The PFT data show that the sediment samples from the center of APC cores contained generally <0.1 µL seawater/g sediment (Table T2), which corresponds to an entrainment of <50 prokaryotes/g sediment. This is a maximum estimate that assumes that prokaryotes can follow the flow of seawater in the same way as a PFT tracer molecule. In contrast to microspheres, PFT can travel through very small pore spaces and is found in the laboratory air and on the hands of anyone who has handled a core liner. Therefore, although its presence at high concentrations in a sediment sample (>0.1 ng PFT/g sediment) indicates seawater contamination, it is not a sufficient indication that microorganisms from the drilling fluid have contaminated the sample.
Figure F3 shows the number of microspheres detected plotted against the potential seawater contamination as determined from PFT concentrations for the 34 samples for which PFT concentrations and bead counts were done on the same sediment sample. These 34 samples were chosen because they represent a wide range of measured PFT concentrations, permitting the potential seawater contamination as determined by PFT concentrations to be related to the estimations of particulate contamination as determined using the bead deployment. In several cases, samples thought to be contaminated with seawater based on PFT concentrations contained no microspheres (Fig. F3, data points along the y-axis). On the other hand, the absence of PFT from a sample is a good indication that contamination by drilling water and also by prokaryotes has not occurred. Of the four samples that remained below the detection limit of 0.01 µL seawater/g sediment (0.01 ng PFT/g sediment), three contained no detectable beads (Fig. F3).
In general, the concentrations of the easily diffusible PFT tracer, indicating the extent of seawater contamination and influx of dissolved compounds, are somewhat correlated with microscopic counts of visible microspheres. The number of fluorescent microspheres (2 x 1011/20-mL bag) that are deployed is equivalent to the number of prokaryotes in ~400 L of seawater (assuming 5 x 108 bacteria/L); however, their deployment does not produce a uniform dispersion in the drilling water, but could result in microsphere concentrations at the drill bit that temporarily exceed natural seawater concentrations of prokaryotes by an unknown factor. In Figure F3, with the exception of one point, the samples taken from the periphery of APC cores where drilling water is expected and observed (shown in red) show a positive correlation, demonstrating that the beads end up at a concentration of ~1000 beads/mL seawater after being deployed. This represents a large dilution for the concentration of beads found in the initial deployment bag but is similar to the ratio of 50 prokaryotic cells/0.1 µL seawater predicted for the drilling fluid. This indicates that the microsphere bag deployment can ultimately produce a concentration of microspheres in the fluid similar to actual prokaryotic concentrations, suggesting that the microsphere test is a sensitive measure of prokaryote contamination in sediment samples. In one periphery sample with a very high PFT concentration, however, no beads were detected (Section 201-1227A-7H-2 [outside]) (Table T5). This result inspired the change in the method of bead deployment, which we believe resulted in a more reliable and more uniform bead deployment.
Figure F3 also shows a difference between the relationship of PFT results and bead results between center subcores of APC cores and of XCB cores. Whereas subcores from XCB cores show uniformly high PFT values and are highly variable in their bead concentrations, subcores from the center of APC cores all have bead counts <85 beads/g sediment in spite of variable PFT values. Although 85 beads/g may represent contamination of as many as ~85 microbes/g sediment, in some cases these low numbers are unreliable because they are based on the observation of only a few beads (in a laboratory environment with background contamination).
In summary, although PFT can always be found in sediment samples taken from the edges of cores, microspheres can be absent, indicating an inhomogeneous distribution of microspheres along the core liner. At this point, without knowing the factors that control the final concentration and distribution of microspheres in the drilling water, one should consider the beads as only a semiquantitative measure of contamination. The presence of multiple microspheres is a strong indication that contamination by microbe-sized particles has occurred; without further data, their absence alone cannot confirm that a sample is uncontaminated.
For most biogeochemical process rates, organic biomarker studies, and microscopic cell counts, the potential presence of small numbers of contaminating prokaryotes (<50 prokaryotes/g sediment in APC cores) (Table T2; Fig. F3) as indicated by PFT and microsphere data is not a problem. Low-level contaminants are highly relevant if there is a method-inherent risk that contaminating populations are amplified and selected for by specific enrichment in culture media or by nucleic acid techniques such as the polymerase chain reaction (PCR). However, even enrichment cultures may not be affected by low levels of contaminating cells (<50 prokaryotes/g) because most surface marine microorganisms (from which the contaminating cells are derived) are currently unculturable and a portion of all cells do not survive the shock associated with being brought into culture media.
The potential seawater concentrations for APC sediment cores show considerable variability, with standard deviations frequently exceeding the averages (Table T2). The spottiness of contamination, as determined by PFT and microsphere data, does not allow guarantees that a specific sample material, cored and treated correctly, will be free of contaminants. Although it may not always be possible, it is best that a sample used for microbial cultivations and slurry preparation be checked for contamination directly, preferably by both chemical tracer and microsphere techniques.
The identification of microbial isolates or phylotypes from potentially contaminated slurries can help to decide whether contamination has occurred, if the isolates are identical to frequently cultured seawater bacteria. Under some drilling and sampling conditions, contaminations are harder to exclude; for example, XCB coring could be unavoidable in a particular location. To double check cultivated isolates and molecular phylotypes under such problematic conditions, potential seawater contaminants could be screened for comparison from supernatant water of the mudline core or from the outer core liner.