739 



ourselves, that in a great space of time the 

 effects of these variations mutually compensate 

 each other : that in a long succession of years 

 the means of the temperature of the humidity, 

 and of the barometric pressure, differ so little 

 from month to month ; and that nature, not- 

 withstanding the multitude of partial perturba- 

 tions, follows a constant type in the series of 

 meteorologic phenomena. Great rivers unite in 

 one receptacle the waters, which a surface of 

 several thousands of square leagues receives. 

 However unequal may be the quantity of rain, 

 that falls during several successive years in 

 such or such a valley, the swellings of rivers, 

 that have a very long course, are little affected 

 by these local variations. The swellings repre- 

 sent the mean state of the humidity, that reigns 

 in the whole basin ; they follow annually the 

 same progression, because their commencement 

 and their duration depend also on the mean of 

 the periods, apparently extremely variable, of the 

 beginning and end of the rains in the different 

 latitudes, through which the principal trunk and 

 it's various tributary streams flow. Hence it 

 follows, that the periodical oscillations of rivers 

 are, like the equality of temperature of caverns 

 and springs, a sensible indication of the regular 

 distribution of humidity and heat, which takes 

 place from year to year on a considerable extent 

 of land. They strike the imagination of the 

 3 b 2 



