In western North America, pollen data from highland lakes are often used to reconstruct vegetation on the adjacent lowlands. Plant macrofossils and pollen from packrat middens now provide a means to evaluate such reconstructions. On the basis of pollen diagrams from the Chuska Mountains, H. E. Wright, Jr., A. M. Bent, B. S. Hansen, and L. J. Maher, Jr., ((1973), Geological Society of America Bulletin, 84 , 1155–1180) arrived at conservative estimates for late Pleistocene depression of highland conifers. In their interpretation, a proposed slight depression of 500 m for lower tree line precluded expansion of Pinus ponderosa into elevations now in desertscrub. Instead, it was suggested that pinyon pine and Artemisia occupied the lowland plateaus. Packrat midden records on either side of the Chuskas fail to verify this model. Early Holocene middens from Chaco Canyon, New Mexico, and a terminal Pleistocene midden from Canyon de Chelly, Arizona, show that blue spruce, limber pine, Douglas fir, dwarf juniper, and Rocky Mountain juniper expanded at least down to 1770 m elevation Neither Colorado pinyon nor ponderosa pine was found as macrofossils in the middens. Artemisia pollen percentages are high in the terminal Pleistocene midden, as they are in the Chuska Mountain pollen sequence, suggesting regional dominance by sagebrush steppe. Of 38 taxa identified, only 3 are shared by middens dated 11,900 and 3120 yr B.P. from Canyon de Chelly, indicating a nearly complete turnover in the flora between the late Pleistocene and late Holocene. Although corn was previously thought to have been introduced to the Colorado plateaus after 2200 yr B.P., the midden dated 3120 yr B.P. contains pollen of corn and other indicators of incipient agriculture.