We studied the effects of graded hypoxia on mechanical performance of cardiac ventric-ular muscle by producing controlled, stepwise decreases in partial pressure of oxygen (Po2) in the medium bathing the kitten papillary muscle preparation and in the perfusate of the Langendorff-prepared rabbit heart. For kitten papillary muscle at 30°C and with stimulation rate at 30/min, maximum rate of contraction (+dT/dtmM) and maximum rate of relaxation (-dT/dtm.*) were 184 ± 10 inN/mm2per sec and 162 ± 12 mN/mm2per sec, respectively, during control conditions with Po2at 634 ± 7 mm Hg. Step decreases in Po2from 634 mm Hg produced decreases in steady state -dT/dtmu, that were significantly greater than corresponding decreases in +dT/dtma», except at the lowest Po2. When Po2(mm Hg ± SE) was 411 ± 10, 218 ± 4, and 92 ± 3, steady state +dT/dtmMvs. -dT/dt-, (expressed as%of pre-hypoxia control value ± SE) were: 97 ± 4 vs. 85 ± 7, 76 ± 5 vs. 59 ± 6, and 47 ± 5 vs. 28 ± 4, respectively. When the lowest Po2of 34 ± 6 mm Hg was achieved, considerable shortening of the duration of the mechanical cycle occurred, and values for +dT/dtm«x and -dT/dtma, (expressed as % of pre-hypoxia control value ± SE) of 28 ± 7 and 21 ± 7, respectively, were not significantly different. Graded hypoxia similarly affected left ventricular isovolumic pressure developed by the coronary perfused rabbit heart. In both preparations, changes in relaxation relative to changes in contraction during progressive hypoxia were biphasic: decreases in maximum relaxation rate were disproportion-ately greater than decreases in maximum contraction rate with intermediate hypoxia, but the proportionality was restored when severe hypoxia produced a decrease in cycle duration.Circ Res 47: 516-524, 1980