much as 45 percent. One is tempted to imagine studies in humans where the effect of heating on flavor is separated from its direct thermic effect. Heating a basically unpalatable food would presumably suppress intake substantially if it brought out the aversive flavor as well as added unwanted heat to the system. Presumably a study in which the animal or person was offered a choice between hot and cold versions of the same food might help to disambiguate these results.

Although there is obviously not an invariant connection between environmental heat and environmental light, some of the hottest environments, especially desert areas, are notable for the intensity of the light. This fact becomes relevant, perhaps, in conjunction with recent work on seasonal affective disorder (SAD), a variant of clinical depression that is seasonal in nature and, more specifically, responsive to "light therapy" (for example, exposure for a period of hours to a bright [2500 lux] full-spectrum fluorescent light). The connection between SAD and appetitive disorders has been remarked on repeatedly (for example, Rosenthal et al., 1986; Wurtman, 1988). Specifically, the depressive phase is associated with overeating, carbohydrate craving, and weight gain. Periodic exposure to bright light produced weight loss in SAD patients, although this effect was accompanied by a decrease in their surprisingly high resting metabolic rate (Gaist et al., 1990). It is tantalizing to imagine that bright sunlight might contribute to the appetite suppression observed in hot environments; however, there is essentially no evidence that normal control subjects' appetites are affected by light exposure. Rats show a transient decline in appetite when exposed to constant light (Dark et al., 1980); but rats are nocturnal feeders, so the extended presence of light would be expected to disrupt feeding briefly, independent of profound physiological changes. The relevance to humans of studies of rats' reactions to extra light is probably negligible.

The discussion above regarding thermic effects of different macronutrients suggests that, in the heat, there should be a relative suppression of the already relatively suppressed (Drewnowski et al., 1989) protein preference/intake. Johnson and Kark (1947), in their survey of wartime military nutrition, found that "regardless of environment, the percentage of proteins voluntarily chosen from the rations was practically constant." Edholm et al. (1964) concur. In mice exposed to hot and cold environments, the only