SUMMARYOscars are often subjected to a combination of low levels of oxygen and fasting during nest-guarding on Amazonian floodplains. We questioned whether this anorexia would aggravate the osmo-respiratory compromise. We compared fed and fasted oscars (10-14 days) in both normoxia and hypoxia (10-20 Torr, 4 h). Routine oxygen consumption rates (Ṁ O2 ) were increased by 75% in fasted fish, reflecting behavioural differences, whereas fasting improved hypoxia resistance and critical oxygen tensions (P crit ) lowered from 54 Torr in fed fish to 34 Torr when fasting. In fed fish, hypoxia reduced liver lipid stores by approximately 50% and total liver energy content by 30%. Fasted fish had a 50% lower hepatosomatic index, resulting in lower total liver protein, glycogen and lipid energy stores under normoxia. Compared with hypoxic fed fish, hypoxic fasted fish only showed reduced liver protein levels and even gained glycogen (+50%) on a per gram basis. This confirms the hypothesis that hypoxia-tolerant fish protect their glycogen stores as much as possible as a safeguard for more prolonged hypoxic events. In general, fasted fish showed lower hydroxyacylCoA dehydrogenase activities compared with fed fish, although this effect was only significant in hypoxic fasted fish. Energy stores and activities of enzymes related to energy metabolism in muscle or gills were not affected. Branchial Na + uptake rates were more than two times lower in fed fish, whereas Na + efflux was similar. Fed and fasted fish quickly reduced Na + uptake and efflux during hypoxia, with fasting fish responding more rapidly. Ammonia excretion and K + efflux were reduced under hypoxia, indicating decreased transcellular permeability. Fasted fish had more mitochondria-rich cells (MRC), with larger crypts, indicating the increased importance of the branchial uptake route when feeding is limited. Gill MRC density and surface area were greatly reduced under hypoxia, possibly to reduce ion uptake and efflux rates. Density of mucous cells of normoxic fasted fish was approximately fourfold of that in fed fish. Overall, a 10-14 day fasting period had no negative effects on hypoxia tolerance in oscars, as fasted fish were able to respond more quickly to lower oxygen levels, and reduced branchial permeability effectively.