Patterns of variation in bone size and shape are used to reconstruct many aspects of hominin biology, including ecogeographic adaptation, life history, and functional morphology. Bone thickness, or robusticity, is particularly useful for inferring behavior due to the substantial phenotypic plasticity of bone tissue. Exercise-induced strains can stimulate periosteal modeling, so bone thickness often serves as a proxy for individual loading history. However, strain is only one of numerous genetic and environmental influences on bone growth. The potential effects of physiological factors, such as hormones and growth factors, on bone-strain interactions are largely unexplored.

In this thesis, I test the hypothesis that the hormone estrogen affects exercise-induced changes in skeletal robusticity. Estrogen (E2) is of particular interest for understanding bone-strain interactions because it mediates longitudinal and periosteal bone growth, as well as bone’s mechanosensitivity via the estrogen receptor alpha (ERa). I develop a new model for interactions of estrogen, ERa, strain, and bone growth. The prediction is that greater estrogen bioavailability during skeletal growth will increase mechanosensitivity and therefore osteogenic responses to mechanical loading, while reduced estrogen exposure will decrease mechanosensitivity and diminish osteogenic responses to comparable mechanical loads.

Tests of the estrogen-strain model included controlled experiments in sheep and mice and a comparative study in humans. In sheep, cortical bone growth in exercised, high-E2 animals was 6-27% greater than in exercised animals with lower E2 levels, or sedentary animals regardless of E2 dose. In mice, E2 treatment was associated with 23- 34% increases in mandibular bone growth in response to masticatory strains. In young women, both E2 and exercise influence bone bending strength. Women with the highest E2 levels in the first postmenarchal year had 8-12% greater femoral neck bending strength vs. lower-E2 women, and the most physically active women had 9-11% greater femoral neck bending strength vs. less active women.

These results support the hypothesis that E2 alters osteogenic responses to mechanical loading. The implications of the estrogen-strain model for inferring human behavior from patterns of skeletal robusticity are discussed, along with some testable hypotheses about estrogen’s role in human evolution, particularly in mediating skeletal vs. reproductive development.