MUTATIONS 119 



Experiments in the production of double flowers. 1 After 

 remarking that mutations occur as often among cultivated as 

 among wild plants, De Vries drops the caution that in all experi- 

 mentation of this order hybridism must be carefully guarded 

 against. 2 He observes, too, that white varieties seem compara- 

 tively old, as they are common in the wild state, while double 

 flowers are rare in the wild state and correspondingly recent, 

 indicating their origin under cultivation, and thus making the 

 matter of doubling a favorable character with which to conduct 

 investigations upon mutation. 



In the experiments upon peloric toadflax nothing new was 

 attempted. The object was to repeat what nature was known 

 often to have done, but so to control conditions as to " be there " 

 when it happened next time. 



In this experiment, however, De Vries determined to attempt 

 a new mutation, that is, to try to secure double flowers where 

 they had never been observed in nature. He accordingly chose 

 the corn marigold (Chrysanthemum segetum), common in the 

 grain fields of central Europe, and its cultivated variety, grandi- 

 florum. The number of ray florets is variable in both, but is, 

 on an average, thirteen in the wild and twenty-one in the culti- 

 vated. This indicated the latter as the more favorable for the 

 experiment, and it was therefore chosen; but it is far from pure, 

 for many of its heads have as few as thirteen rays. Only six 

 out of the first lot of three hundred plants reached an average 

 of twenty-one, and these were selected as the foundation. 



The seeds of each of these were sown separately. Five gave 

 proof of being still mixtures with the wild form and were re- 

 jected. The offspring of the sixth plant averaged twenty-one 

 ray florets, and after counting some fifteen hundred heads the 

 two plants were selected whose secondary heads made the best 

 showing. The progeny of these plants also averaged twenty-one, 



1 De Vries, Species and Varieties, etc., pp. 489-515. 



2 If a new form is a mutant it will "breed true " to itself in the great majority 

 of cases and perforce hybridize with the original stock. If, on the other hand, it 

 is an ordinary hybrid, it will not breed true, but will observe the principle of 

 Mendel's law (to be discussed later), by which a certain definite percentage is of the 

 original types. Thus it is comparatively easy to ascertain whether a new type is the 

 product of a single race by mutation or of two races by hybridization. 



