500 Conclusion. 



0. lata and another besides, 0. nanella', both were repre- 

 sented by several examples (p. 224). I sowed seeds, 

 collected in the field, again in 1889 on a larger scale and 

 again got the same two forms, and yet a third form, 

 hitherto unknown, O. rubrmervis (p. 304). Later on I 

 found the two former (0. lata and 0. nanella) in the 

 field as well (1894). 



This rapid succession of discoveries decided me to 

 practically give up the experimental sowings of the other 

 species and to investigate Oenothera Lamarckiana as 

 minutely as possible. Two methods of investigation pre- 

 sented themselves. On the one hand, observation in the 

 field, together with a yearly sowing of seeds collected 

 there. On the other, the cultivation in the garden of 

 families of plants through many generations. I adopted 

 the latter without however neglecting the former. And 

 in this connection I wish to particularly emphasize the 

 point that my cultures are nothing more than a repe- 

 tition of what occurs in nature. My idea was merely 

 to follow the natural process of the origin of new species 

 as accurately as I could, by excluding, wherever necessary, 

 the sources of error and uncertainty which are the un- 

 avoidable result of free fertilization by insects. 



I have made observations in the field every year since 

 1886; I thus witnessed the origin of new forms, the 

 majority of which, however, perished. They were es- 

 sentially the same as those in my cultures. There is no 

 reason for supposing that any forms arose, in my garden, 

 which would not also have arisen in nature under suffi- 

 ciently favorable circumstances. In nature there is not 

 room enough for all seeds to germinate, still less to give 

 rise to adult plants: so that the rarer and weaker sorts 

 perish; whereas in my garden they are transplanted and 



