Recent in vitro investigations have suggested that GABA is involved in the development of the mammalian central nervous system. To evaluate the roles of GABA in neurogenesis in vivo, we generated mice lacking both the isoforms of glutamic acid decarboxylase (GAD), GAD65 and GAD67, by mating GAD65- and GAD67-mutant mice generated by homologous recombination in this laboratory. Similar to GAD67-deficient mice, the GAD65/67-deficient mice did not survive after birth because of cleft palate. We thus analyzed these mice at the fetal and newborn stages. GABA was scarcely detectable in the GAD65/67-deficient brains, indicating that the GAD-independent GABA synthetic pathway was not active. The activity of ornithine decarboxylase, which is possibly involved in such a pathway, did not increase with the GAD deficiency. Histological and immunohistochemical studies of the GAD65/67-deficient brain did not reveal any discernible disorders of histogenesis. The discrepancy between the results of previous in vitro investigations, performed mostly on rat tissue, and those of the present analysis on mutant mice may be attributed to the different species used or to the possibility that other mediators can compensate for GABA functions in vivo.