The net exchange of glucose and lactate across the leg and the splanchnic bed and the arterialdeep venous (A-DV) differences for these substrates in the forearm were determined in healthy subjects during 3-3.5 h of leg exercise (bicycle ergometer) at 58% maximum O(2) uptake and during a 40-min post-exercise recovery period. Leg glucose uptake rose 16-fold during exercise and throughout the exercise period exceeded splanchnic glucose output. The latter reached a peak increment (3.5 times basal) at 90 min and fell by 60% during the third hour. As a result, blood glucose declined 40%, reaching frank hypoglycemia (blood glucose, <45 mg/dl) in 50% of subjects at 3.5 h. Splanchnic lactate uptake rose progressively during exercise to values four times the basal rate at 3 h in association with a rise in arterial lactate to 1.5 mM. There was, however, no significant net output of lactate from the legs beyond 90 min of exercise. In contrast, the A-DV lactate difference in the forearm became progressively more negative throughout exercise, reaching values three times the basal level at 3.5 h. The rise in arterial lactate during exercise was proportional to the elevation in plasma epinephrine, which rose ninefold. During recovery, splanchnic lactate uptake rose further to values six times the basal rate, whereas lactate output by the legs was no greater than in the basal state. The A-DV lactate difference in the forearm became even more negative than during exercise, reaching values four times basal. During exercise as well as recovery, forearm uptake of blood glucose could account for no more than 25-67% of forearm lactate release. Leg glucose uptake during recovery was threefold to fivefold higher than in the basal state in the face of plasma insulin concentrations that were 60% below basal and in association with a respiratory exchange ratio of 0.7. We conclude that (a) during prolonged leg exercise at 58% maximum O(2) uptake an imbalance between splanchnic glucose production and leg glucose utilization results in a fall in blood glucose that may reach hypoglycemic levels in healthy subjects; (b) there is a marked increase in the uptake of lactate by the splanchnic bed that cannot be attributed to increased output of lactate from the exercising legs; (c) lactate is released by forearm muscle and, together with other relatively inactive muscle, may be an important source of the increased lactate turnover during and after prolonged leg exercise; (d) the increasingly negative A-DV lactate difference in the forearm cannot be accounted for by uptake of blood glucose, suggesting the breakdown of glycogen in forearm muscle during and after leg exercise; (e) increased glucose uptake by the legs in association with hypoinsulinemia during recovery suggests an increase in insulin sensitivity that permits glycogen repletion in previously exercising muscle in the absence of food ingestion; and (f) the evidence for increased lactate output in the forearm and augmented glucose uptake in the legs during recovery raises the possibility that after leg exercise glycogen stores are decreasing in muscle that was relatively inactive (e.g., that of the forearm) while increasing in the previously exercising leg muscles.