To determine if the age-related decline in male sex behavior is correlated with hormonal factors, a longitudinal study was conducted. Sexually experienced males were given mating tests every 2 months from 7 through 27 months of age. To study possible relationships between changes in behavior and alterations in hormone levels, blood samples were taken before and after these bimonthly tests. At 23 months, cross-sectional studies were also conducted comparing results to those obtained in 5-month-old males. Significant changes in mating behavior first appeared at 11 months; mount latency, intromission latency, ejaculation latency, postejaculatory interval, and intercopulatory interval were increased. Similarly, detectable decreases in testosterone (T) also occurred at this age. A significant decline in luteinizing hormone (LH) was not seen until 19 months. Correlational analyses revealed small (r less than or equal to -0.29) but significant negative correlations between T and parameters of mating behavior with age. When each age was examined separately, no significant correlations appeared. Plasma T was not predictive of behavioral performance. At 23 months, cross-sectional studies revealed deficits in mounting and penile reflex behavior but ejaculatory reflex capacity was unimpaired. At 28 months, males were decapitated. Only T levels showed a significant effect of age; estradiol, prolactin, and LH were unaffected when compared to 5-month-old males. The data suggest that although there are small and significant negative correlations between circulating testosterone and parameters of mating behavior with advancing age, it is unlikely that the observed decline in testosterone is the primary cause of the age-induced behavioral deficits. It is likely that the major causal factor(s) involves non-hormone-dependent changes within the CNS.