404 
KIVER EMBANKMENTS. 
it, but the thawing of the walls of the tunnel rapidly enlarged 
it, and before the lake was half drained, the barrier gave way 
and the remaining 500,000,000 cubic feet of water were dis¬ 
charged in half an hour. The recurrence of these floods has 
since been prevented by directing streams of water, w T armed 
by the sun, upon the ice in the bed of the valley, and thus 
thawing it before it accumulates in sufficient mass to threaten 
serious danger. 
In the cases of diversion of streams above mentioned, im¬ 
portant geographical changes have been directly produced by 
those operations. By the rarer process of draining glacier 
lakes, natural eruptions of water, which would have occasioned 
not less important changes in the face of the earth, have been 
prevented by human agency. 
The principal means hitherto relied upon for defence 
against river inundations has been the construction of dikes 
along the banks of the streams, parallel to the channel and 
generally separated from each other by a distance not much 
greater than the natural width of the bed.* If such walls are 
high enough to confine the water and strong enough to resist 
its pressure, they secure the lands behind them from all the 
evils of inundation except those resulting from infiltration; 
but such ramparts are enormously costly in original construc¬ 
tion and maintenance, and, as we have already seen, the filling 
up of the bed of the river in its lower course, by sand and 
gravel, involves the necessity of occasionally incurring new 
expenditures in increasing the height of the banks, f They 
* In the case of rivers flowing through wide alluvial plains and much 
inclined to shift their beds, like the Po, the embankments often leave a 
very wide space between them. The dikes of the Po are sometimes three 
or four miles apart.— Baumgakten, after Lombardini, Annales des Ponts et 
Ohaussees , 1847, ler s<§mestre, p. 149. 
t It appears from the investigations of Lombardini that the rate of ele¬ 
vation of the bed of the Po has been much exaggerated by earlier writers, 
and in some parts of its course the change is so slow that its level may be 
regarded as nearly constant.— Batjmgarten, volume before cited, pp. 175, 
et seqq. 
If the western coast of the Adriatic is undergoing a secular depression, 
