72 HANS TEDIN 



F, (see pag. 76 — 77 and tables 17 — 20) it is at once observed, that purple 

 colour is conditioned not only by two factors but by three. The pre- 

 sumtive factor for rose adopted from previous investigations (White's 

 and Tschermak's factor A, mentioned before) is not a simple factor 

 but composed of two factors. One of these, which I call A, gives 

 light purple, and this one is also the real fundamental factor for the 

 flower colour in Pisum aruense. A second factor, my B factor, gives 

 rose together with A. Then there is an additional third factor G 

 (White's and Tschermak's purpling factor B) which together with A 

 gives violet, while all the three together give purple. B and C are, 

 therefore, both alone or coexisting w'ithout effect (cryptomeric or 

 masked) as soon as A is absent. 



Thus the different flower colours in peas show the following con- 

 stitution in the homozygous states according to my results so far 

 obtained. 



Light purple: AAbbcc. 



Rose: AABBcc. 



Violet: AAbbCC. 



Purple: AABBCC. 



White: aaBBCC, aabbCC, aaBBcc and aabbcc. 



No difference between homo- and heterozygotes of the respective 

 flower colours has been observed. 



The whites crossed with the homozygous purple (AABBCC) by 

 Mendel and other workers must have had the constitution aaBBCC 

 as a segregation in Fn in purple and white in the ratio 3 : 1 invariably, 

 took place. And the whites used by various workers when crossing 

 with the homozygous rose (AABBcc) must have been constituted in 

 the same manner as the segregation in Fa resulted in purple, rose and 

 white in the ratio 9:3:4, thus necessitating a Fi constituted AaBBCc. 

 These observations favour the belief that the common white flowered 

 varieties of the peas usually if not always (most of them have not yet 

 been investigated in this respect) contain both B and C in homozygous 

 state. This circumstance would get quite a natural explanation if 

 we, following the opinion of Lock (1. c), suppose that the whites 

 have originated from purple flowered plants by the »loss» of the 

 colour factor A, a supposition quite hypothetical at present. 



It is of very great interest to find that the hilum of the violet 

 flowered individuals have an abnormal structure, and these plants 

 show also a very poor development of the seeds. A normally developed 



