This study examined the effects of progressive exercise to fatigue in normoxia (N) on muscle sarcoplasmic reticulum (SR) Ca2+ cycling and whether alterations in SR Ca2+ cycling are related to the blunted peak mechanical power output (POpeak) and peak oxygen consumption (Vo(2 peak)) observed during progressive exercise in hypoxia (H). Nine untrained men (20.7 +/- 0.42 yr) performed progressive cycle exercise to fatigue on two occasions, namely during N (inspired oxygen fraction = 0.21) and during H (inspired oxygen fraction = 0.14). Tissue extracted from the vastus lateralis before exercise and at power output corresponding to 50 and 70% of Vo(2 peak) (as determined during N) and at fatigue was used to investigate changes in homogenate SR Ca2+-cycling properties. Exercise in H compared with N resulted in a 19 and 21% lower (P < 0.05) POpeak and Vo(2 peak), respectively. During progressive exercise in N, Ca2+-ATPase kinetics, as determined by maximal activity, the Hill coefficient, and the Ca2+ concentration at one-half maximal activity were not altered. However, reductions with exercise in N were noted in Ca2+ uptake (before exercise = 357 +/- 29 mumol(.)min(-1.)g protein(-1); at fatigue = 306 +/- 26 mumol(.)min(-1.)g protein(-1); P < 0.05) when measured at free Ca2+ concentration of 2 muM and in phase 2 Ca2+ release (before exercise = 716 +/- 33 mumol(.)min(-1.)g protein(-1); at fatigue = 500 +/- 53 mumol(.)min(-1.)g protein(-1); P < 0.05) when measured in vitro in whole muscle homogenates. No differences were noted between N and H conditions at comparable power output or at fatigue. It is concluded that, although structural changes in SR Ca2+-cycling proteins may explain fatigue during progressive exercise in N, they cannot explain the lower POpeak and Vo(2) (peak) observed during H.