19Previous research has shown that non-human animals exhibit an inverted-U pattern of sweet 20 preference, with consumption increasing across moderate levels of sweetness and then 21 declining for high levels of sweetness. In rodents, this pattern reflects an avoidance of the 22 postingestive effects of consuming energy-dense sugar solutions (conditioned satiation). 23Here, we examined whether humans also adjust their preferences to compensate for the 24 anticipated energy content / satiating outcomes of consuming sweetened foods. In two 25 experiments (each N = 40), participants were asked to taste and imagine eating small (15 g) 26 and large (250 g) portions of five novel desserts that varied in sweetness. Participants 27 evaluated the desserts' expected satiety, expected satiation, and expected sickliness. A 28 measure of estimated energy content was also derived using a computerized energy 29 compensation test. This procedure was completed before and after consuming a standard 30 lunch. Across both experiments, results confirmed that participants preferred a less sweet 31 dessert when asked to imagine eating a large versus a small portion, and when rating the 32 dessert in a fed versus fasted state. We also obtained evidence that participants anticipated 33 more energy from the sweeter desserts (even in Experiment 2 when half of the participants 34 were informed that the desserts were equated for energy content). However we found only 35 partial evidence for anticipated satiation-expected sickliness was related systematically to 36 increases in sweetness, but expected satiation and expected satiety were only weakly 37 influenced. These findings raise questions about the role of sweetness in the control of food 38 intake (in humans) and the degree to which 'sweet-calorie learning' occurs in complex 39 dietary environments where sweetness may actually be a poor predictor of the energy 40 content of foods. 41