H. M. Leake 43 



of the seed, and hence the surface area and " density," will, in part at 

 any rate, be determined later by the nutrition supplied to the ovule 

 throughout development. It seems, therefore, advisable to use the 

 number of fibres arising from a single seed in preference to the figure 

 obtained for the density for the fibres. 



From such a priori considerations, there is reason to suppose that 

 the ginning percent actually recorded will be the resultant of at least 

 the four characters : 



(1) Volume of seed, 



(2) Specific gravity of seed, 



(3) No. of fibres arising from a single seed 

 which may be simple characters, and 



(4) Weight of the individual fibres, obviously a complex character. 

 A further consideration will show that the problem is probably even 



more complex. The seed and lint are developed in a closed cell of the 

 fruit, each cell of which contains a number of seeds, frequently eight. 

 According to the conditions of environment, the nature of the tissue, 

 and, consequently, the space in which the seed must develop, will vary. 

 It is probable, therefore, that, during the course of development an 

 unknown, but appreciable, effect is produced bj' the mutual pressure 

 thus brought into play. Conclusive evidence of the exact extent of such 

 nmtual pressure is not so far available. It is clear, however, that the 

 effect will not directly influence the ginning percent but that such effect 

 as it may produce will be indirect and through one or other of the four 

 characters already noted above. The importance of a recognition of this 

 nmtual pressure lies in the fact that the actual values obtained for the.se 

 characters may not represent the true potential capacity of the jjlant. 

 This has already been noticed by Balls'. The present discussion, which 

 concerns the factors directly controlling the ginning j^ercent, is not 

 affected therebj', and the problem as it now stands may be defined as 

 the determination of the degree to which each severally of the four 

 characters given above influences the ginning percent and of the extent 

 to which they, in combination, account for the range of variation recorded 

 in that character. 



Methods. In a preliminaiy note of this nature it is not possible to 

 give in detail the precautions it has been found necessary to adopt. 

 The main outlines of the method only can be indicated. The unit used 

 as a sample is the seed-cotton derived from a whole and not diseased 



' Balls, W. L., The iottoii plant in F.yijpt, p. 170. 



