126 VARIATION 



plants. The types that arose at different times have already 

 been described, but considerable interest and no little profit 

 attaches to the details of the experiment, especially in re- 

 gard to the order and manner of the appearance of the new 

 types. The following is abstracted from the experimenter's 

 account of one of these four experiments, running through 

 eight generations. 1 



Beginning in the fall of 1886 he took nine large rosettes of 

 O. Lamarckiana from the field and planted them in the garden. 

 The second generation was sown in 1888 and flowered in 1889. 

 The seed produced fifteen thousand seedlings, of which ten 

 were divergent at once, five lata and five nanella. No inter- 

 mediates appeared. "They came into existence at once," says 

 De Vries, " fully equipped, without preparation or intermediate 

 steps. No series of generations, no selection, no struggle for 

 selection was needed. It was a sudden leap into another type, 

 a sport in the best acceptation of the word." 2 



The third generation of ten thousand plants showed three lata 

 and three nanella, besides one rubrinervis. 



Growing expert in detecting mutants at an early stage, he 

 discovered 334 young plants out of 14,000 of the fourth gener- 

 ation (1895). This is about 2.5 per cent. Of these 176 were 

 oblonga, 73 lata, 60 nanella, 15 albida, 8 rubrinervis, I scintil- 

 lans, and I gigas. 



The larger number and wider range of mutants discovered 

 this year are to be ascribed to growing skill in detecting them 

 at an early age. Manifestly such immense numbers must be 

 greatly reduced at the earliest possible date, and without doubt 

 some good forms were overlooked in the earlier generations. 

 After this (fourth) generation the number of seedlings was 

 greatly reduced, with the effect of reducing the number of 

 mutants and also the chances of the rarer forms appearing at 

 all ; indeed, gigas never appeared again, and scintillans not after 



1 De Vries, Species and Varieties, etc., pp. 549-556. 



2 It may occur to the student to object to the conclusion on the ground that 

 the parent stock taken from the field may not itself have been pure. If, however, 

 the stock had been in any sense hybrid, the departures should have been, accord- 

 ing to Mendel's law, more than ten ; but not in this or in later generations did 

 either parent stock or mutant behave like a hybrid in this respect. 



