Studies of athletes in sports favouring one limb (tennis, jumping, etc.) have shown large side-to-side differences in bone strength, muscle cross-sectional area (CSA) and strength in favour of the dominant limb. In sub-elite tennis players, bone strength differences correlated with grip strength and muscle CSA differences(1), supporting the idea of a strong influence of the muscle on bone. To examine an elite population, forty-one national-level tennis players (26m, 15f, mean age 13.4±1.7yrs) were recruited. Bone strength parameters were examined in both arms from pQCT scans at 4% (R4) and 60% (R60) distal-proximal radial length and 35% (H35) distal-proximal humeral length. Muscle CSA (MuscA) was also examined at R60 and H35. Peak force (Fpeak) and power (Ppeak) during a power press-up on a force platform and grip strength (GS) in both arms was measured, along with details of participant’s training history. Data were examined using paired T-tests to locate side differences, univariate ANOVA to examine age/gender effects and linear regression to examine the muscle-bone relationship – data shown as mean +/- SD. Large side differences (in favour of the racquet arm; P < 0.001) were found in MuscA at R60 (20.2±6.6%) and H35 (10.7±5.3%). At R4, total CSAs (Ar.tot) of both radius and ulna were greater (23.3±13.4% and 13.6±28.3% respectively; P < 0.001) in the racquet arm. Radial and ulnar bone mineral density (vBMD.tot) was also greater (15.9±10.8% and 9.1±14.3%) in the racquet arm. These size and density differences resulted in higher racquet arm total bone mineral content (vBMC.tot) in both radius and ulna (39.6±20.5% and 23.7±34.5%; P < 0.001). However, cortical BMC (vBMC.ct) side differences at R60 radius and ulna (19.2±8.8% and 13.8±7.3%) and H35 humerus (39.2±12.9%) were made up almost entirely of a greater cortical bone CSA (Ar.ct) (radius 18.9±8.1%, ulna 15.1±9.0% and humerus 39.8±13.5%) – all P < 0.001, with no significant difference in BMD. Compared to age-matched reference data(2) R60 Ar.tot was 18.3±8.7% greater and MuscA 14.1±18.3% greater than average in the racquet arm (P < 0.01), values in the non-racquet arm were not significantly different than average. Racquet arm muscle-bone ratio was lower in the ulna (3.2±9.0%; P < 0.01), radius (5.6±9.6%; P < 0.001) and humerus (20.2±7.5%; P < 0.001). There were strong correlations between MuscA and Ar.ct in both forearms and upper arms (Figure 1) (P < 0.001). Ppeak (13.0±11.4%), Fpeak (4.9±7.1%) and GS (24.2±26.9%) were all higher in the racquet arm (all P < 0.001). These results show an association of participation in elite-level tennis with side differences in bone strength, muscle size and force/power production. Whilst both arms showed a strong muscle-bone relationship, side differences in these relationships show that other factors aside from muscle size dictate exercise-induced bone adaptation.