18 



CIRCULAR NO. 128, BUREAU OF PLANT INDUSTRY. 



screw with a uniform motion, and the breaking strain is read on the 

 dial in tenths of grams. Twenty fibers are thus broken, one at a 

 time. The breaking strain of each is recorded, showing the varia- 

 tion, and the average is determined as the tensile strength of the 

 sample. 



While there is much variation in every sample, it has been found 

 by numerous trials that the average breaking strain of 20 fibers is 

 approxunately the same as that of a larger number. This is espe- 



FiG. 1. — Fiber tester, with a capacity of 0.1 to 50 grams , used for deterimning 

 the strength of cotton fibers. 



cially true of seed cotton, where it is possible to take one fiber from 

 each of 20 different seeds. Furthermore, it is found that the fibers 

 taken from midway on the side of the seed are more imiform than 

 those at either end. Those at the pointed end are most variable. 



The results from a large number of tests of samples from nearly 

 all of the prominent varieties in the seven different groui)s of Ameri- 

 can Upland cotton, and also Sea Island and Eg^^ptian cottons, are 

 condensed in Table I. 



|Cir. Il-'S] 



