The best general hypothesis for the control of 'spontaneous' tidal and daily patterns of behaviour in coastal animals postulates an endogenous physiological pacemaker system which generates approximate periodicity, together with environmental adjustment of the clock(s) to local time. Free-running endogenous rhythms of circatidal, circadian, circasemilunar and circalunar periodicity have been demonstrated in a number of species in constant laboratory conditions, in some cases clarifying hitherto poorly understood aspects of the behavioural repertoire of animals in the sea. Entrainment of circatidal rhythmicity has been demonstrated using cycles of simulated tidal variables such as temperature, hydrostatic pressure, salinity and wave action. The crab Carcinus shows increased locomotor activity after changes of salinity (halokinesis); responses to 34% salinity entrain the endogenous clock, but responses to salinities above or below 34% are purely exogenous and do not persist in constant conditions after entrainment. Phase responsiveness of circatidal rhythms to pulses of tidal variables has been demonstrated in several species; phase response curves show marked differences from those of circadian rhythms. The endogenous basis of tidal and diel behaviour in marine molluscs and crustaceans involves matching spontaneous rhythms of neuroelectrical activity. Also, in decapod crustaceans a peptidic neurodepressing hormone (NDH) modulates neuroelectrical and behavioural rhythmicity. NDH is produced rhythmically in the eyestalk neurosecretory complex, perhaps partly under the control of other clock components elsewhere in the CNS. The physiological basis of circasemilunar, lunar (and annual) rhythms of behaviour has not been studied, but studies of synchronization of these rhythms have been undertaken. In some localities it has been shown experimentally that light intensities equivalent to moonlight are sufficient to entrain such rhythms. In other localities where moonlight is a less reliable cue the relative timing of tidal and daily variables has been shown to be important. So far there is no evidence that synchronization is achieved by absolute differences between tidal variables at neap and spring tides.