The aim of the study was the histomorphometric comparison of the osteogenic potential of beta-tricalcium phosphate (beta-TCP) alone or in a calcium sulfate matrix. Three round defects, 10 mm (diameter) x 5 mm (depth), were created on each iliac crest of 4 dogs. The defects were divided into 3 groups. Ten defects were filled with beta-TCP in a calcium sulfate (CS) matrix (Fortoss Vital; group A), 10 defects were filled with beta-TCP alone (Fortoss Resorb; group B), and 4 defects were left ungrafted to heal spontaneously (group C). All defects were left to heal for 4 months without the use of a barrier membrane. Histologic evaluation and morphometric analysis of undecalcified slides was performed using the areas of regenerated bone and graft remnants. All sites exhibited uneventful healing. In group A sites (beta-TCP/CS), complete bone formation was observed in all specimens, graft granules dominated the area, and a thin bridge of cortical bone was covering the defect. Group B (beta-TCP) defects were partially filled with new bone, the graft particles still dominated the area, while the outer cortex was not restored. In the ungrafted sites (group C), incomplete new bone formation was observed. The outer dense cortical layer was restored in a lower level, near the base of the defect. The statistical analysis revealed that the mean percentage of new bone regeneration in group A was higher than in group B (49.38% and 40.31%, respectively). A statistically significant difference existed between the 2 groups. The beta-TCP/CS group exhibited significantly higher new bone regeneration according to a marginal probability value (P = .004 < .05). The use of beta-TCP in a CS matrix produced significantly more vital new bone fill and preserved bone dimensions compared with the use of beta-TCP alone.