206 Scientific Proceedings, Royal Dublin Society. 



llie water runs over it and escapes, some of it into the lagoons and the rest 

 into the canadones, through which it eventually makes its way to the rivers. 

 Thus the essential factor in the production of floods is not rain, but snow, and 

 these floods are always found in the spring of the year, when the snow melts. 

 The snow which falls on the plains and lowlands is also most productive of 

 floods. If the climatic conditions as they exist at present became changed in 

 such a way that the winter snows were increased to four or five times their 

 present amount, the spring floods would be correspondingly increased, and, 

 instead of having a de]Dositing action, as they have at present, they would 

 first clear all the recent soil out of the valley, and next all the gravel and 

 sand, and finally the bed-rock would be cut into, so that the river valley 

 would once more become deepened, and probably new terraces would come 

 into existence. This sequence of events, namely, a change from a period of 

 mild climate such as exists at present to one when the snow-fall was con- 

 siderably increased and the spring floods were hugely augmented, probably 

 occurred several times in the past, and it is likely that all the terraced river 

 valleys were cut down intermittently in this way. It seems, then, that the 

 present spring floods are only the shrunken remnants of what once existed 

 on a grand scale, and that all the great river valleys and canadones were 

 originally cut out by the action of spring floods, due to the melting of large 

 masses of snow and ice. 



We are now in a position to adopt certain rough divisions of the past 

 history of Southern Patagonia as far back as our story goes. We may 

 divide it into, first, a pre-shingle period, when wind was the dominant 

 eroding force ; second, a shingle period, which probably was identical with 

 the first great advance of the ice — no river valleys or cafiadones existed 

 then ; third, a long period of erosion by water, which was subject to con- 

 siderable oscillations due to alternations in the climatic conditions ; and, 

 fourth, a genial period, which is still going on, and which has probably 

 existed for some thousands of years. 



VII. 



Erosion subsequent to the outpotiring of the Lava. 

 The surface of a river terrace is covered with sand and shingle, very 

 'similar to that which is found on the pampa, of about the same thickness, and 

 distributed practically as evenly. If, however, a canadon or gully which cuts 

 into one of these terraces be examined, it will be found that this shingle- 

 layer is somewhat diflerent. A cutting running into one of the terraces near 

 the mouth of the Coyle river shows marked current bedding, the pebbles 

 being sorted out into irregular layers, and the spaces between being filled 



