Osteocalcin, the gamma-carboxyglutamic acid-containing protein, which in most species is the predominant noncollagenous protein of bone and dentin, has been postulated to play roles in bone formation and remodeling. Recently, genetic studies showed that osteocalcin acts as an inhibitor of osteoblast function. Based on von Kossa staining and measurement of mineral apposition rates in tetracycline-labeled bones, osteocalcin knockout animals were reported to have no detectable alterations in bone mineralization. To test the hypothesis that, in addition to regulating osteoblastic activity, osteocalcin is involved in regulating mineral properties, a more sensitive assay of mineralization, Fourier transform infrared microspectroscopy (FT-IRM) was used to study thin sections of femora of 4-week-, 6-month- (intact and ovariectomized), and 9-month-old wild-type and osteocalcin-knockout mice. FT-IRM spectra provided spatially resolved measures of relative mineral and carbonate contents, and parameters indicative of apatite crystal size and perfection. No differences were detected in the mineral properties of the 4-week-old knockout and wild-type mice indicating that the mineralization process was not altered at this time point. Six-month-old wild-type animals had higher mineral contents (mineral:matrix ratios) in cortical as compared with trabecular bones; mineral contents in knockout and wild-type bones were not different. At each age studied, carbonate:phosphate ratios tended to be greater in the wild-type as compared with knockout animals. Detailed analysis of the phosphate nu1,nu3 vibrations in the spectra from 6-month-old wild-type animals indicated that the crystals were larger/more perfect in the cortical as opposed to the trabecular bones. In contrast, in the knockout animals' bones at 6 months, there were no differences between trabecular and cortical bone in terms of carbonate content or crystallite size and perfection. Spectral parameters of the cortical and trabecular bone of the knockout animals resembled those in the wild-type trabecular bone and differed from wild-type cortical bone. In ovariectomized 6-month-old animals, the mineral content (mineral:matrix ratio) in the wild-type cortices increased from periosteum to endosteum, whereas, in the knockout animals' bones, the mineral:matrix ratio was constant. Ovariectomized knockout cortices had lower carbonate:phosphate ratios than wild-type, and crystallite size and perfection resembled that in wild-type trabeculae, and did not increase from periosteum to endosteum. These spatially resolved data provide evidence that osteocalcin is required to stimulate bone mineral maturation.