Some population-based studies on male aging measure testosterone and cortisol in saliva instead of serum, but very few measure estradiol and dehydroepiandrosterone sulfate (DHEA-S), suggesting further testing is needed for reliability and comparative validity. In addition, the effects of interview stress and circadian hormone secretion need to be checked. In a pilot study on the overall sexual capacity of aging men, 48 randomly selected, healthy, heterosexual, cohabiting men aged 50-80 years, from Mannheim, Germany, and 50 from the State College, Pennsylvania, USA, were administered a standardized interview covering medical biography, present and previous life and work, marriage and emotional status. Two saliva samples were collected from each subject for measurement of testosterone, cortisol, estradiol and DHEA-S levels before and after the interview, and each subject completed a confidential self-administered questionnaire on intercourse, masturbation, orgasm, fantasies, libido and arousal. Questionnaires, hormone measurement techniques and the survey protocol had been extensively pretested. Prior to the pilot study, the kits for measuring testosterone and DHEA-S in saliva were checked for comparative validity against established measuring techniques in serum in 31 cases for testosterone and in 24 different cases for DHEA-S. These 55 cases underwent clinical diagnosis and were not otherwise involved in this study. The cases had been referred to the Andrology Unit of the University Hospital, Marburg, for reasons unrelated to this study. Given the biological differences for both steroids between their presence in blood and in saliva, a perfect correspondence between the two values was not expected and was not observed. The correlations obtained, however, support the assumption that all statistical relationships between testosterone and DHEA-S values in serum and clinical, as well as behavioral, variables reported to date may be replicated for testosterone and DHEA-S values in saliva.