544 Length of the Fruit in Oenothera Lamarckiana. 



to any considerable difference in the results. The sig- 

 nificance of free crossing by insects is usually very much 

 exaggerated. The pollination of the evening primrose 

 is chiefly done by humble-bees; although Pliisia gamma, 

 Agrotis segetum and allied moths participate to a small 

 extent in my garden. These insects, especially the humble- 

 bees, usually visit all the flowers on the same stem one 

 after the other, so that there is a great amount of self- 

 fertilization taking place. 



In 1892 I placed the separate cultures at some dis- 

 tance from one another, and separated them by brush 

 wood, so that frequent crossings between them were 

 practically impossible. Artificial pollination in these ex- 

 periments has the great disadvantage that one has to 

 choose the seed-parents whilst they are in flower, that 

 is to say, a long time before the full development of their 

 fruits ; the choice is therefore not nearly as free as when 

 pollination is left to insects. 



The following description of my experiments shall 

 be condensed as much as possible. I shall begin with the 

 long-fruited race. 



The seed-parents in 1892 were two plants with a 

 mean fruit-length of 42.6 to 43.0 mm. The seed-parents 

 in 1893 were three plants whose mean fruit-lengths w^ere 

 37.0, 37.0 and 41.0. In the first year therefore there was 

 a considerable advance on the fruit-length of the plants 

 chosen in 1891 (32.6) ; in the second year however a 

 slight retrogression which was brought about by the 

 necessity, referred to just now, of carrying out the se- 

 lection before the fruits were ripe. 



The crop in both years was dealt with and measured 

 in exactly the same way as in 1892. There resulted the 



