By the Rev. William Williamson. 379 



should the winter prove mild, but that of January is almost 

 certain to do well, though the roots will be much smaller 

 than those raised from the autumnal sowing * Another 

 point to be attended to is not to sow the seeds too deep ; the 

 least covering will be sufficient. I need scarcely mention 

 the necessity of a frequent and careful weeding by hand, dur- 

 ing the time the plants are growing. When the foliage de- 

 cays, which will be early in June, the ground ought to be 

 entirely pared to a sufficient depth and afterwards sifted, that 

 all the roots may be certainly saved, for the smallest mostfre- 

 quently produce the best flowers. By these methods, I have 

 often raised many thousands in one year ; and by continual 

 selection of the best seed, I now have them in such perfec- 

 tion, that out of many thousand new seedling roots, which I 

 bloomed last year, I had only one absolutely single one, that 

 is, with five petals, but I had many full double flowers of 

 great beauty. 



The semidouble flowers 1 continue to bloom for several 

 years; as I find that I cannot appreciate their real qualities 

 under two or three years, for though they are only semidou- 

 ble when tbey first blow, they may afterwards become double 

 by skilful cultivation. These flowers become double in 

 two ways ; first, by the multiplication of their petals ; such 

 are obtained at once from seed, and seldom or ever vary 

 from what they were originally : secondly, by the conversion 



Note.—* March IS, 1821. I sowed a large bed, containing several perches, on 

 the 22d September last, the seed vegetated in the latter part of October, and the 

 young plants continued in a growing state, till they were entirely cut off by the 

 frost in the latter part of the year : the other half I sowed as soon as the weather 

 admitted in January last; they are now just appearing, and likely to do well. 



VOL. IV. 3 C 



