OBJECTIVE: To study the role of alimentary sensory stimulation on the thermic effect of food. SUBJECTS: Nine male healthy volunteers (age: 20 ± 34 y, body mass index (BMI): 17.4 ± 25.3 kgam 2 ). DESIGN AND MEASUREMENTS: Four experimental situations were investigated on different days. Subjects: 1) ate a four course meal containing 2582 kJ (meal), 2) saw, smelt and tasted the same courses (alimentary sensory stimulation), 3) saw, smelt and tasted non-alimentary substances (non-alimentary sensory stimulation), 4) received directly into the stomach the previous meal (tube feeding). Energy expenditure (EE) was measured over a 20 min rest period, then for 110 min. RESULTS: The changes over midday rest EE were: meal 12.0%; alimentary sensory stimulation 3.2%; nonalimentary sensory stimulation 72.6%; tube feeding 5.7%. The increase in EE was higher after the meal than after either alimentary sensory stimulation (P`0.01) or tube feeding (P`0.01); the increases after the last two were, in turn, greater than after non-alimentary sensory stimulation (P`0.05). Only after alimentary sensory stimulation, was the increment in EE signi®cantly correlated with BMI (r 70.700; P`0.05). CONCLUSIONS: The early phase of postprandial thermogenesis therefore depends on both sensory and metabolic events. Furthermore, the sensory component was more pronounced in the lean subjects.