Despite intensive study, the relation between insulin's action on blood flow and glucose metabolism remains unclear. Insulin-induced changes in microvascular perfusion, independent from effects on total blood flow, could be an important variable contributing to insulin's metabolic action. We hypothesized that modest, physiologic increments in plasma insulin concentration alter microvascular perfusion in human skeletal muscle and that these changes can be assessed using contrast-enhanced ultrasound (CEU), a validated method for quantifying flow by measurement of microvascular blood volume (MBV) and microvascular flow velocity (MFV). In the first protocol, 10 healthy, fasting adults received insulin (0.05 mU. kg(-1). min(-1)) via a brachial artery for 4 h under euglycemic conditions. At baseline and after insulin infusion, MBV and MFV were measured by CEU during continuous intravenous infusion of albumin microbubbles with intermittent harmonic ultrasound imaging of the forearm deep flexor muscles. In the second protocol, 17 healthy, fasting adults received a 4-h infusion of either insulin (0.1 mU. kg(-1). min(-1), n = 9) or saline (n = 8) via a brachial artery. Microvascular volume was assessed in these subjects by an alternate CEU technique using an intra-arterial bolus injection of albumin microbubbles at baseline and after the 4-h infusion. With both protocols, muscle glucose uptake, plasma insulin concentration, and total blood flow to the forearm were measured at each stage. In protocol 2 subjects, tissue extraction of 1-methylxanthine (1-MX) was measured as an index of perfused capillary volume. Caffeine, which produces 1-MX as a metabolite, was administered to these subjects before the study to raise plasma 1-MX levels. In protocol 1 subjects, insulin increased muscle glucose uptake (180%, P < 0.05) and MBV (54%, P < 0.01) and decreased MFV (-42%, P = 0.07) in the absence of significant changes in total forearm blood flow. In protocol 2 subjects, insulin increased glucose uptake (220%, P < 0.01) and microvascular volume (45%, P < 0.05) with an associated moderate increase in total forearm blood flow (P < 0.05). Using forearm 1-MX extraction, we observed a trend, though not significant, toward increasing capillary volume in the insulin-treated subjects. In conclusion, modest physiologic increments in plasma insulin concentration increased microvascular blood volume, indicating altered microvascular perfusion consistent with a mechanism of capillary recruitment. The increases in microvascular (capillary) volume (despite unchanged total blood flow) indicate that the relation between insulin's vascular and metabolic actions cannot be fully understood using measurements of bulk blood flow alone.