Calcium phosphate ceramics, beta-calcium pyrophosphate (Ca2P2O7), beta-tricalcium phosphate (Ca3(PO4)2), hydroxyapatite (Ca10(PO4)6(OH)2) and tetracalcium phosphate (Ca4(PO4)2O), were prepared. The calcium:phosphorus ratios and microporosities were 1 (31.6%), 1.5 (1.6%), 1.66 (1%) and 2 (34.6%) respectively. Samples (15 mm x 10 mm x 2 mm), abraded with No. 2000 alumina powder, were implanted into the tibial metaphysis of mature male rabbits. Failure load, when an implant detached from the bone or the bone itself broke, was measured. At 10 wk after implantation, the failure loads in beta-calcium pyrophosphate, beta-tricalcium phosphate, hydroxyapatite and tetracalcium phosphate were 31.65 +/- 9.90 N, 72.81 +/- 19.01 N, 49.49 +/- 17.25 N and 43.22 +/- 14.99 N respectively. At 25 wk after implantation, the values were 47.04 +/- 14.90 N, 71.34 +/- 19.50 N, 69.09 +/- 16.17 N and 62.03 +/- 18.62 N respectively. Histologically, bone bonding behaviour of calcium phosphate ceramics did not vary with the calcium:phosphorus ratio, as observed by contact microradiogram, Giemsa surface staining and scanning electron micrograph-electron probe micro analysis. There was no intervening soft tissue at the interface of bone and ceramics. Hydroxyapatite or tricalcium phosphate are used as bone substitutes. However, their mechanical strength is insufficient for weight-bearing and they are used as bone filler. This study showed that the apparent insignificance of strict calcium:phosphorus ratio with respect to the biological results greatly simplifies processing of calcium phosphate ceramics for clinical application. In clinical application, calcium phosphate ceramics with different Ca:P can be used as bone fillers for bone defects or bone cavities under non-weight-bearing conditions.