Prenatal exposure to the antiepileptic drug valproic acid (VPA) has been associated with the formation of spina bifida aperta, meningocele, and meningomyelocele in the human. Until now, a direct relationship between VPA application and spina bifida has not been experimentally demonstrated. VPA was known only to induce exencephaly in mice, a defect of the anterior neural tube. Maximal sensitivity toward production of this defect was on day 8 of gestation (plug day = day 0). The closure of the posterior neuropore occurs later in the development of mice than the closure of the anterior neuropore. To investigate whether there is a direct relationship between VPA application during pregnancy and induction of spina bifida in mice, we administered various doses of the drug on day 9 of gestation, at three time intervals (at 0, 6, and 12 hr). This administration of VPA produced spina bifida aperta and spina bifida occulta in mice. High doses of VPA (3 × 450 and 3 × 500 mg/kg) induced a low rate of spina bifida aperta in the lumbosacral region. High incidences of spina bifida occulta, a less serious form of spina bifida, were induced with lower doses. This malformation was demonstrated in double-stained fetal skeletons by measurements of the distance between the cartilaginous ends of each vertebral arch. The occurrence of this defect and its localization was dose-dependent. The lumbar region was affected by all doses investigated (3×300, 3×350, 3×400, 3×450, and 3×500 mg/kg). The sacral/coccygeal region was affected additionally, but with higher doses (3×400, 3×450, and 3×500 mg/kg). A comparison of the results obtained with day 16 and 17 control fetuses showed that the pattern of gaps present in the lumbar and sacral region of the spinal cord in treated groups was drug-specific and not related to a developmental delay. Our results indicate that multiple administrations of VPA on day 9 of gestation in mice result in a low incidence of spina bifida aperta and a high incidence of spina bifida occulta, and provides a relevant model for the study of human spina bifida defects.