A 35-year prospective study was undertaken in 126 former college students to determine the predictive value of psychophysiological patterns previously recorded in response to repetitive laboratory stress experiments. Detailed health information has been obtained in 116 (92.1%) of these subjects. The emotion of "severe anxiety" expressed in one or more of the prior tests appeared to be a reliable marker for increased susceptibility not only to coronary heart disease but to overall future illness. This form of pathological anxiety, moreover, was frequently shown to be linked to marked conflict about hostile impulses. Contrariwise, neither anger-in nor anger-out was found to be associated with a higher incidence of subsequent disease. Failure to express emotion was observed in a variety of subjects who as a group exhibited no predisposition to sickness in later life. Psychological Mastery was predictive of favorable prognosis, but Physiological Mastery, contrary to expectations, did not show statistically significant advantages in that regard. Thus, the construct of "Mastery" itself as a determinant of prognosis was not fully supported by the findings in the present study. Cardiovascular hyperreactivity could not be confirmed as a major biologic mechanism responsible for cardiovascular disease. Such hyperresponses were common in association with "anger-in" without evidence of increased susceptibility to cardiovascular disease or other forms of illness. Further research is needed to identify pathophysiological pathways that may be activated by the emotion of severe anxiety in mediating its apparent relationship with total morbidity and mortality over time.