Do good theories always come from good data?

According to conventional wisdom concerning the scientific method, good theories come from good experimental data, and bad theories from bad experimental data. Yet the history of the physics of foraging is a remarkable counterexample. To illustrate this, we briefly recount one of the important scientific investigations in the field, published in Nature in 1996. The original study of wandering albatrosses [390] inspired dozens of other studies, yet later required correction due to its spurious data.

Lévy flights of the wandering albatross

The albatross can fly great distances, at exceptional speeds. There are significant differences among species of albatross [402]. Wandering albatrosses in southern Georgia can sustain a speed in excess of 100 km/h by taking advantage of the local wind field [284]. They frequently fly 500 km per day, with an upper limit in the range 750–950 km per day. Phillips et al. [284] report that one gray-headed albatross circumnavigated the Southern Ocean in only 46 days. Because of their great mobility and large size, we decided to focus on the albatross (instead of, e.g., the sparrow) in our original study. The foraging strategy of the wandering albatross [403] stands apart from that of other seabirds [401]. Weimerskirch et al. [404] studied the distribution of prey encounters for wandering albatrosses and reported results that strongly suggest a foraging strategy that differs from those of most seabirds.