92 HEREDITY AND VARIATION 



In 1887, a year after his discovery of the potato-field, De Vries 

 found two well-defined new forms a short-styled 0. brevistylis 

 and a beautiful smooth-leaved 0. Icevifolia distinguishable from 

 the parent 0. lamarckiana in many details. He hailed these as 

 two new " elementary species," * and he applied one of the crucial 

 tests of specific or subspecific rank : Did they breed true ? He 

 found that this was so ; from their self-fertilised seeds similar 

 forms arose. Neither of the two new forms was represented in the 

 herbaria at Leyden, Paris, or Kew ; neither had been described in 

 the literature of Onagraceae. They seemed to be distinctively 

 new. It is interesting to note that in 1887 there were few ex- 

 amples of these two new elementary species, and that each 

 occurred on a single plot in the field. The impression conveyed 

 was that each had arisen by a sudden mutation from the seed 

 of an individual parent. 



The next chapter in the famous investigation began with a 

 transference of samples of the new forms and the parent stock 

 partly as plants and partly as seeds from the potato-field at 

 Hilversum to the botanic garden at Amsterdam. 



The three stocks gave rise under cultivation to many thousands 

 of individuals, which bred true along certain lines, and yet gave 

 rise to other new forms. In short, De Vries had found a plant in 

 process of evolution. The predisposition to mutability which 

 remains a mystery was present, De Vries gave it scope, and 

 like the primeval gardener he had the pleasure of giving names to 

 a crop of new creations which emerged before him. From each 

 of his three samples there arose several distinctive groups which 

 if they had been found in nature would have been reckoned as 

 distinct species of evening primrose. But the most interesting 

 feature was the apparent abruptness in the origin of the ne\v 



* By an " elementary species " is meant simply a group of individuals 

 which agree with one another and differ from other groups in a certain 

 number of characters, normally constant through successive generations. 



