Lung parenchyma is normally considered to be isotropic, that is, its properties do not depend upon specific preferential directions. The assumption of isotropy is important for both modeling of lung mechanical properties and quantitative histologic measurements. This assumption, however, has not been previously examined at the microscopic level, in part because of the difficulty in large lungs of obtaining sufficient numbers of small samples of tissue while maintaining the spatial orientation. In the mouse, however, this difficulty is minimized. We evaluated the parenchymal isotropy in mouse lungs by quantifying the mean airspace chord lengths (Lm) from high-resolution histology of complete sections surrounded by an intact continuous visceral pleural membrane. We partitioned this lung into 5 isolated regions, defined by the distance from the visceral pleura. To further evaluate the isotropy, we also measured Lm in two orthogonal spatial directions with respect to the section orientation, and varied the sample line spacing from 3 to 280 microm. Results show a striking degree of parenchymal anisotropy in normal mouse lungs. The Lm was significantly greater when grid lines were parallel to the ventral-dorsal axis of the tissue. In addition the Lm was significantly smaller within 300 microm of the visceral pleura. Whether this anisotropy results from intrinsic structural factors or from nonuniform shrinkage during conventional tissue processing is uncertain, but it should be considered when interpreting quantitative morphometric measurements made in the mouse lung.