The effect of strength training and endogenously elevated hormone levels (plasma testosterone, growth hormone (GH) and cortisol) was studied in 16 young untrained males, divided into an arm only training group, A, and a leg plus arm training group, LA, in order to increase circulating levels of anabolic hormones. Both groups performed the same one-sided arm training for 9 weeks, twice a week. Group A trained only one arm (AT), the contralateral arm serving as control (AC), whereas group LA additionally trained their legs following the training of the one arm (LAT), with the contralateral arm serving as control (LAC). In spite of the attempt to match the two groups, the initial isometric arm strength was 20-25% lower for group LA compared to group A (significant for the arm to be trained). Isometric strength increased significantly in LAT and LAC by 37% and 10%, respectively, while the 9% and 2% increases in AT and AC, respectively, remained insignificant. Isokinetic strength increased at one out of three velocities tested for the trained arm relative to the untrained arm in both group A and group LA (P<0.05). Functional strength increased significantly by 20% in LAT, 18% in LAC, 19% in AT, and 17% in AC. Hormonal responses were monitored during the first and last training sessions. Resting hormone levels remained unchanged for both groups. However, during the first training session plasma testosterone as well as plasma cortisol increased significantly in group LA but not in group A. Plasma GH rose in all exercise tests, except during the last test in group LA, but was significantly higher in group LA than in group A in the first training session. In conclusion, a larger relative increase in isometric strength was found in the group having the highest hormonal response. However, due to the initial difference in isometric strength caution must be taken with the interpretation of this finding, which may only indicate a possible link between anabolic hormones and muscle strength with training.