244 



of drawing, but they are not so after hardening and drying. The cross- 

 section usually becomes quite irregular. This indicates a condition of 

 internal strain which acts to lower the breaking strength of the thread. 

 Coarse threads would be subject to greater strains than fine threads, and 

 therefore we should expect them to show a smaller tensile strength than 

 the fine threads. This is in accord with the results of experiment as 

 shown by the data of Table I. 



TABLE I. 



Cross Section in 



Breaking Strength 



Tensile Strength in 



sq. mm. 



in grams. 



Kgm. per sq. cm. 



.000835 



27.6 



3,305 



.000919 



21.5 



2,340 



.001002 



21.1 



2,106 



.001334 



37.2 



2,778 



.001670 



35.0 



2,096 



.002610 



58.2 



2,230 



.004524 



96.5 



2,133 



.006729 



138.1 



2,052 



.013920 



211.4 



1,519 



.035900 



400.1 



1,114 



. 130300 



800.0 



614 



. 264900 



2 , 850 . 



1,076 



.608900 



5 , 600 . 



919 



1 . 709200 



6,100.0 



357 



Inasmuch as the error in measuring the breaking strength of any par- 

 ticular thread was relatively small and all strains tended to decrease that 

 strength as measured, I have included in Table I only maximum readings; 

 that is, readings which gave the greatest values of the tensile strength for 

 the several sizes of threads. Average readings gave results some twenty 

 per cent lower. 



Curve A of Fig. 1 is a plot of some of the individual maximum read- 

 ings of Table I. It will be observed that the measured tensile strength in- 

 creases very rapidly as the threads are made thinner. A similar increase 

 takes place in wires and glass threads, and is attributed to a "skin effect." 

 This increase is shown in Curve B, Fig. 1, which represents the results ob- 



