Summer half-year (April-September) mean temperatures are reconstructed across western North America between 1600 and 1982. The reconstructions, ultimately in the form of gridpoint anomaly time series, are produced using a principal-components regression technique to relate variability in a network of up to 53 high-elevation maximum latewood-density chronologies to a number of important temperature principal-component amplitude series. The reconstructions are of good quality over the area between 35° and 55°N but are subject to large uncertainty north of 55°N, particularly prior to 1750. Four regional time series, the average of between two and six gridpoint series—British Columbia and the Pacific Northwest, California, eastern Rockies and northern High Plains, and the Southwest deserts—plus one more-extensive western United States series are presented and described. Examples of individual-year and decadal-mcan anomaly maps are illustrated, and the results of preliminary spectral analyses of the r... Abstract Summer half-year (April-September) mean temperatures are reconstructed across western North America between 1600 and 1982. The reconstructions, ultimately in the form of gridpoint anomaly time series, are produced using a principal-components regression technique to relate variability in a network of up to 53 high-elevation maximum latewood-density chronologies to a number of important temperature principal-component amplitude series. The reconstructions are of good quality over the area between 35° and 55°N but are subject to large uncertainty north of 55°N, particularly prior to 1750. Four regional time series, the average of between two and six gridpoint series—British Columbia and the Pacific Northwest, California, eastern Rockies and northern High Plains, and the Southwest deserts—plus one more-extensive western United States series are presented and described. Examples of individual-year and decadal-mcan anomaly maps are illustrated, and the results of preliminary spectral analyses of the r...