Two groups of beagles, accustomed to spacious group housing, were subjected to social and spatial restriction and studied for manifestations of chronic stress with a time interval of 7 weeks between the groups. The change from outside group housing (the control period) to individual housing in small indoor kennels resulted in sustained decreases in urinary adrenaline/creatinine and noradrenaline/creatinine ratios for the total group. Urinary dopamine/creatinine and noradrenaline/adrenaline ratios were statistically unaffected. Socially and spatially restricted dogs that had experienced pleasant weather during the control period showed (a) increased salivary and urinary cortisol concentrations, (b) a diminished responsiveness of the pituitary-adrenal axis to a sudden sound blast or exogenous CRH, (c) intact plasma ACTH and cortisol suppressions after dexamethasone administration, and (d) increased concanavalin A induced lymphocyte proliferations. When social and spatial restriction was preceded by a control period during which the weather was bad, these physiological responses were either augmented (lymphocyte proliferation), or offset (salivary and urinary cortisol), or directed oppositely (CRH-induced ACTH and cortisol responses). Together with the previously presented behavioral observations, these data suggest that bad weather conditions during spacious outdoor group housing induced early stress that attenuated the negative appraisal of the subsequent period of social and spatial restriction. In comparison to male dogs, bitches showed increased HPA responses to a sound blast or exogenous CRH. Their increased attenuations of the ACTH and cortisol responses to CRH after 5 weeks of restricted housing indicates that bitches are not only more susceptible to acute stress, but also to chronic housing stress. It is concluded that the quality of circumstances preceding a period of affected well-being determines the magnitude and even the direction of the behavioral and physiological stress responses. Basal salivary and urinary cortisol measurements are useful for the assessment of chronic stress, and of poor welfare in dogs. The use of urinary catecholamine, peripheral leucocyte, and lymphocyte proliferation measures requires further investigation.