Pregnant Sprague-Dawley CD rats were administered 0, 100, or 200 mg/kg of phenytoin on days E7-18. Litters were reduced to 12, balancing for sex. Mean (+/- S.E.) maternal serum concentrations of total phenytoin 1 hr after dosing on E18 were 15.1 +/- 3.1 and 20.9 +/- 4.3 micrograms/ml in the PHT-100 and PHT-200 groups, respectively. Determinations of unbound concentrations revealed the drug to be 89% serum protein bound in both phenytoin groups. Maternal phenytoin concentrations in rats are, therefore, comparable to those seen therapeutically in humans with epilepsy. The PHT-200 group had elevated early postnatal mortality, while the PHT-100 group did not differ from controls. Phenytoin induced the typical dose-dependent increase in preweaning square-field locomotor activity. When this effect was compared to a new circular open field it was found that this device clearly distinguished phenytoin's effects. Phenytoin offspring also showed the typical dose-dependent abnormal circling behavior. Phenytoin offspring exhibited large dose-dependent increases in errors in a complex water maze, an effect which persisted even when rats exhibiting abnormal circling were excluded from the analyses. Offspring were also assessed for ability to locate a hidden vs. visible platform in an open swimming tank. Controls and PHT-100 offspring showed large improvements in performance when the hidden platform was made visible, but the PHT-200 offspring did not. Finally, offspring were assessed for working memory in an appetitive radial-arm maze. Both phenytoin groups exhibited impaired performance as measured by the number of reinforcements obtained in the first 8 arms visited.(ABSTRACT TRUNCATED AT 250 WORDS)