iv THE COTTON FIBRE 85 



submitted to an expert * for grading. The first pickings 

 from all were classified as uniform. The second pickings, 

 ripened during the rise of the water-table, were progres- 

 sively graded ; the top terrace gave the strongest lint, and 

 each step down was a stage weaker. These second pickings 

 came from the bolls which had ripened during the rise 

 of the water-table. 



Another set of data came from an F. 2 of Egyptian x 

 American. The lints were classified by the grader into 

 groups, according to their strength. The amount of 

 shedding per plant in this family had been computed 

 up to a certain date and found to range from 100 per 

 cent, to 19 per cent, of the flowers opened, with a mean 

 at 70 per cent. Grouping those plants which possessed 

 Egyptian lint according to their strength, the following 

 figures were obtained : 



Strength. Mean Shedding. 



Weak 70-2 per cent. 



Medium t>8'7 



Strong 60-6 



In this case the cause of variation in shedding is 

 constitutional, and the correlation is hence more striking 

 than when the variations were due to environment only. 



This same family was dissected in the same manner for 

 other characters, but no other correlations were obtained, 

 except with regard to lint- weight (mean weight of lint per 

 seed). This ranged from 0'021 g. to O'OSO g. with the 

 mean at 0'036 g. Classifying the Egyptians again by 

 strength we found : 



Strength. Mean Lint weight. 



Weak ... 0-0348 g. 



Medium OO358 



Strong 0-0373 



Thus the heavier lint is the stronger. Further, since 

 lint-weight and seed-weight are closely correlated, we find 



* Mr. H. C. Thomas, of the National Bank of Egypt. 



