Dynamics of fish shoals: Identifying key decision rules

Research output: Contribution to journalArticle

68 Scopus citations

Abstract

Social aggregations of fish are extremely common in nature. Pitcher (1983) defines a social aggregation of fish as a shoal, with a highly polarized shoal constituting a school. The ultimate causes of fish shoaling have been extensively studied and are well established, with the main causes being protection from predators and enhanced foraging ability. The proximate mechanisms by which groups are maintained are less well understood. Several models exist that examine these decision rules. A common theme to these models is the existence of a region at large nearest-neighbour distances (NND), where an individual will move towards its nearest neighbour, and a region of repulsion at small NND, where the individual will move away from its nearest neighbour. Recent models (Gueron et al., 1996; Couzin and Krause, 2003) suggest that shoaling decisions may not be instantaneous, instead being hierarchical as fish assess the presence or absence of one or more neighbours in zones of attraction, repulsion and neutrality. Here a mixed shoal of creek chubs (Semotilus atromaculatus) and blacknose dace (Rhinichthys atratulus) was examined in the field through video recording and digitization of images. Shoal dynamics were examined both under normal, undisturbed conditions, and in the presence of a simulated predator. By tracking the movements of individual fish over time, we find evidence for both attraction and repulsion zones. But three other novel features emerge as well. First, between regions where focal fish are attracted to, or are repulsed by, neighbours, there appears to be a 'neutral zone', where no consistent response occurs. Second, changes in NND at far distances, more so than at close distances, are affected not only by the position of nearest neighbours, but by the orientation of their velocity vectors as well. Third, the effect of increased fear level of fish induced by the appearance of simulated predators is a decrease in both the average NND and the sizes of the stress, attraction and neutral behavioural zones.

Original languageEnglish (US)
Pages (from-to)555-565
Number of pages11
JournalEvolutionary Ecology Research
Volume6
Issue number4
StatePublished - May 1 2004

All Science Journal Classification (ASJC) codes

  • Ecology, Evolution, Behavior and Systematics

Fingerprint Dive into the research topics of 'Dynamics of fish shoals: Identifying key decision rules'. Together they form a unique fingerprint.

  • Cite this