OBJECTIVE To establish normative data for muscle performance during isokinetic horizontal abduction and adduction of the shoulder in elite junior tennis players. METHODS Thirty six tennis players were evaluated (23 male, 13 female; mean age 14 years (range 12-18)). An isokinetic dynamometer was used to test the shoulder horizontal abductors and adductors at 60 and 180 degrees/s. Absolute and relative peak torque (PT and PT/BW), total work (TW), endurance ratio (ER), and the ratio of the peak torque between horizontal abductors and adductors (HAB/HAD ratio) were recorded. Data were compared for the dominant and non-dominant shoulders, horizontal abductor and adductor muscles, and between players grouped according to age. RESULTS The dominant shoulder was significantly (p<0.05) stronger than the non-dominant shoulder in all variables except ER and HAB/HAD ratio. The abductors were significantly (p<0.05) weaker than the adductors in all subjects. The type of backhand (one handed or two handed) did not influence the strength of the shoulder horizontal abductors on the dominant side. The number of years of tennis practice had an effect on muscle strength as evaluated by absolute data (PT and TW) but not relative measurements (PT/BW and TW/BW). CONCLUSION The findings confirm that horizontal abduction and adduction are stronger in the dominant shoulder of junior tennis players. The clinical relevance of these findings is not established, and more studies are needed to compare tennis players with athletes from other sports and non-athletes.