Critical (<30min) and prolonged (>60min) swimming speeds in laboratory chambers were determined for larvae of six species of Australian freshwater fishes: trout cod Maccullochella macquariensis, Murray cod Maccullochella peelii, golden perch Macquaria ambigua, silver perch Bidyanus bidyanus, carp gudgeon Hypseleotris spp. and Murray River rainbowfish Melanotaenia fluviatilis. Developmental stage (preflexion, flexion, postflexion and metalarva) better explained swimming ability than did length, size or age (days after hatch). Critical speed increased with larval development, and metalarvae were the fastest swimmers for all species. Maccullochella macquariensis larvae had the highest critical [maximum absolute 46Â·4cms-1 and 44Â·6 relative body lengths (LB) s-1] and prolonged (maximum 15Â·4cms-1, 15Â·6 LB s-1) swimming speeds and B. bidyanus larvae the lowest critical (minimum 0Â·1cms-1, 0Â·3 LB s-1) and prolonged swimming speeds (minimum 1Â·1cms-1, 1Â·0 LB s-1). Prolonged swimming trials determined that the larvae of some species could not swim for 60min at any speed, whereas the larvae of the best swimming species, M. macquariensis, could swim for 60min at 44% of the critical speed. The swimming performance of species with precocial life-history strategies, with well-developed larvae at hatch, was comparatively better and potentially had greater ability to influence their dispersal by actively swimming than species with altricial life-history strategies, with poorly developed larvae at hatch.