Glucose and lactate metabolic rates were evaluated for cultures of cord blood (CB) mononuclear cell (MNC), peripheral blood (PB) MNC, and PB CD34(+) cell cultures carried out in spinner flasks and in T-flasks in both serum-containing and serum-free media. Specific glucose uptake rates (q(gluc), in micromoles per cell per hour) and lactate generation rates (q(lac)) correlated with the percentage of colony-forming cells (CFC) present in the culture for a broad range of culture conditions. Specifically, the time of maximum CFC percentage in each culture coincided with the time of maximum q(gluc) and q(lac) in cultures with different seeding densities and cytokine combinations. A two-population model (Q(lac) = alpha[CFC] + beta([TC] - [CFC]), where [TC] is total cell concentration; Q(lac) is volumetric lactate production rate in micromoles per milliliter per hour; alpha is q(lac) for an average CFC; and beta is q(lac) for an average non-CFC) was developed to describe lactate production. The model described lactate production well for cultures carried out in both T-flasks and spinner flasks and inoculated with either PB or CB MNC or PB CD34(+) cells. The values for alpha and beta that were derived from the model varied with both the inoculum density and the cytokine combination. However, preliminary results indicate that cultures carried out under the same conditions from different samples with similar initial CD34(+) cell content have similar values for alpha and beta. These findings suggest that it should be possible to use lactate production data to predict the harvest time that corresponds to the maximum number of CFC in culture. The ability to harvest ex vivo hematopoietic cultures for transplantation when CFC are at a maximum has the potential to speed the rate at which immunocompromised patients recover.