245 



tained by drawing out glass threads and determining their tensile strength 

 by the method already described. That the increase in the case of glass 

 threads was due in 



4900 



3GCO 



part only to the "skin 

 effect" was indicated 

 by the fact that 

 the tensile strength 

 m e a s u r e ments for 

 threads of different 

 sizes were made more 

 nearly uniform by 

 carefully annealing 

 the threads. This was 

 done by hanging the 

 threads in a verti- 

 cal iron tube with a 

 small weight at- 

 tached to the lower 

 end of each thread 

 to keep it straight. 

 The entire tube was 

 then brought to a 

 temperature slightly 

 below the melting 

 point of glass and 

 maintained at that 

 temperature for one 

 hour, after which it was allowed to cool slowly in the asbestos jacket sur- 

 rounding it. 



An attempt was made to anneal the gelatine threads by hanging them 

 in a moist atmosphere in an enameling oven. Various oven temperatures 

 were tried with little permanent effect on the threads. The author in- 

 clines to the view that the "skin effect" does not account for the larger 

 values of the tensile strength shown by the finer gelatine threads, but that 

 it is due chiefly to the cause already suggested — the greater Tiniformity in 

 flip finer threads. 



fooo 

 /(ffm. 

 p?r 

 SlJCTn. 





















































































































1 



















































































































r \ 













































































K \ 







































* 



























































































^ 



\ 



t\l 



t 



Yi 







































■li 



Q 



Lc 



lS 



=> 







- 

























































u 



\r 



'/ 



"X 



~S 



71 











a. 



'1 1 



-Ytf 







































r 





























































































































































,01 S(j. mm. .on 



pt t .1. 



.03 



.04 



