^*Department of Psychology, University of Toronto, Toronto, Ontario, Canada

Correspondence: C. Peter Herman, Department of Psychology, University of Toronto, 100 St. George Street, Rm 4020–4th Floor, Toronto, Ontario M5S 3G3 Canada. E-mail: herman@psych.utoronto.ca

In their clever experiment, Wansink et al. (1) provide a dramatic illustration of the importance of visual cues in the control of food intake. As they put it, aphoristically, "people use their eyes to count calories and not their stomachs." Thanks to the Rube Goldbergesque imperceptibly self-refilling bowl, the experimental participants—through the clear evidence of their own senses—were misled into believing that they had eaten less than they really had. As a result, they consumed a full 73% more soup than did participants who ate from a normal soup bowl, without realizing it and without feeling any fuller. The ancient notion of the "wisdom of the body," in which calories are regulated automatically through hormonal/neuronal mechanisms and negative-feedback loops, simply cannot accommodate the fact that trick soup bowls can so easily fool the eater.

Wansink et al. (1) argue that, rather than keeping track of how much we eat—monitoring mouthfuls or spoonfuls or even calories, we turn this difficult and annoying task over to proxies. For instance, we, or at least 61% of us, according to the supplementary data of Wansink et al., are habitual plate-cleaners; in effect, we accept the portion on our plate as an appropriate amount to eat, and, therefore, we eat the full portion, monitoring our intake only to the extent of stopping when the plate is empty. Because our portions have grown larger (2), we eat more, without particularly realizing it, thereby doing our part to promote the current obesity epidemic. The bottomless bowl study raises some perplexing questions. Importantly, what exactly was the visual cue that the experimental participants used to terminate eating? Not the empty bowl; none of them got that far; even the normal bowl was replenished by the experimenter when only 25% of the soup remained. The eaters ended up, on average, eating (or leaving) one-half, which Wansink et al. suggest may have been their target when eating from a large (18-oz) bowl. However, clearly several people in the normal bowl condition ate at least 75% of the soup—triggering a refill—so the one-half mark was not their original target. When do people target the entire portion and when do they target only one-half? Will they change their minds partway through the meal if they realize that consuming a full portion is not possible?

Finally, it may be argued that, in this study, there was inadequate time for satiety cues to accumulate and feedback to stop intake. If the meal had been interrupted and later resumed, perhaps participants would have received more reliable information from their brain–gut axis about how sated they really were. Possibly—but in the real world, we eat our meals without the sort of interlude that might encourage useful internal feedback. Accordingly, the size of the original portion dictates our intake, and when that portion is excessive, our intake will be excessive too. If only our bowls and plates were programmed to imperceptibly empty themselves!