Reduced folate derivatives participate in numerous reactions of bacterial intermediary metabolism. Consequently, the well-characterized enzyme implicated in the formation of tetrahydrofolate--dihydrofolate reductase FolA--was considered to be essential for bacterial growth. However, comparative genomics has revealed several bacterial genome sequences that appear to lack the folA gene. Here, we provide in silico evidence indicating that folA-lacking bacteria use a recently discovered class of flavin-dependent thymidylate synthases for deoxythymidine-5'-monophosphate synthesis, and propose that many bacteria must contain uncharacterized sources for reduced folate molecules that are still waiting to be discovered.