From the earliest days of geology it has been recognized that differences between ancient glacial limits and modern ones constitute strong evidence of Quaterary climate change. However, the conversion of geologically demarcated glacial limits into quantitative paleoclimatic conditions has remained very difficult because glaciers respond strongly to both changes in temperature and changes in precipitation, and thus a unique climate state cannot be inferred. Advances in glacial modeling may help to overcome this ambiguity. We have developed a two-dimensional glacier model intended for paleoclimate analysis. It calculates snow accumulation based on precipitation-energy balance. The spatial pattern of snow accumulation drives a two-dimensional glacier flow model that can be tuned to match ancient glacial limits. We have used the model to examine the differences in sensitivity of large glaciers (terminus much below the ELA) and small glaciers (terminus close to the ELA) to decreases in temperature and increases in precipitation. (An energy-balance model is necessary because of the profound effects of terrain shading on small glaciers.) Small glaciers turn out to be more sensitive to changes in temperature because small temperature increases will raise the ELA enough to greatly reduce the glacier accumulation area, whereas large glaciers suffer proportionally much smaller changes in accumulation area and are hence able to integrate the positive effects of increases in precipitation. By simultaneously modeling the limits of large and small glaciers, the combination of temperature and precipitation responsible for the observed limits of both can be uniquely calculated. This approach should be widely applicable to paleoclimate reconstruction in areas of past mountain glaciation.