A. St Clair Caporn" 269 



It should be emphasized that whenever an insufficient number of 

 plants forjudging was present in an ^3 row of which the parental colour 

 was doubtful, or roguing was suspected, these cases were not included 

 in the count; so that the 27 flusheds, 8 streakeds, and 123 non- 

 coloureds contain no individuals about which any uncertainty exists. 

 Furthermore, although no particular note was made regarding the 

 matter, very probably the grains from which these 180 or so F^ plants 

 were derived all came off a single F^ plant. For, in spite of poor 

 tillering power, plants of later generations with a fair amount of room, 

 such as those at the end of rows, were observed with five or six well 

 filled ears; whereas among the ^1 individuals the encouragement to 

 tiller is considerably greater, since they are always purposely well 

 spaced. 



Apparently, then, one, and only one, heterozygous combination gave 

 rise to the -^2 generation. This conclusion is supported by the sim- 

 plicity of the ratio : it is too even to be a fortuitous mixture of different 

 F2 segregation systems. For 27 : 8 : 123 is, considering the not very 

 large number of plants involved, a reasonably close approximation to 

 3 : 1 : 12, the actual expectation being 30 : 10 : 118. 



In view of the full purple colour of the F^ grains, the minor pro- 

 portion of flushed forms in the F^ generation is very striking. No less 

 so is the behaviour of the 123 non-coloureds, of which 



111 threw non-coloureds only, and 

 12 „ „ „ -f- streakeds. 

 [In the 12 non-coloureds absence of colour was apparently a dominant, 

 as the streakeds made up the smaller part of the offspring.] 



Now if the ratio 3 : 1 : 12 be expressed in the form 48 : 16 : 192, 

 the third term can be conveniently split up on a 1.5 : 1 basis so that 

 the whole may be restated thus, 



48 : 16 : (180 -f 12). 



On applying this to the F^ results the comparison with expectation 

 reads as follows : 



Observed. 27 flusheds : 8 streakeds : (111 -|- 12) non-coloureds. 

 Expected. 30 „ : 10 „ : (111 -f- 7) 



While, therefore, the probability cannot be pressed to the verge of 

 certainty, there is still ground for believing that the two kinds of non- 

 coloureds in the F^ generation were to each other as 1-5 to 1. 



To speculate on the Mendelism of these figures at this stage were 



