The critical swimming speed (Ucrit) of rainbow trout parr (Oncorhynchus mykiss) and three life stages of Galaxias maculatus, larval (whitebait), postlarval inanga and adult inanga, were tested at temperatures from 5oC to 25oC. All fish were swum at their acclimation temperature under normoxic conditions to determine the optimal aerobic exercise temperature. To determine whether acclimation affected swimming ability, trout parr acclimated to either 10oC or 20oC were swum at 20oC and 10oC, respectively. The potential effect of mild hypoxia (75% saturation) on trout parr and whitebait was also examined at 10oC, 15oC and 20oC, and also tested separately and in combination were the effects of mild hypoxia and severe anaemia on the prolonged swimming ability of trout smolts at temperatures from 10oC to 20oC. For all trout experiments, blood samples were taken from non-exercised and exercised fish by acute caudal venepuncture to determine haematological responses to both acclimation and exercise. Under normoxic conditions, Ucrit max for trout parr (7.0 0.5 cm fork length) was calculated to be 5.8 body lengths per second (BL s-1) at 15.1oC, but declined at lower and higher temperatures. This result implies that swimming performance was limited by temperature below 15oC, whereas performance at higher temperatures was limited by oxygen availability. In support of this hypothesis, mild hypoxia (75% saturation) had no effect at 10oC or 15oC but caused a significant reduction in Ucrit at 20oC. However, fish acclimated at 20oC showed an adaptive elevation in oxygen carrying capacity due to an increase in mean erythrocyte volume and haemoglobin content. Furthermore, acclimation to 20oC improved warm water swimming performance. Trout parr acclimated to 10oC performed significantly worse than fish acclimated to 20oC when swum at 20oC. However, trout parr acclimated to 20oC performed as well as fish acclimated to 10oC when swum at 10oC. Following exercise, haematocrit was elevated under both normoxic and hypoxic conditions. However, the primary cause of this apparent increase in oxygen carrying capacity was splenic release of erythrocytes under normoxic conditions, whereas stress-induced erythrocytic swelling contributed to the observed increase in hypoxia. This contrasting response was most pronounced at 10oC. Larval whitebait (4.7 - 5.0 cm total length (TL)) also showed a temperature dependence of prolonged swimming ability with Ucrit max calculated to be 5.1 BL s-1 at 17.7oC. Hypoxia significantly reduced Ucrit at 15oC and 20oC, lowering the optimal aerobic temperature to 13.9oC and reducing Ucrit to 4.2 BL s-1. Mild hypoxia therefore had a more pronounced impact on inanga whitebait than trout. Postlarval inanga (3.9 - 4.0 cm TL) performed poorly at higher temperatures with Ucrit max of 5.6 BL s-1 at 9.4oC indicating an ontogenetic change in swimming ability, possibly resulting from a developmental shift in red muscle kinetics or a greater dependence on anaerobic muscle. Adult inanga (5.5 - 6.8 cm TL) prolonged swimming ability showed similar temperature dependence to that of inanga whitebait but lower relative swimming speeds due to their larger size. The dramatic decline in performance exhibited by juveniles at warmer temperatures was not apparent in adults. Ucrit max for adults was 4.0 BL s-1 at 18.3oC. The critical swimming speed of trout smolts, subjected to mild hypoxia (6.8 mg