Behavioural modulation of predation risk: moonlight avoidance and crepuscular compensation in a nocturnal desert rodent, Dipodomys merriami

The temporal and spatial distribution of above-ground activity in Merriam's kangaroo rats, Dipodomys merriami, radio-tracked during 10 winters varied in relation to phases of the moon. At the full moon, animals were more likely to be found in their day burrows at scheduled hourly radio fixes, and when they emerged, they stayed closer to home. At partial moon phases, above-ground activity was preferentially allocated to hours when the moon was down. As predicted from an extension of Rosenzweig's (1974, J. Mammal., 55, 193-199) model of optimal above-ground activity, the suppression of nocturnal activity at the full moon was partially offset by relatively great activity at dusk and dawn, and losses to nocturnal versus diurnal predators were distributed differently over the lunar cycle. Gross predation rates did not differ significantly among moon phases, but the kangaroo rats' relatively crepuscular activity pattern at the full moon both reduced their vulnerability to nocturnal predators, as compared with the new moon phase, and raised their vulnerability to diurnal shrikes. In various domains ofbehavioural decision making, animals act as if the risk of predation were a cost to be weighed against expected benefits (Lima & Dill 1990). An example is moonlight avoidance. Nocturnal rodents that forage in relatively open habitats respond to moonlight by reducing activity outside their nests or burrows, and by shifting such activity towards areas of relatively dense cover (e.g. Lockard & Owings 1974; Price et al. 1984; Bowers 1988; Wolfe & Summerlin 1989). Artificial moonlike illumination elicits similar responses (Lockard 1975; Kotler 1984; Brown et al. 1988). Although these behavioural responses have been assumed to reduce predation risk, no study of moonlight avoidance has incorporated data on actual predation events in the wild. Neither have researchers hitherto assessed the moon's influence on the spontaneous travels of nocturnal rodents in natural habitats, relying instead on measures of activity in enclosures and on visits to artificial food sources. Radio-tracking is a technique by which natural predation events can be detected and related to the behaviour of radio-tracked prey *Present address: Michael Brandman Associates, San Diego, CA 92130, U.S.A. #Present address: Department of Psychology, University of Utah, Salt Lake City, UT 84112, U.S.A. 0003 3472/92/070001+09503.00/0 (Daly et al. 1990), and by which the moon's effects on behaviour can be assessed without experimental interventions. Kangaroo rats (Dipodomys, Heteromyidae) are predominantly granivorous, burrow-dwelling residents of western North American arid and grassy habitats. Kangaroo rats forage more in the open spaces between perennial shrubs than do sympatric granivorous rodents such as pocket mice (Lemen & Rosenzweig 1978; Thompson t 982; Price & Brown 1983), and their bipedal locomotion (Nikolai & Bramble 1983; Thompson 1985) and hypertrophied middle ear cavities (Webster 1962; Webster & Webster 1971) have been interpreted as adaptations to a high level of predation risk in their preferred foraging environment. Crucial decisions for kangaroo rats include the amount of aboveground activity to be undertaken each night and its distribution in time and space (Rosenzweig 1974; Behrends et al. 1986a, b). Daly et al. (1990) analysed 50 predation deaths among 176 radiotracked Merriam's kangaroo rats, D. merriami, and found a strong association between recent surface travels and the risk of predation in both sexes, those animals who were most mobile being most at risk. Rosenzweig (1974) proposed that the benefits of above-ground activity are likely to exhibit 9 1992 The Association for the Study of Animal Behaviour