Between 1995 and 1998, eighty revision total knee arthroplasties were done for the primary reason of advanced polyethylene wear. The primary arthroplasties prosthesis that failed included thirty-four mobile bearing knees and forty-six fixed bearing knees. In thirty-four Low Contact Stress (LCS) mobile bearing knees, osteolysis was identified intraoperatively in sixteen knees (forty-seven per cent). There were varying of fixation methods included nine cemented, four cementless and three hybrids. In forty-six fixed bearing knees, osteolysis was identified intraoperatively in six knees (thirteen per cent). The fixation methods of prostheses included two cemented and four cementless. The incidence of osteolysis was statistically significant difference between the mobile bearing and fixed bearing knees (p< 0.02). Both scattering electron microscope (SEM) and light scattering analysis were used to examine the UHMWPE wear debris collected from tissue sample. The particle size analyzed by light-scattering is coincident with the measurement by SEM. The major type of wear debris extracted from failed knee prostheses is granular shape. There are more granular wear debris appear in the mobile bearing knees than in the fixed bearing knees. The particle size of UHMWPE wear debris with osteolysis was significantly smaller than that without osteolysis. The high rate of osteolytic lesions in mobile bearing knee (LCS) is well illustrated in our result that a lot of fine UHMWPE wear debris generated in the Low Contact Stress knee. The result also illustrates that there is no relationship between fixation methods and the third body wear that associate with osteolysis