Social groups are structured by the decisions of their members. Social insects typically divide labour: some decide to stay in the nest while others forage for the colony. Two sources of information individuals may use when deciding whether to forage are their own experience of recent task performance and their own physiology, e.g. fat reserves (corpulence). The former is primarily personal information; the latter may give an indication of the food reserves of the whole colony. These factors are hard to separate because typically leaner individuals are also more experienced foragers. We designed an experiment to determine whether foraging specialisation is physiological or experience based (or both). We invented a system of automatic doors controlled by radio-tag information to manipulate task access and decouple these two sources of information.
Our results show that when information from corpulence and recent experience conflict, ants behave only in accordance with their corpulence. However, among ants physiologically inclined to forage (less corpulent ants), recent experience of success positively influenced their propensity to forage again. Hence, foraging is organised via long-term physiological differences among individuals resulting in a relatively stable response threshold distribution, with fine-tuning provided by short-term learning processes. Through these simple rules, colonies can organise their foraging effort both robustly and flexibly.