The use of semen and reproductive histories in epidemiologic studies is reviewed. Traditional parameters of semen quality are easily measured. A significant change may be positive proof of an effect on the reproductive tract and may occur within two or three months of the start of exposure. Conversely, semen specimens are difficult to obtain. The use of an appropriate reference group may be hard to ensure. A change in semen quality need not lead to abnormalities of reproduction. Past effects of exposure may not be detectable if exposure has been reduced or eliminated. Reproductive histories are readily obtained for fertility analysis. A reduction in fertility may be evidence of impaired health. Past effects of exposure are easily determined. For the detection of reduced fertility, however, two or three years of exposure may be needed. A change in fertility does not necessarily indicate an effect on the reproductive tract. The analysis of fertility requires considerable technical expertise. Since data for fertility analysis may be gathered in the workplace with little effort, suitable information should be obtained during employee medical examinations wherever potential exists for effects on reproduction.