Social Transfer of Predation Risk Information Reduces Food Locating Ability in European Minnows (Phoxinus phoxinus) Jo ¨ rgen I. Johnsson & L. Fredrik Sundstro ¨m Section of Animal Ecology, Department of Zoology, Go ¨ teborg University, Go ¨ teborg, Sweden Introduction Many species of animals aggregate into groups for a number of different reasons related to, for example, predation, foraging, mating and energy conservation (Krause & Ruxton 2002). One advantage of joining a group is the potential to acquire information, such as food location, from other group members (Greene 1987). The ability to learn new feeding routes can be sex specific, depend on motivation, or vary with the predisposition to be innovative (Laland & Reader 1999a,b; Sundstro ¨ m et al. 2004). However, relying upon social cues to find food may be disadvanta- geous if others do not know the best way (Laland & Williams 1998), so it is not surprising that indivi- duals preferentially associate with successful foragers (Dugatkin & Wilson 1992) even if they are of another species (Sasva ´ri & Hegyi 1998). While social information of food location would appear most beneficial in complex environments, such environments also make predator detection more difficult. Predators may use structural complex- ity to avoid detection, so foragers may have to remain vigilant at all times of activity. Accordingly, several studies indicate that animals adjust their foraging effort in response to perceived predation risk (Lima & Dill 1990; Abrahams & Healey 1993; Anholt et al. 2000). Social information transfer can reduce predation risk by enhancing predator detec- tion (Kenward 1978), or by altering movement deci- sions (Pollock et al. 2006). Reader et al. (2003) showed that observer guppies in the presence of pre- trained demonstrators learned the correct choice of two possible escape routes to avoid being trapped in a trawl net. This choice remained even after removal of demonstrators. Furthermore, recent experiments Correspondence Jo ¨ rgen I. Johnsson, Department of Zoology, Go ¨ teborg University, Box 463, SE-405 30 Go ¨ teborg, Sweden. E-mail: jorgen.johnsson@zool.gu.se Received: July 6, 2006 Initial acceptance: August 7, 2006 Final acceptance: August 15, 2006 (K. Reinhold) doi: 10.1111/j.1439-0310.2006.01296.x Abstract Little is known about how food location ability of animals is affected by social information of predation risk. This question was therefore addressed in an experimental study where naı¨ve ‘observer’ European minnows (Phoxinus phoxinus) were allowed to search for food in a maze in the presence or absence of a predator (Salmo trutta). Observer min- nows were accompanied by conspecific demonstrators which had previ- ously been trained in the same maze either in the presence or absence of the predator. Observer minnows were most likely to locate food when the predator was absent both during their trial and during the pre-training of demonstrators. When demonstrators had been trained with predators, observer success in locating the food was halved, although they were never exposed to predation risk themselves. When observers were exposed to predation risk their probability of locating food was further reduced regardless of the experience of their demon- strators. Our results show that predation risk can affect the foraging abil- ity of minnows both directly and indirectly through social information from conspecifics. We conclude that social information may influence and constrain individual behavioural decisions, especially in rapidly changing environments where private information is often insufficient. Ethology Ethology 113 (2007) 166–173 ª 2007 The Authors 166 Journal compilation ª 2007 Blackwell Verlag, Berlin