104 Mr. H. L. Callendar on the 



[I have been informed that a very high authority has 

 brought forward as an argument against Mr. Pickering's 

 methods the fact that a spring (or lath) may be pinned down 

 so as to coincide to any desired degree of accuracy with any 

 number of experimental points, and trust that I may be 

 excused if I attempt to answer the objection. The use of an 

 unlimited number of applied forces (one at each pin) is open 

 to the same theoretical objections as would apply to an equa- 

 tion containing an unlimited number of constants. As 

 Mr. Pickering used for drawing his curves a part of the lath 

 throughout which no applied forces were acting, two distinct 

 curves of the kind used by him would meet at each pin, 

 and a discontinuity would exist, which can be shown by 

 mathematical analysis to involve an abrupt change in the 



d 3 v 

 value of ~k, i. e. there would be a break at these points. Rapid 



da . dry 



changes too in the value of -~ might take place if the applied 



forces were large, or the flexural rigidity of the lath small. 

 In fact the pinning-down process simply masks any real 

 breaks which may exist by substituting a (probably much 

 larger) number of others. It would even seem to be theoreti- 

 cally possible that experiments as to the minimum number of 

 applied forces required might indicate the points at which the 

 breaks occur.— E. H. H.] 



XVII. On the Construction of Platinum Thermometers. By 

 H. L. Callendae, M.A., Fellow of Trinity College, Cam- 

 bridge *. 



THE great superiority of the platinum-resistance thermo- 

 meter over other instruments for measuring temperature 

 lies in its comparative freedom from change of zero. Provided 

 that the wire is pure to start with, and that it is protected 

 from strain and from contamination, its resistance, when once 

 annealed, is always very nearly the same at the same tem- 

 perature. 



This statement, which I made in 1886, has been received 

 in some quarters with surprise and incredulity. The evidence 

 on which it has been rejected appears to rest mainly on two 

 well-known facts. Firstly, that the Siemens electrical pyro- 

 meters have always shown large and continuous changes of 

 zerof; and, secondly, that platinum wires when used as 



* Communicated by the Author, 

 f British Association Keport, 1874. 



