60 



ROYAL SOCIETY OF CANADA 



In this arrangement the application of one stress in no way affects 

 the application of the other, and an apparatus of this kind was used by 

 the author for some time and it worked very satisfactorily. It is, 

 however, too sensitive for ordinary use, as the freedom of motion of 

 a specimen suspended in this way is so great that every application 

 of a load disturbs the equilibrium of the test bar with reference to the 

 supporting beam, and it takes some time for the vibrations to die out; 

 moreover, since the weigh lever must be levelled before a reading can 

 be taken, some means must be provided to effect tbis after each applica- 



tion of the load, and this was done by a right and left-handed nut 

 forming part of the link H. This serves very well for experiments up 

 to the yield point of a bar, but it is not a convenient arrangement if 

 the twist is to be carried further. 



In order to overcome these abjections a second apparatus was 

 designed in which the lever F and the adjusting link H were replaced 

 by a worm wheel, secured in a suitable casing and free to rotate about 

 an axis coincident with the centre line of the lever F, and this axis 

 was secured in fixed bearings. This arrangement permitted a test to 



