We have investigated the effects of a standardised mixture of hydroxyethylrutosides (HR, Venoruton), a mixture of five of its main components (M) and each of the five components separately (7-mono-HR, 7,4'-di-HR, 7,3',4'-tri-HR, 5,7,3',4'-tetra-HR and 7,3'4'-tri HQ) upon the permeability of single perfused capillaries and venules in the mesenteries of pithed frogs. 2. In each experiment, the hydraulic permeability (Lp) of a single perfused microvessel and the effective osmotic pressure (sigma delta pi) exerted by macromolecules across its walls were estimated by a microcclusion technique, first during control perfusion and then in the presence of a known concentration of test substance. 3. HR, M and 7,4'-di-HR reduced Lp in a similar concentration-dependent manner over the range of 1 microgram ml-1 to 1 mg ml-1 (maximum reduction was to 40% of control Lp at 1 mg ml-1). At perfusate concentrations greater than 1 mg ml-1, these substances reduced Lp to a lesser extent. While the four other test substances reduced Lp significantly when their perfusate concentrations equalled or exceeded 100 micrograms ml-1, they were all less potent than 7,4'-di-HR. 4. The reduction in Lp induced by the mixture of flavonoids was only slightly reversed by subsequent perfusion with flavonoid-free solutions. 5. When permeability was increased by perfusing with protein-free solutions, both HR and 7,4'-di-HR reduced and then reversed the increase in Lp in a concentration-dependent manner over the range of 1 microgram ml-1 to 100 micrograms ml-1. None of the other component flavonoids was effective in restoring Lp under these conditions.