Job strain, a high level of psychological demands combined with a low level of decision latitude, has been hypothesized to induce mobilization of energy and inhibition of anabolism. In the present project this hypothesis was tested using four repeated observations every third month in a group of 44 men working in six widely different occupations. On each occasion scores of self-reported demands and decision latitude were calculated for every participant. An earlier report has shown that systolic blood pressure during work hours--an indicator of mobilization of energy--increased with increasing job strain (ratio between demands and decision latitude). Blood samples were drawn in the morning at the work site. For each man the plasma testosterone levels--representing the general level of anabolic activity--on the two occasions with the worst strain (ratio between demands and decision latitude) were compared with the plasma testosterone levels on the two occasions with the least strain. The results indicated that total plasma testosterone (but not free testosterone) levels increased when strain diminished in sedentary but not in physically demanding work. Subjects with a family history of hypertension showed a greater decrease in testosterone levels than others when job strain increased.