Journal of Comparative and Physiological Psychology 1966, Vol. 61, No. 1, 1-4 NOVEL FOOD PREFERENCES IN THIAMINE-DEFICIENT RATS1 WILLARD RODGERS AND PAUL ROZIN University of Pennsylvania Thiamine-deficient rats choosing between a novel diet and a familiar diet invariably showed a marked preference for the novel diet. If the novel diet contained thiamine, this preference was maintained over a 10-day period. If the novel diet was deficient in thiamine, and thiamine was added to the familiar diet, Ss switched from an initial novel diet preference to a familiar diet preference after 3-4 days. Control rats showed no preferences for either diet. The exclusive initial ingestion of a novel diet may facilitate the development of a sustained, learned preference. with eating that particular diet. (6) Thiamine-deficient rats injected with large quantities of thiamine from 12 hr. to 10 days before being offered a choice between a deficient and a thiamine-containing diet nevertheless show a strong preference for the thiamine diet, even though the thiamine ingested orally presumably has no effect (Rozin, 1965). (c) No preference is observed when thiamine is in one of two bottles of water, in spite of the fact that thiamine produces the same physiological effects in the rat whether ingested in water or food (Rozin, Wells, & Mayer, 1964). The situation, therefore, remains unclear, insofar as the explanations of thiamine preference (as innate and as learned) seem equally inadequate. It may be, however, that while neither explanation is correct in itself, there is some truth in each. Perhaps there is an instinctive feeding pattern in the thiamine-deficient rat which maximizes the possibility of learning. An analogous explanation has been offered for the reaction of wild rats to poisoned foods (Rzoska, 1953). Rats that have been poisoned sometimes develop a reluctance to eat new foods. When they finally try a new food, they frequently take only a few nibbles and then leave the food for a period of time before eating more. This sampling method would apparently maximize the possibility of associating ill effects with a particular food. It is possible that a similar, but reversed, response to new foods is present in the thiamine-deficient rat. Our casual observations indicated that these rats re1 This research was supported by National Sci- sponded readily to novel foods, and sugence Foundation Grant GB-1489. * gested the experiment that is reported here. That thiamine (vitamin Bl) deficient rats prefer foods containing thiamine has been known for years, but very little is known about the mechanism of the preference. Some of Richter's work (e.g., Richter, Holt, & Barelare, 1937) suggests that many preferences based on dietary deficiencies may involve innate recognition of the needed component. One difficulty with this explanation is the failure of an immediate response to thiamine to appear reliably in deficient rats. Furthermore, if thiamine is initially paired with a distinctive flavor and afterwards switched to a diet with a different flavor, thiamine-deficient rats continue to prefer the diet with which the thiamine was originally associated (Harris, Clay, Hargreaves, & Ward, 1933; Scott & Verney, 1947). Most investigators believe that thiamine preference is learned: ingestion of thiamine-rich foods makes the thiamine-deficient rat feel better. This explanation of thiamine hunger by a simple reinforcement mechanism also runs into a number of difficulties, (a) Ingestion of thiamine would not have any beneficial effects for at least 10 min., and probably not for 30 min., and learning does not ordinarily take place with such a long delay of reinforcement. Furthermore, a normal rat in a preferencetesting situation frequently samples from the dietary choices available. Unless there is a mechanism operative in the deficient rat which results in an exclusive choice of the thiamine-containing diet, thiamine's beneficial effects would not be associated WILLARD RODGERS AND PAUL ROZIN DEFICIENT GROUP NOVEL DIET 1 THIAMINE IN NOVEL DIET THIAMINE IN FAMILIAR DIET recorded daily and also frequently during the first 8 hr. on the first day. The general procedure used was identical to that of Rozin, Wells, and Mayer (1964). The experiment was repeated with four additional groups of three Ss, using a second novel diet very different from both the first novel diet and the standard diet so as to test the generality of the results. RESULTS The results of the basic experiment are shown in Figures 1 and 2. Deficient rats in both groups showed an unequivocal, immediate, initial preference for the novel diet, regardless of whether or not it contained thiamine. The preference appeared clearly within 15 min. of first presentation. Very FIG. 1. Preference for first novel diet by thia- little, if any, familiar diet was eaten in the mine-defioient rats over a 10-day experimental first 8 hr. The avid response of the deficient period. [Preference is expressed as: (novel diet rats to the novel food was in marked coneaten/total diet eaten) X 100 for each day.] Each trast to their anorexia preceding presentablock represents the daily preferences of one rat tion of the novel diet. The deficient rats for 10 days. which were given a choice between a novel diet with thiamine and the standard diet METHOD without thiamine maintained the prefFour groups of three rats (Charles River, female, CD Sprague-Dawley strain) were fed a erence for the novel diet throughout the standard synthetic thiamine-deficient diet for the 10-day choice period. The deficient rats first 3 wk. of the experiment.2 All Ss were 21-day- which were given a choice between a novel old weanlings at the beginning of the experiment. diet without thiamine and the standard Six Ss (in the two control groups) were injected diet with thiamine ate very small amounts with thiamine subcutaneously throughout the experiment, so that they never became deficient; of food (about 5 gm/day) during the first however, the amount of food given to them each few days, since they were eating the novel day during the first 3 wk. was restricted so as to deficient diet almost exclusively. They keep their weight equal to that of the deficient showed a sharp change in diet preference on rats. At the end of 3 wk. on deficient diet, the uninjected & were losing weight rapidly, a clear sign of thiamine deficiency. Each S in the four groups was then offered a choice between the standard (familiar) diet and a novel diet for 10 days. For one group of deficient Ss and one control group, thiamine (20 jug/gm) was added to the novel diet but not to the standard diet; for the other two groups, thiamine was added to the standard diet but not to the novel diet. Food intakes of both diets were 2 All diets were mixed in the laboratory from standard synthetic components and contained 1% thiamine-deficient vitamin mix in sucrose (General Biochemicals) and 4% Hegsted salt mix. The standard diet contained, in addition, 3% Mazola, 2% cod liver oil, 65% sucrose, 25% casein. The first novel diet contained 10% peanut oil, 45% sucrose, 10% starch, 15% casein, and 15% soy protein. The second novel diet contained 10% Mazola, 40% sucrose, 15% casein, and 30% gelatin. In thiamine-containing diets, the thiamine concentration was always 20 /ig/gm. PAIR-WEIGHED CONTROL GROUP NOVEL DIET 1 THIAMINE IN NOVEL DIET THIAMINE IN FAMILIAR DIET 10 FIG. 2. Preference for first novel diet by control rats, plotted as in Figure 1. NOVEL FOOD PREFERENCES the fourth choice day, selecting the standard diet with added thiamine on each of the last 7 days, and showing a marked increase in total daily food intake. The rats in the control groups showed no significant preference for either the standard or the novel diet throughout the 10-day choice period. In the repetition of the experiment with another novel diet, a large initial preference for the novel diet was again shown by both groups of deficient rats, whereas a slight avoidance of this diet was shown by the two groups of control rats (Figures 3 and 4). During the 10-day choice period, each control group showed a three-to-one preference for the standard diet. The deficient group whose novel diet contained thiamine showed a two-to-one preference for the novel diet. The deficient group whose standard diet contained thiamine showed a four-to-one preference for the novel diet on each of the first 3 days, but on the last 6 days a preference for the standard diet. DISCUSSION This experiment shows that thiamine-deficient rats change their feeding patterns and invariably prefer novel diets. The preference for a new diet may reflect either an increased desire for novel foods or an acquired aversion to the familiar diet. Such a preference for new foods by deficient rats DEFICIENT GROUP NOVEL DIET 2 THIAMINE IN NOVEL DIET 25 THIAMINE IN FAMILIAR DIET 28 FIG. 3. Preference for second novel diet by thiamine-deficient rats, plotted as in Figure 1. (Rat 29 was almost dead by Day 5 as he ingested very little thiamine, so the experiment was terminated.) PAIR-WEIGHED CONTROL GROUP NOVEL DIET 2 THIAMINE IN NOVEL DIET THIAMINE IN FAMILIAR DIET 31 34 FIG. 4. Preference for second novel diet by control rats, plotted as in Figure 1. in the natural situation would maximize the likelihood of encountering the needed substance. Given that this initial preference for novel diets is innate, the question remains as to what maintains the preference, or what causes the rats to revert to the familiar diet when thiamine has been added to it. It is here that a learning mechanism may have its function. When the deficient rat is confronted with a choice between a novel and a familiar diet, it initially prefers the novel diet. If this novel diet contains the needed thiamine, the beneficial effects of ingesting this diet may reinforce the novel preference. The fact that the rat eats almost exclusively from the novel diet, in conjunction with the distinctive characteristics of .this diet, makes it more reasonable to assume that the beneficial effects of the ingested thiamine will reinforce the particular response of eating this diet. In the first few hours after the first meal, the effects of previously ingested thiamine should result in gradual recovery of the deficient rat. Thus recovery would be taking place during the subsequent meal periods when the rat is again consuming the novel diet. If the novel diet contains no thiamine, a preference for the familiar diet develops after 3-4 days. The mechanism underlying this shift is still unknown, but it is probably some process independent of the novelty of the diet. WILLARD RODGERS AND PAUL ROZIN In most specific hunger experiments, the deficient rats have been offered a choice between the old, familiar diet and a slightly novel, new, vitamin-rich diet. On the basis of the results of the present experiment, it appears that the standard specific hunger paradigm fails to indicate whether the needed substance is preferred because of its novelty or because of its specific properties. In other words, it is not clear that most specific hungers have really been shown to be specific. Adequate controls have indicated true specificity in sodium hunger. (Nachman, 1962), but similar controls have not been used in most other cases. These experiments raise several interesting questions. For instance, is the needed component (in this case thiamine) more effective than an arbitrary substance as a novel stimulus? Must the thiamine be part of the novel diet to maintain the initial preference, or would injections of thiamine be just as effective? Do deficiencies in other substances also result in a novel diet preference? Is the failure of a specific hunger for thiamine to appear in a liquid choice situation related to the absence of a novelty response to liquids? Does this novelty preference explain the preference for thiamine diets by rats previously recovered from thiamine deficiency? Is the change in response to novelty limited to foods, or does it extend to other aspects of the environment? REFERENCES HARRIS, L. J., CLAY, J., HARGHEAVES, F., & WARD, A. Appetite and choice of diet. The ability of the vitamin B deficient rat to discriminate between diets containing and lacking the vitamin. Proc. Roy. Soc. London, Ser. B., 1933, 113,161-190. NACHMAN, M. Taste preferences for sodium salts by adrenalectomized rats. J. comp. physiol. Psychol, 1962, 55, 1124-1129. RICHTEH, C. P., HOLT, L. E., & BARBLARB, B., JR. Vitamin B craving in rats. Science, 1937, 86, 354. ROZIN, P. Specific hunger for thiamine: Recovery from deficiency and thiamine preference. J. comp. physiol. Psychol., 1965, 59, 98-101. ROZIN, P., WELLS, C., & MAYER, J. Specific hunger for thiamine: Vitamin in water versus vitamin in food. J. comp. physiol. Psychol., 1964, 57, 78-84. RZOSKA, J. Bait shyness, a study in rat behavior. Brit. J. anim. Behav., 1953, 1, 128-135. SCOTT, E. M., & VERNEY, E. L. Self-selection of diet: VI. The nature of appetites for B vitamins. J. Nutr., 1947, 34, 471-480. (Received March 1, 1965)