66 GENETICS 



bered having seen it before and dismissing it as 

 pathological. Because of its poverty in chlorophyll 

 it is a mutant which probably would not maintain 

 itself successfully in nature, but it breeds constant 

 under cultivation. 



Oblonga always bred true with the exception of 

 throwing an albida in 1895 and a rubrinervis in 1899. 



Of rubrinervis over 2000 invariably bred true, 

 while nanella bred true in over 20,000 offspring, with 

 but three exceptions when oblonga characters appeared. 



Lata, since it produces only female flowers and so 

 cannot be self -fertilized, had constantly to be crossed 

 back with the parent lamarckiana, when it produced 

 from 15 to 20 per cent lata and 80 to 85 per cent 

 lamarckiana. 



Finally, scintillans which appeared at three separate 

 times proved constant only in its inconstancy be- 

 cause it invariably produces a heterogeneous progeny. 

 The 1895 plant gave 53 per cent lamarckiana, 35 

 per cent scintillans, 10 per cent oblonga, and 1 per 

 cent lata. One of the 1896 plants gave 51 per cent 

 lamarckiana, 39 per cent scintillans, 8 per cent 

 oblonga, 1 per cent lata, and 1 per cent nanella, while 

 another 1896 plant gave only 8 per cent lamarckiana, 

 but 69 per cent scintillans, 21 per cent oblonga, and 

 2 per cent of nanella and lata together. 



These seven elementary species are distinguished 

 from each other by features which are unmistakable 

 even to the uninitiated. The old-time systematist 

 would undoubtedly have regarded them as distinct 

 species. 



