Glass, and Metal Wires subjected to Constant Pxdl. 517 



The relations of the constants a and b to the stretching 

 force were then found hy plotting the curves of stretching 

 for the different loads in exactly the same way as is shown 

 for the 500 gram load. The following results were ob- 

 tamed : — 



Load. 



0. 



b. 

 •170 



b 



load 



•00176 



100 grams 



(VS 



200 „ 



147 



•385 



•001925 



300 „ 



235 



•580 



001933 



400 „ 



334 



•800 



•002000 



500 „ 



462 



1006 



•00201 



These numbers show that b is nearly proportional to the 

 load. 



The effect of superposing two loads was then investigated, 

 but this showed that the effect was by no means the sum of 

 the effects which each would produce separately. E. g., sup- 

 pose the stretch due to the first load be given by 



# = «! + &! log t l9 



ti being the time which has elapsed since the first load was 

 put on, and suppose the effect of the second load be given by 



aj=a 2 + 6 2 log t 2 , 



t 2 being the time since the second load was put on. 

 the total stretch should be 



l c = a l + a 2 4- b x log t l + b 2 log t 2 . 



Then 



No such relation was found to hold however, and the only 

 conclusion which was drawn from these experiments is that 

 when t 2 is small compared with t x —t 2 the indiarubber stretches 

 almost as if the second load were the only one acting, and 

 when t-> is large compeared with t l — t 2 the stretch is the same 

 as if both loads had been put on at the same time. This 

 shows that unless the duration of an experiment is small 

 compared with the time which has elapsed since the india- 

 rubber was last -trained, an effect will be produced by that 

 previous strain. 



The alteration in pull required to keep a piece of india- 

 rubber stretched to a certain length was next investigated. 



