72 BOTANY 



facts are in the meantime accumulated which might 

 never have been sought for otherwise. Sometimes the 

 results of the experiments have seemed at first very start- 

 ling and difficult to explain. For instance, in the course 

 of Mendelian work, one experimenter had two races 

 of Stocks, one with white flowers and one with cream 

 flowers. These were crossed in the usual way, and all 

 outside pollen carefully kept from them. The result- 

 ing offspring were not white, nor cream, but a brilliant 

 reddish-purple. At first sight this would look as if 

 something was wrong with the laws the experiment 

 set out to test, but in reality it indicated the inter- 

 play of other pairs of characters which affected the ones 

 that were for the moment under investigation. Work 

 such as this leads on through an endless chain of ex- 

 periment, hypothesis, theory, and again, and all the 

 way along, experiment. 



Experimental work on these lines is, of course, done 

 also by zoologists, but for many of the problems plants 

 afford more convenient working material. Care at 

 the time of pollination and in the collecting of seeds 

 are the main things in plant breeding. There are few 

 of the complicated pieces of apparatus required for 

 such work as are necessary for experimental physiology, 

 and, consequently, for a botanist cut off from the big 

 institutions experimental breeding offers one of the 

 most profitable fields of research. In modern experi- 

 ments often thousands of specimens are grown all of 

 one kind, and their pedigrees are kept for generation 

 after generation. 



Modern research in experimental breeding of plants 

 received an enormous stimulus and a new direction 

 from the work of Hugo de Vries, whose book on " The 

 Mutation Theory " appeared so recently as 1901. The 

 essential difference between the work and theories of 



