The objective of this study was to analyze the efficacy and correlation between clinical and histologic parameters used to evaluate oral implants. After extraction of the premolars and a healing time of 4 months in 16 Dutch goats, four Branemark implants were placed in the maxillary left and right premolar regions. After a healing time of 6 months, followed by another 4 months with the permucosal abutments, the goats were sacrificed and the jaws were block-resected. Before histologic preparation, long-cone radiographs were made and Periotest scores of the implants were recorded. Bone level measured histomorphometrically were found to be 0.85 mm more apically, compared to that measured radiologically (P = .001). Furthermore, statistically significant correlations (P > .02) were not found between the Periotest values of the calcium-phosphate-coated and uncoated implants for (1) the first thread in contact with bone, or (2) with the total number of threads in contact with bone. It was concluded that the radiologic data overscored the real marginal bone level around screw-shaped oral implants, and that the Periotest device is neither able to discriminate between the first thread nor between the total number of threads in contact with bone.