DISCUSSION : FORESTS, RESERVOIRS, AND STREAM FLOW 533 
the 1907 floods. Applying this to the 1904 record at Middletown, mr. Chitten- 
which is well inside the water-shed, a precipitation of between 20 and 
22 in. is indicated at that point. A similar use of the Mt. St. Helena 
record would have raised the figure. 
In short, the writer went at the matter in what seemed to him a 
direct practical way. It was not a question of isohyetal lines at all, 
but of trying to find out what actually happened; and he believes that 
engineers will agree with him, on the basis of the data presented, that 
the precipitation on those two water-sheds for the period in question 
was about in the proportion of 20 in. on Puta Creek to 15 in. on 
American River. Mr. Labelle arrives at a similar result by taking the 
precipitation for a full month. 
But the writer is not at all particular as to the exact figure. He is 
willing to take in the records of Cache Creek (but not under any cir- 
cumstances those of the plain). Add in Lakeport and Guinda, and the 
record will still be about 18 in. Even if it were as small or some- 
what smaller than that of the American River, it would not affect the 
principle involved. For some cause the Sierra water-shed, much 
larger in area, sent down a much greater percentage of run-off than 
did Puta Creek. The writer believes this was due in part to the dis- 
tribution of snow on the Sierra slopes caused by the presence of the 
forests. In any event, the restraining effect of the forest totally failed 
to materialize. 
And here the writer would like to ask why there should be so much 
stumbling in this matter. If momentary prejudice is dropped, and the 
question is looked at from a direct common-sense point of view, is not 
what happened the most natural thing in the world? Is it possible 
that the general result should have been other than what it was? 
Snow always has the effect of transferring the run-off from precipita- 
tion to a period subsequent to that on which the precipitation occurred. 
If the snow melts when additional precipitation is falling, the net 
result is to bring two storms together, and the run-off must necessa- 
rily be greater than that from either storm alone. So far as forests 
affect the distribution of the snow as to increase the rate of melting, 
so far they intensify the run-off. 
The writer hopes that Messrs. Pinchot and Leighton will not waste 
any further sympathy upon him on account of this “unfortunate” 
illustration. 
One more “unfortunate” of Mr. Pinchot’s is found in Fig. 2, Plate 
XL, in regard to which it is insinuated that the writer selected a level 
tract of ground to illustrate the absence of erosion resulting from 
timber destruction. The writer would not notice this matter at all if 
it were not that it affords an opporunity to point out one of the chief 
sins of the forestry propaganda—the use of its photographs. As to the 
picture in question, it was taken several years ago by an artist from 
