The loads are weighed oxi the hirge universal testing machine in another i)art of the room. 

 This is done by liaviug both machines connected up to the same ])nni]), liiocking the weighing 

 machine so that the load on its plunger is transmitted to the scales and weigiiing beam, and then 

 l)umping into both machines. The plungers are of exactly the same diameter; they have similar 

 leather cup jtacking, and hence the error of this method is simi)ly the ditterence in the friction of 

 the two plungers in their packing rings. To test tlu^ accuracy of this method, and to deteiiuine 

 the error, if any, at auy time, a nest of calibrating sjjrings (shown on Plate ii) was made and tested 

 first on the Emery machine at the United States Arsenal, at Watertowu, Mass. The loads were 

 found which corresiionded to given detiections, or in other words the stress diagram of these 

 springs up to a 30,00l>-pound load, which correspoiiils to a little more than one inch elastic detlec- 

 tion. By repeating this test on Hie l()(),Oii()-ponnd universal or weighing machine, and then on 

 the large beam inachine, and plotting the stress diagrams obtained from each, not only can these 

 machines be compared with each other, but both can be com]tared or calibrated with tiie lunery 

 machine at the Watertowu Arsenal. 



In Fig. 5 the three curves corres])onding to the three machines are given. Tlu;^y are so nearly 

 coincident that it is sliown tiiat not only is the universal l(H».(t(K> pound lliehle uuichine correctly 



]"i(i 



BcHoction in incln-a. 

 -StaiHl:iri1izin<r tests with faliliriitini; 



spriuy 



graduated, but that the method used of weighing the loads on the beam machine by means of the 

 universal machine results in no appreciable error. This test can be applied at any time, and jnoof 



