JLCT. XVIII. OF FLOWERS. 583 



fliould be coo), or kept fo by occafional watering : 

 The beginning of May, however, is not too late. 



Thin the young plants on the feed beds a little, foon 

 after they appear, to about an inch, and again to three 

 or four inches afunder, and keep them well weeded. 

 They may either thus remain till autumn, to be planted 

 out where they are to blow : or if they grow too ftrong 

 and crowding, let every other be drawn in Jurnmer, 

 (chu'fing a moift time, if poflible) and planted out 

 wider into nurfery beds for ufe in autumn, or the fol- 

 lowing fpring: The latter feafon will do for final 

 planting, though the former is belt, as the roots get 

 eltablifhed in the ground ; when if moved in the fpring 

 they meet with a check. It is belt, if poflible, to 

 tranfplant with earth about the roots ; but ihorten all 

 itraggling fibres, and cut off dead and rambling leaves. 

 In jevtre winters, thofe moved in autumn (it not very 

 «?arly) are fometimes killed, and therefore a few may be 

 referved to fpring, in cafe of fuch an accident ; when 

 being moved with good balls of earth, they will not be 

 much checked. If planted late (as November) let them 

 have fine well broke earth about the roots, ^and lay 

 mofs, old bark, or fmali Hones round them for protec- 

 tion from froft. Some of the Peremiials might form 

 another clafs, and be called Biennial-Perennial, &c» 

 from their uncertain continuance. 



3. OF PERENNIALS. 



This clafs (as has been obferved) is very numerous* 

 and the plants are propagated, many of them by their 

 roots, according to their nature, as fibrous, bulbous, &c. 

 fome by layers, fuckers, offsets, flips, cuttings, and a 

 few by yW only. All forts (bearing feed) are occa- 

 fionally propagated this way, for ?iezu varieties, or to 

 produce fiver plants, as thofe from feed generally prove, 

 with rdpect to fbength, fymmetry, and flowers. It 



happens, 



