ivith Temperature in Wires under Stress. 317 



the month, satisfactory measurements were made as to the 

 relative expansion of the heavily and lightly loaded wires. 

 At the end of the month of April I ceased to have the neces- 

 sary supply of steam, and the experiments therefore came, for 

 the time, to an end. I have therefore erected a wooden pro- 

 tection round the lower end of the tube, and the wires are left 

 hanging exactly as they were, and will remain so till the 

 middle of October, when I propose to continue the experiments. 



It now remains for me to describe the arrangement for the 

 final comparison between the expansions of the lightly loaded 

 and heavily loaded wires. This consists of three parts : a 

 multiplying lever, a fixed scale, and a movable scale (see 

 figs. 1, 2, & 3). The multiply in g-lever is one of Sir William 

 Thomson's aluminium levers, used in his milliampere balances. 

 The lever and supports were in fact taken from a disused trial 

 instrument, and the knife-edges and supports were inverted 

 and adapted to the purpose in hand. One of the supports (a) 

 of the lever is attached to the lightly loaded wire, and the 

 other {b) to the heavily loaded wire (PI. IX. figs. 1 & 2), and 

 the free end of the lever moves over the movable circular scale 

 (mn, fig. 1). Supposing the two wires to expand and con- 

 tract equally, the lever would simply be carried down and 

 up parallel to itself. But if both wires expand and contract 

 together, but one more than the other, the lever will be 

 carried down and up bodily, and will be tilted as well. 



The method which I have adopted up to the present of 

 comparing the two wires has been to take the lightly loaded 

 wire as the standard ; and, using the fixed scale 6' t and a 

 Quincke microscope-kathetometer, to lower or raise the zero 

 of the movable scale m n by means of a screw, h h! (shown en- 

 larged, fig. 3), by as much as the knife-edge of the support a 

 has been lowered or raised by the expansion or contraction of 

 the lightly loaded wire which bears it. The reading of the 

 pointer on the movable scale then gives the difference in ex- 

 pansion or contraction between the two wires. 



I am not yet prepared to give a definite numerical state- 

 ment : and as I have so far only had a single satisfactory set 

 of experiments, I feel that even such results as I have ob- 

 tained require confirmation. There seems, however, to be no 

 doubt that there is a measurable difference between the ex- 

 pansion by heat of the heavily loaded and lightly loaded 

 wires, the expansion of the heavily loaded wire being the 

 greater. 



The corresponding result for gold, platinum, and palladium 

 wires must be carefully determined by special experiments, 

 and must be taken into account in the deductions to be made 



