1. With the advent of single-photon emission computerized tomography, controversy has arisen with regard to the significance of gravitational influences on regional pulmonary perfusion (Qr) in the supine versus prone postures. We investigated the dorsal-ventral distributions of Qr in prone (n = 5) and supine (n = 5) normal subjects, as assessed by single-photon emission computerized tomography after intravenous injection of technetium-99m-labelled macroaggregated albumin at end-tidal expiration. Reconstructed serial (one pixel thickness) coronal sections were traced on a computer screen, to yield the encompassed radioactive counts and number of pixels per each image. Coronal section data (expressed as mean radioactive counts/pixel) were expressed in a 'profile' and normalized to the maximum coronal section of each lung (% maximum). 2. Coefficients of variation and linear regression slopes for the prone versus supine profiles for left and right lungs were not statistically different (unpaired Student's t-test). The coronal section with maximum Qr was identified in the more dependent lung regions and, hence, affected by gravity. 3. We conclude that, in contrast to previous canine models, which have suggested postural differences in dorsal-ventral perfusion gradients, in normal man gravity primarily determines the non-dependent to dependent distribution of Qr. We speculate that interspecies differences in physiology may be teleological and related to the different perfusion demands of the quadruped compared with upright man.