At a given age, bone mass represents the difference between the maximal amount of bone mineral mass gained during growth (peak bone mass) and post-menopausal and/or senile bone loss. Twins and parents-offspring models have shown a strong inheritance of peak bone mass and it is now known that familial resemblance in various bone mass constituents is detectable well before the pubertal growth spurt. Recent developments in the molecular epidemiology of osteoporosis have focused on the association between areal bone mineral density and common polymorphisms in several candidate genes. Among them, vitamin D receptor (VDR), estrogen receptor and collagen-1 -alpha-1 (COL1A1) genes have been the most extensively investigated. Although controversial results have been reported and osteoporosis cannot be predicted by any single polymorphic gene marker, recent advances in this field have emphasized the complexity of bone mineral mass determination, because of gene-gene and gene-environment interactions.