Three types of surfaces for external fixation pins were compared. One hundred and eight stainless-steel tapered 5/6-millimeter pins were divided into three groups: thirty-six pins remained uncoated (Group A), thirty-six were plasma-sprayed with hydroxyapatite (Group B), and thirty-six were plasma-sprayed with titanium (Group C), The pins were implanted in the left tibia of eighteen sheep, with each sheep receiving six pins from the same group, A unilateral fixator then was assembled on the pins. The medial aspect of the mid-part of the tibial diaphysis was exposed, and a five-millimeter-long cylinder of bone was removed so that load would be borne by the bone-pin interfaces. Six weeks after the procedure, radiographs demonstrated rarefaction of twenty-nine pin tracks in Group A, fifteen in Group B, and thirty in Group C (p = 0.021 for Group A compared with Group B and p = 0.016 for Group B compared with Group C), The mean final insertion torque land standard deviation) was 4360 +/- 1050 newton-millimeters in Group A, 3420 +/- 676 newton-millimeters in Group B, and 3740 +/- 643 newton-millimeters in Group C, With the numbers available, no significant differences could be detected among these values. The mean extraction torque was 253 +/- 175 newton-millimeters in Group A, 3360 +/- 1260 newton-millimeters in Group B, and 1720 +/- 1030 newton-millimeters in Group C (p = 0.002 for Group A compared with Group B, p = 0.017 for Group A compared with Group C, and p = 0.03 for Group B compared with Group C), The extraction torque was significantly lower than the corresponding insertion torque in both Group A (p < 0.001) and Group C (p = 0.003); no significant difference could be found, with the numbers available, in Group B (hydroxyapatite-coated pins), At sixty times magnification, direct contact was seen along a mean of 16 +/- 9 per cent of the bone-pin interface in Group A, 30 +/- 12 per cent of the interface in Group B, and 28 +/- 15 per cent of the interface in Group C (p = 0.042 for Group A compared with Group C), However, at 10,000 times magnification, direct bone-pin contact was found only in Group B. CLINICAL RELEVANCE: The better fixation and greater interface strength of the hydroxyapatite-coated tapered pins should lead to a substantial decrease in the frequency of loosening and infection and consequently to a remarkable improvement in the clinical results of external fixation.