For reproducible measurements of right ventricular (RV) volume and function, it may be important to use a consistent method to identify end systole (ES). We determined whether a significant difference exists between RV volumes measured using varying criteria from previous studies to define the timing of ES. In three normal subjects and nine patients with congenital heart disease, we measured RV volume from 3D reconstructions generated from 12 short and long axis magnetic resonance images (MRI). Cine frames analyzed included two frames before and three frames following ES, which we determined as the frame in which chamber area was most frequently minimum. ES coincided with onset of aortic valve closure in ten of 12 subjects; complete closure occurred 1 frame later. The tricuspid valve began to open 1-2 frames after ES, and completely opened 2-4 frames after ES. RV volume was unchanged between ES and the frame following. However, ES volume differed significantly from volume measured 1 or 2 frames before ES and from volume measured 2 or 3 frames following ES, although these volume differences lay within the range of observer variability. The time of minimum RV area in the 4-chamber view agreed closely with that of ES (intraclass correlation coefficient = 0.962). We conclude that minimum RV area in the 4-chamber view is a convenient marker of use for ES, and that aortic valve closure or onset of tricuspid valve opening could also be used, being unlikely to result in clinically significant errors.