Let us consider the experimental data available regarding omnivores other than man. The rat, in
particular, has quite extraordinary capacity for learning about food. It was in rats that the capacity
for aversion learning, an apparently quite unique form of learning through conditioning, was first
demonstrated. A single experience is sufficient to provoke in rats durable aversion to a food
associated with gastro-intestinal upset, even if this occurs several hours after ingestion (Garcia,
Ervin & Koelling, 1966). Rats are also found to have a remarkable capacity for innovative
prudence. When confronted with several new foods, a rat behaves in fact like the most careful
scientific experimenter, varying only one parameter at a time: it tries only one unknown food at a
time, and in a small quantity (see Rozin, 1976). In a sense, therefore, the rat resolves the
omnivore's paradox through its remarkable learning capacities, and more generally through its high behavioural flexibility. It is capable of minimizing the risks of nutritional innovation,
particularly by acquiring aversions.