278 Inheritance of Glume Length and of Colour in Oats 



writer it seems not at all impossible that streaking is dependent on an 

 entirely different factor (or factors) from flushing, and that apparent 

 reversal of dominance in the non-coloureds throwing streakeds may be 

 due to an inhibitor affecting 'streak' alone. For I cannot inhibit 

 ' flush ' as in maize, because, as has just been remarked, no non-coloured 

 forms throw flusheds, nor do any streakeds \ But whether or not this 

 same factor I affects flushing to any extent in an indirect manner is 

 not clear. It plainly has no visible action in the Ft^ generation, where 

 the purple flush is fully dominant; but the remarkable deficiency of 

 full coloured forms in the F^ generation is a puzzle to explain otherwise. 

 That, and the lack of any F^ segregation with the F^ ratio, is the snag 

 on which every theory so far launched seems fated to founder. 



Stabilisation. 



Several pure lines have been extracted during the experiment. 

 Most are full coloured purples with long glumes. That no pure long- 

 glumed flusheds appeared in the F2 generation is to be attributed to 

 the very small number of pure flusheds produced rather than to any 

 coupling of ' short glume ' with ' flush.' The glume distribution in the 

 intermediate No. 20 disproved the latter suggestion and no difficulty 

 was experienced in obtaining them in numerous F^ cultures. Outwardly 

 similar, genetically they are very varied. Three pure flushed lines 

 which were bred from the \'2F : SS : IN group, for instance, may or 

 may not be alike. They may give ^1 mono- or di-hybrids on crossing, 

 they may mask a streak even as a black oat will contain, but not show, 

 hypostatic grey, or they may be pure for the flushed character alone. 

 The same remarks apply to the pure purples of the 27F : 9S : 28iV 

 section in Table IV, although no ^''4 cultures of them were actually 

 isolated in quantity. 



No streaking occurred in the SF : IN and 9^ : *7N groups, and, 

 assuming that in both cases the colour is due to the interaction of 

 the factors C and P, then the extracted pure purple races must all 

 have the formula CCPP. 



As for the various pure cultures obtained from No, 20, all that can 

 be said of them is simply that they are pure flusheds, but whether alike 

 or different and of exactly what nature can only be decided from their 

 behaviour after being crossed with whites of known constitution. 



Pure streakeds, both long- and short- glumed, have also been stabilised, 

 and lastly non-coloured Elobonis. 



^ An extremely doubtful exception which may or may not contain a stray, and of 

 unknown parentage, is No, 34. But see remarks on p. 273. 



