Glutaraldehyde tanning of carotid arteries was used to develop a model for studying the effects of compliance on arterial graft performance, independent of other graft parameters. Canine carotid segments were filled with dilute phosphate-buffered glutaraldehyde (0-0.5%, pH 7.4), maintained at physiological pressure, and then immersed in either saline or 10.0% glutaraldehyde for up to 1 hr. After rinsing with saline, compliance was measured in vitro. All vessels which were immersed in 10% glutaraldehyde exhibited a significant reduction in compliance compared to native artery control [C = 11.8 +/- 1.3 (mean +/- SEM), % radial change/mm Hg X 10(-2), measured at 100 mm Hg], but maximum stiffness (C = 1.1 +/- 0.3) required that the lumen be specifically exposed to at least 0.025% glutaraldehyde in addition to simple immersion of the vessel segment in 10% fixative. Exposing the artery to 0.5% glutaraldehyde internally, without immersion of the entire structure, caused a decrease in compliance similar to that obtained after immersion in 10% glutaraldehyde, with only saline present in the lumen. Matched pairs of stiff and compliant grafts were generated by exposing the lumen to 0.025% glutaraldehyde and immersing the vessels in 10% fixative or saline, respectively. Light and scanning electron microscopy, internal reflection spectroscopy, and measurements of critical surface tension revealed nearly identical wall morphology and lumenal surface chemistry for these matched pairs. Differential tanning of the internal and external surfaces of carotid arteries thus provides a good model of arterial prostheses, wherein a substantial compliance mismatch can be studied without the complicating influences of varying diameter or differing flow surface properties.