DISCUSSION : FORESTS, RESERVOIRS, AND STREAM FLOW 501 
It would be, indeed, a most welcome fact, if it were true, that, by Mr. Chitten- 
introducing cultivation into the arid regions, rainfall could be in- 
creased; but it would be less welcome if it had to be accompanied by 
another fact, that the settlement of the humid regions reduced pre- 
cipitation. 
Mr. Leighton takes exception to the writer’s statement that seasons 
of drought are more to be dreaded than seasons of flood. He states 
that the flood damages for the year 1907 amounted to $260 000 000. 
He does not say whether this means direct damages, or whether it is 
made up in part of indirect or inferential losses, such as interference 
with transportation, temporary depreciation of real estate, and the 
like. The figure looks pretty large for direct damages alone. Nothing 
is said, either, as to the great compensating effect of fertilization of 
bottom land due to general overflow. It is well known that the effects 
of such overflows in increased yield of crops are noticeable for several 
years after their occurrence. Their aggregate value is undoubtedly 
large, and constitutes a material offset against direct damages. In 
many parts of the world these overflows are a regular reliance, and 
one of the evil results of the levee system, wherever used, is that it 
cuts off this means of periodic renewal of the soil. But taking the 
above figures as representing actual and uncompensated loss for this 
one year (they certainly do not represent the average annual loss), 
let them be compared with the effects of the shrinkage of crops through 
drought. The value of the agricultural products of the United States 
for the year 1908 was more than $7 750000000. A skrinkage of a trifle 
more than 3% would equal the above flood loss, and it is well known 
that a not excessive drought will cause a shrinkage far in excess of 
this amount. Famines, the world over, and throughout recorded his- 
tory, have generally resulted from widespread deficiency of rain, par- 
ticularly where such deficiency has extended through several years. 
The constantly increasing losses from floods, with the settlement 
and development of the country, are often cited as evidence that 
floods are increasing. But the fallacy of such reasoning is strikingly 
illustrated by a single example. In 1844 occurred the greatest flood 
on record in the Missouri and Kaw Rivers at their confluence. The 
damage was insignificant, because there was little property there. The 
next greatest flood was in 1903, and the damage approximated 
$30 000000. Kansas City had been built in the path of the flood. 
Care should be taken, in comparisons of this sort, not to ascribe to 
increasing flood heights or frequencies losses which are due to Man’s 
operations in placing property in the way of floods. ; 
Messrs. Pinchot and Willis have referred to the cooler status of 
the forests as inducing precipitation, Professor Willis citing the evi- 
dence of aeronauts that the air is cooler over forests than in the open. 
But the inconclusive character of evidence like this is apparent at a 
