FORESTS, RESERVOIRS, AND STREAM FLOW 257 
baskets (the great drifts), and give it the maximum possible exposure 
to the melting influences whenever these shall arrive. As a general 
rule, these snow storms are followed by warm southerly winds and 
rains—the rains frequently heavy in themselves—and rain and snow 
join hands, two storms in one, and rush down to the ocean in tremen- 
dous freshets and floods. The Skagit River, the largest in Washing- 
ton except the Columbia, and a very considerable stream, has been 
known to rise 1 ft. per hour for 16 hours, and this where the stream 
has a fall of 4 ft. to the mile and carries off its floods very rapidly. 
Fig. 1, Plate XL, taken on another stream with only 480 sq. miles of 
water-shed above it, shows the terrific power of these streams that come 
down from the most densely wooded and perfectly protected water- 
shed in existence. The great flood of 1906 in this section, was a per- 
fect demonstration, not only of the vast intensifying effect of forests 
upon floods due to snow-melting, but of the utter helplessness of the 
forest bed, when saturated with long rains, to restrain floods. 
The same effect was very manifest in the great flood of 1907 in 
the valley of the Sacramento River, California. The tributaries on the 
east side come down from the densely wooded slopes of the Sierras; 
those on the west side from the bare or sparsely wooded slopes of the 
Coast Range. If the forest theory be true, these smooth western slopes 
should send down a greater flow for the same precipitation than the 
eastern slope. Exactly the reverse seems to have been the case. For 
the period, March 17th-26th, the precipitation on the Puta Creek water- 
shed, on the west side (805 sq. miles), averaged 22.7 in. The maximum 
resulting run-off per second per square mile for one day was 39.1 cu. ft. 
Directly across the valley, on the Sierra slope, the precipitation on the 
American River water-shed (2.000 sq. miles), averaged 14.6 in. for the 
same period, and the maximum daily discharge was 48.7 cu. ft. per sec. 
per sq. mile. Considering the fact that unit run-off for the same condi- 
tions is always less, the greater the water-shed, this result is quite 
remarkable. It is undoubtedly due to the action of the Sierra forests 
on sow-melting, and again illustrates the inability of forests to exer- 
cise any restraining influence upon great floods.* 
*In the paper, ‘“‘The Flood of March, 1907, in the Sacramento and San Joaquin 
River Basins, California,” by Messrs. Clapp, Murphy and Martin, published in 
Transactions, Am. Soc. C. H., Vol. LXI, p. 281, the authors say, ‘In the Sierras 
the greater part of the precipitation is normally in the form of snow, and the mag- 
nitude of floods depends largely on the rate of melting. A heavy warm rain on 
deep, freshly fallen snow produces a maximum run-off.” 
