Appendix 2. 



ANALYSIS OF THE CAUSES OF RAINFALL WITH SPECIAL RELATION 



TO SURFACE CONDITIONS. 



By George E. Curtiss. 



Tlie possibility of clianging the amount of raiufall by bnraan agency has in late 

 years received a great deal of popular attention, and an undercurrent of hypothesis 

 with regard to it has gradually set in, Avhich itself has stimulated scientific inquiry. 

 The question has been directed most frequently to the effect of forests and to the 

 special inquiry as to whether forestatiou increases and deforestation decreases the 



rainfall. 



In this inquiry the statistical method has been extensively employed, and all the 

 existing rainfall data bearing upon the question and some that are entirely irrele- 

 vant have been discussed. Special raiufall observations have also been inaugurated 

 for the purpose of determining the relation of forests to rainfall. Without giving a 

 resume of these various investigations it is sufficient to state that no definitive con- 

 clusion has as yet been derived from them. The one conclusion which the statistical 

 results seem to yield is that, if forests affect the rainfall the amount of effect has, iu 

 most cases, not been greater than the amount of probable error in the observations 

 themselves, and, therefore, the statistics give no assurance that the effect is not an 

 error of observation. This, however, is a result of importance, for it serves to de- 

 limit, for the regions to which the statistics apply, a maximum value which the sup- 

 posed effect of forests has not exceeded. Tliis maxinmm value is, iu most cases, but 

 a small fraction of the total rainfall, an amount too small to be of any considerable 

 hydrographic or economic importance. Another reason for the unsatisfactory re- 

 sults of the statistical investigations is, that they have seldom been com1)inod with 

 a rational explanation of the process by which a change in the raiufall may be 

 brought about, and, consequently, they have not helped to clarify the misty meteoro- 

 logical conceptions which are current thereon. To do this is the object of this 

 paper. 



If in any region the rainfall is increased by a forest cover it must be brought about 

 either (1) l)y an increase in evaporation, which increase must be precipitated over 

 the same region, or (2) by a diversion to the forest area of rain that would other- 

 wise fall in some other locality. 



Let us now analyse the causes and conditions of ra in fall, so far as they arc now 

 understood, and see if the results will not materially aid us in determining how far 

 and in what cases each of these methods of increasing the rainfall can be operative 

 and efficient. I say iu what cases, for one of the most important objects of this an- 

 alysis will be to make clear that an effect that obtains in one climate may be entirely 

 absent iu another. Ideas as to the causes of rain have been greatly simplified by re- 



187 



