More than 200 samples of pooled fetal bovine sera (FBS) were tested to determine their capacity to support primary immune responses by cultured mouse spleen cells. Less than 10% of the FBS samples were fully supportive; the majority of the samples were moderately to very deficient. Several lines of evidence suggest that bacterial contamination during the processing of sera plays a major role in determining which sera are supportive. Serum samples known to have been temporarily contamined during processing were strongly supportive. Samples of FBS which were likely never to have been contaminated were deficient. Bacteria from a specific lot of supportive serum converted a very deficient serum into one which was supportive. Several mechanicsms by which bacteria could effect sera are discussed.