The purpose of this study was to assess and compare eccentric (ECC) and concentric (CONC) torque output of the quadriceps and hamstring muscles and to analyze the effect of gravity effect torque (GET) correction on the calculation of the hamstring/quadriceps peak torque quotient (H/Q quotient). Twenty female subjects performed maximal voluntary CONC and ECC contractions of the quadriceps and hamstring muscles at five isokinetic lever arm velocities from 60 degrees/s to 360 degrees/s. Peak torque was measured and corrected for GET. Mean ECC torque did not significantly change with increasing ECC velocity for either the quadriceps or hamstring muscles (P greater than 0.05). Mean CONC torques were significantly lower than the corresponding ECC torques (P less than 0.05) and decreased with increasing CONC velocity. At each test velocity, the CONC H/Q quotient was significantly lower than the corresponding ECC H/Q quotient (P less than 0.05). Mean H/Q quotients did not significantly change with increasing velocity for either the CONC or ECC tests (means: 0.46 and 0.57; P greater than 0.05). Mean H/Q quotients not corrected for GET significantly increased with increasing velocity for the CONC (0.61 to 0.78; P less than 0.05), but not ECC tests (0.66 to 0.71; P greater than 0.05). The results indicate that the ECC torque-velocity curve is essentially level for both quadriceps and hamstring muscles. The present findings point strongly toward the necessity of correcting for GET when calculating both CONC and ECC H/Q quotients.(ABSTRACT TRUNCATED AT 250 WORDS)