The frequency distributions of mineral apposition rate (MAR) and mineralizing surface (MS), measured separately on the intracortical, endocortical, and cancellous surfaces in 46 normal subjects and 79 patients with postmenopausal osteoporosis, indicated that MAR has a finite lower limit of 0.3 μ/day (uncorrected for section obliquity) but that MS has no finite lower limit. We conclude that in the absence of labels MAR, and indices derived from it, must be treated as missing values, but that MS and indices with MS in the numerator should be allowed to take values of zero. To avoid infinite values for indices with MS in the denominator, we propose that osteoid mineralization rate (the reciprocal of mineralization lag time) and osteoblast vigor (the reciprocal of formation period) be used instead. For surfaces with genuine single labels (SL) but no double labels, we propose that MS is calculated as SL/2 and that for MAR either the lower limit of 0.3 or the mean measured value from other surfaces be used for calculating derived indices.