Until now, right atrial (RA) volume calculation by means of two-dimensional echocardiography (2-DE) has only been attempted in a single plane: the apical four-chamber view. Our study reports a new method for RA volume calculation using two intersecting 2-DE views. For this purpose, silicone rubber casts of 19 human necropsy hearts were obtained and thin-walled natural rubber moulds of the RA casts were prepared. Totally filled with and immersed in water, the moulds could be visualized in the apical four-chamber view and an additional 2-DE plane, the latter corresponding to the subcostal view in vivo. In this view the vertical extension of RA could be estimated. Areas and lengths of RA were determined in the respective planes, and RA volume was calculated by applying the formula, area x length, to two intersecting planes. Finally, volume of the silicone casts was determined angiocardiographically (Angio) using a biplane method (30 degrees RAO, 40 degrees LAO-40 degrees hepatoclavicular). The true RA volume was 106 +/- 23 ml (mean +/- 1 SD) as determined by water displacement. Using Angio an excellent correlation was found: the calculated volume amounted to 106 +/- 23 ml; the difference was 5.5 +/- 4.8 ml (n.s.); Angio vol = 0.93 true vol + 7.77; r = 0.95; SEE = 7.4 ml. Volume determination from the apical four-chamber view of 2-DE using a monoplane disk method resulted in a mean volume of 62 +/- 17 ml. The mean difference to the true RA volume was 44 +/- 16 ml (p less than 0.001). When volume calculations were made using the biplane method, a value of 105 +/- 22 ml resulted. The mean difference to true volumes was 7.4 +/- 4.8 ml: y = 0.84x + 15.88; r = 0.91; SEE = 9.4 ml. In an in vivo study endsystolic RA volumes were calculated in a normal adult population (n = 40) from the same intersecting planes as in vitro. A normal value of 38 +/- 6 ml/m2 was found. In vivo validation using Angio showed a slightly higher normal value of 43 = 7 ml/m2. Thus, 2-DE is highly accurate in determining RA volume. In the in vitro as well as in the in vivo study the results of monoplane calculations are clearly inferior to a method which also takes account of the vertical extension of RA.