SUMMARY1. Alechanical factor(s) associated with the initiation of eccentric contractioninduced muscle injury were investigated in isolated rat soleus muscles (n = 180; 42 protoc ols with 4-6 muscles per protocol). Five eccentric contractions were performed with 4 min between contractions. Three levels of peak eccentric contraction force (100, 125 and 150% of pre-injury maximal isometric tetanic tension. P0), length change (041. 0 2 and 0 3 muscle length, Lo) and lengthening velocity (0(5, 1P0 and 1P5 Lo/s) were utilized. Force was varied with stimulation frequency (10-150 Hz). The eccentric contractions were initiated at muscle lengths of 0(85 or 0(90 Lo.Following the fifth eccentric contraction, the muscle was incubated in Krebs-Ringer buffer for 60 min. Peak isometric twitch tension (PT), PO, maximal rate of tension development (+dPI/dt), maximal rate of relaxation (-dP/dt), and creatine kinase (CK) release were measured prior to the five eccentric contractions and at 15 min intervals during the incubation period. Total muscle [Ca2+] was measured after 60 min incubation.2. The mean (+S.E.M.) initial decline in P0 for the muscles performing the most injurious protocol was 13-6 + 4-8 % (n = 6); P0 in control muscles immediately following performance of five isometric contractions was elevated 1P2 + 1 0 % (n = 8). These means were different at probability. p = 0(005. Mean [ATP] G. L. WARREN AND OTHERS control muscles but the elevation was unrelated to any of the four mechanical factors. 4. These data support the hypothesis that eccentric contraction-induced injury is initiated by mechanical factors, with muscle tension playing the dominant role. They also demonstrate that specific mechanical factors differentially affect the various injury criteria, i.e. reductions in contractile performance were most related to produced forces, and CK release was most related to lengthening velocity.