THEOKIES OF DKIFT. 91 



the agents, and they differ only in the proportion and manner in which they have oper- 

 ated. One class consider the whole country to have been overspread for a long time with 

 a sheet of ice of great thickness, in the form of a glacier, which urged its way southerly, 

 scoring the rocks and transporting the bowlders. The access of heat to melt the glacier 

 produced abundant currents of water, by which deposits of finer materials were made 

 upon the drift, or sometimes in the midst of it. 



We are unable to adopt these views ; first, because all known glaciers are confined to 

 valleys, though at their head they may be connected with extensive fields of ice, capping 

 the summits of the mountains : secondly, because no known glacier is more than 50 or 60 

 miles wide (the great glacier called Humboldt, in Greenland, described by Dr. Kane, is of 

 this width), whereas the ancient American glacier must have been at least 2500 miles wide, 

 and have spread over all the mountains as well as valleys, and often have been obliged to 

 move up hill as well as over a level surface : thirdly, because in our country we have two 

 and probably three prominent directions to our drift, and it is difficult to see how one 

 glacier would have moved in so many directions, especially as the most usual course of 

 the striae in New England does not follow a valley, but crosses over mountains obliquely. 



We have no doubt, however, that a part of what we call drift phenomena in New Eng- 

 land was produced by glaciers, such as we have described as once connected with the 

 Green Mountain range. But the main features of drift we impute to icebergs and ice- 

 floes, as the continent was gradually sinking beneath the ocean. All will admit that it 

 was depressed 500 or 600 feet, as late as the drift period, because to that height we find 

 in Champlain valley and in Canada sea shells in the regular deposits, and we shall 

 endeavor to show that we have other evidence that it went down 2000 feet, or more than 

 sufficient in fact to bring icebergs over the tops of most of our mountains. Their effects, 

 in connection with that of shore ice on the rocks at the bottom of the ocean, could not be 

 distinguished from those of glaciers, both having gravel and bowlders frozen into their 

 under side to act as gravers. Some icebergs in the Atlantic have been seen that extended 

 2500 feet below the surface, deep enough to reach the bottom anywhere, if we suppose the 

 ocean to have risen 2000 feet above its present level. Moreover, such a depression of the 

 continent would bring a northerly or northwesterly oceanic current over it, which would 

 bear along icebergs and icefloes, as is now done in the Atlantic. These would depress 

 the temperature so that glaciers might form on the mountains as they rose out of the 

 ocean. 



Two circumstances make it probable that the main features of drift were produced by a 

 sinking rather than a rising continent : One is, that in this way we can explain the occur- 

 rence of striae on the stoss side of mountains, running up hill. For, suppose an iceberg to 

 strand on the northern side of a hill which was sinking ; as the hill went down, however 

 slowly, the iceberg would be urged by the southerly current behind it higher and higher up 

 its side, and at length over it, pressing hard upon the ledges beneath and scoring the sur- 

 face. The other circumstance is, that many of the phenomena of modified drift indicate 

 that most of this was deposited while yet the continent was beneath the ocean, and that the 

 valleys formed in it were the result of drainage, which implies an elevation of the land. 

 In many cases we can show that old river beds were filled by modified drift during the 

 sojourn of the continent beneath the ocean, and if so it must have been done before the 



