Abstract: 1. Diel vertical migrations (DVM) are typical for many cold-water fish species such as Pacific salmons (Oncorhynchus spp.) and coregonids (Coregonus spp.) inhabiting deep lakes. A comprehensive recent overview of DVM in freshwater fish has not been available, however.
2. The main proximate trigger of DVM in freshwater fish is the diel change in light intensity, with declining illumination at dusk triggering the ascent and the increase at dawn triggering the descent. Additional proximate cues are hydrostatic pressure and water temperature, which may guide fish into particular water layers at night.
3. Ultimate causes of DVM encompass bioenergetics efficiency, feeding opportunities and predator avoidance. None of these factors alone can explain the DVM in all cases. Multi-factorial hypotheses, such as the âantipredation windowâ combined with the thermal niche hypothesis, are more likely to explain DVM. It is suggested that planktivorous fish move within a layer sufficiently well illuminated to capture zooplankton, but too dark for predators to feed upon the migrating fish. In complete darkness, fish seek layers with a temperature that optimises bioenergetics efficiency. The strength of each factor may differ from lake to lake, and hence system-specific individual analyses are needed.
4. Mechanistic details that are still poorly explored are the costs of buoyancy regulation and migration, the critical light thresholds for feeding of planktivorous and piscivorous fish, and predator assessment by (and size-dependent predation risk of) the prey fish.
5. A comprehensive understanding of the adaptive value of DVM can be attained only if the behaviour of individual fish within migrating populations is explicitly taken into account. Size, condition and reproductive value differ between individuals, suggesting that migrating populations should split into migrants and non-migrants for whom the balance between mortality risk and growth rate can differ. There is increasing evidence for this type of partial DVM within populations.
6. Whereas patterns of DVM are well documented, the evolution of DVM is still only poorly understood. Because experimental approaches at realistic natural scales remain difficult, a combination of comprehensive data sets with modelling is likely to resolve the relative importance of different proximate and ultimate causes behind DVM in fish.