56 



the greater proportion of the potash taken up by the plants is 

 directly returned to the soil, but where practically the whole of 

 the produce is removed from the land it is probable that partial 

 potash-exhaustion may take place in the course of a succession 

 of crops of sugar cane. 



Several of the new varieties appear to be able to utilise the 

 nitrogen in the deeper layers of the soil to better advantage than 

 the Bourbon cane does, and this is a matter of great importance 

 with regard to the economical production of sugar from the sugar 

 cane. 



THE RELATION OF FORESTS TO STREAM 



FLOW. 



By James W. Toumey.* 



Collaborator, Bureau of Forestry, U.S.A. 



Introduction. 

 For the purpose of the present discussion " forest" must be un- 

 derstood to mean a growth of trees sufficiently dense to form a 

 fairly unbroken canopy of tops, not a scattered growth of low, 

 round-headed trees with bushes and herbage constituting the domi- 

 nant types of vegetation. 



Forests of this kind do not occur in the United States where the 

 mean annual precipitation falls below i8 to 20 inches, except on 

 restricted areas where unusual conditions prevail. The line of 

 separation between the great eastern forest area and the plains 

 approximately coincides with a north and south line marking a 

 mean annual rainfall of 20 inches. The streams which rise in the 

 Rocky Mountains and flow eastward are bordered by forests for 

 long distances into the plains, where the annual rainfall is much 

 less than 20 inches. These forests, however, are not so much a 

 result of the rainfall in the regions where they occur as of surface 

 and seepage flow from adjacent regions. The mesquito forests of 

 the desert regions of southern Arizona, where the mean annual 

 rainfall is but 8 to 12 inches, are made possibly by the seepage 

 and surface waters from the adjacent mountains. 



The question of the exact relation which exists between forests 

 and stream flow has long been under discussion. The broad fact 

 that a relation exists is indeed indisputable. Forest destruction 

 always produces a change in the character of the run-off. But the 

 scientific determination of all the causes which produce this effect, 

 and of their relative importance is a difficult and complicated 

 matter. In spite of the fact that for many years European forest 

 experiment stations have been carrying on observations, mea- 



* From " Yearbook of the U. States Department of Agriculture, 1903." 



