THE INHERITANCE OF FLOWER COLOUR IN PISUM 71 



sion of flower colours it also happens that a stout basal branch, which 

 can be followed to the surface of soil, is mistaken for an individual 

 plant. Because of my absence at certain times it has not always been 

 possible to control the work as strictly as could be wished. Therefore, 

 there is no doubt that just in the errors made in the marking 

 we will find the most important cause to the fact that the numerical 

 relations (especially of the descendants of the particular plants) 

 not always agree with the theoretical ratio as well as could be 

 wished. The number of the individuals has often been small, and this 

 fact explains some cases of nonconformity. Last year the marking 

 was made especially difficult by a very irregular and uneven develop- 

 ment of the plants. 



It is evident that the marking of the plants would be both easier 

 and more accurate, if they were sown more sparsely, in two or still 

 better in only one row with greater distance between the plants in the 

 row. But both space and expenses are saved by the way of sowing 

 used in my cultures, and the development of the plants is further 

 kept within more normal limits. If some care is taken, no difficulty 

 is met with in separating the plants without tearing them, when 

 harvesting. 



This paper deals exclusively with the flower colours. The in- 

 vestigations are not yet completed. (This year a great F. and various 

 Fl crosses will be sown). Nevertheless I have for several reasons 

 thought it better not to put off the account of the results already 

 obtained. I hope, however, in course of time to have occasion to come 

 back to these questions in connection with an account of an investiga- 

 tion already started on the genetic relations between the flower colour 

 and leaf axils and certain characteristics of the seeds. 



3. THE RESULTS OF THE INVESTIGATIONS. 



As far as can be observed from the investigations already carried 

 out the mode of inheritance of the flower colours in Pisiim is very 

 simple, as mentioned before, although different from the one already 

 known. Crosses between 01001 (homozygous light purple) and com- 

 mon whites give without exception purple flowered Fi, and F2 segre- 

 gates in purple, violet, rose, light purple and white in the ratio 

 27 : 9 : 9 : 3 : 16 (table 1). From what is said in the above and consi- 

 dering the segregation of the progenies in F3 (tables 2 — 16) as well 

 as the results of the back-crosses and other crosses with whites from 



