The fixation of the femoral stem in a total hip arthroplasty was studied in sixty patients (sixty-four hips) with use of roentgen stereophotogrammetric analysis. The hips were randomly stratified on the basis of the age, sex, and weight of the patient; the roentgenographic quality of the bone; and the reason for the operation (primary or secondary osteoarthrosis). The hips were then randomly assigned to one of three types of fixation of the femoral component: insertion with cement, hydroxyapatite coating, or porous coating. Examinations with roentgen stereophotogrammetry were done as long as two years after the operation. Micromotion of the prostheses was evaluated in terms of subsidence or proximal migration and rotations and translations of the proximal-lateral (shoulder) and distal (tip) parts of the prostheses. The clinical results at the two-year follow-up evaluation did not differ significantly between the groups with regard to the Harris hip score or the pain score (p > 0.05 for both; Wilcoxon rank-sum test). There were small or no differences in rotations and micromotions of the shoulder and the tip of the prostheses. Increased subsidence of 0.1 to 0.2 millimeter was recorded for the cemented and the porous-coated prostheses (p = 0.002 and p = 0.02, respectively; Wilcoxon rank-sum test). Thus, proximal hydroxyapatite coating seems to enhance the early fixation of the stem. Conventional roentgenography revealed an increased number of radiodense lines surrounding the porous-coated prostheses. Distal hypertrophy of the femoral cortex was found mainly around the hydroxyapatite-coated prostheses. Proximal resorption of bone and heterotopic ossification occurred to approximately the same extent regardless of the type of fixation used.