To demonstrate fat embolism during total hip replacement, thirty-four unselected patients were studied with an ultrasound probe over the femoral vein; twenty-one of them also had serial analyses of control and femoral venous blood for fat. Computer energy-density spectrum analysis of the sounds recorded by the ultrasound probe were performed; the fat emboli were counted in the blood samples by two histological techniques; and the plasma was analyzed for triglycerides before and after filtration. Characteristic sounds ("chirps") were heard with the Ultrasound probe during insertion of the femoral component in all patients. The individual chirps lasted two to five milliseconds and occurred every ten to fifteen milliseconds. The mean duration of activity was 4.2 minutes. Rarely, a few chirps were heard during seating of the acetabular component, reaming, or insertion of the cement into the femur. Blood samples during peak ultrasound activity showed: (1) seventy-nine globules per high-power field (control, 1.7) by the cryostat test; (2) 360 globules per high-power field (control, 6.8) by the millipore filter test; and (3) a mean drop in triglycerides after filtration of 27.8 milligrams per 100 milliliters (control, + 1.2). There was no clinical or laboratory evidence of the fat-embolism syndrome. Venting the femoral medullary canal during placement of the femoral component reduced the amount of fat emboli.