F. L. EnCtLkdow and J. P. Shelton 199 



Upbringing by such a ""foster mother'' might have some predetermining 

 influence and might be the cause of "shift" [for a fuller consideration 

 see Engledowd). pp. 127-8]. This explanation is based upon a belief 

 that there is a fairly close relationship between the weight, composition, 

 etc. of the mother seed, and the attributes (glume-length included) of 

 the resulting plant. To test this relationship liecame a second object of 

 investigation. 



Many morphological and economic attributes of plants may be 

 metrically represented in a number of different ways. Cereal plants 

 illustrate this point. As generally grown a jilaut has a main axis and a 

 number of axillary shoots or tillers. To obtain an expression of the 

 glume-length, rachis-length, ratio of grain to straw etc. for a population 

 of any variety, it is possible to confine observation to one ear-bearing 

 stalk per plant. The largest, the first formed, or a random ear may be 

 chosen: or every ear of the plant may be included and the observations 

 for the whole plant be averaged. Labour is minimised if only one ear 

 per plant be observed but the available number of observations is 

 increased by the inclusion of every ear. Differing sets of circumstances 

 have led sometimes to the one practice and sometimes to the other and 

 it is clear that the justification of each in its own circumstances, mu.st 

 re.st upon the relationships which prevail among the tillers of the indi- 

 vidual plant. These relationships formed the third part of the enquiry. 



§ II. Material and Method. 



The strains of Polish and Kubanka were the ones used in the glume- 

 length investigation. They are both old and carefully kept pure lines. 

 The seed for each variety was obtained from forty plants of the 1919 

 harvest. The main ear of every one of these plants was measured for 

 rachis-length and glume-length and all of the grains were weighed and 

 separately labelled. Sowing (in 1920) was completed in one day and 

 both germination and growth were good and of as great uniformity as 

 is usually attainable under the conditions of careful experiment. At 

 harvest (1920), every ear of every plant was separately collected, and 

 later on its rachis-length and glume-length were determined. Previously 

 it had been usual to measure only one glume per ear [see Engledow(i), 

 pp. 111-2], but for this investigation both glumes of the "median" 

 spikelet of each side of the ear {i.e. four glumes per ear) were measured. 

 The average of these four measurements is, hereinafter, referred to as 

 "glume-length." 



Journ. of Agric. Soi. xii 1'^ 



