308 



and, before they discharge themselves over 

 the lowest parts, wear and undermine their 

 higher banks? a distinction, which does 

 not exist in what are called imitations of 

 rivers. Do not the marks of such floods 

 on the higher banks remain after the river 

 has retired into its proper channel, that is, 

 nearly to the height of the lower banks? 

 but even on a supposition of its never 

 overflowing, and never sinking, the same 

 thing would happen in some degree ; for it 

 does happen in stagnant water, and must 

 wherever there are any steep banks ex- 

 posed to the usual effects of rain and 

 frost. 



The image in Claudian is extremely 

 poetical, and no less pleasing in reality ; 

 the passage relates, however, to a small 

 rivulet, not to a river: but supposing it 

 did relate to a river, are we thence to infer 

 that according to the poet's meaning, no- 

 thing but grass ought any where to be in 

 contact with the water, and that the turf 

 must every where be regularly sloped down 

 to it ? that there must be no oilier image ? 



