DISCONTINUOUS VARIATION. 61 



marckiana. De Vries classified them as follows: 350 

 0. oblonga; 229 0. lata; 158 0. nanella; 56 0. albida; 

 32 0. rubrinervis; 8 0. scintillans; and 1 0. gig as. Of 

 these new " species/ 7 oblonga, albida, rubrinervis, 

 nanella, and gigas remained absolutely constant in sub- 

 sequent generations, when crossed among themselves, 

 or self -fertilised, in the case of 0. gigas. 0. scintillans 

 was not nearly so constant, the offspring yielding only 

 about a third of the parent form, and the rest of them 

 being Lamarckiana, oblonga, and lata. 0. albida bred 

 quite constant, but the plants were weak, and not very 

 fertile. 



De Yries looks upon his sports as true species, and 

 not varieties, for he says that varieties differ from their 



/ tj 



parent species in only one or two characters, whilst 

 species differ from their nearest allies in almost all their 

 characters. Thus in comparison with the parent form, 

 0. Lamarckiana, gigas was stronger and albida was 

 weaker, both forms having broader and shorter leaves. 

 The flowers of gigas were larger, those of rubrinervis 

 darker yellow, those of oblonga and scintillans smaller, 

 and those of albida paler. The cuticle of albida was 

 rough. The bosses on the leaves of lata were increased, 

 and on those of scintillans diminished. The forma- 

 tion of pollen was increased in rubrinervis and dimin- 

 ished in scintillans. The seeds of gigas were larger, 

 and those of scintillans smaller; those of rubrinervis 

 more abundant, and those of lata more scantv. 



' u 



By artificial selection De Yries obtained in one in- 

 stance * what he regards as a true mutation. This was 

 in the case of Linaria vulgaris (yellow toadflax). 

 * " Die Mutationstheorie," p. 552. 



