When bone implants have to be examined in situ ground sections are required. Histomorphometric measurements are usually performed on two-dimensional sections, causing biased results when they are wrongly extrapolated to 3D without any knowledge of stereology. Unbiased results can only be obtained using stereological principles. The aim of the study was to describe an unbiased design for evaluating boneto-mplant contact (BIC) and peri-implant bone density (BD-i) in three-dimensions. The unbiased design was based on a fixed axis vertical random sampling technique. Three bone-implant blocks were collected from 3 rabbits. Four sections were obtained from each animal using a fixed axis vertical random sampling technique. The BIC was estimated by creating a stereological method based on a systematic test line set. The BD-i was estimated using a design based on a systematic point set. The efficiency of the systematic sampling was evaluated for each sampling level: Coefficient of error CE( i w ) for the systematic test line set, CESyst between the 4 rotated sections, CENoise for the reproducibility. These variances were compared to the biological variation (CV) between animals. The mean CE( i w ) was 5 to 6%; the CESyst was 7.8% for the BIC estimates and 5.7% for the BD-i estimates; the CENoise was 5.8% for the BIC and 7.7% for the BD-i. The CV was 19% for the BIC estimates and 24% for the BD-i estimates. These results demonstrated that the stereological technique used in the present study was a very efficient method to obtain unbiased estimates of BIC and BD-i in 3D on 2D implant-bone sections.