We used M-mode echocardiography to measure left ventricular dimensions in diastole (Dd) and systole (Ds) and to assess ventricular performance by computing the percent dimensional shortening (%AD) and the normalized rate of dimensional shortening (Vd) during isometric and isotonic exercise in normal subjects. In 27 subjects, isometric handgrip exercise at 50% of maximum grip until fatigue produced a significant increase in Ds (33 ± 3.4 (SD) vs 30.6 i 3.7 mm, p < 0.001), and a reduction in %AD (34 ± 4 vs 39 ± 5%,p < 0.001) and Vd (1.15 ± 0.15 vs 1.28 0.19 sec-',p < 0.001). Handgrip exercise at 15% of maximum grip produced similar but less marked changes in the 27 subjects, and acute pressure loading with phenylephrine caused similar but more marked changes in 10 of the subjects. In the 20 subjects who performed at least 12 minutes of supine bicycle exercise, Ds decreased significantly (25.6 ± 4.0 vs 31.7 ± 2.8 mm, p < 0.001) and %AD increased (49 ± 6 vs 36 ± 5%,p < 0.001). We observed similar results in the 12 subjects also studied during upright bicycle exercise. Dd was smaller in the upright position but unchanged during either isometric or isotonic exercise. We conclude that: 1) end-diastolic left ventricular size is maintained during isometric exercise and moderate dynamic exercise, even in the upright position; 2) isometric exercise leads to a mild decrease in left ventricular shortening, whereas dynamic exercise results in marked increases in shortening; this difference may be related to the relatively greater increase in blood pressure than in heart rate during isometric exercise; and 3) M-mode echocardiography can be successfully accomplished in selected subjects during various forms of exercise.