Prolonged stress inhibits the hypothalamus-pituitary-gonadal (HPG) axis and reduces plasma testosterone (T). However, enhanced secretion of luteinizing hormone (LH) and T has been documented during the initial stages of acute stress in mammals. This study assayed the effect of short-term stress on plasma T and corticosterone (B) in juvenile, pubertal, and adult White Leghorn cockerels. Stress was induced by brief physical restraint of caged juvenile (7 weeks), pubertal (17 weeks), and adult (40 weeks) cockerels, as well as 40-week-old adults reared together in a room lined with wood shavings (group reared). Blood was sampled immediately before restraint (0 time), at the end of a 10-min restraint period, and at 30, 60, and 180 min after 0 time. Restraint resulted in an initial increase in plasma T in all groups, along with a rise in B. Whereas B generally reached its peak level at the end of the restraining period, T peaked 20 min later. The maximum increase of T and B relative to prestress levels (T and B ratios) was similar in all groups, with median T ratio reaching 1.25-1. 5-about half that of the B ratio. Thus, the extent of T and B response to short-term stress was not influenced by basal levels of T, which were highest in adults, and basal levels of B, which were higher in caged adults than in group-reared adults. Injection of ACTH did not induce a greater increase in plasma T than in sham-injected controls. Further, the elevation of T in response to stress was extinguished in castrated adults, indicating that T is secreted from the testes rather than the adrenals in response to stress. When the same regime of blood sampling was applied to adults not subjected to restraint, the T ratio rose by up to 11 times. It can therefore be stipulated that T response depends on the type of stress applied, a factor that should be considered when investigating androgen levels in plasma.