90 
PROFESSOR HUGH L. CALLENDAR OK 
of platinum-silver annealed at a red heat, and compensated with similarly annealed 
platinum. The box measured 15” X 6'5" X 3‘5". The bridge-wire was of platinum- 
silver, nearly 34 centims. long, hut only 5 centims. on either side of the middle 
were actually used in the measurements. The scale was of brass, divided by 
Troughton and Simms to half millimetres, with a vernier reading to a hundredth of 
a millimetre. To secure this order of accuracy in the readings, the contact piece, 
consisting of a short length of the same platinum-silver wire with a nearly sharp 
edge, was rigidly fixed to the sliding piece carrying the vernier. The bridge-wire 
was stretched at such a height as to clear the contact edge by about '01 inch, and 
contact was effected by pressing down the bridge-wire on to the contact wire by 
means of an india-rubber finger. This finger was provided with a screw adjustment, 
so that the contact could be set and held at any desired point. To keep the tension 
& 0 0 
| w uy " | 
Fig. 2. Compensated Resistance Box. 
of the bridge-wire constant, which is most important in accurate work, the wire was 
stretched between parallel bars of brass and iron in an intermediate position corre¬ 
sponding approximately to its coefficient of expansion. Connection was made to the 
galvanometer through a similar wire stretched parallel to the bridge-wire, in order to 
eliminate thermal effects at the sliding contact. All the connecting wires throughout 
the box were accurately paired and compensated, and the thermometer connections 
were made by means of mercury cups at PP' CCk Screw terminals were used in the 
galvanometer and battery circuits only, where changes of resistance are immaterial. 
The ratio coils were adjusted to equality by the method of interchanging, and were 
tested and compensated for equality of temperature-coefficient before being fixed in 
their places in the box. They were not wound up together, as is usual in apparatus 
of this class, but were merely fixed side by side, as it is most important to secure the 
most perfect insulation of the ratio coils for delicate differential work. 
