THE RIVER FLOODS 265 



the savage sound and motion. Others keeping 

 near the ground glided behind separate groves, and 

 brought them forward into relief with admirable 

 distinctness; or, passing in front, eclipsed whole 

 groves in succession, pine after pine melting in their 

 gray fringes and bursting forth again seemingly 

 clearer than before. 



The forms of storms are in great part measured, 

 and controlled by the topography of the regions 

 where they rise and over which they pass. When, 

 therefore, we attempt to study them from the val 

 leys, or from gaps and openings of the forest, we 

 are confounded by a multitude of separate and ap 

 parently antagonistic impressions. The bottom of 

 the storm is broken up into innumerable waves 

 and currents that surge against the hillsides like 

 sea- waves against a shore, and these, reacting on 

 the nether surface of the storm, erode immense 

 cavernous hollows and canons, and sweep for 

 ward the resulting detritus in long trains, like the 

 moraines of glaciers. But, as we ascend, these 

 partial, confusing effects disappear and the phenom 

 ena are beheld united and harmonious. 



The longer I gazed into the storm, the more 

 plainly visible it became. The drifting cloud de 

 tritus gave it a kind of visible body, which ex 

 plained many perplexing phenomena, and published 

 its movements in plain terms, while the texture of 

 the falling mass of rain rounded it out and ren 

 dered it more complete. Because raindrops differ 

 in size they fall at different velocities and overtake 

 and clash against one another, producing mist 

 and spray. They also, of course, yield unequal 



