We first constructed full stoichiometry, including cell synthesis, for glucose mixed-acid fermentation at different initial substrate concentrations (0.8-6 g-glucose/L) and pH conditions (final pH 4.0-8.6), based on experimentally determined electron-equivalent balances. The fermentative bioH(2) reactions had good electron closure (-9.8 to +12.7% for variations in glucose concentration and -3 to +2% for variations in pH), and C, H, and O errors were below 1%. From the stoichiometry, we computed the ATP yield based on known fermentation pathways. Glucose-variation tests (final pH 4.2-5.1) gave a consistent fermentation pattern of acetate + butyrate + large H-2 while pH significantly shifted the catabolic pattern: acetate + butyrate + large H-2 at final pH 4.0, acetate + ethanol + modest H-2 at final pH 6.8, and acetate + lactate + trivial H-2 at final pH 8.6. When lactate or propionate was a dominant soluble end product, the H-2 yield was very low, which is in agreement with the theory that reduced ferredoxin (Fd(red)) formation is required for proton reduction to H-2. Also consistent with this hypothesis is that high H-2 production correlated with a high ratio of butyrate to acetate. Biomass was not a dominant sink for electron equivalents in H-2 formation, but became significant (12%) for the lowest glucose concentration (i.e., the most oligotrophic condition). The fermenting bacteria conserved energy similarly at similar to 3 mol ATP/mol glucose (except 0.8 g-glucose/L, which had similar to 3.5 mol ATP/mol glucose) over a wide range of H-2 production. The observed biomass yield did not correlate with ATP conservation; low observed biomass yields probably were caused by accelerated rates of decay or production of soluble microbial products.