Digitization of periosteal and endosteal contours of bone cross section is one of several methods used for calculating long bone cross-sectional geometry (CSG). In this study, invasively obtained bone histological samples were used to calculate intra-observer measurement error for CSG parameters. Intra-observer error was measured based upon repeated measurements of cross-sectional areas (total area [TA], cortical area [CA]) and moments of area (Imax, Imin, Imax/Imin, J) in B6CBA mice (n = 17). Cross sections were cut at 50 % of the biomechanical length of the left tibia and the samples were further processed for CSG and histological analysis. Intra-observer error was measured to estimate the accuracy of the digitization method. Accuracy of the tested digitization method was expressed by mean difference (MD), mean absolute difference (MAD), and limits of agreement (LA). The results confirm our assumption that intra-observer error decreases with the number of repeated measurement events. Thus, the error can be minimized by acquiring experience in the section digitizing. Our results also show that TA, CA and polar moment of area (J) are more susceptible to intra-observer error than are Imax, Imin and Imax/Imin.