The *ransvcrse tests were carried out with the Witkstced 100-to 

 machine by menus of a specially designed iirrnpgcmcnt shown in the 

 photograph on the opposite page. 



By this arrangement the two ends are gradually forced downwards 

 while the centre is supported upon the addle suspended from the 

 lever of the machine. Thus the two halves of the beam are really 

 equivalent to two cantilevers loaded at the ends. By means of a very 

 simple device, the pressure can be increased t-o regularly as to ensure 

 an absolute equality in these end loads. 



Figures 1 and 2 show the device employed to keep the pressure on 

 the ends of the beam always normal to the suiface. The spherical 



joint allows the bearing to revolve, and by means of the prismatic slot 

 any form of benring surface may bo introduced. 



The formula used in calculating the skin-streuths and co-efficients of 

 elasticity have been deduced by means of the ordinary theory of flexure 

 which is bused upon assumptions which actual experience shows to be 

 far iroro being true. These assumptions are : 



