C. Barus — Resistance of Stressed Glass. 



345 



per tube, of which the external and. internal diameters were 

 2j0 1 = , 53 cm 2|0 !! = , 38 cm respectively. The table gives the oscilla- 

 tions for the divers loads P ; the corresponding absolute de- 

 crement dP of P, and the relative value of this decrements in 

 terms of P. 



The amount of variation here given for glass is somewhat 

 smaller than was found at 200° above. In the last case, how- 

 ever, the data are less accurate, and definite statements can not 

 be made. In Table I the values for the 2d apparatus are 

 smaller than for the first, a circumstance obviously depending 

 on the tubes chosen, but which I will also leave without 

 further comment. 



Table I. — Resistance of stretched glass at 100°. 



Tubes. 



P 



Maximum 

 oscillation. 



6R 



10 3 x SR/R. 



Method. 





kg. 



cm. 



ohms. 





Differential Galvanometer. 



I and II 



6 



1-05 



— 21000 



-2-8 







10 



1-63 



—33000 



—4-4 







15 



2-20 



— 44000 



-5-8 







19 



2-90 



— 58000 



-7-7 





I and II 



6 



•87 



-17000 



-2-2 







10 



1-50 



— 29000 



-3-8 







15 



2-17 



—42000 



— 5-4 







19 



291 



-56000 



-7-3 



v 



III and IY 



6 



•29 



— 6000 



— -9 







10 



•74 



— 15000 



-2-2 







15 



1-05 



-21000 



— 31 





in and IV 



6 



•33 



— 7000 



—1-0 







10 



•76 



-15000 



-2-3 







15 



1-11 



— 22000 



— 3-3 





IV 



5 



•56 





— 1-2 



Bridge. 





10 



1-33 





— 2-8 





This table proves conclusively, that within the given limits 

 of variation, the resistance decrement experienced by glass is 

 proportional to the applied stress. For the given conditions 

 (2/> 1 =-53 cm 2 i o 2 =40 0m ) it is as high as 380/10" per kilo stress, and 

 is not below 210/10° per kilo stress. Since the section q= -10 cm2 

 nearly, it follows that the mean relative variation of resistance 

 due to stretching is about 30/10 9 per gram load, per square 

 centimeter of section. Mr. H. Tomlinson* who investigated 

 the effect of stretching metals, finds that for steel, iron and 

 brass the total variations are only about 1/15 as large as this, 

 and of the opposite sign. §11. 



1. When the temperature is sufficiently constant, for instance 

 in the case of a steam bath, experiments may be made with a 

 single tube. Let a bridge adjustment be so arranged that 



* H. Tomlinson: Proc. Roy. Soc, xxv, p. 451, 1876; id. xxvi, p. 401, 1877. 



