110 



SCIENCE 



[Vol. LV, No. 1411 



different types of gametes, wliicli in combina- 

 tion reproduce the heterozygous Lamar chiana 

 complex. Renner in 1914 presented this point 

 of view, after studies on seed sterility in sev- 

 eral species of CEnothera, and the conception 

 of impure Oenothera species was rather fully 

 discussed in my paper "The test of a pure 

 species of CEnothera" published in 1915. Thus 

 certain workers with QDnothera were fully 

 aware of the possible significance of gametic 

 and zygotic mortality in relation to problems 

 of CEnothera genetics some j'ears before 

 Morgan and Muller in 1918 discussed the 

 findings of balanced lethals in Drosophila. 

 Renner deserves particular mention as an 

 investigator quick to bring the facts of gametic 

 and zygotic sterility into relation with the 

 peculiarities of CEnothera breeding. As the 

 result of his studies and those of other investi- 

 gators we have reason to feel confident that 

 most of the tenotheras that have been the sub- 

 ject of experimental study are impure species, 

 that is to say, heterozygous in their genetical 

 constitution. 



I am, nevertheless, confident that pure spe- 

 cies of CEnothera do exist but it will require 

 much patience in observation, in cytological 

 analysis, and in experimental crossing to 

 establish them. The most promising form in 

 my experience is a line of CEnothera francis- 

 cana, which has almost perfect pollen and 

 produces seed about ninety per cent, viable. 

 This line I have selfed for eight generations 

 without finding a single departure from the 

 type. The last generation, grown during the 

 past summer, was a culture starting with 1,425 

 seedlings from seeds experimentally forced to 

 complete germination, a germination per- 

 centage of 87.3 per cent. In this large culture 

 1,373 plants survived the vicissitudes of the 

 season, a loss of onlj' 52 plants mostly as 

 seedlings. This culture was large enough to 

 biing out variants if present in the propor- 

 tions thrown by Lamarckiana, which for some 

 variants is as high as one per cent., but the 

 culture gave no exception to the type. Also, 

 crosses have been made with biennis, muricata 

 and grandiflora and, when franciscana was the 

 pollen parent, the results have been uniform 

 T\ generations, indicating that the pollen 



grains of franciscana are all alike in genetical 

 constitution. Finally, a cytological study of 

 pollen formation now in progress by my 

 former student R. E. Cleland shows a regular 

 pairing of chromosomes during meiosis in 

 contrast to the loose association of chromo- 

 somes characteristic of the same stage in 

 Lamarckiana and such other Oenotheras as have 

 been studied with the exception of a race of 

 grandiflora. Thus the evidence of high fer- 

 tility, uniform progeny when selfed, uniform 

 Fj generations when used as the pollen parent, 

 and regularity of chromosome pairing dui-ing 

 meiosis all point to the genetic purity of this 

 race of CEnothera franciscana. I present this 

 line as the purest CEnothera material known 

 and safer than the race of grandiflora that I 

 selected twelve years ago and which satisfied 

 fairly well the tests of a pure species except 

 that it threw occasional weak dwarfs. This 

 isolation of an apparently pure species of 

 CEnothera is a matter of satisfaction and of 

 some importance for the future of genetical 

 studies in this group of plants since in the past 

 we have had no standard material of unques- 

 tioned purity with which forms could be 

 mated in tests of cross breeding. My appar- 

 ently pure race of CEnothera franciscana is 

 vigorous, easily grown in cool latitudes, and 

 has a long flowering season, qualities impor- 

 tant for experimental work, and I confidently 

 offer it to students of CEnothera as a plant 

 worthy of their attention. 



The interpretation of the breeding behavior 

 of CEnothera Lamarckiana on the hypothesis 

 of its impure germinal constitution has 

 received important and most substantial support 

 from the investigations of Muller on material 

 of Drosophila which led to his theory of bal- 

 anced lethals. The condition of balanced 

 lethals results when two different lethals are 

 present, the first in one chromosome and the 

 second in the other chromosome of a pair. 

 Thus each lethal is present in a single dose 

 and the genetical constitution is therefore 

 heterozj'gous for each lethal but the two lethals 

 are in different chromosomes of a synaptic 

 pair. Since the lethals operate when in double 

 doses close breeding in such a race will result 

 in a succession of generations repeating the 



