Studies using gravimetric analysis of lungs of frozen animals have suggested that the differences in pulmonary microvascular pressure between non-dependent and dependent lung do not influence the formation of regional pulmonary edema. We wondered if the inability to detect variation in regional extravascular lung water (EVLW) was due to the slow freezing process and, therefore, reassessed the distribution of EVLW in vertically suspended isolated perfused dog lungs with a radioisotopic technique that does not require freezing. Total lung water (TLW), blood or intravascular lung water (IVLW), and EVLW were measured in absolute quantities using a positron camera and the positron-emitting isotopes C15O as a blood label and H2(15)O as a total lung water label. Mean isotopic TLW in 17 lungs that were normal or moderately edematous (wet:dry ratio < 7) was 142 +/- 9 (SE) ml compared to the gravimetric estimate of 148 +/- 7 ml (r = 0.92) and isotopic EVLW was 64 +/- 6 ml compared to the gravimetric estimate of 70 +/- 6 ml (r = 0.8). Analysis of the distribution of regional isotopically measured EVLW in the 17 lungs in various states of spontaneous edema formation revealed a small non-dependent to dependent, gravity-related increase in percent regional EVLW compared to percent regional TLW, which did not vary with the degree of edema in the lung. Serial measurements of absolute regional EVLW in four lungs during spontaneously developing edema also failed to show a disproportionate increase in accumulation of EVLW in any lung zone. Thus, despite the wide variation in microvascular hydrostatic pressure between top and bottom of the vertical isolated lung, edema formation seems to be uniform.