The air-righting reflex has been used for many years to assess reflex integrity in rodent developmental neurotoxicity investigations. However, little refinement of the technique has occurred. We describe two methodological improvements: (a) an improvement in the method of positioning rats for air-righting, and (b) a stop-action photographic method to capture the rat's mid-air performance. We also compare results obtained using a visual scoring method to the newly developed photographically based scoring method. Prenatal phenytoin exposure has been shown to induce marked delays in air-righting development (14), therefore, phenytoin was used as a positive control treatment. Pregnant rats were administered 200 mg/kg phenytoin in propylene glycol or propylene glycol alone by gavage once/day on E7-18. Offspring were tested in the same apparatus twice and scored for air-righting success either by direct observation or photographed and the photographs scored subsequently. Rats were administered 6 trials per day (3 trials with each method) on days 16-24 of postnatal development. Detailed analyses of the two methods included subdividing phenytoin animals into those exhibiting the neurological abnormality of circling later in life and those that did not. Both methods revealed that phenytoin animals were delayed in air-righting development compared to controls and both methods revealed that phenytoin-circlers were more impaired than phenytoin-noncirclers. Advantages of the photographic method were that it provided a more precise method of scoring and a permanent record of the animal's response. One disadvantage was that it did not distinguish groups quite as well as the visual method.