Intravital microscopy was used in a preparation of rat cremaster muscle that was isolated from its normal blood supply and externally perfused with a physiological solution, thus allowing exclusion of circulating converting enzyme, renin, and angiotensinogen. The arterioles studied were classified as second-, third-, and fourth-order arterioles with mean diameters of 60.5, 29.9, and 14.8 microns, respectively. Topical administration of 1 nmol/mL angiotensin I or 1 nmol/mL tetradecapeptide renin substrate induced marked vasoconstrictions (i.e., 38.5%, 61.5%, and 90.1% and 25%, 34%, and 88% for second-, third-, and fourth-order arterioles with angiotensin I and tetradecapeptide renin substrate, respectively). The angiotensin converting enzyme inhibitor quinapril significantly inhibited the vasoconstrictions caused by either angiotensin I or tetradecapeptide renin substrate. Almost no vasoconstriction was found when angiotensinogen-rich renin-free plasma containing either 2.45 nmol/mL of angiotensinogen or 1.2 micrograms/mL renin was administered. Conversely, these two compounds induced significant constrictions in cremaster muscle preparations in which normal blood perfusion (and thus circulating renin and angiotensinogen) was left in place. We concluded that, in skeletal muscle, 1) the microvascular network is a very effective site of local angiotensin converting enzyme activity and consequently an important target site of angiotensin converting enzyme inhibitors; 2) the effects of tetradecapeptide renin substrate are very different from those of angiotensinogen from plasma and suggest that a large part of the effect of tetradecapeptide renin substrate was due to its nonspecific hydrolysis; and 3) at the microvascular level, circulating renin and angiotensinogen are more effective in inducing arteriolar constriction, in the presence of their substrate or associated enzyme, than local renin and angiotensinogen.