This paper reports an experience over ten years with cementless total hip arthroplasty in a series of 2688 personal cases commencing in 1975. Stress is laid on the results over a period greater than 5 years and up to 10 years as much on the long-term effects on the host bone as on the implant itself. The causes of failure of biological anchorage by endosteal osteogenesis are analysed. Various phenomena, particularly the reaction of the cortices to stress, several cases of bony resorption with debris due to wear and fatigue fractures of the stem of the prosthesis have been seen. These and the problems posed by the occasional need for extraction have led to the progressive development of the prosthesis, resulting in 1983 in a new type of anchorage with a polarised surface and a better-balanced prosthesis in relation to stress and extractability.