Predator-prey relations in fish: a battle of intellects

Until the 1970's research on predator-prey interactions in fish was rare. But in the last few decades, interest in the topic has grown explosively. By now there are thousands of articles on the topic. One aspect that has received special attention is learning. Research makes it increasingly clear that fish are not the anti-social, memory-impaired blobs they were once thought to be, but that they are intelligent creatures that form social relationships with their shoal mates and perform sophisticated techniques to outwit their predators and / or prey (Laland et al 2003; Bshary et al 2002). Many fish species live in groups or shoals. Shoals provide their members with advantages, such as a choice of breeding partner, as well as disadvantages, such as food competition (Krause & Ruxton 2002). When shoaling fish detect a predator they aggregate, forming a more compact shoal (e.g. Templeton & Shriner 2005: guppies; Ferrari et al 2005: fathead minnows; Brown et al 2004: glowlight tetras; Magurran & Pitcher 1987: European minnows). Also, several studies find that fish living in predator-dense areas spend more time aggregated than their conspecifics in less predator-dense areas. So it seems protection from predation is a major advantage of shoaling (Seghers 1974: guppies; Magurran 1986: minnows). In this review I will focus on how this protection by shoaling arises. I will show that shoaling does not always protect, but may expose their members to predation instead. In addition I will discuss various techniques employed by predators to deal with shoaling prey. Finally I will show how these predatorprey interactions are shaped by learning. In the course of this review it will become clear that the theories commonly used to explain the anti-predator benefits of shoaling are difficult to prove with empirical data. In their attempts to do so, researchers use experimental set-ups, so tightly controlled, that their naturalness becomes doubtful. Also, the number of species used in research is limited, casting doubt on the universality of our current knowledge. The text is accompanied by examples from various studies. These either illustrate the issues discussed or add an extra dimension to them.