Pregnant Sprague-Dawley CD rats were administered phenytoin by gavage on days 7-18 of gestation in doses of 0 or 200 mg/kg. The offspring were tested at various ages to determine the duration of postnatal dysfunction and its replicability and generality compared to previous experiments. Phenytoin offspring had increased newborn (5.2%) and preweaning (16.7%) mortality compared to controls (0% and 3.1%, respectively), and an 8.5% reduction in average body weight at 28-70 days. No weight differences were significant at other ages. Phenytoin offspring showed increased activity on multiple tests, swam slower in a straight channel, committed more errors and took more time in the Cincinnati water maze, startled less, and had longer latencies on the Morris hidden platform test. Among phenytoin offspring 42.3% exhibited the abnormal circling defect previously described (14,17). Consequently, data were reanalyzed in terms of circlers, noncirclers, and controls to determine the contribution of this effect to the dysfunctions observed. Circlers accounted for the differences in open-field activity, figure-8 ambulation, hole-board horizontal locomotion, straight channel swimming time, water maze retention errors, tactile prepulse startle inhibition, and some trials of the Morris test. Circlers and noncirclers differed from one another and from controls on measures of figure-8 rearing, water maze errors and times, and some trials of the Morris test, with circlers more affected than noncirclers. Circlers and noncirclers did not differ from one another, but both differed from controls, on measures of early locomotion, hole-board vertical activity, and unmodified startle amplitude. Circling was hypothesized to reflect an underlying vestibular defect, however, the data also support the view that phenytoin has effects beyond those accounted for by possible vestibular effects.