Psychological stress is thought to undermine host resistance to infection through neuroendocrine-mediated changes in immune competence. Associations between stress and infection have been modest in magnitude, however, suggesting individual variability in stress response. We therefore studied environmental stressors, psychobiologic reactivity to stress, and respiratory illness incidence in two studies of 236 preschool children. In Study 1, 137 3- to 5-year-old children from four childcare centers underwent a laboratory-based assessment of cardiovascular reactivity (changes in heart rate and mean arterial pressure) during a series of developmentally challenging tasks. Environmental stress was evaluated with two measures of stressors in the childcare setting. The incidence of respiratory illnesses was ascertained over 6 months using weekly respiratory tract examinations by a nurse. In Study 2, 99 5-year-old children were assessed for immune reactivity (changes in CD4+, CD8+, and CD19+ cell numbers, lymphocyte mitogenesis, and antibody response to pneumococcal vaccine) during the normative stressor of entering school. Blood for immune measures was sampled 1 week before and after kindergarten entry. Environmental stress was indexed with parent reports of family stressors, and a 12-week respiratory illness incidence was measured with biweekly, parent-completed symptom checklists. The two studies produced remarkably similar findings. Although environmental stress was not independently associated with respiratory illnesses in either study, the incidence of illness was related to an interaction between childcare stress and mean arterial pressure reactivity (beta =.35, p <.05) in Study 1 and to an interaction between stressful life events and CD19+ reactivity (beta =.51, p <.05) in Study 2. In both studies, reactive children sustained higher illness rates under high-stress conditions, but lower rates in low-stress conditions, compared with less reactive peers. Stress was associated with increased rates of illnesses, but only among psychobiologically reactive children. Less reactive children experienced no escalation in illness incidence under stressful conditions, suggesting that only a subset of individuals may be susceptible to the health-altering effects of stressors and adversity.