Plots of the first and second derivative of the oxygen profiles, which give hints on the oxygen flow and consumption rate indicated that most of the oxygen consumption took place in the upper millimeters of the sediment. Even where oxygen penetrated deeper than 5 cm, more than 90 % of the respiration acitivity was located in the upper 5 mm.

The oxygen measurements were complicated by two facts. Firstly, there was bioturbation caused by macrofauna at several sites. This caused an increase of oxygen penetration around the channels formed (Fig. 2). Secondly, there was some scatter in the oxygen electrode signals. No noise was observed during measurements in the water column. In the upper sediment layers, however, the signal became noisy after about one minute. A possible explanation for this phenomenon is that the alkali produced at the electrode tip (O₂ + 4 e^- + 4 H⁺ ® 2 H₂O) caused some precipitations at the electrode surface. These were wiped off when the electrode was moved

Fig.1 Oxygen profiles at stations #79 and #82 (A). (B) Oxygen profile and first and second derivatives at station #85.

Fig. 2 Core taken at station #78 showing macrofaunal bioturbation (A). (B) Oxygen penetration into the sediment within a borrow hole, and 1.5 cm off.

deeper to the next level. At low oxygen concentrations (and corresponding low alkali production by the electrode) the signal stayed stable.

The pH values measured in the water above the multicorer cores were 8.2 to 9.1, while in the upper sediment layers they were 0.2 to 1.2 units lower (Tab. 1). Two hours after sampling and after having measured the oxygen profile, the pH in the water was usually decreased by about 0.5 pH units, probably due to release of buffering compounds from the sediment.

Free sulfide was not detected in any of the multicorer cores except those from the Urania basin. Thus, there was no oxygen-sulfide interface in the other sediments.

Total numbers of bacterial cells reached values of about 10⁸ per cm³ in all sapropel layers of the gravity core #69-2 (Fig. 3A). With the exception of the sediment surface, bacterial counts in layers of low C_org were much lower. No significant differences in the frequency of dividing cells was observed, however. A characteristic pattern of microbial activity emerged when the glucose degradation rates and ectoenzyme activities were determined. The highest rate of glucose mineralization was detected at the sediment surface. Below, significant microbial activity was detected in all 4 sapropel layers present, while the values in low-C_org intermediate layers were at or below the detection limit (Fig. 3B).

Fig. 3 (A) Total numbers of bacterial cells (bars), frequency of dividing cells (circles) and (B) glucose degradation rates in sapropels and intermediate layers of gravity core #69-2.

The vertical differences in protease activity (Fig. 4C) were very similar to those observed for glucose degradation. Phosphatase activity (Fig. 4A) reached maximum values in all sapropel layers. By comparison, the activity of ß-glucosidase (Fig. 4B) was one to two orders of magnitude lower and did not show the distinct vertical pattern of the other two ectoenzymes.

Our experimental results clearly indicate that the sapropel layers harbor a microbial community which, despite of the high age of the surrounding sediment matrix, still exhibits measurable and significant physiological activity. The fact that rates were significantly lower in the sediment layers between the sapropels indicates that a contamination of the older layers by sediment from above did not occur during sample preparation. Because of the distinct maxima in bacterial cell numbers and physiological activity within the sapropels, a supply of carbon substrates solely by pore water transport appears unlikely. Rather, sapropels themselves must provide favorable conditions for the growth and survival of chemotrophic bacteria.

Fig. 4. (A) Ectoenzyme activity of alkaline phosphatase, (B) ß-glucosidase, and (C) protease in sapropels and intermediate layers of gravity core #69-2.

At station #76 a CTD profile was recorded down to 3520 m (Fig. 5A). Water samples above, in, and below the chemocline were obtained (Fig. 5B). Furthermore, sediment samples were taken from a multicorer.

Fig. 5 (A) CTD profile at 3460 to 3480 m at the urania basin. (B) pH and sulfide concentrations in water samples (3400 to 3520 m).

A sharp chemocline had already been detected at a water depth of 3470 m (MEDRIFF consortium 1995). The temperature increased by more than 2 °C within 2 m. Concomitantly, the oxygen concentration dropped to zero, and the fluorescence signal increased by a factor of ten. The conductivity record stopped at 58 mS/cm due to range overflow. In the water samples, a pH shift from 8.5 to 6.7 was found. Furthermore, free H₂S was detected at concentrations of up to 24 mM. If the electrode data are verified by chemical analysis at home, our values would represent the highest free sulfide concentration ever found in seawater (Luther et al. 1990; Henneke et al. 1991). In the sediment only 3 mM free sulfide was detected.

Within the free water column of the Urania deep, maximum activities of all three ectoenzymes were detected in the chemocline (Fig. 6), indicating a zone of increased bacterial activity at the oxic/anoxic interface. Activities of alkaline phosphatase and protease were lower in the surface sediment layer at this site compared to surface sediments in core #69-2. Similarly, the degradation rate of glucose was lower by one order of magnitude (0.362 nM Glucose h^-1). In contrast, ß-glucosidase activity in Urania basin sediment was increased by a factor of 1000.

Fig. 6. Ectoenzyme activity of alkaline phosphatase, protease and ß-glucosidase in the free water column and surface sediment of the Urania basin. The chemocline was positioned at 3470 m depth. Note the logarithmic scale of enzyme activity values.

Our measurements on bord have revealed elevated bacterial cell numbers and microbial activity in up to 124,000 year old sapropel layers of the Mediterranean. In our forthcoming work we are going to analyze the bacterial communities by microbiological, molecular biological and geochemical approaches. We have the opportunities to compare sapropel layers of different ages with layers of low organic content. The anoxic water body in the unique Urania basin shows conditions that might be compared with those during development of the sapropels in some aspects. Furthermore, the sharp chemocline with extreme sulfide concentrations will probably harbour a community of sulfide-oxidizing bacteria. We are going to analyze the presence of those bacteria in the water and in the sediment, where the conditions do not allow growth by sulfide oxidation with O₂. This might allow to differentiate between inactive bacteria from the water body and populations growing in the sediment.

• The molecular biological studies will include aseptical extraction of the total DNA, followed by PCR amplification of 16S rDNA sequences and denaturing gradient gel electrophoresis in order to to analyze the composition and diversity of the bacterial communities. PCR methods for the selective amplification of specific groups of bacteria will also be used. The free DNA extracted electrophoretically from a series of sediment samples will be used to differentiate between living bacteria and fossil residues. A similar approach will be used for the phylogenetic characterization of microbial communities present in the chemocline of the Urania basin, using the samples of bacterial cells concentrated on membrane filters.

• The microbiological investigations will be based on the isolation in pure culture of the bacteria present in the highest MPN dilutions and in the various enrichment cultures. The strains obtained will be compared phylogenetically with the composition of the whole bacterial community. Furthermore, a physiological characterization of bacterial strains will provide more insight into the culturability of deep-sea bacteria from ancient sediment layers. The 16S rDNA sequences of the isolates will be determined and compared with the sequences present in the total and free DNA extracts.

• The geochemical characterization will include a chemical analysis of inorganic sulfur and organic carbon compounds and an analysis of the stable isotopic compositions. This will hopefully more precisely determine the role of bacteria in the transformation of these compounds within the sediment layers. Our set of pore water samples will be used to determine low molecular weight compounds (carbohydrates, amino acids) present using high sensitivity HPLC techniques (PAD and OPA). The sulfur compounds to be analyzed chemically and isotopically include sulfate, FeS, FeS₂, S⁰, H₂S. Furthermore, bacterial membrane lipids and diagenetically altered lipids will be extracted and identified.