time of 78 W*m -2 (hours before dusk) 5 4 3 2 1 time of retreat (hours before dusk) 5 4 3 2 1 06:00 12:00 18:00 number of animals above ground 0 10 20 30 illuminance (Lux) 10 3 10 4 10 5 * rain time (hours before dusk) 7 6 5 4 3 2 1 log irradiance (W*m -2 ) 0 1 2 3 Solar eclipse: During the eclipse, light intensity decreased by more than two log units (fig. 4). From 12:16 - 13:16 on average 16.3 animals were active above ground, similar to either the hour before (15.4) or the hour after the eclipse (16.4). Four animals with radio transmitters were above ground for most (156 - 160 min, 20 min absence by rain) of a three-hour period around the eclipse. Light intensity in the afternoon and during a solar eclipse: effects on retreat underground in a diurnal mammal Kamiel Spoelstra, Roelof Hut, Arjen M. Strijkstra, Gerard J. F. Overkamp and Serge Daan Zoological Laboratory, Haren, the Netherlands www.biol.rug.nl/animal_behaviour/ e-mail K.Spoelstra@biol.rug.nl Introduction European ground squirrels (Spermophilus citellus) in their natural habitat emerge from their burrows ~4 hours after twilight at dawn and disappear ~3 hours before twilight at dusk 1 . Entrainment of their endogenous circadian rhythm apparently does not require the perception of twilight. Here we investigate whether more subtle changes in light intensity trigger retreat behaviour. Conclusion Both afternoon and eclipse data show that neither a specific light intensity nor a specific intensity decrease acts directly as a signal for retreat. A general decrease in light intensity in the afternoon may partly act as a signal for afternoon retreat. Apparently this will be phase dependent since a similar decrease in light intensity around the middle of the day does not induce retreat behavior. Methods 4 male and 5 female ground squirrels in an enclosure in Haren, the Netherlands (53 0 10’N) were equipped with light sensitive radio transmitter collars 1 , enabling determination of above ground presence. During 66 precipitation free days, time of retreat was recorded. Concurrent light intensity measurements in 60’ intervals were obtained from Groningen airport Eelde, at 6-km distance. In addition, we exploited the solar eclipse of August 11, 1999 above a field population of European ground squirrels near Vienna, Austria (48°18’N) 2 . On August 10, 11, and 12, ground squirrels above ground were counted every 5 min in a 1 Ha area. The presence of five individuals equipped with light sensitive radio transmitter collars was recorded every 10 min, along with light intensity. Results Afternoon light decrease: Retreat occurred at an average light intensity of 78 W*m -2 (fig. 1). For all animals retreat occurred within a range of less than two log units. fig. 2 shows retreat time against the time of day at which 78 W*m -2 was reached. Although the timing of 78 W*m -2 is not the only determinant of retreat, the later this intensity occurs the later the animals retreat into their burrows (p=0.017). In fig. 3 the slope of the radiation curve at the time of retreat is plotted. Clearly, retreats do occur at a wide range of decreases in light intensity, and not when a particular rate of change is reached. retreat (hours before dusk) 5 4 3 2 1 dusk slope irradiance curve (log W*m -2 *h -1 ) -2 -1 0 1 Fig.4. Number of ground squirrels active and concurrent light intensity (Lux), observed in a 1 Ha focal area in a population near Vienna, Austria. Thick horizontal lines indicate presence of 5 animals with radio transmitters above ground. * Denotes timing of the partial (99%) solar eclipse at 12:46 local time, on August 11, 1999 (civil twilight dawn 5:07; civil twilight dusk 20:51). Fig.1. Time of end of activity (dots, hours before dusk) and concurrent radiation (curves, W*m -2 ). Fig.2. Occurrence of 78 W*m -2 (average retreat light intensity) and retreat (hours before dusk) for each subjective day. Dashed line: Y=X (if animals on average would retreat during 78 W*m -2 ). Fig.3. Average decline in light intensity of all subjective days (W*m -2 *h -1 , ±SD) and moments of retreat (hours before dusk) with momentary decline in light intensity (W*m -2 *h -1 ). References. 1. Hut, R.A., B.E.H. van Oort & S. Daan. J. Biol. Rhythms 14, 290-299 (1999). 2. Millesi, E., A.M. Strijkstra, I.E. Hoffmann, J.P. Dittami & S. Daan. J. Mamm. 80, 218-231 (1999).