In Experiment 1, rats were tested in a plus-maze, with or without small ledges on the open arms, after injection with vehicle or chlordiazepoxide (7.5 mg/kg). They were scored either on their first or second exposure to the maze; those scored on trail 2 had received a 5-min undrugged exposure to the maze 24 h earlier. This dose of chlordiazepoxide had a significant anxiolytic effect on trial 1 only in the maze without ledges, and on trial 2 only in the maze with ledges; thus, the presence of ledges differentially affected anxiolytic sensitivity on trials 1 and 2. The results of a factor analysis study (Experiment 2) confirmed that ledges had a differential effect when rats were repeatedly exposed to the maze. Thus, in the maze without ledges, the scores reflecting anxiolytic activity on trial 1 loaded on one factor, whereas the scores from trials 2 and 3 loaded on another independent factor. In the maze with ledges, the scores reflecting anxiolytic activity on trials 1, 2, and 3 loaded on three independent factors. Considering the published evidence and the results of the present study, we suggest that both types of plus-maze may be measuring the same type of anxiety with different sensitivities on trial 1 (e.g., generalised anxiety or fear of open spaces); different types of anxiety on trial 2 (without ledges--phobia/fear of heights; with ledges--not known), and trial 3 in the maze with ledges, yet another type of anxiety. The factor analysis results are also presented for ethological measures on the plus-maze, and for activity and exploration in the holeboard. Based on the factor loadings, a composite measure of anxiety on trial 1 is presented which will increase the sensitivity of the plus-maze to anxiolytic treatments. The measures of motor activity in the plus-maze load on a different factor from those derived from the holeboard, thus cautioning against considering all measures of motor activity as interchangeable.