Isotopically anomalous CH 4-rich gas deposits are found in mining sites on both the Canadian and Fennoscandian shields. With δ 13C CH4 values from -22.4 to -48.5% and δD CH4 values from -133 to -372%, these methane deposits cannot be accounted for by conventional processes of bacterial or thermogenic methanogenesis. Compositionally the gases are similar to other CH 4-rich gas occurrences found in Canadian and Fennoscandian shield rocks ( SHERWOOD LOLLAR et al., 1993). However, the isotopically anomalous gases of this study are characterized by unexpectedly high concentrations of H 2 gas, ranging from several volume percent up to 30 vol%. The H 2 gases are consistently depleted in the heavy isotope, with δD H 2 values of -619 to -659‰ 3H/ 4He ratios in the range of 0.4 × 10 -8 to 4.4 sx 10 -8 indicate that there is no resolvable component of mantle-derived He in these deposits. Based on these results a mantle-derived source for the C-bearing gases is unlikely. Several lines of evidence support an alternative abiogenic origin for the gases from Sudbury, Canada, and Juuka and Pori, Finland. The D-depleted H 2 gas and calculated ∆D (CH 4-H 2) equilibration temperatures of 110-170°C at all three sites are in good agreement with results obtained for abiogenically produced CH 4 and H 2 in ophiolite sequences in Oman and the Philippines. Serpentinization and the hydration of ultramafic rocks are the proposed mechanisms for CH 4 and H 2 production in these ophiolite sequences. The widespread occurrence of serpentinized and altered ultramafic rocks at Juuka and at a number of other sites in both Canada and Finland implies that similar mechanisms may be involved in gas production at at least three sites on the shields. The origin of the gases at the remaining shield sites is discussed. Alternative hypothesis include (1) production of the rest of the shield gases by mixing between abiogenic endmembers and bacterially generated hydrocarbon gas such as identified elsewhere on the Canadian and Fennoscandian shields ( SHERWOOD LOLLAR et al., 1993); and (2) production of a series of isotopically distinct abiogenic CH 4 endmembers at each site due to variability in the isotopic composition of available carbon sources. We cannot conclusively distinguish between the alternative scenarios based on existing data. However, the evidence for an abiogenic CH 4 endmember at at least three sites emphasizes the need for substantial revision of current theories of methanogenesis. Abiogenic processes of CH 4 production may be considerably more widespread than previously anticipated.