274 Inheritance of Glume Length and of Colour in Oats 



behaviour of the known factors, is the fact that almost none of the 

 ^3 generation show the same ratios as the F^ generation." They then 

 instance a ratio of 1843 purples: 188 reds: 545 non-purples obtained 

 from ^1 seeds which is obviously 12 : 1 : 3. Nothing like it appeared 

 in the following generation, but, while admitting later that the de- 

 ficiency of reds is too great to justify this step fully, they seize upon 

 the slender chance of their figures representing a ratio of 9 : 3 : 4 and 

 thence unfold a carefully formulated theory with which most of the 

 other results accord fairly well. 



Now a 12 : 1 : 3 ratio appeared in the F^ generation of this wheat 

 cross, but the opportunity for harmonising it, — or, at any rate, attempting 

 to do so, — with any even remotely similar F^ segregation was absolutely 

 lacking. For in all the subsequent 12 : 1 : 3 ratios the colours of the 

 terms were transposed. In F^, 12:1:3 represented non-coloureds, 

 streakeds, and flusheds respectively; in F^, flusheds, non-coloureds, and 

 streakeds. The nearest approach to the F^ ratio, if one ignore the far 

 too high proportion of streakeds, was in the rogued row, No. 146 ; but 

 because of the adulteration no confidence can be put in it. Further- 

 more the F^, seeds which gave rise to the row differed from the F^ 

 grains in one vitally important point. They were not purple. In any 

 case, however, even were as many as 4 factors involved in the pro- 

 duction of colour, out of 256 plants 16 heterozygotes identical with the 

 Fi type would be expected, — in other words at least 9 out of 158. 



It may be pretty definitely asserted, therefore, that as far as this 

 particular purple x non-purple cross is concerned no F^ ratio arising 

 from purple seeds like those in Fi and resembling the ratio of the F^ 

 generation exists. 



If the full flushed purples yielding offspring free from streaked 

 forms be lifted out of Table IV and separately studied, it will be noted 

 how very much akin in their behaviour they are to certain of the dark 

 maize purples investigated by East and Hayes. The production of 

 flusheds to non-coloureds in the ratios 3 : 1 and 9 : 7 indicates F^ mono- 

 and di-hybrids which are functions of two complementary colour-making 

 factors. When also these data are taken in with those of the remaining 

 full coloured plants in the F^. generation, 4 pure purples out of 27 give 

 further support to this idea. Nevertheless, as East and Hayes point 

 out in a very similar situation \ as soon as these individual ratios are 

 taken in conjunction with the main mass of data, it has to be recognised 

 that " tri-hybrids and tetra-hybrids are possible which give such results." 



^ East and Hayes, p. 85. 



I 



i 



