may have a significant effect over and above seasonal variations. The consumer survey (unpublished, 1991) yielded a strong consensus about the effect of air-conditioning on customer behavior. In the summer, "air-conditioning attracts customers and after sitting in the restaurant, many order 'normally'.... If we have an air-conditioning breakdown, sales drop dramatically." The summer peak in ice cream sales is much more noticeable in street outlets than in mall outlets, although it is unclear whether this is because the malls are air-conditioned in the summer or because they are heated in the winter. One ice cream chain blamed the increasing prevalence of air-conditioning for its slower gain in sales in recent years. It is worth noting here that there is some lingering confusion about whether the effect of air-conditioning on appetite (countering heat-induced appetite changes) depends on acute effects (that is, air-conditioning at the eating site) or more chronic effects (that is, exposure to air-conditioning through much of the day). Presumably the answer to this question depends on whether chronic exposure to air-conditioning counteracts a heat-induced drop in BW set-point.

It is not clear what happens in an environment that is naturally hot much of the time but which cools off dramatically at other times. Such dramatic cooling may occur naturally at nighttime, or even during the day, with the advent of air-conditioning. How responsive is appetite to abrupt shifts in environmental temperature? Does one's appetite pick up when leaving the broiling heat to enter an air-conditioned dining room? Or does the pressure to dissipate heat carry over even in the air-conditioned environment? The same questions can be asked when substituting "cool nights" for "air-conditioning." One study, albeit on rats, speaks at least indirectly to this question. Refinetti (1988) examined feeding in rats that were housed in normal (29°C) or cold (19°C) conditions and fed in a separate chamber that was either normal temperature or cold. Both housing and feeding environment temperatures additively affected appetite; thus, the temperature that obtains when eating occurs does affect eating, but there is also some carryover from the external environment. One possibly significant finding in this study is that animals who went to a cold environment to feed gained much more weight than did animals who remained in a warm environment to feed. The finding suggests that if one spends most of one's time in the heat but eats in an artificially cooled environment, one might end up eating more than needed, with potential problems for heat dissipation when one returns to the hot environment.

Research on the effects of variations in environmental temperature on feeding was stimulated by Brobeck's (1948) hypothesis. He found that rats' food intake dropped precipitously as the environmental temperature rose from 18° to 36°C; these rats, acclimated to a temperature