ENVIRONMENT OF THE BOLL 109 



proportional to the number of fibres tested, and the 

 testing which the grader applies is nearer to " impact " 

 than to "straining." Consequently, what the grader 

 does is to test the resistance to impact of equal weights 

 of lint each time. 



But this being the case, he might be expected always 

 to obtain the same result, and if all the 



Lmi ^ fibres in a sample were alike he would do so. 

 and Strength. 



All the fibres are not alike, however, and so 



the determination of strength by the grader resolves 

 itself very largely into a test for regularity of strength. 



We can corroborate this deduction by comparing the 

 breaking strain variability from fibre to fibre in large 

 samples with the grader's judgment upon them. There 

 may be a high proportion of strong fibres present, but if 

 they are mixed with weak ones the sample is graded as 

 "weak." This comes down finally to whether the tuft 

 of fibres breaks under the grader's hands with a " snap " 

 i.e., all simultaneously or raggedly. 



Having discussed the evidence obtained mainly from 

 the abnormal conditions of a wide-sown, water-shortened, 

 experimental plot, we will now consider the results ob- 

 tained in actual field crop, since it might well be the case 

 that the variations we induced experimentally were far 

 greater than would ever arise in normal cultivation. 



