Two attempts have been made to develop body mass prediction formulae specifically for immature remains: Ruff (Ruff, C.C., 2007, Body size prediction from juvenile skeletal remains. American Journal Physical Anthropology 133, 698–716) and Robbins et al. (Robbins, G., Sciulli, P.W., Blatt, S.H., 2010. Estimating body mass in subadult human skeletons. American Journal Physical Anthropology 143, 146–150). While both were developed from the same reference population, they differ in their independent variable selection: Ruff (2008) used measures of metaphyseal and articular surface size to predict body mass in immature remains, whereas Robbins et al. (2010) relied on cross-sectional properties. Both methods perform well on independent testing samples; however, differences between the two methods exist in the predicted values. This research evaluates the differences in the body mass estimates from these two methods in seven geographically diverse skeletal samples under the age of 18 ( n = 461). The purpose of this analysis is not to assess which method performs with greater accuracy or precision; instead, differences between the two methods are used as a heuristic device to focus attention on the unique challenges affecting the prediction of immature body mass estimates in particular. The two methods differ by population only in some cases, which may be a reflection of activity variation or nutritional status. In addition, cross-sectional properties almost always produce higher estimates than metaphyseal surface size across all age categories. This highlights the difficulty in teasing apart information related to body mass from that relevant to loading, particularly when the original reference population is urban/industrial.