We conducted a 2-year (42 weeks of consecutive training in each year, separated by IO weeks of testing and vacation rime) randomized, controlled trial of weight training in 142 healthy male and female subjects, aged 60 to 80 years. Measurements included dynamic strength, symptom-limited endurance in cycling, treadmill walking and stair climbing, muscle size, and bone mineral density and content of the lumbar spine and whole body. One hundred and thirteen subjects completed the study (57 exercise, 56 control), with a mean attendance of 85% among the exercisers. Muscle strength was unchanged in the control subjects but increased (collapsed across age and gender) from 32% (leg press) to 90% (military press) in the exercisers. Symptom-limited endurance in cycling, treadmill walking, and stair climbing increased in the exercisers by (mean +/- SE) 6.2 +/- 0.8%, 29.2 +/- 7.3%, and 57 +/- 12%, respectively; the only change in the controls was cut unanticipated 33% increase in stair climbing performance during the first year. These values were unchanged in the controls. Cross-sectional area of the knee extensors increased by 8.7 +/- 0.9% in the trained subjects and was unchanged in controls. Measures of whole body, lumbar spine bone mineral density, and lumbar spine bone mineral content were unchanged in the exercisers, but whole body bone mineral content decreased by 1%. In contrast, there were small increases (<4.0%) in bone mineral density among the controls. Long-term weight training proved to be a safe and wed-tolerated mode of exercise for the elderly. Increased strength was associated with muscle hypertrophy in each year, and with increased endurance in cycling, walking, and stair climbing. There were no changes in bone mineral density bat a small reduction in whole body bone mineral content.