A. St Clair Caporn 237 



amount of variation in the relative numbers of the different con- 

 stituents, the six F^_ rows were essentially alike. All included "pure 

 tights," tight-containing types similar to the F^ parents, and one or 

 more of the forms such as have been described as occurring in the 

 original samples of Avena nuda. 



It was found that practically all^ the plants of mixed type picked out 

 of the F^ generation gave rise to F^ rows of the kind just described, no 

 matter what the relative proportions of tight and loose grains. Some 

 even among the panicles sown as " pure tight " and " pure loose" behaved 

 in the same way, these, as has been explained before, being obviously 

 either damaged panicles or cases in which the preponderance of one 

 type of grain was so great as to render classification uncertain unless 

 the whole plant were examined. The number of these imagined 

 " pure " types which split in this way was small, and it has already been 

 shown (p. 236) that in the case of the pure tights, in which the observed 

 figures were too high, the excess over expectation was due simply to 

 these mistakenly placed heterozygotes. 



According to the nature of the paleae the progeny thrown by each 

 of the similarly segregating ^3 rows were placed in the following 

 groups. 



I. Pure tights. — All the grains on the plant enclosed in tough, 

 wholly sclerotised paleae (cf. the respective tight parents). 



II. Tight-containers'^. — One or more " pure tight " paleae, the rest 

 varied. 



III. Hardbacks. — No " pure tight " pales. Pales partly mem- 

 branous, partly hardened, the hard portion varying from a slightly 

 thickened midrib to the stiffening of nearly the whole palea. 



IV. Penulti-looses. — As in III, but the hardening is never found 

 above the lowest palea in any of the spikelets. These verge very closely 

 on the 



V. Pure looses. — All paleae absolutely membranous. 



119 sowings gave rows containing 2445 plants and exhibiting the 

 splitting just described. The following table shows the distribution of 

 the various forms. 



1 Exceptions are the "tight-containers" dealt with subsequently. See Table IV et seq. 



■'' To avoid possible misapprehension it may be stated that this term is devoid of any 

 implication as regards the genetic properties of the plants to which it applies. Most 

 " tight-containers " throw " pure tight " plants among their offspring, but there are some 

 which do not throw any (cf. p. 239). 



