74 HANS TEDIN 



poor development of the seeds on the whole it is not surprising to find a 

 relatively large number of seeds incapable of germinating. I have 

 not been able to find any difference with regard to these pecu- 

 liarities between violet flowered forms in the homozygous and in the 

 lieterozygous states, nor between the seeds of the latter combination 

 of violet coloured individuals giving rise to plants with violet or other 

 colours (light purple or white). 



The abnormal structure of the hilum is, therefore, directly connec- 

 ted with the factorcombination AC without B. I am not able to give 

 any explanation as to the exact nature of this connection, nor can I 

 say anything about other characters possibly involved. Investigations 

 are now being planned to settle these questions. 



4. THE TABLES. 



The results of the investigations are tabulated in tables 1 — 20. 

 They contain all the detail numbers, i. e. the numerical segregation 

 of the different Fi and F. progenies. For the sake of control the 

 field numbers of the different families are also put out (in col. 2). A 

 comparison between the observed and the theoretical segregation num- 

 bers of a rather large number of progenies from different individuals, 

 and especially of those showing the most diverging numerical relations, 

 has also been made, although not put out in the tables. However, the 

 correspondence between the observed and the theoretical numbers 

 have as a rule been satisfactory even in these. There are, nevertheless, 

 exceptions where the deviations are too large. 



Only the comparison between the theoretical numbers and the 

 segregation numbers sumined up from all of the progenies belonging 

 together has been tabulated. 



Table 1. Shows segregation in F^ of three different crosses be- 

 tween 01001 (AAbbcc) and three different whites, two of which are 

 without names, and the third called by its stock book number 0154. 

 All three gave purple flowers in Fi. By self-fertilizing Fi a Fa segre- 

 gation in purple, violet, rose, light purple, and white took place in the 

 ratio 27 : 9 : 9 : 3 : 16, as shown in the table. Consequently, the white 

 parents must have had the constitution aaBBCC. The summed up 

 observed and theoretical numbers agree very well. In all three crosses, 

 however, the number of the purples is somewhat too small, while the 

 number of whites is somewhat too large. When the numbers from 



