TOBRENTS OP THE HIGH ALPS. 23 



the warmth of spring supervene without a gradual preparation there is 

 poured forth in the course of a few days the mass of water accumulated 

 during months. This may be considered one of the principal causes of the 

 violence of certain floods. 



He cites the torrents which proceed from the Col Izoard towards Arvieui, 

 to which reference has already been made, as presenting the most complete 

 and perfect type of the gullet of a bassin de reception. There, as has been 

 stated, more than sixty torrents, within less than 3000 metres, or two miles, 

 precipitate into the depth of the gorge the debris torn from the two flanks 

 of the mountain. 



In the torrents of the second kind the basin, instead of being cut out on 

 the cols of the mountains, is formed by an indentation of their summits, and 

 is hoUowed out in their revers. It is in this kind of basin that it is easiest 

 to trace the disposition to assume the funnel shape so characteristic of 

 these basins, as the eye can take into one glance the entire course of 

 the torrent, aU the points of which are depicted before it. The torrent of 

 Merdinal, at Saint-Cr6pia, may be cited as a type. 



Lastly, in the third kind the basin is reduced to a kind of large bog, 

 hoUowed out by some ravine, and which in the country often bears the 

 name of comie, as for instance the Combes of Puy-Saniere, the torrent of 

 Combe-Barre, the torrent of Comboye. 



It receives no affluents or feeders, and it collects little more, if any, than 

 the waters which faU in the same enclosure as the depression. It is always 

 dug out in the flanks of the mountains and below their summit; but it tends 

 to grow, and it creeps up little by little towards the summit, which it 

 reaches at last. This process goes on with greater rapidity iu grounds 

 subject to rapid disintegration, and thus is formed in the long run many 

 of the torrents of the second kind. And one can, in many cases, foUow 

 the progress and the different phases of the formation of these, from their 

 nascent condition on to their complete development. 



Below the basin of reception, and in continuation of the gullet, is a region 

 in which there is neither any more downfall of earth nor is there as yet any 

 deposit. This is designated the caned d^ecovlement. Of the three parts of 

 the torrent this is the least marked by characteristics, and almost alwayS the 

 least extended. It is the longer the more gentle are the changes of inclina- 

 tion in its bed And this is the reason why it is generally pretty lengthened 

 in torrents of the first kind ; it becomes shorter in those of the second ; and 

 lastly, in those of the third, it reduces itself almost to a vanishing point. 



The canal dCecovlement is always contained between mountains well defined. 

 In fact, when there are no mountains the slope does not suffice to prevent 

 the torrent from spreading itself out; and in doing this it woiid lose 

 velocity and it would cease to be. 



The canal dOcovlement is the only part of the course ia which the torrents 

 do little damage. UnhappUy it is the least extensive. It is here bridges 

 should be located. 



If we could artificially prolong this channel to its confluence with the 

 river, maintaining throughout its slope, its section, and its course, we would 

 stop the ravages. And this is the problem in the embankment of torrents. 



The lUs de dejections, or beds of deposit, at the mouth of the torrent next 

 demand attention. The aspect of many of these is suggestive at first sight 



