1. Animal studies of the effects of early exposure to CNS agents devoid of a major teratogenic potential must assess possible deviations from normal behavioral development in both a stage-specific and a behavior-specific fashion; several experiments on prenatal benzodiazepine (BDZ) exposure are reviewed, illustrating such an assessment strategy and discussing caveats on experimental designs and statistical analysis. 2. The offspring of mouse dams treated in late pregnancy with oxazepam (15 mg/kg p.o. twice daily on days 12-16) show a mild and reversible impairment in somatic and neurobehavioral development which is unlikely to be responsible for a series of other more specific changes. 3. The treatment produces a selective reduction of locomotor activity and amphetamine hyperactivity at the end of the second postnatal week, as well as a selective impairment of active avoidance at the young adult stage, in the absence of similar changes in scopolamine hyperactivity and passive avoidance. 4. The treatment also prevents the appearance at 28 days of morphine hyperactivity and of rebound hyperactivity after muscimol depression, without modifying the developmental profile of pain reactivity and of morphine and muscimol analgesia. 5. Young adult females previously exposed to oxazepam in utero show a marked enhancement of maternal aggression towards male intruders; mother-pup interactions are also modified, leading either to reduced or to exaggerated maternal care as a function of fostering procedures. 6. Overall, several effects of prenatal BDZ exposure appear to be amenable to modifications in monoaminergic system functions and/or to an accelerated development of GABAergic mechanisms; some of the changes in social and parental interactions, however, point to subtle modifications in the balance between different components of the fear-defensive repertoire, possibly due to an altered stimulus reactivity by mechanisms which are still poorly understood.