BRITTLE RACES OF OENOTHERA LAMARCKIANA. 705 



daily on the flower buds, and with broader foliage. I have called 

 this form pallida. Table VII records the results of the repeated self- 

 fertilizations in the race of tarda. 



The percentages are averages from the separate countings of the 

 progeny of the seed-bearers of the previous year. In the third genera- 

 tion I fertilized ten specimens, in order to know whether some of 

 them might have a uniform offspring. This was not the case. The 

 amount of pallida differed between 20 and 30 per cent, with an 

 average of 27 per cent. 



From table VII we may deduce that the pallida is split off after 

 the formula: tarda = (tarda + pallida), which should give, after self- 

 fertilization, 25 per cent of pure tarda, 50 per cent of tarda x pallida, 

 and 25 per cent of pure pallida. The first combination must result in 

 barren grains, since no living specimens of it were observed. The 

 empty seeds amounted to 2—5 per cent in the harvest of 1917, 1—9 

 per cent for eight of the ten plants of 1923, and 14 and 24 per cent 

 for the two others. On the average the percentage was 8. From 

 this we conclude (8) that the lethal factor was that of the laeta, the 

 same as in the parent scindens and the grandparent Lamar ckiana, 

 but which fails in most of the original mutated gametes of scindens. 



The offspring of the type tarda x pallida continued the race, 

 while the pallida must be uniform and constant, having only recessive 

 characters. In 1922 I fertilized two specimens of it, and had in the 

 two following years the second and third generations, with 60 and 

 44 flowering plants, all of which were true to type. From this it is 

 clear that no lethal factors could occur in pallida. Besides this, some 

 other mutant strains without lethals are known. They are often used 

 in crosses. They are combined in table VIII. 



Table VIII. 



Homogeneous strains derived from 0. Lamarckiana. 

 A: laeta B: velutina 



1. decipiens, tough 3. blandina, tough 



2. deserens, brittle 4. pallida, brittle 



The group is thus seen to be completed by our new strain. 



In 1922 I made some crosses with 0. tarda of the second genera- 

 tion, in which I combined it with 0. blandina and with different 

 species, and had in 1923 a culture of 60 flowering plants for each of 

 them. All of these 480 plants showed the stature and characters of 

 the velutina from the corresponding crosses of Lamarckiana, which 



45 



