The purpose of this study was to highlight the issues related to thickness-based muscle size evaluation that is commonly done in field studies. The cross-sectional area, thickness (the vertical distance from the upper end of the elbow flexors to that of the humerus) and width (the horizontal distance from the left to the right end of the elbow flexors) of the elbow flexors at levels from the reference site (60% of the upper arm length from the acromial process of the scapula to the lateral epicondyle of the humerus) to 5 cm distal to it were determined in 11 young men using magnetic resonance imaging, both at rest and during isometric elbow flexion at 30% of maximal voluntary contraction. During 30% of maximal voluntary contraction, the thickness increased but the width decreased at each measurement site compared with those at rest. This was possibly due to difference in muscle slackness between both conditions. The correlation coefficients between the thickness and cross-sectional area for the elbow flexors were significantly lower at rest ( r =0.551–0.856) than during 30% of maximal voluntary contraction ( r =0.711–0.922). The present findings indicate that the thickness-based muscle size measurement at rest includes errors owing to the slackness of the resting muscles.