254 Mr. H. Tomlinson on the Effect of Change of 



produced by loading was comparatively small, and was mea- 

 sured by the usual mirror lamp-and-scale arrangement, the 

 mirror being attached to the brass block at the end of the 

 wire. 



It then occurred to me that a similar arrangement might 

 be used for testing the effect of permanent extension and 

 compression on the thermal expansibility of a metal. A wire 

 was suspended in the axis of an air-chamber, consisting of two 

 concentric brass cylinders enclosing an annular space between 

 them 6 millim. thick. The length of the air-chamber was 

 120 centim. and its outer diameter 10 centim. When the 

 required amount of permanent torsion had been imparted to 

 the wire, steam from a boiler was admitted into the annular 

 space through a tube soldered into it near its lower extremity, 

 and passed out through a similar tube near the top, where it 

 was condensed in a vessel of cold water. After a period 

 ranging from \ hour to 2 hours, when the permanent un- 

 twisting which invariably accompanied the heating had ceased 

 to become sensibly greater, the position of the light on the 

 scale was noted and the action of the steam stopped. The 

 annular space was now filled with cold water, and in about 

 ten minutes afterwards the position of the light on the scale 

 was again noted. The difference between the two readings 

 was a measure of the amount of temporary twist or untwist 

 produced by the change of temperature, the last being, on the 

 average, £5° C. In several cases the heating and cooling 

 were repeated with little or no variation in the amount of 

 temporary twist or untwist. The length of each wire was 

 nearly 120 centimetres and the diameter 1 millimetre. 



The results of the experiments will be found in the following 

 Table, in which are given the amount of temporary twist or 

 untwist in scale-divisions, the amount of permanent untwist, 

 and the load hanging on the wire during the experiment. 

 The wires were all well annealed before being permanently 

 twisted. 



We may gather from the Table the following particulars : — 



(1) Rise of temperature from 15° C. to 100° C. produces 

 in permanently twisted iron, aluminium, and silver temporary 

 untwist, showing for these metals less thermal expansibility 

 in the direction of permanent extension than in the direction 

 of permanent compression. With copper and platinum, on 

 the contrary, rise of temperature produces temporary twist. 



(2) The amount of temporary twist or untwist is in all cases 

 small compared with the amount of permanent untwist pro- 

 duced by rise of temperature. It is also absolutely very small*. 

 * 100 scale-divisions only represent torsion through, an angle of 2° 20'. 



