276 



Atlantic. I discuss here the possibility only of 

 executing works to which there will certainly 

 be no need to have recourse. 



The expence of water for feeding a canal in- 

 creases, with the extent of the nitrations, the 

 frequency of passages, or of the lockages {ex- 

 clusee *) and with the size of the chambers of 

 locks, but not with their number. The facility 

 of collecting an enormous mass of rain waters 

 within the tropics, is beyond what the engineers 

 of Europe can imagine. When Lewis the 14th 

 ordered the gardens of Versailles to be embel- 

 lished, Colbert was made to hope that the rains 

 would furnish, on a surface of 1 2,700 hectares 

 of plains which communicated with ponds and 

 reservoirs, 9 millions of cubic toises of water 'f-. 

 Now the rains in the vicinity of Paris amount 

 annually only to from 19 to 20 inches, while 

 within the torrid zone in the New World, above 

 all, in the region of the forests, the quantity is 

 at least from 100 to 112 inches J. This irn- 



* The exclude is the successive filling of the lock to en- 

 able the boats to ascend or descend in a canal, at the point of 

 partition. 



t Only ^ could be collected ; the remainder was lost by 

 filtrations, and it became necessary to construct the machine 

 of Marly : Huene de Pommeuse, sur les canaux navigables. 

 Supplement, p. 45. 



X See above, Vol. ii, p. 248, 344, 743. The mean quan- 

 tity of rain that falls annually at Kendal, on the western 

 side of England, is 57 inches • at Bombay 72 and 10<> 



