New paleobotanical data suggest that in the late Eocene the erosion surface which capped the Front Range, Colorado was 2.2--3.3 km in elevation, which is similar to the 2.5-km present elevation of surface remnants. This estimated elevation casts doubt on the conventional belief that the low-relief geomorphic surface was formed by lateral planation of streams to a base level not much higher than sea level and that the present deeply incised canyons must represent Neogene uplift of Colorado. Description of the surface, calculations of sediment volume, and isostatic balance and fluvial landsculpting models demonstrate that while the high elevation of the erosion surface was due to tectonic forces, its smoothness was mostly a result of climatic factors. A sediment balance calculated for the Front Range suggests that from 2 to 4 km of material were eroded by the late Eocene, consistent with fission track ages. This amount of erosion would remove a significant portion of the 7 km of Laramide upper crustal thickening. Isostatic modeling implies that the 2.2--3.3 km elevation was most likely created by lower crustal thickening during the Laramide. A numerical model of fluvial erosion and deposition suggests a way that a late Eocene surface could have formed at this high elevation without incision. A humid climate with a preponderance of small storm events will diffusively smooth topography and is a possible mechanism for formation of low-relief, high-level surfaces. Paleoclimate models suggest a lack of large storm events in the late Eocene because of cool sea surface temperatures in the equatorial region. Return to a drier but stormier climate post-Eocene could have caused the incision of the surface by young canyons. By this interpretation, regional erosion surfaces may represent regional climatic rather than tectonic conditions. ¿ American Geophysical Union 1994