Seventeen volunteers performed unilateral strength-training of the quadriceps with high-resistance, low-repetition, dynamic exercise, thrice weekly for an average of 5 weeks. Both before and after the training period, bilateral measurements were made of isometric quadriceps strength, quadriceps cross-sectional area (by ultrasound scanning), and thigh circumference. There were no significant changes in the untrained thighs. The trained quadriceps increased their isometric strength by more than they changed their cross-sectional area (mean increments = 15% and 6% respectively). Quadriceps hypertrophy was underestimated by measurements of thigh circumference and could not be predicted from them. We conclude that studies of localized muscle growth require direct measurements of the size of the muscle(s) concerned. Nevertheless, these may still underestimate the improvements in strength produced by high-resistance training.