A recently published theory of deep glacial erosion of the Canadian Shield (White, 1972) is shown to be consistent with the present distribution of Phanerozoic impact craters on the shield. The anomalous preservation of early Phanerozoic craters on a Precambrian surface can be explained by assuming: (1) that an impact crater can be preserved over hundreds of millions of years only if it forms on a stable area and is soon buried below sediments, and (2) that the sedimentary layers covering the shield craters were only recently removed by Pleistocene glacial erosion. This erosional history carries the implication of a particular relation between size, age, and location of the shield craters. Although the data on the craters are insufficient to statistically choose between White's theory and the traditional view of glacial erosion, the pattern implied by White's theory is not definitely violated by any of the 14 known craters. Using this theory, the lack of craters from the first and last 100-m.y. periods of the Phanerozoic can be attributed to erosional processes, and evidence for variation in the rate of infall of meteorites onto the earth is indicated only for one anomalous period some 350 m.y. ago. This analysis will be proved correct or not as further impact craters are discovered and dated, for it presents a consistent theory of crater preservation on the earth that prohibits, for example, the presence of small Paleozoic craters on Precambrian shield areas not exposed to Pleistocene glaciation. Independent support of White's theory would be provided by the success of this analysis and would indicate that determining the terrestrial rate of impact cratering will be very difficult.