After approximately 6 mo without access to a receptive female, eight old (19 to 25 yr) and eight middle-aged (9 to 15 yr) males were given nine daily 10-min tests of sexual behavior with receptive females. Three ovariectomized, estrogen-primed females served as partners. Blood samples were taken from the males at 0900 and 2100 h on 2 consecutive days before and after the series of nine behavioral tests. Serum levels of testosterone (T) were measured by means of radioimmunoassay. There were no significant differences between serum T levels after 6 mo of sexual deprivation and after the nine tests of sexual behavior, nor did serum T levels differ between middle-aged and old males. Serum T levels were significantly higher in the 2100-h samples than in the 0900-h samples for both groups. The nocturnal increases in serum T, both absolute amounts and percentages of increase, did not differ between middle-aged and old males. Repeated daily testing had little effect on the major components of male sexual behavior. Middle-aged males outperformed old males; middle-aged males had higher rates of contracting the female, mounting and intromitting. They ejaculated earlier and had a greater percentage of tests with ejaculations than did old males.