Tidal volume measurements based on the sum of volume displacements of the rib cage (RC) and abdomen (Ab) are limited in accuracy when changes in posture occur. To elucidate the underlying sources of error, five subjects performed spinal flexion-extension isovolume maneuvers and then performed Konno-Mead isovolume maneuvers at different lung volumes while erect, with the spine fully flexed, and at intermediate degrees of spinal flexion. RC and Ab dimensions were measured with respiratory inductance plethysmograph belts, and spinal flexion was assessed by a pair of magnetometers measuring the xiphi-Ab distance (Xi). RC and Ab volume-motion coefficients (alpha and beta, respectively) were calculated from the slope (-beta/alpha) of the Konno-Mead isovolume lines. We found that 1) spinal flexion with constant lung volume mainly increases the RC dimension, thereby displacing the Konno-Mead isovolume lines, and 2) spinal flexion decreases the -beta/alpha by decreasing beta. The error related to displacement averaged 28.4 +/- 15% of vital capacity, whereas the error related to changes in beta averaged 14 +/- 6% (SD). The systematic relationship of these errors with the degree of spinal flexion provides a mechanism whereby the addition of Xi to RC and Ab displacements significantly (P less than 0.001) improves volume estimates.