238 CAUSES OF INUNDATIONS. 
have just alluded does not occur. The banks of the rivers and 
smaller streams in the North American colonies were formerly 
little abraded by the currents.* Even now the trees come down 
almost to the water’s edge along the rivers, in the larger forests 
of the United States, and the surface of the streams seems 
liable to no great change in level or in rapidity of current. 
* In primitive countries, running streams are very generally fringed by 
groves, for almost every river is, as Pliny, Wat. Hist., v. 10, says of the 
Upper Nile, an opifex silvuarwm, or, to use the quaint and picturesque language 
of Holland’s translation, ‘‘ makes shade of woods as he goeth.”’ 
+ A valuable memoir by G. Doni, in the Rivista Forestale for October, 1865, 
p. 438, is one of the best illustrations I can cite of the influence of forests in 
regulating and equalizing the flow of running water, and of the comparative 
action of water-courses which drain wooded vaileys and valleys bared of 
trees, with regard to the erosion of their banks and the transportation of 
sediment. 
‘“¢ The Sestajone,” remarks this writer, ‘‘and the Lima, are two considerable 
torrents which collect the waters of two great valleys of the Tuscan Apennines, 
and empty them into the Serchio. At the junction of these two torrents, 
from which point the combined current takes the name of Lima, a cnri- 
ous phenomenon is observed, which is in part easily explained. In rainy 
weather the waters of the Sestajone are in volume only about one-half those 
of the Lima, and while the current of the Lima is turbid and muddy, that of 
the Sestajone appears limpid and I might almost say drinkable. In clear 
weather, on the contrary, the waters of the Sestajone are abundant and about 
double those of the Lima. Now the extent of the two valleys is nearly equal, 
but the Sestajone winds down between banks clothed with firs and beeches, 
while the Lima flows through a valley that has been stripped of trees, and in 
great part brought under cultivation.” 
The Sestajone and the Lima are neither of them what is technically termed 
atorrent—a name strictly applicable only to streams whose current is not 
derived from springs and perennial, but is the temporary effect of a sudden 
accumulation of water from heavy rains or from a rapid melting of the snows, 
while their beds are dry, or nearly so, at other times. The Lima, however, ina 
large proportion of its course, has the erosive character of a torrent, for the 
amount of sediment which it carries down, even when it is only moderately 
swollen by rains, surpasses almost everything of the kind which I have obsery- 
ed, under analogous circumstances, in Italy. 
Still more striking is the contrast in the régime of the Saint-Phalez and the 
Combe-d’Yeuse in the Department of Vancluse, the latter of which became 
subject to the most violent torrential floods after the destruction of the woods 
of its basin between 1823 and 1833, but has now been completely subdued, and 
