of Cyclically Varying Temperature. 285 



figure, and by means of the screw-gland D is forced into all 

 the interstices — the instrument thus being made tight under 

 the pressure. Four leads are used, two being joined by the 

 measuring-wire, the other two being joined directly together*. 



Description of the Resistance- Measuring Apparatus. 



In order to readily eliminate the changes of resistance of 

 the leads, an equal-arm slide-bridge was employed. The 

 connexions are shown diagrammatically in fig. 2. 



In the arm A are placed the main leads ; in the opposite 

 arm B are the compensating leads in series with the resistance- 

 box. In the arms C and D are two carefully adjusted equal 

 coils, each having a resistance of about 8 ohms. The bridge- 

 wire is of platinum-iridium, and is half a metre long ; its 

 resistance is 0*2980 ohm at 15° C. This wire was calibrated 

 by Carey Foster's method, and, for the degree of accuracy 

 required (0*001 ohm), was found to be of sensibly equal 

 resistance per unit length throughout. 



The galvanometer, a very dead-beat astatic instrument, had 

 a resistance of about 100 ohms. The magnet system and 

 mirror were suspended by a quartz fibre 1*5 inches long, and 

 was damped by a specially large mica damper very closely 

 enclosed. 



The scale was placed 4 feet from the mirror, so that a 

 change of resistance of 0*0005 ohm could be readily detected. 



The resistance-box was an ordinary plug series box with 

 platinoid coils. It was, in the first instance, calibrated by 

 Mr. E. Wilson, of King's College ; but after considerable 

 use so much trouble was given and, indeed, so many experi- 

 ments spoiled through the inefficiency of the plugs, that links, 

 screwed up by a screwdriver, were substituted for them. 

 Although the links are not so convenient to work with as the 

 plugs, the results have been in every way satisfactory since 

 this change. I then re-calibrated the box, and the two deter- 

 minations were in fair agreement. Below I give a table of 

 the values of the coils used during the experiments : — 



Nominal value 



Calibrated valu 



e 





of coil. 



at 11° C. 





ohms. 



ohms. 





0*5 



0*5063 " 







1 



1-0042 







2 



2*0020 



> 



In terms of my coil, 



2 



2-004 





C.L.C. 417. 



5 



5-013 







* For permission to use the compensators (the invention of Mr. Cal- 

 endar), 1 am indebted to the Cambridge Scientific Instrument Company. 



