Precise quantification of left ventricular (LV) cavity dimensions assumes great importance in clinical cardiology. Pediatric guidelines recommend the left parasternal short axis (PSA) imaging plane for measuring LV cavity dimensions, while measuring from the long axis (PLA) plane is the convention in adult echocardiography. We sought to compare measurements obtained by two-dimensional (2D) and M-mode (MM) techniques in the two imaging planes. Healthy subjects were prospectively recruited for research echocardiography. Complete 2D, spectral and color flow Doppler examinations were performed in a non-sedated state. All subjects had structurally and functionally normal hearts. LV cavity dimensions were obtained in PLA and PSA views using 2D and MM yielding four measurement sets for each subject: PLA direct 2D; PLA 2D-guided MM, PSA direct 2D, PSA 2D-guided MM. A commercially available ultrasound system (Vivid E9, GE) was used and data stored digitally for subsequent analysis (EchoPAC BT11, GE). Acquisition and measurements were made by a single observer from at least three consecutive cardiac cycles, and averaged for each of the four categories. The study cohort consisted of 114 subjects (mean age 9 years, range 1-18; mean BSA 1.1 m(2), range 0.42-2.6). The smallest estimate of LV end-diastolic dimension (LVED) was obtained by PLA 2D, with larger estimates by PLA MM, PSA 2D, and PSA MM. Largest estimates of LV end-systolic dimension (LVES) are by 2D methods, with smaller estimates by both MM techniques. The smallest shortening fraction (SF) was by PLA 2D; other methods yielded larger SF. Temporal resolution is limited in 2D methodology and may account for the smaller LVED, larger LVES and smaller SF observed. Long axis methodology may predispose to off-center or non-perpendicular data acquisition and the potential for dimensional underestimation, particularly in diastole. Consistency in method for assessment of LV dimensions in children is an important factor for serial comparisons.