“Recognizing from personal observation the full identity of character of the loess of northern China, Europe and the Missouri Valley, I am obliged to reject my own explanation of the origin of the Chinese deposits, and to believe with Richthofen that the true loess, wherever it occurs, is a sub-aerial deposit, formed in a dry central region, and that it owes its structure to the formative influence of a steppe vegetation. The one weak point of Richthofen’s theory is in the evident inadequacy of the current disintegration as a source of material. When we consider the immense area covered by loess to depths varying from 50 to 2,000 feet, and the fact that this is only the very finest portion of the product of rock-destruction, and again that the accumulation represents only a very short period of time, geologically speaking, surely we must seek a more fertile source of supply than is furnished by the current decomposition of rock surface. It seems to me that there are two important sources: I. The silt brought by rivers, many of them fed by the products of glacial attrition flowing from the mountains into the central region. Where the streams sink away, or where the lakes which receive them have dried up, the finer products of the erosion of a large territory are left to be removed in dust storms. II. The second ... source is the residuary products of a secular disintegration.” The evidence presented by Pumpelly for the eolian origin of loess—structure, texture, composition, fossil content and topographic position—is complete, and to him belongs the credit for the correct interpretation of the Mississippi valley deposits. Unfortunately his contribution came at a time when the geologists of the central States were intent on tracing the paths and explaining the work of Pleistocene glaciers, and the belief was strong that loess was some phase of glacial work. Its position at the border of the Iowan drift so obviously suggests a genetic relation that the fossil evidence of steppe climate suggested by Binney in 1848[69] was minimized. Students of Pleistocene geology in Minnesota, Iowa, Nebraska, Missouri, although less vigorous in expression, were substantially in agreement with Hilgard (1879).[70] “The sum total of anomalous conditions required to sustain the eolian hypothesis partakes strongly of the marvellous.” The last edition of Dana’s Manual, 1894, and of LeConte’s Geology, 1896, the two most widely used text-books of their time, oppose the eolian theory, and Chamberlin, in 1897,[71] states: “the aqueous hypothesis seems best supported so far as concerns the deposits of the Mississippi Valley and western Europe” (p. 795). Shimek, in papers published since 1896 has shown that aquatic and glacial conditions can not account for the loess fossils, and the return to the views of Pumpelly that the loess was deposited on land by the agency of wind in a region of steppe vegetation is now all but