Based on studies of the interaction of insulin with its receptors in vitro, we calculated that a receptor compartment should be measurable directly in vivo. For this purpose, rabbits were injected intravenously with a labeled insulin that has low affinity for receptors in combination with a radioiodinated insulin that has high affinity for receptors. Plasma concentrations of labeled insulins were measured at selected intervals after injection. Apparent volumes of distribution were calculated by extrapolation of plasma distribution were calculated by extrapolation of plasma disappearance curves; high affinity insulins consistently distributed into spaces that were two-three times greater than those of the low affinity insulins. Injections of unlabeled pork insulin before tracer insulins decreased the distribution space of the high affinity insulin in a dose-dependent manner while having little or no effect on the distribution space of the low affinity labeled insulin. When unlabeled insulin was injected after the tracer insulins, there was an immediate rise in the plasma concentration of the high affinity insulin with only a slight change in the plasma concentration of the low affinity insulin. These results demonstrate that high affinity insulins distribute into a body compartment which has many properties of the insulin receptor previously studied in vitro. This receptor compartment: (a) recognizes insulins based on their biological potencies; (b) is saturated by elevated concentrations of insulin; and (c) insulin bound to receptors is in equilibrium with free hormone in plasma. Further, the bound to free ratios for hormone, calculated from these data, suggest that in vivo greater than 50% of the extrapancreatic insulin is bound to receptors during normal physiological states.