The effect of anticipated learning on the choice of food sources in the fish gambusia affinis

Sagiv Shifman*, Dan Cohen

*Corresponding author for this work

Research output: Contribution to journalArticlepeer-review


Animals usually prefer the less variable food sources of equivalent mean yields. However, we expect this preference to decrease or even to reverse if the animals can learn to associate a particular cue such as color, shape or location with the distribution or quality of the more variable food. Learning this association would enable the animal to preferentially exploit the potentially rich food sources and these decrease the effective variance and increase the effective mean of the variable food source. In this study we tested these predictions. We performed three different types of experiments in which the fish Gambusia affinis were offered a choice between two food sources. In one source, both the food and the cues were always constant, and in the second source the food and/or the cues were variable. We found, as expected, that with uniform cues in both sources, the fish preferred a constant source over a source with variable food distribution. In a second experiment, in which the fish could learn to associate between the food distribution and the variable cues of the variable source, they made equal foraging efforts (number of visits and flakes eaten) at the variable and the constant sources. In a third experiment, the fish were offered one constant source and a second source that was variable in both food and cues, but without correlation between them. The fish made equal foraging efforts at the two sources during the first two days of this experiment, while subsequently they preferred the constant source.

Original languageAmerican English
Pages (from-to)287-298
Number of pages12
Issue number3
StatePublished - Mar 2001


Dive into the research topics of 'The effect of anticipated learning on the choice of food sources in the fish gambusia affinis'. Together they form a unique fingerprint.

Cite this