The precision of measuring polyethylene penetration in total hip arthroplasties has been studied using phantom models and retrieved cups, but not on radiographs obtained in vivo. We used Hip Analysis Suite Hip32 computer software to measure two-dimensional penetration in two groups of anteroposterior pelvic radiographs taken in vivo. We determined the measurement error for precision, the effect of position change on precision, and compared direct and indirect measurements. The influences on precision with respect to followup time were measured using a separate group of radiographs. Precision was 0.89 to 1.16 mm and decreased as more variables were added. The direction of wear was toward the contact surface of the cup in greater than 50% of patients with more than 5 years followup. Negative wear severely affected precision in 40% to 50% of patients. Individual penetration measurements were unreliable when total penetration was low (< 1.1 mm). The two-dimensional computer technique has limitations when measuring such low wear. Measurement errors were caused by slight radio-graphic distortions of the femoral head and acetabular component and by blurred edges on some clinical radiographs.