In humans, the offspring of maternal cocaine misusers are known to have subtle cognitive and motor impairments in later life. It was therefore hypothesized that such exposure in animals would also affect the morphological structure of the brain. This possibility was investigated by exposing rats to cocaine between embryonic day 15 and postnatal day 6. Samples of the cocaine-exposed and control rats were killed for examination at 22 and 150 postnatal days of age. Stereological procedures (the Cavalieri principle together with the physical disector method) were utilized to estimate the total number of pyramidal and granule cells in defined regions of the hippocampal formation. At 22 days of age, the control offspring had about 373 000 pyramidal cells whereas the cocaine-treated animals only had about 310,000 cells in the CA1 + CA2 + CA3 region. By 150 days of age the values were about 396,000 and 348,000, respectively. The differences between age-matched groups were statistically significant. There were about 626,000 and 687,000 dentate gyrus granule cells in the 22-day-old control and cocaine-treated groups, respectively. By postnatal day 150 the control rats had about 832,000 granule cells whilst the cocaine-treated rats had about 693,000. There was a significant main effect of age as well as group-age interaction in this measure. These results show that even moderate exposure to cocaine during the late gestation and early postnatal period in rats is a potent teratogen and can markedly influence the development of neurons in the hippocampal formation.