4 



BOTANICAL GAZETTE [july 



of flowers on the main axis (DeVries /. c. 2:416). The heredity- 

 coefficient of a mutantj when the large earlier flowers are used, appears 

 to be a fairly constant quantity and to be usually or at least frequently 

 near the Mcndelian ratio, though it may often fluctuate in individual 

 crosses of flowers from o to about 50 per cent. Thus the heredity- 

 coefficient of O. lata is often 21-24 P^^ cent., and that of O. nanella is 

 practically the same. This percentage is said to remain about the 

 same from generation to generation and to be independent of the 

 pollen parent; but in crosses of single flowers the percentage may 



r 



vary widely. Thus 39 crosses of O. Lamar ckiana flowers were made 

 with O. nanella pollen, and the seeds from the different flowers sown 

 separately. The proportions of O. nanella obtained, that is the 

 heredity-coefhcient, varied from i to 48 per cent,, the average for the 

 total number of 3768 seedlings being about 17 per cent. (DeVries 



/. c. 2:408V 



The proportions of the parents appearing in a cross also fluctuate 

 from other causes. In crossing O. Lamar ckiana wdth O. nanella 

 pollen, in one experiment when 90-100 seedlings were produced from 

 one ovary the proportion of O. nanella was 44 per cent., but when the 

 number of seedlings produced from an ovary was only 20-40, the 

 proportion of O, nanella was reduced to 10 per cent- The quantity 

 of pollen used in a cross also influences the offspring, and the propor- 

 tions may be varied by experimental methods. Thus in one experi- 

 ment (DeVries /. c. 2:417), all of the 4-8 lobes of the stigma of O. 

 Lamarckiana flowers were removed except one lobe or a small portion 

 of one. They were then rather sparingly pollinated with O. nanella 

 pollen (one flower being used to pollinate several), and of 1593 seed- 

 lings produced from 8 O. Lamarckiana plants whose flowers were 

 thus treated, all were O- Lamarckiana. In another experiment all 

 of the stigma was removed but one lobe, and of 1687 seedlings ob- 

 tained only about 2 per cent, were O. nanella. Other experimental 

 means of vary^ing the results of a cross need not be mentioned here. 

 But that such results can be obtained is of the greatest interest and 

 suggests a line of inquiry as to the nature of the factors determining 

 what forms will appear in a cross. This may also have a bearing on 

 the determination of the chromosome number, though just what the 

 connection may be cannot be guessed at present. 







it" 



^ 



i 



