COMPARISON OF RESISTANCES. 



that these differences arose from the determination of the tempera- 

 tures. Fleming* resumed this investigation in 1881 ; he found 

 considerable differences from the preceding determinations, and the 

 deviations sometimes amounted to croon. It seems difficult to 

 decide whether such large differences are due to a real alteration 

 of the standards of resistance, or simply to the difficulties of exactly 

 determining the temperatures. 



926. RESISTANCE BOXES. In comparative experiments it is 

 necessary to have a series of resistances, the values of which increase 

 regularly. They usually consist of coils placed in the same box, and 

 provided with keys by which they can be introduced into a circuit. 



In order to prevent any action of these coils on galvanometers, 

 and to reduce the effects of induction to a minimum (760), the wire, 

 after being doubled, is wound on the coil. These wires should be 

 carefully insulated, and when once the coil is constructed, it is well 

 to surround it with a layer of paraffine. 



The two ends of the wire are soldered to stout pieces of brass, 

 with a small interval between them. They are cut away on each side, 

 so as to allow of the introduction of a plug, which thus connects 

 them by a conductor, the resistance of which may be neglected. 

 This mode of connection is practical and convenient. 



The resistance of the plugs cannot always be disregarded, and 

 experiment shows that it may amount to as much as o'oooi of 

 an ohm.f 



The coils are arranged in succession, so that each piece of brass 

 connects two successive pieces. When a plug is removed, the resist- 

 ance of the corresponding coil is brought into the circuit ; by putting 

 in the plug it is removed. 



If, instead of soldering the wires directly to the pieces of brass 

 themselves, they are joined by the intervention of small copper rods, 

 the wire of two successive coils should not be soldered to the same 

 rod, but each wire should have a rod of its own, so that the resist- 

 ance introduced by removing two plugs is exactly the same as is 

 obtained by removing each plug in succession. These rods, as 

 usually employed, have a certain resistance which cannot be neg- 

 lected in accurate measurements. 



In verifying these coils, it is desirable that each of the pieces of 

 brass should have a hole, in which a special plug, provided with a 

 binding screw, can be introduced. This is a very convenient plan 



* British Association Report for 1883, p. 41. Southport. 

 t DORN. Wiedemann's Annalen, Vol. xxn., p. 558. 1884. 



