1. The vigilance states (waking, slow wave sleep (SWS) and paradoxical sleep (PS)) and motor activity were continuously recorded by telemetry in unrestrained rats. The animals were first maintained for 2–3 days under 12-h light — 12-h dark (LD 12∶12) control conditions, and then exposed for 6–9 days to one of the 3 following schedules: skeleton photoperiods (SP), continuous darkness (DD) or continuous light (LL). The SP-schedule consisted of continuous darkness except for 2 daily 20-min light-periods situated at the beginning and at the end of the previous 12-h light-phase. 2. The period (τ) of the circadian activity rhythm increased to approximately 25 h in LL, but remained generally close to 24 h in SP and DD (Table 1). The amplitude (A) of the circadian rhythms was more attenuated by LL than by the two other schedules (Table 2). The value for A was generally larger for PS or motor activity than for waking and SWS. 3. The daily amount of motor activity was significantly increased in DD (Table 3). Activity tended to be increased also in SP, and PS in both SP and DD, while SWS tended to be reduced during these schedules. The daily amount of PS as well as the percentage of PS relative to total sleep (PS/TS) were significantly increased in LL, while the amount of waking was significantly reduced. This increase of PS to 117.3% of the control level was mainly due to a more frequent occurrence of PS-episodes (Table 4). 4. In SP and DD, motor activity and waking were higher during the circadian phase of low activity (ρ) than during the light-phase of the control days, and lower during the circadian phase of high activity (α) than during the dark-phase of the control days (Fig. 5). The demarcation of α andρ were distinct in SP and DD, and less well defined in LL (Figs. 1–3). Nevertheless, basic features of the circadian patterns persisted under all experimental conditions. They include (a) a skewed distribution of activity and waking, with a maximum towards the end of α, and a low level at the beginning ofρ (Fig. 5); (b) an increasing trend of PS/TS duringρ (Fig. 5); and (c) a decreasing trend in the duration of SWS-episodes duringρ (Fig. 6). 5. We conclude that the circadian patterns of the vigilance states and motor activity as well as the differences in the distribution of SWS and PS, are mainly due to endogenous processes, and are influenced only to a minor extent by the 24-h LD-cycle.