Behavioural Adjustment of Population Density to Available Food by Juvenile Atlantic Salmon

The distribution and density of every species of animal is at least loosely controlled by the distribution and abundance of its food. Particular attention has been paid recently to the distribution of salmonids as affected by selection of the site of the fish's feeding territory (Chapman & Bjornn 1969) and to their numbers as affected by varying abundance of food in different streams (Egglishaw 1967). Evidence that food supply does more than permit or prevent a population of salmonids from inhabiting a stream is as yet equivocal. Mason & Chapman (1965) found that the biomass and numbers ofjuvenile coho salmon (Oncorhynchus kisutch Walbaum) remaining in two stream channels from which they could emigrate were greater in the channel which had a greater food supply. However, since there was no replicate in which the varying rations were switched between channels, the possibility remained that the channel with the greater food supply retained more fish because of a greater number of hiding places, better cover, etc. In comparing two streams, one of which had three times as much food as the other, Egglishaw (1967) found that there were fewer trout in his food-poor stream, but more salmon, so that the total density of salmonids differed little (about 0 45 fish/m2 in the poor compared to 0 55 fish/m2 in the rich stream). Biomass differed even less. The streams were different in other respects besides food, however, which, in this case, may have masked the effect of food supply on density. Le Cren (1965) suggests that while a population of fish may be limited by food in the long run, in some instances other regulatory mechanisms usually reduce the importance of food supply as a regulatory factor. Effects offood supply on behaviour, in contrast to those on population density, have been reasonably well defined. In laboratory experiments with young salmon (Salmo salar L.), provision of food in two daily feedings was followed shortly afterwards by an increase in aggression which subsided gradually 30 min later (Keenleyside & Yamamoto 1962). Symons (1968) found when aggression was measured after immediate effects of feeding had waned, that fish deprived of food for 18-66 h were more aggressive than those receiving an abundance of food. Together these results indicate that, in those natural habitats having a continuous abundant supply of food, young salmon are probably less aggressive than in habitats where they are constantly hungry and meals are sporadic. Chapman (1962) has postulated that aggression of some territorial fish may cause the emigration of others with inferior fighting ability. Presumably an increase in aggression would speed this emigration process. Symons (1968) also suggested that the increase in aggression associated with food-deprivation might, in a natural environment, result in an