We have examined ten tissue capsules from patients ranging from five months to nine years of mammary implantation. Contrary to published reports of polyurethane foam "fragmentation" or "disappearance" in the capsules evaluated, the polyurethane foam was still present and embedded in the surrounding tissue capsule. The foam was nearly always invisible by gross observation, or manual palpation. Only after enzymatic digestion of the tissue capsule did the foam become clearly visible as continuous sheets. ESCA analyses show that explanted foams are devoid of nitrogen peaks. Only carbon, oxygen and silicone signals are observed. The same foams do show nitrogen peaks (due to urethane linkages) when probed by FTIR. Since ESCA only analyzes the first 40-50 Angstroms of a surface, we believe that a "protective coating" composed of soft segments has formed. Beneath this "coating" the original polyurethane composition is still present as evidenced by FTIR analysis. Three possible explanations are advanced: (1) The surface hydrolysis, which takes place within the soft segment of the polyurethane polymer, results in the formation of oligomer(s). These oligomers, devoid of urethane linkages, appear to protect the polymer from further bioresorption, by significantly retarding the rate of additional surface hydrolysis. (2) Chain cleavage occurs in the soft segment producing a hydrophilic polyester chain end which orients into the interfacial area. These chain ends then produce a skin effect which increases the distance from the surface to the hard segments, or urethane-containing linkages. (3) Macromolecular motion in the soft segment phases of the polymer could be reorienting under the influence of the in vivo environment, thus producing a surface layer or "coating" which is predominantly soft segment in composition. Regardless of which of the three hypotheses proves to be most plausible, we interpret the data as showing that the polyurethane foam cover undergoes very slow bioresorption, even after 9 years of human implantation. The data further suggests that the in vivo surface of the polyurethane foam cover is biocompatible and interfacial interactions with inflammatory cells are downregulated or reduced because of the apparent biocompatibility of the material.