Mr. Todd. 
344 DISCUSSION: FORESTS, RESERVOIRS, AND STREAM FLOW 
that the Ohio and its tributaries were comparatively low. Great 
damage was caused to unleveed areas on account of the lateness of 
the season. Numerous very disastrous floods have been caused by the 
streams entering from the western slopes, culminating with stages in 
the Ohio and tributaries well below the flood level. Table 6 shows a 
few floods of the above character: 
TABLE 6. 
HieHest Gauce Reapines, IN FEET. 
. Flood 
Stations. stage, in 
ect ise, | 1892. | 1908. 1904. 1908. 
Cincinnati.... 50. 39.8 43.2 16,0 44.4 44.1 
Nashville.... 40. 89.3 38.8 26.8 37.6 19.3 
Chattanooga . 33. 27.0 13.7 11.6 16.9 12.2 
St. Louis. 30. 25.6 35.8 38.0 25.3 24.8 
CATO sisisic saci s cee ciaeeiceiees 2a 45. 45.3 48.3 43.4 49.1 45.1 
Memphis 33. 84.2 34.6 33.0 39.2 35.3 
Little Rock (Arkansas River). 22. 19.2 31.2 26.2 29.4 25.9 
Arkansas City. ............06. 42. 45.4 50.0 45.6 49.0 49.9 
Vicksburg ves ccaneveves cee vane 45. 44,2 48.4 43.2 46.9 47.9 
The flood of 1892 was the greatest on record along a large portion 
of the lower valley up to that date, the serious balance to this flood 
was added by heavy rains over the Arkansas and Central Valleys. 
The floods from the western slopes, when unrestrained, are always 
the cause of greater damage than the earlier floods out of the Ohio, 
usually materializing during the latter part of May or June, catching 
the growing crops at just the time when if destroyed they cannot be 
started successfully again after the overflow recedes. Hence it is 
evident that the Ohio is not the sole cause of damaging floods in the 
lower valley; and, even if they are held well down to below the danger 
or overflow stage, disastrous overflows may still occur below Cairo. 
It cannot be denied that the reservoir system as proposed by Mr. 
Leighton, if executed as designed, would have some effect on the flood 
flow for a certain combination of excessive rainfall conditions; but, 
unfortunately, just the combination taken into Mr. Leighton’s calcula- 
tions may happen only once in a decade. The advocates of the reservoir 
system neglect to take into account the great natural reservoir effect 
of the stream beds, and they overlook the fact, fully recognized by all 
close observers of flood phenomena, that a very general rainfall over a 
tributary water-shed, may not cause any higher gauge readings, at 
lower stations, than the same amount of rainfall occurring over only 
a portion of the water-shed. In the case of the general rainfall the 
lower streams are backed up and act as reservoirs until the upper 
streams run out, and thus the flood is passed often with little, if 
any, higher gauge readings, but of somewhat longer duration. The 
reservoirs could only be of service as eliminating those floods originat- 
