A. 8t Clair Caporn 26/ 



The F., and F.. Generations. 



Glumes. 



183 F^ plants were examined. As is usual in T. polonwum crosses, 

 it was found that, while certain ears were clearly long- or short- 

 glumed, there was a very large proportion which verged on these 

 e.xtremes, as well as those which were obviously intermediates. The 

 glume lengths were therefore measured and plotted in the form of a 

 curve (Table II and Chart II). 



The curve (dotted), it will be seen, falls into three periods, indi- 

 cating that segi-egation, although indistinct, has nevertheless taken 

 place. Professor Biffen' has already explained the real significance of 

 such a curve in other crosses of this kind. It results from the over- 

 lapping of three separate curves representing the lengths of homo- 

 zygous short, homozygous long, and heterozygous short-long glumes. 

 Where the curves overlap there is a piling up effect equal to the sum 

 of the values of the constituent curves at these points. Thus : — 



The triple nature of this type of compound curve is more plainly 

 recognisable when a greater number of glume lengths is available for 

 measurement. The dot-and-dash curve on Chart II represents the 

 glume lengths of 151 F^ individuals derived from 10 heterozygous F. 

 ears, and when added to the first curve, produces the large continuous- 

 line curve in which the three periods are far more distinct. It must 

 be admitted that the third region is far from smooth. This is due to its 

 diffuse character. By adopting the smoothing line {sm), however, one 

 can indicate the real position. The point P marks the summit of the 

 overlapping curves — always very pronounced where a high curve overlies 

 a very low one. 



1 Biffen, B. H., Juurn. Aijr, .S'c, 1905, 



18—2 



