The relationship between energy density (ED) of food and drink consumption ad libitum and energy intake (EI) was analysed. EI was taken as average daily EI over the long term, and as EI during a single meal. Moreover, the distribution of EI over three ED categories was analysed. Average daily EI was related to ED of the food and drinks when ED was strongly influenced by specific macronutrients. When ED was strongly influenced by the weight of water, it was not related to EI. During a meal subjects monitored mainly weight, and to a lesser extent, the energy content of the food ingested. Therefore, covertly manipulated ED of a meal affected EI directly. The impact of ED on EI was modulated by dietary behaviours such as restraint. Overt manipulation of ED for 6 months showed that EI was adjusted to a decreased but not to an increased ED in dietary-unrestrained subjects, and that EI was adjusted to an increased but not to a decreased ED in dietary-restrained subjects. Knowledge of ED was shown to lead to an inverse relationship between portion sizes and ED during a meal. Average daily EI consisted of a distribution of EI over the three different categories of ED, so that obese women ate more of foods with a high ED and less of foods with a low ED compared with normal weight women (and nutritional guidelines). In conclusion, ED affected daily EI by means of macronutrient specific effects. EI from a meal with an unknown ED can become inversely related to EI through learning or conditioning. Therefore, the effect of ED on EI during a single meal observation cannot be extrapolated directly to the 24 h effect on EI. With regard to the treatment of obesity, a conscious decreased consumption of foods high in ED and an increase in consumption of low-ED food is necessary to decrease and subsequently maintain body weight, particularly in subjects with a sedentary lifestyle.