ournal of Comparative and Physiological Psycholog y ,)67, Vol. 64, No. 2, 237-242 SPECIFIC AVERSIONS AS A COMPONENT OF SPECIFIC HUNGERS 1 PAUL ROZIN University of Pennsylvania Rats deficient in thiamine develop an aversion for the deficient diet, which may account for their potentiated feeding response to new foods. The spe- cific aversion to their familiar deficient diet is demonstrated by (a) spillage of this diet (normal rats may spill highly unpalatable foods), (b) redirected feeding responses (chewing inedible objects) in the presence of deficient diets, and (c) after recovery on a new diet, an avoidance of the familiar de- ficient diet even when they are food deprived with no other food available. It is suggested that specific aversions play an important role in most specific hungers, and that specific hungers and adaptive responses to poisoning are two aspects of the same basic phenomenon. Rodgers and Rozin (1966) have made some observations on thiamine-deficient rats that help to remove some of the tradi- tional objections to a learning interpreta- tion of thiamine-specific hunger. They re- ported that thiamine-deficient rats showed strong, often 100% preferences for any new diet offered to them, in opposition to the standard deficient diet. Furthermore, such typically anorexic rats avidly consumed the new diets. The exclusive, continued in- gestion of a new diet over a period of many hours would facilitate the rat's identifica- tion of the recovery-producing diet. Inges- tion of the new diet during the onset of recovery might bridge the delay of rein- forcement "gap," because the rat would be continuing to eat the new diet for hours, and, in the meantime, recovering as a re- sult of thiamine ingested in the initial meals. Rozin and Rodgers (1967) dem- onstrated novel-diet preferences in pyri- doxine- or riboflavin-deficient rats, and Rodgers (1967) demonstrated the same phe- nomenon in rats deficient in a number of minerals, including calcium. In addition, Rozin and Rodgers (1967) showed that the preference for thiamine-rich foods by rats recovered from deficiency (Rozin, 1965) was nonspecific; recovered rats preferred 1 This research was supported by National Sci- ence Foundation Grants GB 1489 and 4372. The author wishes to thank Willard Rodgers, Robert Rescorla, Martin Seligman, and Richard L. Solo- mon for their valuable comments and advice, and Barbara Marks and Deborah Donnelly for their technical assistance. any new food to the food on which they became deficient. In order to account for the appearance of a novelty response in the deficient or recovered rat, Rozin and Rodgers (1967) suggest that either the state of deficiency triggers an innate neo- philia, or that the deficient rat develops an aversion to the deficient diet and thus shows a strong preference for any new food. It seems entirely reasonable to be- lieve that the deficient diet could develop aversive properties, as its ingestion is most certainly associated with feelings of ill- ness; in fact, the anorexia of thiamine de- ficiency could be considered a consequence of the aversion. There appears to be a certain arbitrari- ness in debating whether the deficient diet is aversive or the new diet is highly de- sired. It is certainly unreasonable to as- sume that, in any choice of two diets, the less preferred is aversive, since obviously the choice may be between two highly de- sired substances. However, animals often show distinctive behaviors associated with conflict, withdrawal, and disgust when placed in aversive or noxious situations. Furthermore, if a food were aversive, one would expect that under most circum- stances an animal would prefer nothing (no food) to eating it (see Irwin, 1961). If a hungry rat refuses to eat a nutritious, nor- mally palatable food under normal caging conditions, it is reasonable to suggest that the food is aversive. Because evidence for an aversion is de- 237