Seven Vasculour d grafts, five Gore-Tex grafts and seven Solco-graft, 8 mm by 30 cm, implanted as thoraco-abdominal bypasses in dogs. Sixteen were retrieve at two months. Graft size was assessed angiographically in representative dogs of each group prior to sacrifice. Initial examination of the retrieved specimens revealed pannus extensions at both anastomoses on all graft types, and a variable midgraft appearance from smooth and glistening to partially thrombus coated. Histological assessment of graft segments revealed widely differing healing responses on the three graft types. Tissue attachment to both graft surfaces and varying degrees of tissue incorporation were observed in the Vasculour D and Gore-Tex prostheses. The Solco-graft specimens showed no evidence of tissue incorporation of the wall at two months. Tissue on the graft surfaces remains loosely attached and unincorporated. Pannus extensions were thickest on the Solco-graft. Variably thin fibrin is scattered over the luminal surface of the three graft types. Vascularization of the wall particularly the internal capsulae was noted in the Vasculour D grafts. Although biological responses in the dog are difficult to extrapolate to man, the types of comparison testing allowed by the use of the thoraco-abdominal by-pass procedure can provide valuable information of degree of dilation, relative tissue attachment, tissue incorporation into various types of prostheses and kidney infarcts study. The greatest value of the thoraco-abdominal by-pass lies in graft sizes and length without special regard for patency: assessment of graft size post-implantation and structural changes is then made possible.