substantial variation was in carbohydrate (starch) intake, which was strongly elevated by cold and suppressed by heat, but not below the level for casein-and lard-supplemented diets (Donhoffer and Vonotzky, 1947). Two restaurant chains reported a decided shift away from sandwiches toward salads in the summer (unpublished data from consumer survey, 1991). Ice cream consumption, which peaks in the summer (unpublished data from consumer survey, 1991), seems to be an exception to the rule of heat-suppressed carbohydrate consumption; however, most observers regard the summer appeal of ice cream to reside in its coldness more than its sweetness. In fact, one ice cream chain reported that while ice cream sales tended to rise with increases in environmental temperature from 72° to 82°F, above 82°F customers switch to more thirst-quenching products (for example, ices and "light" ice creams).
Rothwell and Stock's (1986) casual observations suggested that food selection of rats offered a cafeteria diet was unaffected by variations (from 24° to 29°C) in environmental temperature. As already shown, however, the ability of protein to differentially suppress appetite has probably been overstated; accordingly, it should not be surprising if heat has little observable effect on macronutrient preferences. Ashworth and Harrower (1967) reported that proportional nitrogen loss from sweating is lower than normal in acclimatized tropical workers, again suggesting no need for supplemental protein in the diet, a conclusion seconded by Collins et al. (1971) and Weiner et al. (1972). Still, it would seem a reasonable precaution to maintain protein intake at or slightly above a nutritionally desirable minimum in hot environments, especially before full acclimatization has been achieved.
Donhoffer and Vonotzky (1947) cite well-known seasonal changes in thyroid activity as a possible mediator of the heat-induced differential suppression of carbohydrate intake. How thyroid activity might control qualitative aspects of appetite and whether humans are susceptible to such seasonal variations remain obscure.
The consumer survey yielded fairly strong data regarding shifts in food preference in the summer. As noted above, the restaurant chain specializing in salads had its sales peak during the summer. Not surprisingly, so did the ice cream parlors. Seasonal shifts in consumer preferences, however, may not be driven entirely by physiology. A number of retailers indicated that intense promotional activity of different types of foods occurs on a seasonal basis; conceivably, some of the seasonal shift in preferences actually represents conformity with expectations or financial and social pressures. One hamburger chain claimed that "consumption patterns are marketing driven";