Estimation of lung volumes by conventional methods in sick infants is technically difficult and is the subject of controversy. In this study, we compared both thoracic gas volume (TGV), measured with an infant whole body plethysmograph, and functional residual capacity (FRC), determined by the nitrogen washout technique, to planimetric measurements of anteroposterior chest radiographs in 26 infants with bronchopulmonary dysplasia (BPD). The tidal volume (TV) of each patient was added to TGV and FRC because these were measured at the end of expiration whereas chest radiographs were taken at the end of spontaneous inspiration. The regression equations expressing the relationships between TGV and right+left lung field areas [A+B], and between FRC and lung areas are expressed as follows: [TGV+TV](mL) = 3.3 mL/cm2 x [A+B] cm2 + 24 mL and [FRC+TV](mL) = 3.5 mL/cm2 x [A+B] cm2 - 13.5 mL. Correlation coefficients of 0.9 and 0.7 for TGV and FRC, respectively, suggest a stronger correlation between TGV and lung areas than between FRC and lung areas. Lung areas measured by planimetry correlate closely with physiological measurement of lung volumes. We conclude that the planimetric method is an inexpensive and reliable technique for estimating lung volumes in young infants with BPD when chest radiographs are available.