Manchester Memoirs, Vol. xlviii. (1904), ISo. 7- 17 



loading. Now if there are four complete sets of ports, i.e., 

 one complete set per quadrant, then four complete cycles 

 of loading are performed per revolution, and hence it 

 becomes very easy to get the highest frequencies of 

 loading. 



The specimen is fixed at .S"^ between the piston and 

 -an arm of the frame F (one end being screwed into the 

 piston, the other into a sleeve held in position relative to 

 the frame F by locknuts M and N, which may, if neces- 

 sary, be also small spur wheels for facility of adjustment) 

 if for experiments under compression, or at S^ between 

 the crosshead H and the frame F if for tension. This 

 crosshead H is connected to the piston by side rods 

 provided with stops to limit the motion when the speci- 

 men breaks and running — preferably on roller bearings — 

 in guides to prevent the piston jambing. A mark on a 

 side rod will enable the experimenter to keep the piston 

 in a constant position — if any adjustment is necessary — . 

 and thus ensure that the pressures attained in the cylinder 

 shall not vary during an experiment. 



In order to determine what pressures the piston is 

 really subject to we may put indicators on the steam 

 chest and exhaust space below kS". 



The lowest pressure registered by the former and the 

 highest registered by the latter give us the limits of 

 loading on the specimen. We may if we like put 

 indicators on the piston itself In this case, in order to 

 secure correct readings, two separate springs should be 

 used, the one for the upper and the other for the lower 

 limit of pressure, and both springs should be restricted 

 by stops to a small range of motion about that corres- 

 ponding to the pressure to be measured. In this way 

 most of the errors of the indicator would be eliminated. 

 If it is desired to have the load-time curve a sine curve 



