Administration of the antiepileptic drug valproic acid (VPA) during early pregnancy can result in a 1-2% incidence of spina bifida aperta, a closure defect of the posterior neural tube in the human. The predominant defect produced by VPA in the mouse is exencephaly, a closure defect of the anterior neural tube. Recent experiments demonstrate that an appropriate dosing regimen (consecutive doses of VPA on day 9 of gestation) can also result in a low incidence of spina bifida aperta, and a high incidence of spina bifida occulta in the mouse as a potential animal model. Relatively high doses and concentrations of VPA are needed in the mouse to produce neural tube defects, the human appears to be more sensitive in this regard. Maximal concentrations and not AUC (area under the concentration-time curve) values correlate with the incidence of neural tube defects in the mouse which could in part be explained by saturation of plasma protein binding, increased free drug available for placental transfer and the embryonic neuroepithelium acting as a "deep compartment". It is likely that the parent drug and not a metabolite is the proximate teratogen. Structure-activity relationships show a strict structural requirement for high teratogenic potency: the molecule must contain an alpha-hydrogen atom, a carboxyl function, branching on carbon atom 2 with two chains containing 3 carbon atoms each for maximum activity. If these two carbon chains are different, then enantiomers are present such as the R- and S-enantiomers of 2-n-propyl-4-pentenoic acid (4-en-VPA), 2-n-propyl-4-pentynoic acid (4-yn-VPA) and 2-ethylhexanoic acid. These enantiomers were synthesized and shown to be significantly different in regard to teratogenic potency. Pharmacokinetic studies indicate that both enantiomers of each compound reach the embryo to the same degree. Therefore, the intrinsic teratogenic activity of the enantiomers differ, suggesting a stereoselective interaction between the drugs and a chiral structure within the embryo, is involved in the mechanism of action. In sharp contrast to the teratogenic effect, the anticonvulsant activity and neurotoxicity of this compound class show broad structural specificity, opening the possibility for development of novel antiepileptic agents with low teratogenic potency such as 2-n-propyl-2-pentenoic acid (2-en-VPA). The molecular mechanism of the teratogenicity of VPA is quite unknown; of the several hypothesis suggested, the interaction of VPA with embryonic folate metabolism is discussed here.