The purpose of this study was to compare the bone mineral density (BMD) of two types of trained male cyclists (n = 30) with recreationally active men (n = 15), aged 20-40 years. Sixteen of the cyclists regularly trained for, and competed in, cross-country mountain bike races. The other 14 cyclists trained and raced on the road. The cyclists had trained an average of 11 +/- 3 hours per week for 8 +/- 4 years. Fifteen recreationally active men volunteered as controls. Dual-energy X-ray absorptiometry (DXA) was used to assess BMD of the proximal femur, lumbar spine, and total body. Anthropometric, muscle strength and power, aerobic fitness, and sex hormone data assessments were conducted on all participants. Mountain cyclists were younger and weighed less than road cyclists and controls. BMD at all sites was comparable among the three groups (p > 0.05). When adjusted for body weight and controlled for age, BMD was significantly higher at all sites in the mountain cyclists compared with the road cyclists and controls. Some anthropometric, physical fitness, and sex steroid variables were predictive of BMD, but of these variables, only total body weight, total body fat, and aerobic fitness were different between the groups. In conclusion, endurance road cycling does not appear to be any more beneficial to bone health than recreational activity in apparently healthy men of normal bone mass. Higher BMD in the mountain cyclists suggests that mountain cycling may provide an osteogenic stimulus that is not inherent to road cycling.