FISHERIES. GAME AXD FORESTS. S15 



these experiments have been carried on. the average rainfall for twenty-one years was 

 30.29 inches. The mean percolation for the same period through a drain gauge forty 

 inches in depth was 15.16 inches, leaving a mean evaporation for the period of 15.13 

 inches. The maximum evaporation for the whole period is 19.63 inches, or 4.50 

 inches more than the mean; the minimum ev^aporation was I 1. 03, or 4.10 inches less 

 than the mean. The maximum run-off through the drain gauge was 26.03 inches in 

 the year 1878-79, when the rainfall was 41.05 inches, the evaporation for that year 

 being 15.02 inches, or within o. i i of the mean. The minimum run-ofT through the 

 gauge was 5.40 inches in 1872-74, when the rainfall was 21.69 inches. We learn 

 from these figures the much greater range of the run-off than of the evaporation ; that 

 is, the maximum run-off was 10.87 inches greater than the mean run-off, and the 

 minimum 9.76 inches less. Hence, the range of the run-off above and below the mean 

 is more than twice the range of the evaporation above and below the mean. If we 

 examine the several tabulations of stream flows in the United States already referred 

 to, we shall find a somewhat similar law running through the whole series, and which 

 again not only illustrates the extreme persistency of the e\-aporation element but 

 shows that evaporation from a drainage area is broadly independent of temperature. 

 Many other interesting drainage experiments have been made in England which 

 cannot be referred to in this place for lack of space. A fairly complete resume of the 

 data may, however, be found in the forthcoming Upper Hudson River storage report. 

 The run-ofT data of a number of foreign streams are also referred to in that re])ort. 



Lack of space necessarily prevents presenting anything like complete figures at 

 this time, and I ma\-, therefore, simply state that as the result of making very 

 extensive tabulations of data I reach the conclusion that the deforestation of a 

 drainage area will, in the -State of New York and in that vicinity, probably decrease 

 the annual water yield of the area from four to six inches. This conclusion I desire 

 again to say is tentati\e. It is merely what a broad study of the data indicates, and I 

 reserve the right to modify it as further data are gathered. At the present time such 

 a tentative conclusion may be drawn from the data at hand. It ma\' also be drawn 

 from the data studied by Mr. Vermeule when the said data are studied broadly. 



We have seen in the foregoing that some, at any rate, of the current views as to the 



relation of forests to stream flow may be properly modified. As a final division of 



this present discussion I desire to lay before you briefly my present understanding of 



, how it is that deforestation may not only affect floods but that it also leads to an 



actual decrease in total run-off. 



In the first place, the classical experiments on forest meteorology which have been 

 made abroad by Dr. Ebermayer and others have shown a considerable decrease in the 

 mean temperature in forests over that in the open. If, then, Mr. Vermeule's 



