Season- and latitude-dependent effects of simulated twilights on circadian entrainment.

Groups of Syrian hamsters were exposed to LD cycles with twilight transitions and photoperiods simulating natural lighting conditions at the summer solstice (SS), equinox, and winter solstice (WS) at 41 degrees N and at the winter solstice at the Arctic Circle (WS 66 degrees N) but with daytime illuminance truncated at 10 lux (LD-twilight). Separate groups were kept under matching rectangular cycles (LD-rectangular). The inclusion of twilights affected several circadian parameters in a season-and latitude-dependent manner. The most striking difference was in the timing of activity onsets, which followed dusk in the presence of twilights but were more closely related to dawn (lights-on) in their absence. Activity offsets and midpoints were also earlier in LD-twilight than in LD-rectangular, with the differences being most pronounced under WS 66 degrees N. In LD-twilight, longer nights resulted in earlier offsets and midpoints, but in LD-rectangular, midpoints were later under long than under short nights while offsets did not vary significantly. In LD-twilight, activity duration (alpha) increased monotonically with increasing nighttime duration, but in LD-rectangular, alpha was shorter under WS 66 degrees N than under WS conditions. These effects of season and latitude observed in LD-twilight were similar to those reported in animals exposed to natural illumination, while those observed in LD-rectangular differed in several respects. The presence of twilights also resulted in lower day-to-day variability in activity onset times (greater precision), supporting the earlier conclusion that twilights increase the strength of the LD zeitgeber. Free-running periods in constant darkness (DD) were shorter in LD-twilight than in LD-rectangular, especially under WS 66 degrees N, raising the possibility that the effects of twilights on the timing of the entrained activity rhythm reflect their effects on the period of that rhythm. Increasing daytime illuminance to 100 lux (WS conditions only) resulted in earlier activity offsets and midpoints and a shorter alpha but had no effect on activity onsets or on subsequent period in DD. These results indicate that exposure to low twilight illuminances alone can account for several of the documented differences between the effects of natural and rectangular light cycles on circadian entrainment.