Human umbilical cord veins treated with various concentrations of glutaraldehyde (GA) and ethanol (ET) were used to replace short segments of the infrarenal abdominal artery in 40 dogs. The patency rates were 62.5% (5/8) in Group I; tanned with 1% GA for 24 hours and preserved in the same solution 87.5% (7/8) in Group II; tanned with 1% GA for 24 hours and preserved in 50% ethanol, 100% (8/8) in Group III; tanned in the same manner as Group II, but preserved in 70% ET; 87.5% (7/8) in Group IV; tanned with 0.5% GA for 24 hours and preserved in 50% ET, and 87.5% (7/8) in Group V; tanned as in Group IV but preserved in 70% ET when examined 14 days to 6 months after implantation. Angiographically, most of the anastomotic lines gradually contracted with the lapse of time because of excessive proliferation of connective tissue into the intraluminal surfaces, although the structural integrity of the graft was well preserved. Microscopically, the irregular fibrin membrane covering the entire inner surface of the graft became smoother and thinner with the lapse of time. The pannus increased gradually both in thickness and width. Scanning electron-micrography of the graft revealed that no endothelial cells were found more than 5 mm from the suture lines even 6 months after implantation. Antigenicity of the graft using a complement-dependent cytotoxicity test was effectively suppressed with glutaraldehyde regardless of whether its concentration was 0.5% or 1.0%. Although human umbilical cord vein tanned with 0.5% glutaraldehyde and preserved in 50% ethanol until used was considered to be useful as an arterial substitute, final conclusion could not be drawn since six grafts followed for six months, the longest follow-up, were stenotic.