The effect of the known anxiolytic agents diazepam and alprazolam and a putative anxiogenic agent, FG 7142, were assessed in a fully automated and computer-integrated two-compartment light/dark apparatus. In addition, psychostimulant drugs (amphetamine, adrafinil, amineptine, and caffeine) were tested to determinate the influence of increasing locomotor activity on the indices of anxiety. Some modifications, such as using a soiled apparatus, have been made from the initial model to reduce any neophobic response to the test situation. These results have been compared to results obtained after cleaning between trials. In addition, strain differences have been assessed by comparing the effect of Swiss mice with the C57Bl/6J strain. The role of each parameter as an index of anxiety is discussed. The time spent in the lit area and exploratory behaviors seemed to be the most reliable parameter for assessing anxiolytic-like activity. Diazepam and alprazolam were found to have an anxiolytic profile. FG 7142 did not demonstrate any intrinsic effect. Amphetamine was reported to be anxiogenic, and amineptine, adrafinil, and caffeine only had a psychostimulant profile. We conclude that the light/dark test may be useful for identifying putative anxiolytic and anxiogenic agents, but an additional test such as an open field or an actimeter test must be performed as a control with regard to the problem of sedation and change in exploration. The Swiss strain of mice has been found a suitable strain to be used in the test.