H. M. Leake 47 



characters, number of fibres per seed, weight of the individual fibres 

 and volume of seed. It follows, therefore, that the determination of 

 the ultimate causes of variation in the ginning percent cannot be made 

 directly. Rather, those must be sought indirectly through their effect 

 on the three characters under consideration and especially on the 

 number of fibres per seed. 



The conclusion just reached opens up a wide field of investigation. 

 Before practical use can be made of the above conclusions in plant 

 breeding it is necessary to determine the true value which can be 

 assigned to any given plant for these characters. In all observations 

 involving multiple characters a considerable fluctuating variability is 

 found in different samples from the same individual, and it is necessary 

 to determine the extent of such fluctuations. The somewhat tedious 

 nature of the determinations militates against any rapid conclusion 

 being arrived at and, though considerable progress has been already 

 made, a full exposition of the results must await further experiment. 

 Sufficient information has, however, been obtained to show that 

 such individual variations are of a magnitude readily distinguishable 

 from the differences found between different races. A single instance, 

 based on determinati<in of the number of fibres per seed, must here 

 suffice. 



There is thus good reason for believing that further investigation 

 will make it possible to elucidate such cases as that given above where 

 two plants, both having a ginning percent of 25 — 26, when crossed gave 

 rise to offspring of which the ginning percent varied from 36 — 18. 

 There is also hope that the ultimate physiological causes of fluctuation 

 in the ginning percent may be traced. 



In conclusion the writer desires to express his great indebtedness 

 to Mr G. Udny Yule and to Dr G. T. Walker, F.R.S., of the Indian 

 Meteorological Dept., for their kindly help, also to the Authorities of 

 the Imperial College of Science and Technology and especially to 

 Prof. J. B. Farmer, F.R.S., for the facilities given him at that College 

 for carrying out these investigations during a period of leave from India, 

 and lastly to his sisters, too numerous to mention individually, for their 

 kindly help in checking the somewhat lengthy calculations. 



' For an account of the types see Leake, Journal of Genetics, Vol. i. No. 3 (1911), 

 pp. 209—211. 



