To facilitate implant osseointegration in the early postoperative period, coating of the implant surfaces with osseoconductive materials, e.g. hydroxyapatite (HA), is being increasingly recommended. It apparently reduces the rate of radiolucent lines and even improves the osseointegration of implants less well suited for cementless anchorage. We analysed HA-coated stems to find out whether newly formed bone adherent to the implant surface such as seen on explanted stems is also seen on radiographs and whether it follows a specific morphological pattern. We also wanted to know whether newly formed peri-implant bone extends over the entire length of the HA-coating, at which point in time it is seen radiographically and whether it expands within the first few years post implantation by radiological evidence. Radiographs of 40 unselected patients after primary total hip arthroplasty (THR), 14 males and 26 females, aged 44.4 to 86.7 years at the time of THR, with a mean age of 67.3 years were available for analysis. Monitor-guided a.-p. and axial views of the stems were obtained in the early postoperative period up to 6 weeks post THR, at a mean follow-up time of 1.3 (1.0 to 1.8) years and 3.0 (1.9 to 3.7) years. Tapered straight stems with a rectangular cross-section made of a wrought Ti6AI7Nb alloy with an HA coating in the proximal third were used. A.-p. views: At one year there were no signs suggesting an increased ongrowth of bone. At 3 years, on average, delicate bony appositions were seen on the implant surfaces in position 1 in three stems, in one of them also in position 7. These looked like a sclerotic zone lacking a sharp demarcation and were in direct contact with the implants without any bone-to-implant gaps. Axial views: At about one year two stems showed delicate bony appositions parallel to the implant surface in the proximal part in positions 8 and 9 as well as 13 and 14. Named "miniscleroses" by us, these structures were confined to the length of the HA coating. At three years these miniscleroses were clearly visible around 11 stems (more than 25 %). Those already seen at one year were much better defined at three years, but still confined to the HA-coated part of the stem. Some of them were poorly demarcated from the adjacent bone and medullary canal and some showed smooth demarcations. The density of the bony ongrowths was either homogeneous or increased from the medullary space or adjacent bone towards the implant surface. All of them were adherent to the implant surface. While not related to the peri-implant cortical bone, they were in contact with trabecular structures. After a mean follow-up time of one year radiolucent lines were seen around four stems in positions 1, 7 and 8. These were up to 1 mm in size around two stems and 2 mm or more in the other two. At 3 years all of the visible radiolucencies had disappeared except for one measuring 1 mm in position 8. The structures we found radiographically apparently reflect newly formed bone along the entire HA-coated implant surface. Most of them were located in positions 8, 9 and 13, 14. They provide visible evidence of osseointegration at osseoconductive surfaces. As they apparently did not have any contact with the peri-implant bone, they appear to be compatible with the bilateral osteogenesis according to Osborn. The effects of the rapid bony ongrowth on HA-coated surfaces and the striking absence of radiolucencies on the long-term outcome are still speculative. But the structures seen may be taken as a sign of improved implant stability by rapid osseointegration and of early sealing of the medullary canal. As a result, wear particles of the articulating surfaces are barred from spreading to the medullary cavity of the femur. This alone argues in favour of using coated implants throughout. However, more studies are needed to shed light on these issues.