DISCUSSION: FORESTS, RESERVOIRS, AND STREAM FLOW 373 
treme floods. In the case of extreme droughts, however, he considers mr. Swain. 
the case, not where the low-water flow from various tributaries arrive 
simultaneously at a given point, which may equally well occur, but on 
the contrary where relatively high water from one arrives at the same 
time as low water from another. 
The author, however, does not really take his own arguments seri- 
ously, for later he states: “In the records of precipitation * * * 
will be found a@ full and complete explanation of every one of the 
floods,” and, “the true explanation is found in them and not at all in 
the presence or absence of forests on the water-sheds.” This is quite 
true, except that the writer and many others believe—and the author 
admits—that there is abundant proof that as a rule forests reduce the 
violence, frequency, and duration of freshets. 
The author says, “it seems certain that forests decrease somewhat 
the total run-off from water-sheds small or great.” This would seem 
to depend upon the relative amount of evaporation from the forested 
land, and also upon whether there is an increased rainfall produced by 
the forests. If the rainfall is the same as without forests, while the 
evaporation is greater—including under the term evaporation, physi- 
eal evaporation from the soil and physiological evaporation from 
the trees themselves—of course the total run-off will be less. Perhaps 
this is true. The writer does not deny that it is possible, although he 
has seen no proof of it to the present time; even if this is true, he be- 
lieves that uniformity of flow is of much more importance than total 
amount of flow, even on the small tributaries which go to make a large 
stream. But if the forest increases the evaporation, it would seem to 
follow that it increases the rainfall. The rainfall is due to the con- 
densation of the moisture in the air. This moisture comes from 
various sources: from the sea, from lakes and streams, or from the 
land. The rainfall depends upon the total amount of this moisture, 
independent of its source. Any cause, therefore, which increases the 
evaporation from a country increases the rainfall. This does not mean, 
of course, that if the writer increases the evaporation from his back 
yard, he increases the rainfall there. Large areas must be considered, 
and the statement is a general one. Regarding this question of rain- 
fall, too, it must be borne in mind that the total amount of water 
reaching the soil is due not only to that which falls in drops from the 
sky, but also to that which condenses either on the ground or on the 
trees. The author refers to dew, which he says is not present in 
forests, but he does not mention the condensation upon the tree trunks 
and leaves, and on the ground (dew aside), which must be greater 
than in the open, because of the lower summer temperature not only of 
the air but of the trees themselves. 
As to whether forests increase rainfall, the difficulties in the way 
of demonstration are obvious, but there is considerable proof that they 
do. The writer has not seen any reference to the results obtained in 
