Thirty-four hips in which there had been prosthetic replacement were selected for study because of the presence of linear (diffuse) or lytic (localized) areas of periprosthetic bone loss. In all hips, there was careful documentation of the anatomical location of the material that had been obtained for histological analysis, and the specific purpose of the removal of the tissue was for examination to determine the cause of the resorption of bone. Specimens from twenty-three hips were retrieved during an operation and from eleven hips, at autopsy. The area of bone loss was linear only in sixteen hips, lytic only in thirteen, and both linear and lytic in five. In all thirty-four hips, intracellular particulate debris was found in the macrophages that were present in the area of bone resorption. All thirty-four had intracellular particles of polyethylene, many of which were less than one micrometer in size. Thirty-one hips had extracellular particles of polyethylene as well. Twenty-two of the thirty-four hips had intracellular metallic debris; in ten, metallic debris was found extracellularly as well. Ten of the sixteen cemented specimens had intracellular and extracellular polymethylmethacrylate debris. In the mechanically stable prostheses--cemented and uncemented--polyethylene wear debris was identified in areas of bone resorption far from the articular surfaces. The number of macrophages in a microscopic field was directly related to the amount of particulate polyethylene debris that was visible by light microscopy. Although the gross radiographic appearances of linear bone loss and lytic bone loss were different, the histological appearance of the regions in which there was active bone resorption was similar. Regardless of the radiographic appearance and anatomical origin of the specimen, bone resorption was found to occur in association with macrophages that were laden with polyethylene debris. In general, the number of macrophages present had a direct relationship to the degree of bone resorption that was seen. We believe that these findings indicate that joint fluid penetrates far more extensively than previously thought, even in a well fixed component, along the interface between the prosthesis and bone and in the periprosthetic tissues; it is often more extensive than is shown by arthrography. We therefore suggest the concept of the effective joint space to include all periprosthetic regions that are accessible to joint fluid and thus accessible to particulate debris.(ABSTRACT TRUNCATED AT 400 WORDS)