2S OF THE FORMATION SECT. Ill* 



ground under cultivation so situated, is desirable; 

 especially for late succession crops. 



The soil that suits general cultivation best, is a 

 loam; rather the red than the black; but there are 

 good soils of various colours, and this must be as it 

 happens : The worst soil is a cold heavy clay, and. 

 the next a light sand; a moderate clay, however, is 

 better than a very light soil, though not so pleasant 

 to work. If the soil is not good ; .i. e. too poor, too 

 strong, or too light, it is to be carefully improved with- 

 out delay. Let it first, at least, be thoroughly broke, 

 and cleaned of all rubbish, to a regular level depth 

 at bottom as well as top, so as to give about eighteen 

 inches of working mould, if the good soil will admit 

 of it; none that is bad should be thrown up for use, 

 but rather moved away. This rule of bottom level- 

 ling is particularly necessary when there is clay be- 

 low, as it will secretly hold up wet, which should 

 not stand in any part of the garden. When a piece 

 of ground is cleared of roots, weeds, stones, &c. it 

 would be of advantage to have the whole thrown 

 into two feet wide trenches* and lay thus as long as 

 conveniently may be. The ground cannot be too 

 well prepared; for when this business is not per- 

 formed to the bottom at first, it is often neglected, 

 and may not conveniently be done afterwards,; so it 

 happens, that barely a spade's depth (or less) is too 

 often thought sufficient to go on with. There is this 

 great advantage of a deep staple, that in the cultiva- 

 tion of it, the bottom may be brought to the top 

 every other year, by double trenching, and being thus 

 renewed, less dung will do, and sweeter vegetables 

 be grown : Tap-rooted things, as carrots and par* 

 sjiips, require a good depth of soil. 



The aspect of a wall designed for the best fruits, 

 ipay be full South, or somewhat inclining to the East % 

 by which if will catch the sun's ravs at. its ris$, 



