1887 .] Mr A. Campbell on Measurement of Peltier Effect. 391 
(low temperature to high) calculated from neutral point 144° C. 
In Nos. 3 and 4 the temperature was kept uniform by means of a 
double steam-jacket surrounding a small copper box in which the 
junctions and thermopile were packed. By this means the tempera- 
ture could be kept very steady and uniform. 
Zinc and Iron. 
The metals here used were ordinary sheet zinc and thin tinplate. 
In order that the junctions might stand a temperature above 200° C. , 
they were soldered with a suitable alloy of zinc and tin. Thin 
strips were cut from the same specimens, and the neutral point was 
found (by heating their junction in oil) to be 1 96°*7 C. The 
directly observed Peltier effect w T as found to vanish about 204° C. 
The temperature, however, was falling slowly at the time, which 
would account for this disagreement. Table III. gives a measure- 
ment taken just before the temperature had fallen to 204°. The 
last column gives the ratio calculated from 204° as neutral point : — 
Table III. 
Low Temperature. 
High Temperature. 
Ratio Observed. 
Ratio Calculated 
from 204° C. 
22°'5 
215 ° 
-9738 
- 9-80 
Nickel and German Silver . 
The peculiar form of the nickel line between 150° and 300° (X 
made it interesting to find whether the Peltier effect between nickel 
and any other metal (or alloy) agrees with the theory between these 
temperatures. German silver was chosen as being an alloy whose 
Peltier effect with nickel ought to vary in a striking manner. 
According to the thermoelectric diagram, their Peltier effect divided 
by the absolute temperature should remain constant till at least 
150° C., and then decrease uniformly till it vanishes at the neutral 
point ; beyond this it should change sign and increase till about 
300° C. (above which it should probably remain constant for some 
distance). 
The usual form of apparatus was used, but in this case the 
