F. L. Engledow 125 



indeed any influence upon glume-length. Some two months before 

 harvest the ear of the wheat plant has fully emerged from the sheath 

 and, whereas its glumes have attained practically full development, the 

 grain is jUst commencing to form — fertilization having, it may be sup- 

 posed, just taken place. Two plants of the same glume-length may thus, 

 if their post fertilization environments be markedly different, develop 

 grains of distinctly different length, whereas under identical conditions 

 they would, presumably, have developed grains of the same length. 



Whatever may be the explanation of the low values of r, it is to be 

 expected that if classifying by glume-length automatically ensures grain- 

 length classification, then the value of r will be the same for P (or K or /) 

 in Fo, F2, and F3. As a test, r was evaluated for Fz.P. The first of the 

 eight pairs of measurements made in every F2 plant (see § II above) was 

 used, i.e. first glume-length and first grain-length. Calculation gave : 



(7^2.^").^=-^ 0-301 ±0090, 

 and it may be noted that : 



{Fo.P).r-{F2.P).r = + 0-09 = p. e. of (F^ .P).r. 

 Thus r glume/grain appears to have reasonable constancy and there- 

 fore although the absolute value of r is not great, a constant relation- 

 ship between glume-length and grain-length is inferred. Owing to the 

 fact that glume development is not contemporaneous with that of grain, 

 the relationship is liable to be concealed by environmental influences, 

 particularly when it is expressed as a coefficient of correlation. 



VIII. An Apparent Relationship between Glume-Length 

 AND THE Development of Hairs on the Glume. 



Both Polish and Kubanka are, from the point of view of ordinary 

 descriptive botany, regarded as smooth-glumed wheats. As a fact each, 

 and in particular Fq . P, exhibits a slight development of hairs on the 

 glumes. In F^ there were found a few plants decidedly hairy and very 

 noticeably different in this respect firom the parents. All plants more 

 hairy than Fq . P were noted and the distribution of their glume-lengths 

 is given in Table XXIX. They numbered 56/494, but the data available 

 do not permit of any particular construction being placed upon this 

 ratio. It will be observed that the increased development of hairs 

 appears to have taken place in association with glumes of lengths below 

 the length range of Fq . P. This fact and the impression gained from an 

 examination of the glume hairs of great numbers of Fq. P plants, is in 

 agreement with the finding of Backhouse (2) that " length of glume has 



Journ. of Gen. x 9 



