A probe for the measurement of dissolved H-2 in anoxic methanogenic environments was developed. The probe was based on the diffusion of dissolved H-2 through a silicone membrane into a gas space at the end of the probe. This gas space was flushed with N2 and analyzed gaschromatographically for H-2. The probe was calibrated in aqueous solutions of known H-2 concentrations. The detection limit was about 7 nM H-2, the depth resolution of vertical profiles was about 5-10 mm. The probe could also be used to measure dissolved CH4 and CO. The probe allowed the measurement of in situ concentrations of dissolved H-2 that were unbiased by eventual effects of an extraction procedure on the steady state between production and consumption of H-2. However, it could not be implanted into the environment for periods longer than hours because of dramatically increased H-2 signals due to microbial growth on the silicon membrane. The probe was used to measure vertical profiles of dissolved H-2 and CH4 in cores of the deep sediment of Lake Constance and of anoxic paddy soil. The vertical CH4 profiles measured with the probe compared well with those measured by extraction. In the lake sediment, concentrations of H-2 increased with depth and reached values in the range 10-60 nM H-2 within the methanogenic zone, below about 2-5 cm depth. In the anoxic paddy soil, H-2 concentrations also increased with depth reaching values of 40-100 nM H-2. Dissolved CO was about 5-20 nM. The turnover of H-2 was very dynamic as revealed when starch powder was added to the flooding water. Within 1 day, H-2 concentrations increased to 80 muM H-2 in the upper 2 cm of the soil core, but after 5-7 days decreased again to the lower values measured previously.