Behaviours with no apparent adaptive function are sometimes described as play without rigorous testing of alternative explanations. One such behaviour is the repeated dropping and catching of objects by birds. We observed drop-catch behaviour by herring gulls over 3 years at a site where the birds frequently dropped clams to break them on hard surfaces below. We tested unique suites of predictions from three hypotheses, that the drop-catch behaviour is play, that it functions to expose potential kleptoparasites, or that it serves to reposition clams before foraging drops. We tested data from 72 drop-catch series and 504 typical foraging drop series to determine which suite of predictions most closely matched our observations. As predicted if the behaviour were play, drop-catches were performed more by younger birds, not necessarily over a hard substrate, and sometimes with nonfood objects. Clams that were subjected to drop-catches were generally not repositioned or eaten. These results suggest a motivation for drop-catching that is distinct from foraging. Finally, drop-catches were more frequent when it was warm and when there were high winds, also consistent with the play hypothesis. Drop-catch behaviour closely matched the predictions of the play hypothesis and we were able to reject the two alternative explanations.