A growing number of studies suggest that vegetation changes can significantly influence regional climate variations. Herein we utilize a climate model (GENESIS) with a land surface vegetation package to evaluate the potential role of the very large vegetation changes that occurred during the last glacial maximum (LGM). In particular, we focus on the potential response to a significant reduction in the area of tropical rainforest. Simulations employed a global vegetation reconstruction for the LGM and Climate/Long-Range Investigation, Mapping and Prediction (CLIMAP) sea surface temperature (SST) estimates. Results indicate that expansion of dryland vegetation causes a 15--30% additional LGM cooling for Australia (0.4 ¿C) and Africa (0.9 ¿C), respectively. Turnover from conifer to tundra also causes cooling of 2¿--4 ¿C or more in western Europe and Siberia. However, for the largest rainforest area (Amazon Basin), inclusion of realistic vegetation increased modeled temperatures 2¿--4 ¿C and decreased precipitation by 10--35%. These latter results are similar to those obtained with sensitivity experiments of the effects of future Amazon deforestation. Initial assessment of the potential effect of decreased stomatal resistance due to lower ice age CO2 levels indicates little significant response to this effect. Comparison of model-predicted low-elevation LGM temperature changes with estimates from proxy data indicate that inclusion of realistic vegetation estimates for the LGM results in slightly more than 50% agreement between models and data for low-elevation sites in low-mid latitudes. Data at variance with model predictions would appear to be explainable by considering additional changes in vegetation, ice age dust, or a 1¿--2 ¿C cooling below CLIMAP values. This conclusion is at variance with a 3¿--4 ¿C tropical cooling suggested by some studies for explaining estimated land temperature changes during the LGM. In some western European sites model temperatures are colder than proxy data by 2¿--8 ¿C. This model-data discrepancy may be explained by less sea ice in the subpolar North Atlantic than stipulated by CLIMAP, a conclusion consistent with new marine data from that region. ¿ 1997 American Geophysical Union