The geomorphology of Mars

interfering with westward expansion by advocating conservation in water and other natural resources. In its wisdom Congress halved the Survey’s budget. The chief geologist for the Survey, G.K. Gilbert, found that he had no money for field work. Gilbert, who was not one for idleness, began to spend his evenings peering at the moon through the naval observatory telescope in Washington. His study of the moon (Gilbert, 1893) clearly demonstrated that a scientist experienced in the analysis of earth surface processes could make a valuable contribution to the study of other objects in the solar system. Modern geomorphology can still benefit from Gilbert’s example (Baker and Pyne, 1978). One reason for the profound development of geomorphology in the late nineteenth and early twentieth centuries was the intense scientific study of newly discovered terrains in Africa, Asia, Australia, and western North America. With new terrestrial landscapes no longer available to stimulate interest, the modern frontier of geomorphology lies in the comparative study of planetary surfaces. Sharp (1980, 231) has summarized this trend as follows: Planetary exploration has proved to be a two-way street. It not only created interest in earth-surface processes and features as analogues, it also caused terrestrial geologists to look on Earth for features and relationships better displayed on other planetary surfaces. Impact cratering, so extensive on Moon, Mercury, and Mars, is a well-known example... Another is the huge size of features such as great landslides and widespread evidence of large-scale subsidence and collapse on Mars, which suggests that our thinking about features on Earth may have been too small-scaled. One of the lessons from space is to ‘think big’. Interplanetary surficial geology is mainly accomplished with images produced by the sensing devices on spacecraft. These images have their own special qualities (Saunders and Mutch, 1980). The interpreter of landforms on those images relies on analogic reasoning to reconstruct the complex interaction of processes responsible for the observed features. The limitations of this approach are twofold