Some of the main themes in this review are as follows. 1. The notion that non-photic zeitgebers are weak needs re-examining. Phase-shifts to some non-photic manipulations can be as large as those to light pulses. 2. As well as being able to phase-shift and entrain free-running rhythms, non-photic events have a number of other effects: these include after-effects of entrainment, period changes, and promotion of splitting. 3. The critical variable for non-photic shifting is unknown. Locomotor activity is more likely to be an index of some other necessary state rather than being causal itself. This index may be better when tendencies to move are channelled into easily measured behaviours like wheel-running. 4. Given ignorance about the critical variable, quantification of activity may be the best presently available measure of zeitgeber intensity. Therefore, the behaviour during non-photic manipulations must be examined as carefully as the shifts themselves. When no phase-shifting follows manipulations such as IGL lesions or serotonin depletion, if the animals are inactive, then little can be inferred. 5. Lack of information on the critical variable(s) for non-photic shifting makes it problematic to compare data from studies using different non-photic manipulations. However, the presence of locomotor activity (or its correlate) does appear to be necessary for triazolam to produce shifts. 6. Novelty-induced wheel-running in hamsters depends on the NPY projection from the IGL to SCN. It remains to be determined how important NPY is in other species or in clock-resetting by other manipulations, but methods are now available to study this. 7. Interactions between photic and non-photic zeitgebers remain virtually unexplored, but it is evident that photic and non-photic stimuli can attenuate the phase-shifting effects of each other. Interactions are not purely additive or predictable from PRCs. 8. The circadian system does more than synchronize free-running rhythms to the solar day. Its non-photic functions and their interactions with photic inputs probably account for some of the anatomical complexity of circadian circuitry.