To study physiological effects of restricted activity (RA) and subsequent retraining, 10 male mongrel dogs (1-5 years) performed a submaximal exercise endurance test on a treadmill (12 degrees slope, 1.6 m.s-1) during kennel control, after 8 weeks of cage (40 cm-w x 80 cm-h x 110 cm-l) confinement, and after 8 weeks of retraining using the same treadmill protocol 1 h/d for 6 d/week. Compared with control endurance (172 +/- 19 min), endurance decreased to 102 +/- 15 min (delta = -41%, p less than 0.05) after RA and increased to 223 +/- 24 min (delta = +30%, p less than 0.05) after training: the respective final levels and changes in rectal temperature were 41.25 and +2.15 degrees C, 41.60 and +2.70 degrees C (NS), and 41.35 and +2.40 degrees C (NS), respectively. Resting and post-exercise blood glucose and lactate concentrations were unchanged in the three experiments. After RA, resting muscle glycogen was reduced from a control level of 49.9 +/- 4.3 to 34.1 +/- 4.5 mmol.kg-1 (delta = 32%, p less than 0.05) which returned to the control level of 58.4 +/- 3.5 mmol.kg-1 after retraining. Resting plasma FFA levels were unchanged, but the RA post-exercise change was decreased from a control level of +0.400 +/- 0.099 to +0.226 +/- 0.039 mmol.L-1 (p less than 0.05). Neither restricted activity nor training affected glucose tolerance significantly. The results indicated that RA reduces exercise endurance, the effectiveness of exercise thermoregulation, muscle glycogen stores, and the lipolytic response to exercise and to noradrenaline stimulation. All these changes were reversed following 8 weeks of retraining.