162 



EFFECTS OF FORESTS ON RAINFALL. 



least of all. This result agrees entirely with those obtained in pre- 

 vious years, and tends more and more to present itself as a law. 

 The following table has been prepared to make this apparent : — 



"If we represent by 1000 the quantity of water which fell each 

 year at the Ginq-FranohSes, we obtain for the three years during 

 which the observations were made, the following as the proportions 

 thus indicated : — 



1866, ... 1000 — 855 



1867, ... 1000 930 913 



1868, ... 1000 985 791 

 The mean of these gives the proportion noted by M. Marsh. 



The difiFerence is probably not greater, but less, than is the popular 

 belief ; and it is possible it may be represented thus as being greater 

 than it actually is, 



M. C6zanne remarks on this report : — " After what we have seen of 

 the so-called caprices of rain, and the variations which they present 

 in one quarter, and in another of the same city, we must 

 acknowledge that the experiments of M. Mathieu, while furnishing 

 on the whole an argument in favour of the supposition of an action 

 of forests on the phenomenon of the rain, are not absolutely 

 demonstrative. How can we satisfy ourselves that the forests have 

 not acted simply as an obstacle to the pluvial wind 1 If, for example, 

 the agricultural region spread itself out on a level plateau, while the 

 forest region rose on a slope up which crept the pluvial wind, the 

 pluviomfetres placed in these two regions are in no way comparable 

 together. 



"In order that the experiments might have been conclusive it would 

 have been necessary that, viewed in regard to a great number of 

 points, they should have been always accordant. But of this accord- 

 ance in condition we have no evidence. 



" M. Belgrand, comparing the region of the rainfall in the basin of 

 the Grenetifen, which is wooded, with that of the Bouchat, which is 

 not wooded, has ascertained and established that the former 



