27 



nature in hci' lucky varieties and effects, 

 and to co[)3r her as closely as possible, he 

 migiit by way of study, and as a trial how 

 far an imitation could be made to re- 

 semble a beautiful original, take a sort 

 of plan of the ground, independently of the 

 trees, &c. He might then mark out on 

 the sides of the future water, the exact 

 places where the mould which was dug 

 out should be deposited, but without being 

 smoothed or levelled; only directing that 

 each heap, more or less continued and ex* 

 tended in lenoth, should be raised to Cer- 

 tain heights in different parts: all the in- 

 lets and projections might be formed upon 

 the same principle. This, when done, 

 would be the rough ground-work; and 

 would have something of the general shape 

 of what he had admired, but with unavoid- 

 able varieties. Such a state of ground may 

 be compared to the state of a picture when 

 the artiBt has just roughly sketched in the 

 general masses and forms. To a person 

 unused to the process, the whole appears 

 like a heap of confusion, and of dabs of 



