The eigenvector or “empirical orthogonal function” approach is used to determine the dominant precipitation anomaly patterns for the western United States for each month during the last 36 yr. In all months there is enough intercorrelation among monthly precipitation amounts in different parts of the region that at least 45 percent of the total variance can be explained by only three eigenvectors. Usually the most important pattern is one with a single large region of anomalous precipitation, centered in southern California, Arizona, or Nevada in winter and in Washington, Idaho, or Montana in summer. Also important in all months is a pattern with anomalies of opposite sign in the Pacific Northwest and the Arizona-New Mexico-Texas area.