The purpose of this study was to compare the effects of 10 weeks of resistance training with an isotonic bench press machine and 2 types of free-weight bench press exercises on several measures bench press strength. Specificity was investigated by comparing the ability to transfer strength gained from a type of training that differed from the mode of testing. Forty-nine men participated in the study. The subjects completed a pretest on the machine (MB), barbell (BB), isokinetic (IB), and log (LB) bench press to determine baseline strength and completed 10 weeks of training on the MB, BB, or LB. The 3 groups were tested to see whether differential training effects occurred from pre- to post-test scores on the BB, MB, LB, and peak force on the IB. By multivariate analysis, the trial-by-group interaction was not statistically significant. The multivariate and subsequent univariate analyses of variance tests indicated statistically significant effects from pre- to posttest for peak force on the IB test and the BB, MB, and LB. Correlations among the strength tests were high (0.92 > r < 0.97) and moderate between the strength tests and IB peak force (0.62 > r < 0.83). Mean 3 repetition maximum MB strength was 8% higher than BB strength, which was 3% higher than LB strength, indicating differences in the amount of stabilization required to control the resistance. The findings of this study showed that all 3 training groups significantly improved in strength during short-term training on the MB, BB, and LB. These data lend evidence that improved strength after training on the MB, BB, and LB equally transfers to strength gains on any of the 4 modes of testing. These results should be considered when including similar exercises varying in stability into the training program to improve strength. © 2007 National Strength & Conditioning Association.