Design for a Standard of Electrical Resistance. 25 



the wire of the standard, assuming that equilibrium of tem- 

 perature has been attained. If a current is now passed through 

 the coil in order to take a measurement of its electrical resist- 

 ance, the temperature of the wire is raised, and its resistance 

 is altered. 



Other things being equal, the best design of coil is that in 

 which this electrically developed heat is got rid of by diffusion 

 as quickly as possible. The embedding of a coil in a large 

 mass of badly-conducting material like paraffin or ozokerit is, 

 from this point of view, a great disadvantage. 



Sufficient electrical insulation has to be provided ; but this 

 should be achieved without the use of more enveloping insu- 

 lation than necessary. 



The two chief objections to the B.A. form of standard are, 

 however, these : — 



First, it cannot be placed in water with the shell wholly 

 underwater or under ice without short-circuiting the electrodes, 

 and, when used as intended, whilst the narrow or bottom 

 portion of the coil is in the water, the upper and more massive 

 portion is in the air, and therefore may be at a different tem- 

 perature to the bottom portion. Hence arises a doubt as to 

 the actual temperature of the coil of wire. It has to be borne 

 in mind that ihQ limitation of accuracy in such comparisons of 

 standards of resistance is determinedby the difficulty of ascer- 

 taining temperature, and not in the mere measurement of 

 resistance. Uncertainty as to the actual temperature of the 

 wire to the extent of one or two tenths of a degree Centigrade 

 renders nugatory elaborate arrangements for very accurate 

 measurement of resistance. 



Second. The standards, as at present constructed, are liable 

 to another defect. If the standard is being used in melting 

 ice or snow, and therefore cooled to 0° Cent., deposition of 

 dew will take place upon the upper surface, whether the 

 ebonite lid or paraffin-wax surface, through which the copper- 

 rod electrodes protrude. The copper rods are originally lac- 

 quered or varnished, but when the lacquer wears off, any film 

 of moisture so deposited will short-circuit the electrodes and 

 reduce the observed resistance. In comparing standards in 

 melting ice, either then the whole shell must be as far as pos- 

 sible placed under the melting ice, in which case stirring the 

 liquid may splash water on to the surface of the paraffin, or 

 else the shell has to be only partly immersed, in which case 

 ambiguity exists as to the actual temperature of the coil of wire. 



These and some other difficulties, such as that of keeping a 

 rather deep vessel of melting ice at a constant temperature, 

 have impressed on the writer the necessity for modifying the 



