ayid Spouting FumUams. 211 



streams, and great ignorance concerning the quantities of rain^ 

 dew, and snow, in different climates, have given a wide scope 

 to the part played by these internal vapours in the forma- 

 tion of streams and rivers. Thus, for example, it was thought 

 that the basin of the Seine, by which I understand all that por- 

 tion of territory which drains its water into this river, did not 

 receive a quantity of water equal to that which the Seine de- 

 livers to the sea in the same space of time. Perrault and 

 Mariotte were the first experimentally to study the subject, and 

 they found, as is very often the case in such circumstances, that 

 the vague ideas of their predecessors were the very reverse of 

 the truth. According to Mariotte, the Seine every year carries 

 to the sea only a sixth part of the water that falls into its basin, 

 in the shape of rain, dew, and snow. The other five-sixths, 

 then, must either be evaporated so as to form clouds, or be im- 

 bibed by the soil for the nourishment of the vegetation, or last- 

 ly, must penetrate through the fissures of the rocks into the in- 

 terior reservoirs whence the springs issue.* The calculations of 



• A word upon the meteorological facts on which these calculations rest. 

 The quantity of rain which falls at every height, and in every place, can be 

 easily and very accurately ascertained by dh-ect measurement. By meltuig 

 snow in the rain-gauge before it has had time to evaporate, we may also accu- 

 rately determine what quantity of fluid it represents. It is also true that 

 old experiments which have been often repeated, permit us to come to a safe 

 conclusion from the decision of the eye. When the snow has fallen in large 

 flakes, if we measure its depth before any drifting has taken place, we may 

 reckon, when thawed, that the height of the liquid which would result, if the 

 soil were impermeable and horizontal, would not be above a te7ith part of the 

 original depth of the snow. Snow falling in small particles is much more 

 dense, and, in melting, a. fifth part must be allowed for it. Finally, when the 

 snow in falling has been much drifted, the reduction on thawing may be re- 

 garded as Iwo-thirds. 



The water resulting from hail might be usually neglected ; but since it 

 rarely hails without at the same time raining, tlie rain-gauge will accurately 

 supply the joint amount of both. 



As regards dew, Dalton calculates that a depth of water equal to about 

 three inches over the whole surface of the globe, would be a fair annual al- 

 lowance for it. This result is deduced from an experiment of Hales, which 

 might be legitimately generalized so long as it was supposed that dew de- 

 scended in the same manner that rain does. But since the researches of Dr 

 Well.s, — since it has been generally known tiiat dew does not descend at all, 

 but that the air deposits it upon such surfaces as have been previously refri- 

 gerated by the radiation of their caloric into the celestial regions, — and that 



