Passing further north to the Andes, between 39° and 44° south latitude, the glaciation was not so severe, and its records are therefore clearer. The first result of elevation was the cutting of deep canyons by the rivers. This was followed, possibly without much further elevation, by a fall of temperature, which in this connexion may be attributed to the extension of the Antarctic and Tierra del Fuego ice-sheets. Glaciers now developed and spread down the canyons, leaving moraines of great volume and height, associated with all the other criteria of glaciation. The snow-fields from which these glaciers originated lay between 5000 and 6000 feet above the sea, and the snow-line lay at about 3000 feet instead of above 6000 as at present. This glaciation was followed by a long interglacial, during which the glaciers retreated to the highest summits of the Andes. The length of this period is indicated by the fact that the earlier moraines have been eroded to such an extent that they no longer present distinctly the typical features of glacial topography, while the materials of which they are composed are decayed to somewhat the same extent as the older moraines of North America, the granite boulders especially being rotten and friable. This interglacial was followed by a re-development of the glaciers, but to nothing like the same extent as formerly; their moraines are smaller and fresh-looking, indicating that this glaciation was comparatively recent. Still further north, in latitude 20°-25° S., we come to a region of very slight snowfall, where the snow-line lies higher than anywhere else on the face of the earth. The glaciation here was comparatively unimportant, the snow-line descending only 1600 to 2500 feet. Here Keidel found moraines of three glacial advances, and from his description it appears probable that the earliest and greatest was separated by a considerable interval from the two younger, the interglacial between which was short and not characterized by a return to present-day climatic conditions, since during this interval there was very little weathering. Probably we have here to do with two glaciations, of which the second was double. In fact, some writers have described no less than five glacial advances in the Argentine Andes, but most of these are probably merely retreat stadia.