The common practice of describing the histologic distribution of pulmonary lesions from their radiographic patterns is often inaccurate. Recognition of disease entities is more successful if a given pattern is analyzed for its probable causes, rather than by attempting to predict its microscopic distribution. This represents a change in attitude toward the correlation of the radiographic and pathologic findings in disseminated diseases of the lung. The objective of this communication is to enable radiologists to become more secure in recognizing diffuse pulmonary patterns, to understand the shortcomings of this approach, and to stop trying to predict microscopic distribution from the chest radiograph.