

COMPARISON OF PLATINUM AND GAS THERMOMETERS. 45 



considerable, copper was substituted throughout for brass in the new box, 

 the only metals in circuit being copper and manganine. For the platinum-silver 

 bridgr-wire was sulwtituted a manganine strip heavily gilt, placed on edge and 

 >i i etched lietween two adjustable clips. The slider is provided with a fine adjust- 

 ment, which can be manipulated from the outside of the box, without risk of heating 

 tin- x: i lv; 1 1 lometer-contact by repeatedly approaching it with the hand. As in 

 Mr. GRIFFITHS' latest form, the terminals project outside the glass case. The top of 

 tin- lx>x is formed of a heavy slab of white marble 75 centims. long, 30 centims. wide, 

 and 3 centims. thick. For the ordinary form of plug-contacts are substituted heavily 

 gilt forks of forged copper, which can be clamped by powerful steel screws over 

 tongues projecting upwards from the blocks to which the coils are fastened. A 

 general plan of the resistance-box is shown in Plate 1 and the details of one of the 

 contacts in fig. 2. 



VI. THE RESISTANCE-COILS. 



The general scheme of the box connections is almost the same as the one previously 

 described, and may be traced in fig. 1. For the winding, fixing, and annealing of the 

 manganine coils the method described in the official publication of the Physikalisch- 

 Technische Reichsanstalt at Charlottenburg was carefully followed. The specimen of 

 wire used was selected after various tests from several furnished by Messrs. W. T. 

 GLOVER and Co., of Salford, and was double silk covered No. 26 S.W.G. The 

 diameter of several pieces cut from different parts of the bobbin only varied within 

 very narrow limits. In order to simplify the application of the temperature correc- 

 ti->n, the same wire was employed for all the coils except the two lowest. These were 

 of strip manganine, and being originally cut off too wide, could be adjusted till 

 accurate by clipping the edge with shears. 



VII. COIL VALUES ADOPTED. 



In the Callendar-Griffiths resistance-boxes the coils are arranged on the binary 

 scale, and the value of each is determined in terms of the sum of those below it 

 together with a certain length of bridge-wire. Although Mr. GRIFFITHS gives 

 evidence for the utility of this arrangement in general work, it was thought more 

 iinjKtrtant, for the purposes of this research, to have several independent checks in 

 the determination of each coil-value, than that the maximum resistance, measurable 

 with a given number of box-coils, should lie as high as possible. 



Tin- thirteen coils were therefore arranged as follows, the values being expressed ic 

 oli ins : 



40 20 10 4 3 2 1 



A B C D E F G 



02 '05 -1 -2 '3 '4 



N M L K J H 



