Predictions of models for the optimal allocation of time over the dive cycle for divers using aerobic and anaerobic respiration, were tested experimentally on the tufted duck, Aythya fuligula, and pochard, A. ferina. Patterns in the dive cycle were highly correlated with water depth. In both species, foraging time increased with water depth up to 3 m. In the pochard, foraging time subsequently declined before increasing again at around 4 m. Qualitatively similar patterns in foraging time are predicted by the model based on the use of aerobic and anaerobic respiration, the 'mixed metabolism' model. As predicted by this model and previous 'aerobic' models, foraging time decreased in response to increasing substrate depth (used to increase foraging costs), while surface time was not affected. Paddle rates, used as an indicator of foraging costs, however, did not increase. In response to lower water temperature (used to increase the energetic costs of both the travel and foraging phases of the dive), surface times increased as predicted, except at the shallowest depth class. The predicted relationship between foraging time and diving costs is complex, being dependent on water depth and a number of other parameters; the observed trends were non- significant. Ascent duration decreased with decreasing temperature, but descent duration did not change significantly. Estimates of paddle rates and travel speeds suggest that foraging is on average more costly than travelling (ascent and descent combined), diving costs decrease with depth and travel speeds change subtly with depth.