76 



HANS TEDIN 



true of the F^ individuals with light purple coloured flowers sown to 

 Fs (1912). 



Tables 3 — 16 show segregation in F3 of the F2 individuals sown 

 and can, therefore, be said to be a verification of table 2. As is evident 

 from the tables there is generally a good — in most cases a very good 

 — correspondence between the theoretical and observed numbers in 

 the F3 generation. Table 12 and 16, however, show exception. Table 

 12 shows the segregation 3 violet: 1 white in the progenies of F2 in- 

 dividuals with violet flowers and constituted AabbCX. Table 16 gives 

 the segregation 3 light purple: 1 w^hite in progenies of F, individuals 

 with light purple flowers and constituted Aabbcc. In both cases the 

 whites are in excess, and the deviation is greater than three times the 

 standard error. Not taking into consideration those two deviations the 

 whites are often found to be somewhat in excess. I dare not say 

 that the whites on the whole are more branched than those otherwise 

 coloured, but last summer it was found that the person in charge of 

 the marking very often overestimated the number of whites by coun- 

 ting large branches as complete plants. I can give no other explana- 

 tion of the excess of the whites and of other deviations than that 

 of assuming the play of incidental causes. 



Table below, and tables 17 — 20. Back-crosses and other crosses 

 with whites from F2. 



The factorial constitution of the 16 white flowered individuals 

 segregated in Fo of crosses I, II, and III in table 1, and the theoretical 



