In my former writings I have attributed this kind of derangement to two causes; first, the pressure of ice running aground on yielding banks of mud and sand; and, secondly, the melting of masses of ice and snow of unequal thickness, on which horizontal layers of mud, sand, and other fine and coarse materials had accumulated. The late Mr. Trimmer first pointed out in what manner the unequal failure of support caused by the liquefaction of underlying or intercalated snow and ice might give rise to such complicated foldings.* (* See chapter 12.) When "ice-jams" occur on the St. Lawrence and other Canadian rivers (latitude 46 degrees north), the sheets of ice, which become packed or forced under or over one another, assume in most cases a highly inclined and sometimes even a vertical position. They are often observed to be coated on one side with mud, sand, or gravel frozen on to them, derived from shallows in the river on which they rested when congelation first reached the bottom. As often as portions of these packs melt near the margin of the river, the layers of mud, sand, and gravel, which result from their liquefaction, cannot fail to assume a very abnormal arrangement--very perplexing to a geologist who should undertake to interpret them without having the ice-clue in his mind. Mr. Prestwich has suggested that ground-ice may have had its influence in modifying the ancient alluvium of the Somme.* (* Prestwich, Memoir read to Royal Society, April 1862.) It is certain that ice in this form plays an active part every winter in giving motion to stones and gravel in the beds of rivers in European Russia and Siberia. It appears that when in those countries the streams are reduced nearly to the freezing point, congelation begins frequently at the bottom; the reason being, according to Arago, that the current is slowest there, and the gravel and large stones, having parted with much of their heat by radiation, acquire a temperature below the average of the main body of the river. It is, therefore, when the water is clear, and the sky free from clouds, that ground ice forms most readily, and oftener on pebbly than on muddy bottoms. Fragments of such ice, rising occasionally to the surface, bring up with them gravel, and even large stones. Without dwelling longer on the various ways in which ice may affect the forms of stratification in drift, so as to cause bendings and foldings in which the underlying or over-lying strata do not participate, a subject to which I shall have occasion again to allude in the sequel, I will state in this place that such contortions, whether explicable or not, are very characteristic of glacial formations. They have also no necessary connection with the transportation of large blocks of stone, and they therefore afford, as Mr. Prestwich remarks, independent proof of ice-action in the Pleistocene gravel of the Somme.