F. L. Engledow 115 



standards. Table II suggests that an F^ family should be regarded as 

 K if, and only if, no plant in it was of greater glume-length than 

 14*5 mm. Since the accuracy of measurement was only + 05 mm., and 

 the class interval was 10 mm., the value of 155 mm. was finally adopted 

 as an upper limit defining the K type of ^3 family. 



From the records of observations, the serial numbers were noted of 

 all F3 families in which occurred no plant of glume-length gi-eater than 

 15*5 mm. These families were inspected, and every plant of every family 

 proved to be to the judgment of the eye — at this stage fairly practised — 

 of K type in general appearance. Thus far, therefore, the metrical 

 standard was applicable ; moreover it harmonised with eye-judgment. 



For the Polish (P) standard it was natural to turn to Table I 

 (Fo . P distribution) and, for the time being, ignoring the outlying 

 observation at 125 mm., to adopt 19"5 mm. (205— 10) as the lower 

 limit for the P type. ^3 families would thus be accepted as P if, and 

 only if, they contained no plant of glume-length less than 19*5 mm. 

 Certain F^ families were found which, although they failed to pass this 

 numerical test, were obviously composed entirely of plants possessing 

 the unmistakable Polish eye-appearance. The "shift" downwards in 

 glume-length exhibited in F2 had persisted in ^3, and thus had rendered 

 inapplicable the metrical " P " standard derived from the distribution 

 of Fo. P. A fresh metrical standard was essayed in the form of the 

 total range of glume-length variation in ^3 families. This too failed to 

 harmonise with facts of eye-judgment — facts so defined that they could 

 not be ignored. 



After full trial, the metrical standard was abandoned in the case of 

 Polish, and in its stead eye-judgment was adopted. Polish and Kubanka 

 are so well characterised in glume-form that the practised eye has little 

 difficulty in separating the one type from the other and from heterozy- 

 gotes (intermediates). The intermediates proved to be in every one of 

 the numerous glume (and grain) characters, excellent " blends " of the 

 parent forms. 



The standards thus finally adopted were applied, every F3 plant being 

 twice judged. The second judgment accorded with the first and the 

 result was : 



ilUK families = 1237 plants " 

 F^=\261 I „ =2854 „ ■= 5145 plants. 

 ( 101 P ' „ = 1054 „ ) ^ 



Regarding all Pg plants as simple heterozygotes, if in their P3 families 



