Washed sperm suspensions from 64 out of 89 (72%) randomly selected infertility patients produced detectable reactive oxygen species (ROS) compared to 17 out of 67 (25%) prospective semen donors (p < 0.01, Chi-square test). Among patients, the median sperm concentration in ejaculates which yielded sperm suspensions that generated detectable levels of ROS was lower than in those which did not: 36.2 (15.63-57.64) vs. 71.5 (22-108) x 10 6 /mL, respectively (median (interquartile range), p < 0.05, Kruskal-Wallis test). In samples that produced ROS, the basal rate of production and the rates after stimulation with 50 μmol N-formyl met leu phe (N-FMLP) l -1 or with 100 nmol phorbol 12-myristate 13-acetate (PMA) l -1 were significantly and inversely correlated with sperm concentration in the ejaculate (r = -0.43, -0.41 and -0.35, respectively, p < 0.01 Spearman's rank correlation). The rate of ROS production showed no relationship to the motility of spermatozoa in semen, whether evaluated visually or via computer assisted semen analysis. However, there was a significant negative correlation (r = -0.370) between the motile, normal sperm concentration (MNSC) and basal ROS production, and when stimulated with N-FMLP (r = -0.311) or with PMA (r = -0.249) (all p < 0.05). In patient samples that generated detectable ROS, the ability of the spermatozoa to retain motility for 24 h after preparation on a 40/80% Percoll gradient was negatively correlated with basal ROS production (r = -0.310, p < 0.05). ROS production was also related to the outcome of in vitro sperm mucus penetration tests. Unstimulated levels of ROS production showed a significant (p < (0.05), negative correlation with the number of progressively motile spermatozoa present in mucus after 15 (r = -0.379) and 60 (r = -0.362) min. These results suggest that sperm samples with increased ROS tend to have poor semen quality and reduced performance in a number of routine, diagnostic sperm function tests.