In addition, acidified aliquots of interstitial water from between 13 and 25 mbsf emitted a noticeable odor of hydrogen sulfide, a direct product of sulfate reduction. The presence of hydrogen sulfide within this depth range indicates an absence of dissolved oxygen and nitrate, as well as a lack of enough reduced iron and manganese to precipitate all of the reduced sulfur as sulfide minerals. Iron and manganese concentrations lie close to detection limits (2 µM) in the upper 30 mbsf (Fig. F24), except for a few random samples with higher iron concentrations (4-17 µM). Curiously, iron and manganese increase to maximum values (10 µM and 7 µM, respectively) below the sulfate reduction zone at ~36 mbsf, then decrease at greater depths. Normally during early diagenesis, the highest concentrations of dissolved iron and manganese occur near the top of the sulfate reduction zone (Froelich et al., 1979), as observed at Sites 1095, 1096, and 1101. The highly variable iron concentrations measured at Site 1098 could represent an artifact caused by partial oxidation of the sediment before squeezing, incomplete filtration of colloidal particles during and after squeezing, or contamination.

Organic-matter decay has surprisingly little effect on the interstitial water chemistry above the sulfate reduction zone at Site 1098, despite the relatively high organic carbon content (1.0-1.2 wt%; see "Organic Geochemistry") and rapid sedimentation rate (300 cm/k.y.; Leventer et al., 1996). For example, direct byproducts of organic-matter decay, such as alkalinity, ammonium, and phosphate, increase only slightly with depth in the upper 20 mbsf (Fig. F24) and yield weak concentration gradients similar to those observed in the less organic-rich, much more slowly deposited sediment at the rise sites. Within the sulfate reduction zone, however, alkalinity, ammonium, and phosphate increase sharply and approach maximum concentrations (40 mM, 6 mM, and 175 µM, respectively) at slightly greater depths, between 30 and 45 mbsf. Below 40 mbsf, phosphate decreases sharply to a low concentration (28 µM) in the deepest sample. Fluoride decreases slightly with depth in the upper 20 mbsf (to 54 µM), then much more sharply with depth to a minimum concentration (10 µM) at 27 mbsf (Fig. F24), slightly above the base of the sulfate reduction zone and directly within a 3-m interval identified as a turbidite (see "Lithostratigraphy"). As noted at other sites in this region, substantial uptake of fluoride represents a sign of authigenic apatite precipitation (Jahnke et al., 1983; Schuffert et al., 1994). A qualitative comparison of the shapes of the alkalinity, ammonium, and phosphate profiles suggests that phosphate uptake may also occur in the turbidite layer, in support of apatite authigenesis.

We do not entirely understand the apparently low intensity of organic-matter decay in the upper 20 m of the sediment column at Site 1098. Radiocarbon ages from a nearby gravity core indicate that these relatively organic-rich deposits have accumulated rapidly during the last ~6 k.y. (Leventer et al., 1996), thus organic diagenesis by now should have established strong chemical gradients within the interstitial water. The occurrence of well-laminated intervals between 8 and 23 mbsf and the presence of H₂S below 13 mbsf preclude the rather unlikely possibility that extremely vigorous bioturbation or other mixing has maintained a well-irrigated environment to such depths. We note that many of the laminae in the upper 20 mbsf consist of a nearly monospecific assemblage of the resting spores of Chaetoceros diatoms (see "Lithostratigraphy"). These tightly encased siliceous spores protect the internal biomass, perhaps rendering it unavailable for bacterial consumption until dissolution breaches the external casing. Other more readily degradable organic matter must exist, however, in the upper 20 mbsf. Also, this scenario conflicts with the occurrence of a shallow sulfate reduction zone in nearby Site 1099 (see "Organic-Matter Degradation"), where the sediment also contains an abundance of Chaetoceros spores. Finally, we wonder whether the position of the aforementioned turbidite layer precisely within the zone of steepest chemical gradients represents more than a coincidence.

Other inorganic processes, such as dissolution of biogenic silica and precipitation of authigenic carbonate phases, influence the chemical composition of interstitial water at Site 1098 (Fig. F24). Dissolved silica reaches a high concentration (>0.8 mM) at only 1.5 mbsf, increases slightly (to 0.9 mM) with depth in the upper 42 mbsf, then decreases sharply to a low concentration (0.4 µM) in the deepest sample, at 45 mbsf. The silica profile reflects a state of near saturation with respect to biogenic opal in the predominantly diatomaceous ooze that lies above 42 mbsf in Hole 1098C (Kastner et al., 1977) and undersaturation in the calcareous mud encountered below 42 mbsf (see "Lithostratigraphy").

Dissolved calcium and magnesium concentrations do not change significantly from their seawater values above 20 mbsf, but they decrease sharply by equivalent amounts (~10 mM) between 20 and 30 mbsf and remain essentially constant at greater depths. The coincident uptake of calcium and magnesium in a 1:1 ratio at the base of the sulfate reduction zone strongly suggests the possibility of dolomite formation (Baker and Kastner, 1981), presumably promoted by the high alkalinity produced in the sulfate reduction zone. Although examination of smear slides confirms the presence of authigenic carbonate mineral grains in this depth range and below, we did not find any dolomite. A negligible amount of calcium may also precipitate as authigenic apatite, as discussed above. Dissolved strontium remains constant at its seawater concentration (90 µM) in the upper 20 mbsf at Site 1098, decreases slightly with depth to a minimum concentration (72 µM) at 27 mbsf, and then increases sharply to a high concentration (240 µM) in the deepest sample. Either the calcareous mud encountered at the bottom of the hole or some underlying lithology must act as a strong source of strontium to the interstitial water. Potassium remains essentially constant at its seawater concentration throughout the hole and thus does not participate extensively in clay mineral reactions within the shallow sediment section recovered at this site.

We squeezed 12 whole-round core samples for interstitial water at Site 1099 (Table T29). One sample was taken from each core in Holes 1099A and 1099B. Chloride concentrations remain constant with depth (Fig. F25) and indicate an unmodified seawater source for the interstitial water throughout the cored interval.

The interstitial water chemistry profiles at Site 1099 (Fig. F25) reflect conditions of rapid burial and intense decay of organic matter. Dissolved sulfate decreases to <50% of its normal seawater value (28 mM) in the upper 3 mbsf and approaches zero by 10 mbsf, and detectable concentrations of methane are present in all samples from below 3 mbsf (Table T27). The sulfate reduction zone thus extends from near the seafloor to <10 mbsf, much shallower than at Site 1098. Dissolved manganese concentrations remain low (<7 µM) in all samples and show no consistent trend with depth. Dissolved iron concentrations remain near zero in the upper 10 mbsf but increase significantly at greater depths and reach a maximum (30 µM) near 50 mbsf. Unless these iron data represent an artifact of contamination or sample processing, the presence of soluble ferrous iron below the sulfate reduction zone implies that either a ferric iron phase persists to such depths before it reduces or iron sulfide minerals have begun to dissolve. Considering that the sediment at Site 1099 commonly contains as much as 5% pyrite, particularly below 50 mbsf (see "Site 1098 Smear Slides," and "Site 1099 Smear Slides"), significant redistribution of iron must occur after deposition.

Other direct byproducts of organic-matter decay, such as alkalinity, ammonium, and phosphate, reach extremely high concentrations at Site 1099, even in the upper 3 mbsf (Fig. F25). Alkalinity and ammonium increase steadily with depth to maximum concentrations (80 mM and 13 mM, respectively) at 90 mbsf, whereas phosphate reaches a maximum concentration (260 µM) at only 20 mbsf, then decreases with depth to lower concentrations (80-100 µM) below 90 mbsf. Furthermore, the high phosphate concentrations could promote precipitation of authigenic apatite (Jahnke et al., 1983; Schuffert et al., 1994) and account for the low fluoride concentrations (<30 µM) observed in the upper 20 mbsf. Overall, these results illustrate an order of magnitude increase in the effects of organic-matter decay at Site 1099 compared to Sites 1095, 1096, and 1101 on the continental rise. We attribute this large difference to the higher organic carbon concentrations (0.4-1.2 wt%; Table T28) and much faster sedimentation rate in Palmer Deep. As discussed above, however, the upper portion of the sediment column at Site 1098, located only 6 km away, seems anomalous within this context.

Inorganic processes, such as dissolution of biogenic silica, precipitation of authigenic carbonate phases, and perhaps diagenesis of clay minerals, may all influence the chemical composition of interstitial water at Site 1099 (Fig. F25). Dissolved silica remains essentially constant (0.9 to 1.0 mM) and nearly saturated with respect to opal-A (Kastner et al., 1977) over the entire depth range, as expected in these diatomaceous sediments (see "Lithostratigraphy"). Calcium concentrations decrease sharply in the upper 10 mbsf and remain constant at greater depths, most probably because of authigenic calcite precipitation induced by high alkalinity.

The magnesium and potassium profiles show remarkably similar trends with depth, first increasing by ~20% between 3 and 80 mbsf, then decreasing by a few percent in the lower 20 m of Hole 1099B. This represents the only one of our sites where either of these constituents increases above its seawater concentration. Although these changes could relate to diagenetic reactions in the sediment or interaction with basement rocks, this seems unlikely because the Mg/K value does not vary significantly downcore. Chloride concentrations also do not vary significantly downcore; thus, we exclude the possibility that these data reflect mixing with more-saline water, trapped in the restricted basin at Palmer Deep. Alternatively, the magnesium and potassium data could simply reflect a drift in the performance of the ion chromatograph used in their measurement. Strontium remains within ~10% of its seawater concentration (90 µM) but decreases slightly (to 80 µM) below 80 mbsf.