Repetitive isometric tetanic contractions (1/s) of the canine gastrocnemius-plantaris muscle were studied either at optimal length (L-o) or short length (L-s; similar to 0.9 L-o), to determine the effects of initial length on mechanical and metabolic performance in situ. Respective averages of mechanical and metabolic variables were (L-o vs. L-s, all P < 0.05) passive tension (preload) = 55 vs. 6 g/g, maximal active tetanic tension (P-o) = 544 vs. 174 (0.38 . P-o) g/g, maximal blood flow ((Q) over dot) = 2.0 vs. 1.4 ml . min(-1) . g(-1), and maximal oxygen uptake ((V) over dot O-2) = 12 vs. 9 mu mol . min(-1) . g(-1). Tension at L-o decreased to 0.64 . P-o over 20 min of repetitive contractions, demonstrating fatigue; there were no significant changes in tension at L-s. In separate muscles contracting at L-o, (Q) over dot was set to that measured at L-s (1.1 ml . min(-1) . g(-1)), resulting in decreased (V) over dot O-2 (7 mu mol . min(-1) . g(-1)), and rapid fatigue, to 0.44 . P-o. These data demonstrate that 1) muscles at L-o have higher (Q) over dot and (V) over dot O-2 values than those at L-s; 2) fatigue occurs at L-o with high (V) over dot O-2, adjusting metabolic demand (tension output) to match supply; and 3) the lack of fatigue at L-s with lower tension, (Q) over dot, and (V) over dot O-2 suggests adequate matching of metabolic demand, set low by short muscle length, with supply optimized by low preload. These differences in tension and (V) over dot O-2 between L-o and L-s groups indicate that muscles contracting isometrically at initial lengths shorter than L-o are working under submaximal conditions.