The possible causes of delayed muscle soreness which occur 24 to 48 hr after exercise were examined from three different approaches, each designed to test an existing hypothesis. Surface electromyograms were used to evaluate the muscle spasm theory; the possibility of actual muscle cell damage was monitored by the presence of myoglobinuria, while the ratio of hydroxyproline/creatinine (OHP/Cr) in 24 hr urine collection was used as a marker for connective tissue involvement. In the first study, although all volunteers developed muscle soreness 24 and 48 hr after exercise, no change in the EMG activity of the sore muscles was observed. Myoglobin excretion was found in 88% of the subjects who developed soreness. However, in a second study, 92% of the subject who performed both moderate and heavy exercise but did not develop muscle soreness had myoglobinuria. In contrast, during a third experiment subjects on gelatin-free diets showed an increase (P less than .1) in the OHP/Cr between control (.020+/-.001) and 48 hr post-exercise (.002+/-.001, X+/-SE). Soreness resulted in all cases. When the OHP/Cr value is taken for the day of maximal soreness, the post-exercise mean increases to .024+/-.001 and the level of significance rises (P less than .005). These observations support the concept that exercise induced soreness may be related to disruption of the connective tissue elements in the muscle and/or their attachments.