A colony of harbour seals in the Pacific north-west was monitored over two years, concurrent with a variety of environmental variables. Regression models described diel hauling-out activity as: i) a photoperiodic cycle; ii) a function of other environmental factors; or iii) a cycle modified by environmental constraints. Throughout the year, the number of seals on shore followed a diel pattern with a midday peak. Seals hauled-out in lower numbers in winter than in summer, and for a smaller proportion of the day (although for about the same proportion of the photoperiod). During the annual moult, numbers hauled were elevated around the clock, and the midday peak was skewed to late afternoon/early evening. Models that defined hauling-out in terms of environmental factors were significant, but did not fit the data as well as models based on photoperiod. The strongest environmental correlates (such as tidal height) owed much of their explanatory power to artefactual similarities with the photoperiodic cycle. Four general conditions are presented which, if met, should always result in a diel hauling-out cycle with a midday peak. The most fundamental of these involves a proposed ‘cost of immersion’which motivates pinnipeds to haul-out when not foraging. Two likely candidates for such a cost involve risk from aquatic predators and the energetic expense of sleeping while immersed.