206 MASS MUTATIONS AND TWIN HYBRIDS OF 



them were self-fertilized; one yielded a progeny of 50 specimens, 

 which constituted a uniform lot from the first beginning until the 

 end of September. Among these I chose one of the strongest for 

 self-fertilization, and had in 1916 from it a third generation of 58 

 plants, all resembling their parent. Most of them have flowered. 



The other specimens of 1914 yielded seeds, some of which were 

 sown in 1915 and some in 1916. In the first year I had 280 speci- 

 mens, half of which flowered in August and September and were 

 then pulled up, while the remainder flowered for the most part in 

 October. They constituted a uniform lot of widely branched, low 

 plants of pale green color. The culture of 1916 yielded 50 speci- 

 mens, as pure and uniform as the former. 



I crossed the mut. ochracea with the parent species in order to 

 study its hereditary character. I made the crosses in 1915 and got 

 the following progeny in 1916: 



grandiflora ochracea lorea Sum 



O. ochracea x grandiflora 41 18 1 60 



O. grandiflora x ochracea 35 16 51 



Total 76 34 1 111 



Percentage 68 31 1 



The two reciprocal crosses gave evidently the same result, show- 

 ing that both parents are isogamic in respect to their differential 

 character. For this reason I repeated the sowing in 1917, trans- 

 planting the young seedlings after counting them, and determined 

 the percentage of dying individuals besides that of the living och- 

 racea. I found for two crosses of 0. ochracea x grandiflora on May 

 12, 23 per cent dead seedlings and 35—27 per cent living ochracea 

 in a total of 226; means 17 and 31 per cent, together 48 per cent. 

 From the reciprocal cross I had only a small culture of 55 seedlings, 

 among which, however, none died in early youth, while the percen- 

 tage of the living ochracea was 40. The figure for ochracea is smaller 

 ohan the highest one after self-fertilization (44 per cent), but this 

 obviously resulted from a loss of some pale individuals, which died 

 tff in early youth. The figure of 31 per cent was determined in June 

 and should rather be compared with the percentages after self- 

 fertilization determined at the flowering period (15—20 per cent). 

 The lorea seedling was evidently due to a mutation, even as after 

 self-fertilization. 



0. grandiflora mut. lorea. — This mutant is characterized by its 

 very narrow, almost linear foliage throughout its whole develop- 



