White muscle and arterial blood plasma were sampled at rest and during 4 h of recovery from exhaustive exercise in rainbow trout. A compound respiratory and metabolic acidosis in the blood was accompanied by increases in plasma lactate (in excess of the metabolic acid load), pyruvate, glucose, ammonia and inorganic phosphate levels, large elevations in haemoglobin concentration and haematocrit, red cell swelling, increases in the levels of most plasma electrolytes, but no shift of fluid out of the extracellular fluid (ECF) into the intracellular fluid (ICF) of white muscle. The decrease in white muscle pHi was comparable to that in pHe; both recovered by 4 h. Creatine phosphate and ATP levels were both reduced by 40% after exercise, the former recovering within 0.25 h, whereas the latter remained depressed until 4 h. Changes in creatine concentration mirrored those in creatine phosphate, whereas changes in IMP and ammonia concentration mirrored those in ATP. White muscle glycogen concentration was reduced 90% primarily by conversion to lactate; recovery was slow, to only 40% of resting glycogen levels by 4 h. During this period, most of the lactate and metabolic acid were retained in white muscle and there was excellent conservation of carbohydrate, suggesting that in situ glycogenesis rather than oxidation was the major fate of lactate. The redox state ([NAD+]/[NADH]) of the muscle cytoplasm, estimated from ICF lactate and pyruvate levels and pHi, remained unchanged from resting levels, challenging the traditional view of the 'anaerobic' production of lactate. Furthermore, the membrane potential, estimated from levels of ICF and ECF electrolytes using the Goldman equation, remained unchanged throughout, challenging the view that white muscle becomes depolarized after exhaustive exercise. Indeed, ICF K+ concentration was elevated. Lactate was distributed well out of electrochemical equilibrium with either the membrane potential (Em) or the pHe-pHi difference, supporting the view that lactate is actively retained in white muscle. In contrast, H+ was actively extruded. Ammonia was distributed passively according to Em rather than pHe-pHi throughout recovery, providing a mechanism for retaining high ICF ammonia levels for adenylate resynthesis in situ. Although lipid is not traditionally considered to be a fuel for burst exercise, substantial decreases in free carnitine and elevations in acyl-carnitines and acetyl-CoA concentrations indicated an important contribution of fatty acid oxidation by white muscle during both exercise and recovery.