82 Mr. J. J. Guest on the Strength of 



25. Ihe Specimens and Holder s.y-*The specimens upon 

 which the tests were made were formed of tubes (steel, copper, 

 and brass) soldered on to holders, which served conveniently to 

 apply the tension loads and the torques, and to introduce fluid 

 under pressure to the interior of the tube. A convenient 

 size of tube appeared to be 1 ^ in. diameter, with the walls from 

 0*025 in. to 0*036 in. in thickness. All measurements of exten- 

 sion and twist were made upon the same length of 8 in., but in 

 order to provide room for the diametral extensometer which 

 was sometimes used, and to avoid the disturbing effects of 

 the holders upon the distribution of stress and strain, the 

 length of tube free from the holders was a foot. A holder is 

 shown in fig. 2. A is a screw connecting the holder to the 

 jaw of the testing machine, it is adjusted so that the forces 

 applying the torque have no vertical component — the specimen 

 being vertical. B is a cross-bar serving to apply or to resist 

 the torque. C is a passage-way for the oil used to apply the 

 internal pressure. In the upper holder, which was the nearer 

 to the weighing part of the testing machine, the exit was 

 closed by a screwed cap ; sometimes a piece of rubber tubing, 

 doubled over itself to form a packing, was pushed in ahead of 

 the closing plug, and proved to be a very efficient means of 

 preventing leakage. The exit was not closed until fluid 

 appeared at it, in the expectation that should a tube prove 

 unsound, or the pressure get beyond control, no damage would 

 then be done to the strain-measuring apparatus. The lower 

 holder was identical with the upper one, except that, in the 

 place of the plug, a pipe was brazed into it, there being a 

 coupling at the further end of the pipe for connexion with 

 the pressure system. 



2§. The Tension Loads.— The tension loads were applied 

 to the specimen by means of a Ten-ton Wickstead machine of 

 the screw type. The shackles of the machine were fitted with 

 the ball-thrust, or perhaps pull, bearings shown in fig. 2. 

 The " male cone " surfaces C were made spherical in order 

 that the specimen could line itself up to some extent and cause 

 the axes of pull and of the specimen to coincide, and yet 

 allow the specimen freedom to rotate. This action was 

 assisted by gently tapping the specimen at low loads. The 

 method of holding the specimen was by means of split collars 

 I), the screw A of the tube-holder being thrust thrqugh the 

 spherical cone of the ball-bearing so far as to allow the 

 collars to be placed round it and inserted into the spherical 

 cone. This proved a rapid and easy method of firmly securing 

 the specimens. I possessed no conveniences for calibrating 

 the testing machine by the direct application of a large dead 



