Both morphology and behaviour develop during the pelagic larval stage of demersal teleost fishes. Demersal perciform fishes from warm-water habitats begin their pelagic larval stage as plankton but end it as nekton, with behavioural capabilities (including swimming, orientation and sensory abilities) that can influence, if not control, dispersal trajectories. The ontogeny of these behaviours, and the gradual transition from plankton to nekton, are central to understanding how larval fishes can influence dispersal and how behaviour can be integrated into dispersal models. Recent
behavioural research shows that, from about 5 to 8 mm standard length, larvae of warm-water perciform fishes can directly influence dispersal, because they swim in an efficient inertial hydrodynamic
environment, can swim for kilometres at speeds that heuristic models show will alter dispersal trajectories, can swim faster than ambient currents before settlement, can orientate in the pelagic environment and can detect sensory cues (light, sound, odour) that allow orientation. Fish larvae also control their vertical position (which may change temporally, spatially and ontogenetically), allowing indirect influence on dispersal. Most research on larval behaviour relevant to dispersal (i.e. swimming, orientation and sensory abilities) has been done with warm-water perciform species. This invites the question: Will the same be found in cool water or in species of other orders? The hydrodynamic and physiological effects of temperature indicate that larvae in warm water should swim more efficiently and initially at smaller sizes than larvae in cool water. Limited evidence suggests that larvae of perciform fishes are more behaviourally competent and attain morphological and behavioural milestones when smaller (and probably younger) than do larvae of clupeiform, gadiform and pleuronectiform (CGP) fishes. Perciform fishes dominate demersal fish communities in warm water,
whereas CGP fishes dominate in cooler waters. These hydrodynamic, physiological, ontogenetic, phylogenetic and biogeographic factors imply that larval fish behaviour may have more influence on dispersal in warm seas than in cool seas. This hypothesis requires testing. Additional factors that should be taken into account when using behaviour of larvae to produce biophysical models of dispersal are discussed.