Stature estimation methods for adult indigenous humans from the Americas have generally relied on a limited number of regression equations. The available equations, however, are not broadly applicable to the diversity of the populations that lived in the New World prior to European colonization. Furthermore, some equations that have been used were originally derived from inappropriate reference samples, such as the "Mongoloid" group measured by Trotter and Gleser (Am J Phys Anthropol 16 [1958] 79-123). This study develops new stature estimation equations for long bones of the lower limb from a geographically diverse sample of North American archaeological sites. Statures were reconstructed from 967 skeletons from 75 archaeological sites using the revised Fully anatomical technique (Raxter et al., Am J Phys Anthropol 130 [2006] 374-384). Archaeological samples were grouped according to general body proportions, using relative tibia and femur length to stature as guides. On the basis of differences in these proportions, three broad groupings were identified: a high latitude "arctic" group, a general "temperate" group, and a Great Plains group. Sex-specific ordinary least squares regression formulae were developed based on femoral and tibial lengths for each of these groups. Comparisons of the new stature estimation equations with previously available equations were conducted using several archaeological test samples. In most cases, the new stature estimation equations are more precise than those previously available, and we recommend their use throughout most of North America. The equations developed by Genovés for Mesoamerican and US Southwest samples are a useful alternative for these regions. Applicability of the new equations to South American samples awaits further testing.