Bacteria were present in all the samples from Site 1026 (Fig. 2). Total bacterial numbers followed the same general depth trend of other sites (Cragg et al., 1990, 1992, 1996, 1997, 1998; Cragg and Parkes, 1994; Cragg and Kemp, 1995; Cragg et al., 1995a; Cragg et al., 1995b). Bacterial numbers were initially high at the surface (1.47 × 108 cells cm-3), but decreased rapidly to 9.93 × 106 cells cm-3 by 8 mbsf. Below this depth, bacterial numbers remained essentially constant and at ~70 mbsf 1.07 × 107 cells cm-3 were present. Dividing cells were also present in all samples, and their depth distributions generally paralleled that of the total count. Numbers of dividing cells, however, were on average only 12% of the total count. Highest numbers were at the near surface (5.19 × 107 cells cm-3) and decreased to 1.47 × 106 cells cm-3 at ~70 mbsf, the deepest sample.
Bacterial populations were present in all samples; however, at 374, 413, 470, and 509 mbsf, counts did not significantly exceed blanks (Fig. 3). Total bacterial numbers were high at the surface (1.67 × 108 cells cm-3) and must be adapted to the prevailing low temperatures ~2°C (psychrophilic or psychrotolerant). Direct counts subsequently decreased with depth, with some very low concentrations below ~374 mbsf and, in the deepest sample at 565.1 mbsf, counts were 2.16 × 106 cells cm-3. Dividing cells were present above the detection limit in 21 of the 33 samples. The highest value was at the near surface at 1.97 × 107 cells cm-3, constituting ~12% of the total count. The depth distribution of dividing cells generally paralleled the total count. Below 275 mbsf, dividing cells were only significant in three samples at 316, 354, and 429 mbsf, although below 500 mbsf, numbers of dividing cells did increase, coinciding with an increase in the total bacterial population.
Two regions of Site 1027 were found to have consistently elevated bacterial numbers, both statistically (P < 0.05, analysis of variance) and in relation to the general bacterial depth trend at other ODP sites. The first of these was between ~166 and ~273 mbsf with total bacterial numbers maximizing at 1.31 × 106 cells cm-3 (223.84 mbsf). This region of elevated bacterial numbers was associated with a subsurface peak in pore-water ammonia (Fig. 4B) and alkalinity (data not shown, but see Davis, Fisher, Firth et al., 1997). Ammonia and alkalinity formation is usually associated with bacterial degradation of organic matter (Claypool and Kaplan, 1974), and their continued production within the zone confirms the presence of active bacterial populations. The reason for this elevation in bacterial numbers is unclear because although organic carbon concentrations are higher in this zone than at ~100 mbsf, they are similar to those in the remainder of the hole. The second region with elevated bacterial numbers occurs at the bottom of the hole below ~527 mbsf. Bacterial populations in the three samples below this depth (527.84, 547.14, and 565.1 mbsf) are significantly larger (P < 0.05, analysis of variance) than in the zone immediately above (374.24-470.24 mbsf, eight samples) which are all very near or below (at 489.44 and 508 mbsf) the detection limit. The presence and increase in dividing cells, below ~520 mbsf, albeit to just at the detection limit, is coincident with the increased total count and together demonstrates that bacteria are active in this region (Fig. 3). During exponential bacterial growth the numbers of dividing cells increase and, hence, the presence of dividing cells should indicate growing and thus active cells. In water samples the percentage of dividing cells provide a reasonable estimate of bacterial growth rates (Newell and Christian, 1981) but for sediments it greatly overestimates growth (Fallon et al., 1983). Despite this, however, dividing cells in sediments do still indicate the presence of growing and thus active cells, particularly when the numbers of dividing cells significantly increase, as occurs below ~520 mbsf at this site, because stimulation of growth in deep sediment bacteria also results in an increase in the number of dividing cells both in the laboratory (Getliff et al., 1992) and in situ (Wellsbury et al., in press), and numbers of dividing cells correlate with an independent measure of growth, [3H]-thymidine incorporation into bacteria DNA (Wellsbury et al., 1996).
The lower than expected bacterial numbers between ~374 and ~509 mbsf is difficult to explain because the total organic carbon concentration was essentially constant (~0.5 wt%) throughout the lower 300 mbsf of the site (Fig. 4). Total bacterial populations at Site 858 on Juan de Fuca Ridge (Leg 139, Cragg and Parkes, 1994) decreased markedly with depth at temperatures around 30°-40°C, and the temperature at 374 mbsf is within this temperature range. Possibly the combined stress of low organic carbon concentration, lack of sulfate as an electron acceptor (Fig. 4), and elevated temperature inhibits a proportion of the total bacterial population, resulting in "low" bacterial numbers in this region. The number of dividing cells as a percentage of the total population in this region, however, is surprisingly high at ~57% compared to the mean at all other depths (~23%), which indicates that a subset of the total population may actually be more adapted to these mesophilic temperatures (growth optimum 20°-45°C).
The increase in bacterial numbers below ~527 mbsf is associated with increasing pore-water sulfate concentrations from the underlying bedrock (Fig. 4; Davis, Fisher, Firth, et al., 1997). This suggests that deep bacterial populations are stimulated by the presence of a more efficient electron acceptor, sulfate, at thermophilic temperatures (~55°-60°C; Fig. 4). The original sea-water sulfate was removed below 200 mbsf during the first elevated zone of bacterial activity, and subsequent bacterial numbers were lower than expected. Coinciding with this deep elevation of bacterial populations was a marked decrease in the CH4 gas concentration (Fig. 4; Davis, Fisher, Firth, et al., 1997), suggesting that the sulfate enabled CH4 to be used as a deep energy source. Although sulfate-reducing bacteria able to grow on methane have not been isolated, there is strong environmental and laboratory evidence for their involvement in anaerobic methane oxidation (e.g., Kosiur and Watford, 1979; Iversen and Jørgensen, 1985; Jørgensen et al., 1990; Hoehler et al., 1994). Stimulation of deep bacterial populations has been observed at a number of previous ODP sites (Cragg et al., 1990; 1992; 1996; Wellsbury et al., 1997), and these are often associated with increases in deep sulfate and methane oxidation.