74 ON VAUIETIK?, RACES, SUB-SPECJEP. ANU SrEClES. 



doubt that the spuieless specimens which represent it form a group 

 subordinate to the group of spiny specimens constituting^ tiie 

 species, since we are perfectly certain that the latter produced 

 the former. 



We may as well add, that tliere are now at least fifty varieties 

 of Rohinia obtained from seeds and propagated by division ; but 

 these varieties cannot,' it is said, be propagated by seed. 



In short, in the species to which the distinction Beta is appli- 

 cable, we find individual types and varieties. 



Distinction Gamma. 



Individuals may be found in nature in uncultivated places, 

 which may be considered as types of their species ; such are, for 

 example, specimens of the wild carrot. 



Now, by cultivation, plants of the wild carrot may be modified 

 in such a way as to give rise to varieties which may be propa- 

 gated by seed with a tolerable degree of certainty (at all events 

 under determinate circumstances) so that they may be compared 

 with races. 



It is here that the admirable observations of M. Vilmorin 

 will assist us. 



If the seeds of a wild carrot are sown in the spring in a gar- 

 den, an annual, identical with the wild carrot, will be obtained. 

 No appreciable change will be produced if the stem be pinched 

 several times in the course of its growth. \{ ix. sufficient number 

 of seeds are sown in summer instead of in spring, plants will be 

 obtained, the stems of which will not run to flower, and the 

 roots of ivhich will he modified hy the end of autumn. If in 

 the following spring the same plants are again put into the 

 ground, they flower and produce seeds ; their roots are by this 

 time very sensibly changed. 



If the seeds of the specimens with the most changed roots are 

 again sown, carrots will be obtained still more modified than 

 those of the first generation. 



The seeds of the plants of the second generation give a crop 

 of carrots very much modified, and but few of these have run to 

 seed. 



The carrots of this third generation differ from common car- 

 rots ; the flesh of the former being larger, more compact, a little 

 more dry and mealy, with not so strong a taste but with as 

 mucli sweetness as in that of the latter. Most of the roots, like 

 those of the first and second generations, are white or lemon- 

 coloured, whilst others are so orange-coloured and have so stronj^ 

 a taste, tliat M. Vilmorin does not doubt that the seeds of tiiese 

 latter would yield the common carrot ; he preferred, however, to 



