A subsurface microbial environment in the igneous rocks of the ocean crust has been inferred from the microbiology of vent waters. Huber et al. (1990) reported a hyperthermophylic, anaerobic Archaea above an eruption of Macdonald Seamount in the south Pacific that must have been washed from the interior of the volcano. Haymon et al. (1993) and Juniper et al., (1995) described a floc of probable microbial origin on the East Pacific Rise and Juan de Fuca Ridge, respectively. On the Juan de Fuca Ridge, ~1000 kg of floc covered 2.1 × 105 m2 of the ridge axis. This floc was silica and iron in the shape of microbial filaments, coccoids, and rods (Juniper et al., 1995) and microbially produced filamentous sulfur (Taylor and Wirsen, 1997). The floc was attributed to a massive subsurface community that either was dislodged by the eruption or possibly bloomed transiently at the subsurface confluence of oxidized sea water and reduced hydrothermal fluids. The number of microorganisms that appear to inhabit oceanic basalts is quite low (Fisk et al., 1998) and is not sufficient to produce the "snowblower" vents that produce the microbial floc. However, the subsurface organisms may undergo prolific growth when volcanic intrusions crack the crust and increase permeability and at the same time supply heat to drive convective fluid flow, which provides chemical growth substrates and nutrients. Presently, we are not able to evaluate these alternatives because we do not know the relationship between free floating and endolithic organisms at Site 1026.
The particles in the fluid appear to be a mixture of contaminants and material that was present in the formation water. Some particles are clearly contaminants such as the steel rust. Other particles could be derived from the subsurface rocks or from contaminants, such as the barite, TiO2, floc, and clay. Contaminants could be derived from drilling mud, sediment from the upper part of the hole, and the WSTP tool. The origin of the high Cu on the filter and particles is not known, but we suspect that it is from the WSTP. Some of the particles were high in sulfur, which suggests that they could be microbial floc; however, these particles were rare. Rare filaments (Fig. 3C) could be the equivalent to the floc found at hydrothermal vents. Also particles with relatively high silica and iron (Table 3) could be microbial products. The matte found on the filter (Fig. 5) could also be a microbial product (polysaccharide).
Particles on the filters had sparsely distributed fluorescence from the dye Syto-59, which we attribute to microbes attached to the filters. Either these microbes could not be extracted from the filters, or they were present in such small numbers that the molecular techniques used to amplify them were not sufficient.
Our inability to extract and PCR amplify 16S rRNA genes from the Leg 168 particulate sample DNA extraction and PCR effort is not inconsistent with the existence of bacteria in submarine hydrothermal fluids. In earlier experiments using a fluorescent DNA stain, we detected approximately 105 microbial cells per filter in one Leg 168 sample. One round of PCR by our protocol is capable of detecting 103 plasmid-free E. coli cells, which implies that we should have detected amplification products from two rounds of amplification of extracts from the Leg 168 sample. However, it is likely that the potentially slow-growing microbes present in the Leg 168 samples contain both fewer 16S rRNA copies per genome and fewer chromosome complements than the fast-growing E. coli cells used to estimate our theoretical detection limit (Schmidt, 1997). Thus our detection limit for environmental microbes may be higher than our detection limit for E. coli. Although disappointing, it is therefore not too surprising that we were unable to detect 16S rRNA gene sequences from the small number of cells in the Leg 168 particulate sample.
Characterization of contamination of subsurface water samples is important for understanding the biology of the ocean crust. For this reason it is important that an aliquot of formation water be filtered and examined by electron microscope and electron microprobe so that the sources of contamination may be determined. In addition to this characterization of particles on filters, we have four recommendations for improving the water sampling technique. First, the filters that remove clay particles from the entry port of the WSTP should have an opening of at least 2 µm, and the filter should be kept free of shipboard contaminants with a shroud that is removed just prior to the lowering of the tool. Second, the water in the WSTP should be captured in stainless steel or titanium coils and the length of the coils increased so that larger samples can be collected. Third, a method of flushing the tool with sterile, 0.2-µm, filtered water should be devised, and the tool should be flushed immediately before deployment. And fourth, if cells are to be cultured from the water sample, the pressure case should be flushed with nitrogen as it is being closed so that organisms are not exposed to oxygen in the WSTP. Maintaining in situ pressure will also be important during subsampling water and culturing microbes if we wish to study barophilic microorganisms.