The analysis of foraging behaviour with optimization models has been used widely in recent years. Studies on nectar collection in the honeybee have been among the examples cited repeatedly in favour of the approach. These studies (based on the "energy-maximizing" hypothesis) have now been challenged in a recent paper by Varjfi and N6fiez (1991). These authors claim that the observed behaviour can be explained without reference to this hypothesis and even without energetic considerations. Moreover, they maintain that the observations support the "information-exchange" hypothesis, an alternative preferred by the authors. In this paper the arguments used by Varjfi and Nfifiez (1991) are shown to be erroneous in several respects. A recurrent problem results from confusing causal and functional questions. In addition, their analysis fails to distinguish between statistically inferred and normative models, misinterprets the nature of observations and misrepresents the predictions of the energy-maximizing hypothesis. Whereas the information-exchange hypothesis is a plausible alternative, the evidence cited by the authors cannot distinguish between the two competing explanations but rather suggests that energy is important. Varjfi and Nffiez (1991) do not specify the meaning of "information" so that their hypothesis cannot generate specific predictions.