DISCUSSION: FORESTS, RESERVOIRS, AND STREAM FLOW 323 
of several feet; otherwise the saplings would not have withered as Mr. Roberts. 
they did. 
It might be asked, is the forest, in its usual late winter and spring 
condition, with its “miniature reservoirs” full to overflowing, capable 
of holding back more water from later rains than the dry open country, 
as described by the speaker referred to? Or, when it comes to pro- 
tracted droughts, are the forest areas contributing more water to the 
streams than equal areas of deforested country under similar condi- 
tions of soil composition, depth, and slopes? Writing in 1838, in his 
first report to the Monongahela Navigation Company, the late W. 
Milnor Roberts, Past-President, Am. Soc. C. E., the company’s engi- 
neer, states, in reference to the drought of that season, that all the 
tributaries of the river between Brownsville and the mouth of the 
Youghiogheny River, a distance of about 42 miles, were dry at their 
mouths, and again the same thing occurred in 1895 and is practically 
true at this writing.* 
It appears that Colonel Chittenden must be correct in his assump- 
tion that, for the extremes of high floods and low water, as observed 
in our navigable rivers, forests exert no appreciable control. For the 
intermediate, or what might be termed summer, freshets, there is ap- 
parently some reason to think that forests have a measurable influence 
in reducing their height and prolonging their discharge, and this state-’ 
ment may hold, notwithstanding that less rainfall reaches the earth 
in forests than in open country. The fact is perhaps indicated for 
the Ohio in Table 2—for freshets not exceeding 10 ft.—and in Table 
8—for the months of June and July—when the mean of freshets was 2 
ft. higher in the second than in the first period. For the coal-boat 
stage—10 ft. or more—during the first period, the annual duration was 
57.8 days; for the second period, it was 58.9 days. Doubtless the ex- 
cessive rains in recent years will account for the difference in the 
higher stages. Speculations, however, regarding the Pittsburg river 
records in connection with forests, must be received with caution, in. 
view of the fact that in eight of the counties on the Allegheny River 
water-shed, during the decade 1890-1900, there was a decline of more 
than 20000 in the rural population, that is, exclusive of boroughs and 
villages, many one-time mountain farms being now covered with young 
trees or brush. 
Table 5 shows the number of rises in the Ohio at Pittsburg, by 
months, reaching to or exceeding 15 and 20-ft. depths, from 1855 to 
1907, inclusive. The table shows that the winter storms of January, 
February, and March were much more frequent in the second than in 
the first half period. No forest influence appears to be indicated for 
the leafy months, a fact contrary to the common impression. 
*While it is true that the Upper Ohio, about October Ist, 1908, was lower than 
officially known since 1838, it is reported from Hast Liverpool, Ohio, that the “Indian 
picture writing’’ on a rock ledge at that place was not yet completely bared. 
