A/A' IVII.IJAM SIEMENS, J-'.K.S. 125 



A hi(-h proportion is not altered by a general expansion of the 

 metal. A series of very protracted experiments, which it would 

 li;i\r l>een too much to record in this paper, have proved that the 

 conductivity of a metal at a fixed temperature is an exceedingly 

 constant quantity. If copper or iron wire were exposed to intense 

 heat the electrical resistance would be inverted in consequence of 

 the substance diminishing through oxidation ; but that is not the 

 case with platinum at reasonable temperatures, and in measuring 

 f\ti mi'ly high temperature it is necessary not to expose the 

 platinum wire itself to the heat, because it would be partially 

 volatilized. In this case the wire is surrounded by a casing of 

 platinum, and is exposed only to radiation heat from the sides of 

 that casing. It is not even necessary to heat the wire to the 

 ultimate temperature of the furnace, but it suffices to expose it for 

 a given time, say, three or five minutes, to take the reading, and 

 to remove the gauge. In operating thus, we can with this instru- 

 ment measure temperatures reaching to the melting point of 

 platinum itself. "With regard to the second question whether the 

 cylinder of fire clay does itself conduct, I have made the experiment 

 this way. I have taken a clay cylinder with the wire wound upon 

 it, and cut the wire in one place, before exposing the cylinder 

 to an intense white heat. I have therefore measured the amount 

 of current passing through, which current gives the amount of 

 leakage we may expect at a very high temperature. I have thus 

 found that there is indeed conductivity rising with the temperature ; 

 but it is exceedingly small, never reaching more than half per cent, 

 of the conductivity of the wire itself. The case would be different 

 if the pipe clay cylinder were itself exposed to the fire : under 

 which circumstances, a fluxed surface would be formed, which 

 would be conductive. But by excluding the llame from the 

 cylinder, and from the wire, this error has been reduced to an 

 altogether diminished quantity. I forget the third question. 



Mr. Snehts : The method of checking the instrument after it 

 has been in use for some time ? 



Mr. Siemens : That can only be done by a practised electrician, 

 though it would be a desirable thing to do. If the wire was to 

 break, the instrument would cease to indicate altogether ; and in 

 electrical experiments generally, faults, if they occur, are of a very 

 perceptible, and decided nature. For instance, if there was contact 



