A mixed culture chemostat system was used to distinguish between the effects of carbohydrate availability per se and the low pH generated from carbohydrate metabolism on the proportions of bacteria within microbial communities. Nine oral bacteria were grown at pH 7 and pulsed with glucose on ten consecutive days. In one chemostat, the pH was maintained automatically at 7 throughout the experimental period, while in the other, pH control was discontinued for six hours after each pulse. Glucose pulses at neutral pH had little effect on the composition of the microflora. Only the proportions of A. viscosus and V. dispar increased; L. casei and S. mutans remained at low levels (0.2% and 1.0%, respectively). Acetate and propionate were the predominant end-products of metabolism; lactate levels were low. In contrast, when pH was allowed to fall after each glucose pulse, the composition of the microflora altered dramatically. The amounts of L. casei and S. mutans increased both as a proportion of the total count and in absolute numbers, as did V. dispar, whereas the amounts of the other Gram-negative organisms (B. intermedius, F. nucleatum, and N. subflava) and S. sanguis were considerably reduced. Lactate formed a major portion of the metabolic end-products. Successive glucose pulses resulted in both amplified changes in the microflora and a steadily greater rate and final extent of acid production. This is in agreement with the reported shifts in the oral microflora in vivo in response to frequent carbohydrate intake. Analysis of the data strongly suggests that the pH generated from carbohydrate metabolism, rather than carbohydrate availability per se, is responsible for the widely reported shifts in composition and metabolism of the oral microflora in vivo.