To study the role of serum testosterone (T), estradiol (E(2)), and SHBG as regulators of peak bone mass and bone turnover rate in males, a cross-sectional study with data on lifestyle factors collected retrospectively was performed in 204 young Finnish men, 18.3-20.6 yr old. One hundred fifty-four men were recruits of the Finnish Army, and 50 were men of similar age who had postponed their military service for reasons not related to health. Bone mineral content, density, and scan area were measured in lumbar spine and upper femur by dual-energy x-ray absorptiometry. Blood was sampled for determination of serum total and free T, total and free E(2), SHBG, type I procollagen aminoterminal propeptide (PINP), total osteocalcin (TOC) and carboxylated osteocalcin (COC), and tartrate-resistant acid phosphatase 5b (TRACP5b); and urine was collected for determination of type I collagen aminoterminal telopeptide (NTX). Serum sex steroid concentrations did not associate with bone mineral content, scan area, or bone mineral density, adjusted for anthropometric and lifestyle factors at any measurement site. Instead, serum total (r = 0.23; P = 0.008) and free (r = 0.15; P = 0.023) T were positive predictors of serum TOC, whereas serum free E(2) correlated inversely with serum PINP (r = -0.20; P = 0.0039), TOC (r = -0.12; P = 0.086), COC (r = -0.14; P = 0.036), and urinary NTX (r = -0.15; P = 0.041). Interestingly, serum SHBG correlated positively with all the bone markers studied, the correlation coefficients being 0.18 for serum PINP (P = 0.012), 0.24 for TOC (P = 0.0006), 0.24 for COC (P = 0.0005), 0.27 for serum TRACP5b (P < 0.0001), and 0.21 for urine NTX (P = 0.0031). Serum SHBG was also a positive predictor of serum 25-hydroxyvitamin-D level (r = 0.20; P = 0.0036). The correlations of SHBG persisted after adjusting for weight, free E(2), and free T. We conclude that single measurements of serum E(2) and T were not determinants of peak bone mass in this population of young men. However, E(2) and T contributed to bone turnover rate, with serum T increasing bone formation, and serum E(2) suppressing both bone formation and resorption. Moreover, serum SHBG appeared to be an independent positive predictor of bone turnover rate, which also positively associated with serum 25-hydroxyvitamin-D levels.