462 DISCUSSION: FORESTS, RESERVOIRS, AND STREAM FLOW 
Mr. Pinchot. siderable burden of water vapor, and drop it rapidly before and after 
reaching the crest; as the clouds pass eastward over the Sacramento 
Valley, the rainfall diminishes with great rapidity. 
The other rainfall stations on Puta Creek water-shed or in its 
neighborhood have been given their due weight, and the results have 
been worked out in two ways. One is by simply taking the arithmeti- 
eal average of the stations, from which we get the following results: 
Puta Creek, instead of having 22.7 in. of rainfall, has 9.44 in.; while 
the American River has 14.6 in., as given by the author. Calculated 
in this way, it appears that, although the rainfall on the timbered 
American River water-shed—and a very much steeper water-shed— 
exceeded that on the Puta Creek water-shed by 55%, the run-off 
exceeded that of the Puta Creek water-shed by only 23 per cent. 
Calculating by isohyetal lines, we get, instead of 22.7 in. on the 
Puta Creek water-shed, 10.89 in., with the same result, or 14.69 
instead of 14.60 in., for the American River water-shed. From these 
figures, the rainfall on the American River water-shed exceeded that 
on Puta Creek water-shed by 53.3%, while the run-off exceeded it 
by only 23 per cent. 
Colonel Chittenden says: “If the forest theory be true, these 
smooth western slopes should send down a greater flow for the same 
precipitation than the eastern slope. This is exactly what did happen. 
When the facts are correctly stated, the forest theory is absolutely 
sustained. 
Passing to the question of flood height and flood frequency, it is 
perhaps unnecessary to insist that flood height is not the critical factor. 
As far as the influence of the forest is concerned, flood frequency and 
duration and the frequency and duration of low waters are the 
essential considerations. 
The statement is made that “the constantly reiterated statement that 
floods are increasing in frequency and intensity, as compared with 
former times, has nothing to support it.” 
There are no long-time figures of stream flow in the United 
States; but, taking the rivers of which we have definite measures, divid- 
ing the time for those figures in two equal parts, and ascertaining for 
those parts the frequency of floods, the number of days of floods, and 
the number of days of low water, it is found, in by far the majority 
of cases, that floods are increasing in frequency, and low waters like- 
wise. In other words, such records as there are for streams with 
mountain water-sheds, where heavy cutting has been going on, directly 
support the view that the forest has influence on stream flow. 
Figures covering from 15 to 34 years for those streams whose water- 
sheds lie in the Appalachian Mountains and which have been heavily 
cut over within the period of measurement, refute the claim that there 
has been no change in flow during that time. In the report of the 
