In-reservoir microbial removal of the C-25 methyl group from the extended 17 alpha, 21 beta(H)-hopanes (hopanes) generates 25-norhopanes in crude oils from the West Siberia and San Joaquin basins. This C-25 demethylation occurs preferentially among low molecular-weight hopanes (e.g. C-31), while higher homologs are progressively more resistant. Conversion of each hopane to its corresponding 25-norhopane occurred without significant side products and was incomplete al the time of sampling. This is indicated by the match between reconstructed C-31-C-35 hopane distributions for heavily biodegraded oils (sum of hopane parent and 25-norhopane product for each homolog) and those of related, nonbiodegraded oils. C-25 demethylation favors 22S epimers of the C-31 and C-32 hopanes compared to 22R, while the opposite applies to the C-34 and C-35 hopanes, because molecular shapes, dimensions, and volumes vary with stereochemistry at C-22. For example, geometry-optimized C-31 and C-32 22S hopanes from molecular mechanics force field calculations are more voluminous, while the C-34 and C-35 22S hopanes are less voluminous than their 22R counterparts. The C-31 to C-35 hopane 22S and 22R epimers show distinct ''scorpion-'' vs. ''rail-shaped'' conformations respectively, controlled by different 21-22-29-31 and 17-21-22-30 torsion angles. Because 22S epimers of the extended hopanes tend to favor the scorpion conformation, which folds the side chain back toward position C-25, longer side chains may increasingly hinder C-25 from enzymatic attack. Biodegradation can adversely affect the use of %22S/ (22S + 22R) ratios for hopanes to assess thermal maturity. The 25-norhopane ratio improves our ability to distinguish different levels of biodegradation among heavily degraded oils where C-25 demethylation has occurred. Copyright (C) 1996 Elsevier Science Ltd