Two experiments paired flavors with ethanol, sucrose, or ethanol and polycose solutions to investigate learning of flavor preferences by rats. When one solution offered a caloric advantage over the other, a preference was learned for its associated flavor. When one solution offered no caloric advantage over the other, no preference was learned for the associated flavors. Increasing the caloric value of the ethanol while holding the actual percentage of ethanol in solution constant fostered a preference for a flavor associated with the more dense solution. Recent reports involving ethanol intake are examined in light ofthis calorie-based learned-preference mechanism.
421There is abundant evidence for learning about negative postingestional effects of a food and learning to avoid the flavor of that food. For nearly three decades (see Riley & Clarke, 1977, for a bibliography), a copious liturature has been built on studies using a tasty solution or food paired with a sickness-inducing agent (usually lithium) in which animals learn to avoid the sickness-paired taste. This kind of learning is easy to get and hard to extinguish. It seems clear in taste-aversion learning that the taste of a particular food becomes associated with the organismic changes that result from ingestion of the food. This is true even when the food is not the cause of the sickness (e.g., when x-rays are used to induce illness; Garcia, Kimmeldorf, & Koelling, 1955).In contrast to the clear picture that has emerged with conditioned taste aversions, there is at best only sparse evidence for learning about the positive aspects of foods and the flavors associated with them. One line of evidence is in the area of learned taste preferences based on a "medicine effect." Le Magnen, Marfaing-Jallet, and Miceli, (1980) described rats that were dependent upon alcohol and were exhibiting withdrawal symptoms. These rats were given saccharin paired with small doses of alcohol for 2,4, or 8 days. Later, when tested for saccharin consumption, they displayed a positive linear relationship between the number of pairings and saccharin consumption. Le Magnen et al. concluded that a medicine effect was conditioning the learned preference. Zahorik and Maier (1969) offered rats a flavor previously paired with recovery from thiamine deficiency and found a preference for that flavor over other flavors presented in various control conditions. Rozin (1967) showed that rats with experience of sickness tend to prefer familiar flavors over novel ones. That is, they display increased neophobia relative to controls. Zahorik, Maier, and PiesWe wish to thank Moncrief Smith and Hank Sampson for their editorial assistance in the preparation of this manuscript. The authors' mailing address is: Department of Psychology, University of Washington, Seattle, WA 98195.(1974) took this neophobia interaction into account and in a well-designed study offered conclusive evidence that rats really do prefer flavors paired with recovery from illness. In a retardation type of experiment (Zahori...