SHORTCOMINGS OF FOOD CONSUMPTION AS A MEASURE OF HUNGER; RESULTS FROM OTHER BEHAVIORAL TECHNIQUES

I have been tremendously impressed by the work so effectively reported a t the Conference on which this monograph is based. It is interesting to note, however, that almost all of the studies referred to have used as their index of hunger a single technique, namely, the amount of food eaten. If one is interested in weight regulation, or the long-range balance between energy input and output, this method is an entirely appropriate measure. If one is interested in the broader problem of hunger as a drive, or in the complete mechanism of hunger and satiation, this measure has a number of limitations. It can profitably be supplemented by certain other behavioral techniques. A little thought shows us that the amount of food eaten does not necessarily measure the maximum intensity of the hunger. For example, a persistent, or recurring, low level of hunger which keeps the subject nibbling would cause a large consumption of food, while an intense but rapidly satiated hunger would produce a smaller consumption. Furthermore, the total consumption probably depends on the balance between two factors: hunger and the motivation to stop eating. The foregoing type of analysis is supported by a study by Miller, Bailey, and Stevenson3 on rats with the hypothalamic lesions which cause obesity. On ad lib. feeding tests, these animals ate much more food than did normal controls. But a series of behavioral tests showed that they worked less hard for food and were more easily deterred from eating it. Therefore, we concluded that, under certain conditions, the hypothalamic animals can react to food deprivation with a less intense hunger, even though they eat more on an ad lib. diet.* I shall briefly illustrate two of the behavioral measures which we have found useful in this and other studies. The one is the rate of bar pressing aperiodically reinforced by food as originally developed by Skinner6 a t Harvard and extensively used in our laboratory a t Yale. First, hungry animals are trained to press a bar which always immediately delivers a small pellet of food into a little dish below. Then the mechanism is set so that pressing the bar will deliver food only at certain unpredictable intervals. The animals continue working much like a gambler who operates a slot machine in the hope of hitting the jackpot. The rate at which they work seems to be a good measure of the strength of hunger. Furthermore, since the animals get only tiny bits of food infrequently, relatively long tests can be made without appreciably satiating hunger. The other measure involves pitting the aversion produced by quinine against appetite. Animals are presented with a series of tiny samples of food, each of