lOO 



PLANT-BREEDING 



mutability among cereals, and in the third place, the research- 

 es of Nilsson, which have given the leading principle for my 

 suggestion. 



Our first point is now becoming generally recognized. 

 The researches of Jordan and of Wittrock show the existence 

 of races for the species of the genus Viola. Other notable in- 

 stances are those of Draba vcrna, of Capsella Heegeri, of 

 the Xantliium Wootoni, a variety with half the original 

 number of spines on its burs, and many others. 



For myself, I have had opportunities to test the constancy 

 of such elementary forms and in some instances even at the 

 period of their very first discovery. Two local evening- 

 primroses, up to the present time occurring only on a field 

 near Hilversum, where they are growing among the common 

 (Enothera Lamarckiana, have given proof of their absolute 

 constancy in my cultures. They are the O. laevifoHa and 

 O. brevistylis, both of which are still seen to thrive on their 

 small native locality. Other instances are the cruciate 

 form of the ordinary European primrose, (Enothera biennis 

 and an analogous variety of the willow-herb, Epilobium 

 hirsutum. Many other instances could easily be added. 

 The conclusion seems warranted that elementary forms 

 may be found in nearly all systematic species, and are as 

 constant as the latter have always been supposed to be. 

 These facts give a strong presumption that the same rule 

 may hold good for the rye-cultures of Rimpau. 



My second point relates to the question of the part which 

 fluctuating variability and mutabihty may have played in 

 the selection-culture of Rimpau. An exact notion of the 

 first phenomenon, as stated by the works of Quctelet (1870) 

 and Galton (1889) found its way into botanical investiga- 

 tions about the year 1894, or nearly twenty-five years after 

 Rimpau started his pedigree of rye. In his time, therefore, 

 no distinction of this kind could be made, and it is only 



