ZUWACHSGESCHWINDIGKEIT DER FOLLENSCHLÄUCHE 65 



not express the ratio of the gametes formed but of the »gamete 

 representation» in the ovary during fertiHzation. In this case, then, 

 the gamete representation was 6 ß : 5 r. 



In 1918 new experiments were carried out. Some red nerved in- 

 dividuals from the same strain were selffertilized, and the same indi- 

 viduals were back-crossed in both directions with two white nerved 

 plants of common origin. The results obtained are tabulated in table 

 I, pag. 51. The figures in large type (column III) show the result 

 of the crosses rr X Rr. Instead of the 1 : 1 segregation expected segre- 

 gations up to 4:1 are obtained, thus showing a deviation 15 times 

 greater than the one obtained in 1915. The deviation is so great that 

 polymery or reduplication with regard to the factor of red-nervedness 

 could be suspected. But it is at once seen from the table that the 

 reciprocal crosses Rr X rr give the typical mendelian segregation 1:1. 

 Thus the gamete formation is found to be quite normal, while the 

 gamete representation is much disturbed. The difference in the rate of 

 growth of the pollen tubes must have been greater in these crosses 

 than in the 1915 material. The fact that the 1918 experiments were 

 carried out in the latter part of August while the previous crosses were 

 made in the latter part of July or in the beginning of August probably 

 explains the difference of the results. It was found already in 1911 

 that the temperature influences the rate of growth of the tubes. It 

 seems reasonably safe to assume a difference in the influence of the 

 temperature between the R- and the r-tubes in such a way that the 

 slow growth of the r-tubes becomes relatively still more retarded when 

 the year temperature falls. The segregative numbers, which at a 

 cursory glance do not seem to have any resemblance to simple, mo- 

 nohybrid numbers, become explained in this way. 



The calculations of the segregation of the selffertilized red nerved 

 individuals have of course to be based on the gamete representation 

 obtained in the cross rr X Rr. There is also here a competition be- 

 tween R and r, a certation as I have called this process. In these 

 selffertilizing plants we meet, how^ever, other complications discussed 

 in detail in 1915. No RR homozygotes are obtained; all red nerved 

 plants are Rr. All the 92 descendants obtained in my experiments 

 from red nerved plants have segregated. As the combination RR 

 cannot become realized a segregation in the ratio 2 Rr : 1 rr would be 

 expected. However, we get a segregation in the ratio 2,6S : 1, thus 

 almost 3:1. I have tried to explain the matter assuming an incapa- 

 city of the i?-pollen to fertilize the ß-eggs; there is a repulsion be- 



Hercclitas I. 5 



