of the Coefficient of Expansion of Pure Nickel. 629 
the whole length of the wire was 60 ems., the distance of ¢ 
4 or ¢’ from the ends of the specimen was great enough to 
| ensure the existence of a uniform temperature over the 
experimental portion of the nickel. 
During an experiment, the top of the groove was closed by 
a strip of wood containin g two holes cover ed with mica thr ough 
which the scratches could be viewed. The fine platinum 
leads 7,7’ were brought out through small holes hand soldered 
to binding-screws on the base-board. 
| Two separate reading-microscopes of the travelling tyne were 
. used, and it was therefore necessary to know accurately the dis- 
| tance between the two scratches at a particular temperature. 
This was effected by making a comparison of the nickel wire 
with a standard scale, at the temperature of the laboratory. 
The microscopes were read to the thousandth of a millimetre. 
oS 
Measurement of the Temperature. 
The resistance of the portion ¢t’ of the nickel was deter- 
mined by comparing the potential-difference between ¢ and t’ 
Fig. 2. 
La 
Kreatine CIRCUIT 
772. 
NeAASON t’ 2 
Sess a 2 a2) Re S60 00602 

with the potential-difference between two points ona standard 
resistance which was in series with the nickel and carried the 

