208 Scientific Proceedings, Royal Dublin Society. 



all over the country and protect the surface against further erosion. The pampas 

 are sometimes frozen when the water is running freely in the river valleys ; 

 and I may mention a very striking example of this which I myself witnessed. 

 On the north side of the Gallegos river, almost in front of the town, there is 

 a huge canadon which runs back about two and a half miles ; it is very broad, 

 and has several tributary cafiadones, one of which, after rising rather abruptly, 

 ends on the pampa. In the end of June, 1913, I was going to Coyle, and left 

 the river-side at 9 a.m. A misty rain was then falling, and quite a fair- 

 sized little river was running down the canadon ; there was no snow below, 

 and everything was wet and sloppy. The stream continued, although 

 gradually diminishing in size, as I went up, until I got within fifty feet of 

 the top, when it ended in snow and ice ; the mist had by this time turned 

 into sleet, and I was surprised on going about two hundred yards further and 

 reaching the pampa to find that this latter had in its turn changed to 

 snow. I was also surprised to find the pampa dry, hard, and covered with snow 

 and ice, with no sign of thaw. My horses here gave in, and I was compelled 

 to return; and, when I came a half a mile down the canadon, I found every- 

 thing was as before — the same running stream, the same mud, and the same 

 misty rain. It is also a well-known fact that after a severe winter the snow 

 and ice will have completely disappeared from the pampas near the sea fully 

 a month, if not more, before they begin to melt on the same levels sixty to 

 seventy miles inland. 



These facts are very significant, and indicate that when the great ice 

 mantle began to melt it did so first along the sea-coast and in some low-level 

 bay or bight. This bay would be cut out first to a considerable extent, and 

 would eventually mark the starting-point of a river which would cut its way 

 back into the high ground as the ice receded. According to this method of 

 reasoning, it would seem that the shingle covering the river terrace we have 

 been considering came down towards the end of a flood period, and was 

 deposited at a time which came immediately before a phase of mild climate 

 such as exists to-day. All along the Gallegos river valley we find numerous 

 lava-sheets ; some of them are on the pampa, but most of them are found 

 capping the terraces, even down to the lowest. Now these lava-sheets, except 

 where they are broken up by former ice or water action, show very even 

 and compact distribution, and seem to have been poured out over a dry-land 

 surface ; for, had there been floods at the time of the volcanic outpourings, the 

 lava would have been immediately cooled as it met the water, forming huge 

 frizzled-up heaps, and would not have become spread out in thin, hard, compact 

 sheets. Now, except in one instance in the canadon Guer Aike, I have not 

 met with any deposit of shingle on the surface of any of these lava-sheets — at 



