The effectiveness of regular exercise for bone strength into older age, or when begun after puberty is not fully understood. Comparisons of sedentary individuals and athletes may introduce a self-selection bias, whereas study of tennis players allows the non-racquet arm to act as a control for the exercising racquet arm. Therefore, peripheral quantitative computed tomography (pQCT) scans were taken at radius, ulna and humerus mid-shaft, and distal radius in both arms of ninety veteran tennis players (mean age 63.7±11.8y). Large side differences in muscle and bone size and strength – in favour of the racquet arm – were found. The most pronounced findings were a 13±10% higher distal radius bone mass and 23±12% larger cortical area in the humerus (both P < 0.001). There was no age effect on bone mass (all P > 0.2; except humerus where P = 0.055 and R2 = 0.002). Proximal radius and ulna strength in bending and torsion were positively associated with increasing age (all P < 0.05). Despite no age effect on training volume (P = 0.201) many side differences were less pronounced in older players – particularly in the humerus where side differences in bone mass, cortical thickness and stiffness in bending and torsion were 43-47% less in 80- than 40-year olds (all P < 0.01). Gender effects on side difference were found for a number of bone parameters – in all cases differences were more pronounced in men than women. This was most apparent in the humerus where male side differences in most parameters (except cortical BMD and endocortical circumference) were 22-37% greater than in females despite similar training volumes for both genders. Bone strength side differences in players who began playing in adulthood were smaller than those who began during childhood – in particular in total bone CSA, and periosteal circumference at all sites (all P < 0.05). In summary, whilst regular tennis participation in older age is associated with greater racquet arm muscle and bone size and strength these differences are lower at old age and in females. The age-associated decline is likely due – at least in part – to the reduced racquet arm advantage in muscle size and strength at older age. In addition, the effectiveness of exercise for bone when begun in adulthood (i.e. after epiphyseal closure) is less – mainly due to the lack of significant periosteal apposition.