This study assessed the reproducibility of electrically evoked, isometric quadriceps contractile properties in eight people with spinal cord injury (SCI) and eight able-bodied (AB) individuals. Over all, the pooled coefficients of variation (CVps) in the SCI group were significantly lower (ranging from 0.03 to 0.15) than in the AB group (ranging from 0.08 to 0.21) (P<0.05). Furthermore, in all subjects, the variability of force production increased as stimulation frequency decreased (P<0.01). In subjects with SCI, variables of contractile speed are clearly less reproducible than tetanic tension or resistance to fatigue. Contractile properties of quadriceps muscles of SCI subjects were significantly different from that of AB subjects. Muscles of people with SCI were less fatigue resistant (P<0.05) and produced force-frequency relationships that were shifted to the left, compared with AB controls (P<.01). In addition, fusion of force responses resulting from 10 Hz stimulation was reduced (P<.05) and speed of contraction (but not relaxation) was increased (P<0.05), indicating an increased contractile speed in paralysed muscles compared with non-paralysed muscles. These results correspond with an expected predominance of fast glycolytic muscle fibres in paralysed muscles. It is concluded that quadriceps dynamometry is a useful technique to study muscle function in non-paralysed as well as in paralysed muscles. Furthermore, these techniques can be reliably used, for example, to assess therapeutic interventions on paralysed muscles provided that expected differences in relative tetanic tension and fatigue resistance are larger than approximately 5% and differences in contractile speed are larger than approximately 15%.