Porous poly(epsilon-caprolactone-co-L-lactide) (P(CL-co-LA, wt % ca. 5/95) sponges were prepared, coated biomimetically with CaP/apatite, and implanted with noncoated control sponges into rat femur cortical defects and dorsal subcutaneous space. The implants were inspected histologically at 2, 4, and 33 weeks after the operation. All implants were filled with fibrovascular tissue within 4 weeks. The femur implants were partially ossified with compact bone, which in the CaP-coated sponges was less mature and more fragmented. Approximately equal amounts of bone were observed in both types of implants. The polymer induced a mild inflammatory reaction with foreign body giant cells but no accumulation of fluid. Degradation of the polymer was slow; most of it was found intact at 33 weeks in histological samples. Nondegraded polymer seems to prevent complete ossification. Cultured osteoblasts proliferated well on apatite-coated material, whereas only a few cells were seen on noncoated material. Thus CaP/apatite coating helped the attachment of osteoblasts in cell cultures but did not offer any advantage in bone formation over noncoated material in vivo. We conclude that a shorter degradation time of P(CL-co-LA) is needed to create an optimal implant. Furthermore, in vivo experiments seem to be necessary for the estimation of osteopromotive properties of a biomaterial.