Group size and foraging in stoneloach, Noemacheilus barbatulus.

The effects of group size on aspects of stoneloach foraging behaviour were examined in laboratory experiments. Fish were observed under dark conditions, using infra red illumination and video equipment. Experiments were timed to coincide with the stoneloach's natural time of maximum activity, immediately after a period of artificially simulated dusk. Individual stoneloach located a concealed food source more rapidly when group size was increased from one to five fish. Large quantitative differences were observed between fish, and those which were slow to locate food when alone remained poor when conspecifics were present. Over a 900 sec period, individuals in groups of five fish spent a greater proportion of available time searching, and appeared to sample the environment more, when compared with individuals in groups of only two. An increase in sampling may have counteracted the observed reduction in the number of feeding attempts by the fish in the larger groups, which resulted from the increased emphasis on searching. Groups of both six and twelve fish, foraging in tanks comprising two feeding areas each containing a patch, and a central area where no food was available, distributed themselves roughly in accordance with the predictions of an ideal free distribution when patches were equally profitable. When the patches differed in profitability in a ratio of either 2:1 or 5:1, the number of fish found in the low profitability area was as predicted. However, the number found in the high profitability area fell below that predicted, due to the presence of some individuals in the central area of the tanks. It is suggested that the latter individuals were found in the central area as a result of non-aggressive interference between individuals in the high profitability area. Finally, the marginal value theorem was used to predict the optimal patch residence time. The prediction was met in experimental tests at a group size of one, but at group sizes of two or five fish the observed time was significantly lower than that predicted