Abstract Refuge use provides a good model for the study of trade-offs between the benefits of predator avoidance and the costs of lost feeding opportunities. We manipulated the latter costs by subjecting similar-sized three-spine sticklebacks to 2 days of food deprivation followed by a 2-day re-feeding period and recorded asso- ciated changes in body weight and refuge use. Food dep- rivation resulted in a decrease and re-feeding in an in- crease in the duration of refuge use by fish. Emergence times of fish from the refuge were extremely variable (with a ratio of 1:127 between the shortest and the long- est ones) but individual ranks were highly consistent be- tween different days of testing, suggesting that emer- gence times were individually characteristic. Percentage weight change of fish in response to the experimental treatments also showed a high level of inter-individual variation ranging from 0–17%. A significant positive correlation was found between the percentage weight lost and the percentage decrease in emergence time from a refuge after food deprivation and similarly between the percentage weight gained and the percentage increase in refuge use after re-feeding. The relationship between en- ergy turnover and behavioural strategies is discussed. Key words Gasterosteus aculeatus · Foraging behaviour · Refuge use · Weight change Introduction Two of the most important factors that determine refuge use in animals are the intensity of a predation risk and the nutritional state (Dill and Gillett 1991; Sih 1992). Staying in a refuge precludes (or severely limits) forag- ing activity; thus refuge use results in a trade-off be- tween the benefits of predator avoidance and the costs of lost feeding opportunities (Sih 1997). Some recent stud- ies have focused on the importance of the nutritional state as a predictor of emergence time in a number of different taxa (Dill and Gillett 1991: barnacles; Koivula et al. 1995: birds; Dill and Fraser 1997: polychaete worm; Krause et al. 1998: fish). Many of these studies investigated (among other factors) whether and to what degree hiding behaviour was reduced as a function of the duration of food deprivation but little attention was paid to inter-individual variation in the effects of food depri- vation periods on energy reserves and corresponding be- havioural strategies. Krause et al. (1998) reported that food deprivation periods of the same length caused body-length-dependent weight losses in three-spine sticklebacks (Gasterosteus aculeatus) with larger indi- viduals losing a smaller proportion of their initial weight than small ones. This difference in relative weight loss was explained by the difference in mass-specific meta- bolic efficiency, which is greater in larger individuals (Weatherley and Gill 1987; Wootton 1994). As a result of higher relative metabolic costs, smaller fish generally showed shorter hiding times and decreased their hiding behaviour more strongly in response to food deprivation periods than large fish, suggesting a key role of weight loss in influencing refuging behaviour (Krause et al. 1998). Previous experiments have shown that three-spine sticklebacks respond to the introduction to a new envi- ronment by seeking shelter from which they subsequent- ly start to investigate the surrounding area for food (Krause et al. 1998). In this study, we used this proce- dure to examine the effects of food deprivation and re- feeding periods on refuge behaviour in fish of similar J. Krause ( ✉ ) · S.P. Loader · E. Kirkman Ecology & Evolution Group, Centre of Biodiversity and Conservation, School of Biology, University of Leeds, Leeds LS2 9JT, UK e-mail: J.Krause@leeds.ac.uk; Tel.: +44-113-2332840; Fax: +44-114-2332835 G.D. Ruxton Division of Environmental and Evolutionary Biology, Graham Kerr Building, University of Glasgow, Glasgow G12 8QQ, UK acta ethol (1999) 2:29–34 © Springer-Verlag and ISPA 1999 ORIGINAL ARTICLE Jens Krause · Simon P. Loader · Emma Kirkman Graeme D. Ruxton Refuge use by fish as a function of body weight changes Received: 13 December 1998 / Received in revised form: 2 May 1999 / Accepted: 7 June 1999