272 Inlieritance of Glume Lengtli and of Colour m Oats 



The 8 F2 plants with streaked grains were divisible into two classes 

 of 4 each : — 



(1) Pure streakeds. 



(2) Heterozygous streakeds. 



The latter provided a mixed progeny exhibiting a 3 : 1 segregation. 



The next two groups of F3 generations, both derived from non- 

 coloured parents, have already been mentioned. Their chief character- 

 istic is the dominance, or pseudo-dominance, of absence of colour. The 

 first is composed of one quarter streakeds ; in the second the proportion 

 is seven-sixteenths. 



At the base of' the table all the anomalous colour-throwing Fq plants 

 have been collected. That the cause of the anomaly is traceable in 

 several would sufficiently justify their being cast out altogether, but in 

 the interests of exactitude and to avoid criticism on the score of de- 

 liberate omission, they have all been inserted. No. 104 contains only 

 4 plants — too small a number for an accurate determination of the 

 proper ratio. No. 35 has a similar fault. In addition the three or 

 four parental grains were so badly shrivelled and rusted that the colour 

 question had to be left unanswered. The mere germination of the one 

 was in itself a surprise. In three out of the remaining four the facts 

 all point to certain roguing. T. polonicwn and its derivatives are 

 clearly susceptible to more roguing than most wheats on account of 

 the wide gaping apart of the flowering glumes when the anthers are 

 dangling out. It is quite likely that sometimes cleistogamy has not 

 been effected before this occurs, and that very occasionally a stigma 

 receives foreign pollen. In certain cases this roguing and its origin 

 are readily discoverable. For instance, a purple-grained F^ ear with 

 long glumes derived from a long-glumed F2 ear gave an F4, row con- 

 taining 14 plants, all of which were long-glumed and bearded. The 

 exception had slight scurs, intermediate glumes, and purple grain — 

 obviously the result of one of the stigmas on the parent ear receiving 

 pollen from a beardless, short chaffed non-purple belonging to another 

 culture altogether. Something similar happened to No. a in which it 

 would seem that a purple F^ grain was contaminated with pollen from a 

 beardless non-purple. For the F.^ ear was a beardless purple which 

 threw an F.^ row containing 4 beardless plants, of which one was plainly 

 of Squarehead origin. No. a may be a pure purple, and the unravelling 

 of the F^ figures would doubtless result in the settling of this point ; 

 but they are too few. More troublesome to explain is No. 160. Beard- 

 less in the F^, thus indicating roguing of an Fi grain, and of a very 



4 



