Glutaraldehyde is widely used to chemically fix lungs for analysis of pulmonary structure-function relations. Accurate interpretation of observations on fixed tissue requires a clear definition of any artifacts, such as tissue shrinkage, resulting from fixation with glutaraldehyde. Two experimental procedures were used in this study to examine possible shrinkage artifacts resulting from fixation of lung by glutaraldehyde. In the first, isolated perfused dog lungs were rapidly frozen at different transpulmonary pressures. Samples were then freeze substituted at -50 degrees C using 70% ethylene glycol with and without fixatives present. In the second series of experiments, the left lungs of mongrel dogs were fixed by vascular perfusion with glutaraldehyde at different transpulmonary pressures. In both series of experiments any changes in linear dimensions resulting from the fixation procedure were measured. Also, the presence of aldehyde was demonstrated by a positive reaction with Schiff reagent. The results demonstrate that lung tissue fixed either by vascular perfusion or freeze substitution tends to shrink to about the same extent. This shrinkage is reasonably constant at about 9% for transpulmonary pressures of 5 and 15 cmH2O and increases to about 15% when the transpulmonary pressure reaches 25 cmH20.