Stature reconstruction from skeletal remains forms part of the forensic anthropological analysis for the purpose of identification of the individual. The aim of this study was to derive regression formulae for the estimation of total skeletal height, and thereafter to predict stature in South African whites using long bone lengths. The sample comprised 98 white male and 71 white female skeletons from the Raymond A. Dart Collection of Human Skeletons and the Pretoria Bone Collection. For each individual, total skeletal heights and maximum long bone lengths were measured and used to produce univariate regression formulae, with resulting correlations (r) ranging between 0.56 and 0.96. The lowest standard error of estimate (1.75 for females, 1.92 for males) was obtained when the lumbar spine, femur and tibia were used in combination, while the highest SEE (5.21 for females, 5.54 for males) was found when the lumbar spine was used on its own. Recently published corrections for soft tissue additions to obtain living height from total skeletal height make these kinds of formulae more usable, and will reduce the problem of underestimation. The derived formulae are population specific and are designed for use in forensic skeletal analyses of South African whites, but are also generally relevant to theoretical and practical issues in forensic anthropology.