1887.] Mr A. Campbell on Measurement of Peltier Effect. 403 
the Peltier effect in FeArg varies as the absolute temperature, the 
result in equation (1) at once establishes the usual theory. In any 
case, equation (1) perfectly agrees with the received theory. 
Experiments were also made with the four thermopiles inter 
laced in the order FeArg, FeArg, FeZn, FeZn, the alternate 
ones being connected. With this difference, everything was done 
just as in the last described experiments. In Table XII. are the re- 
sults. The current was passed sometimes through the one pair of 
piles (FeArg and FeZn) and sometimes through the other pair. 
These two ways are marked P and Q in the table. The last column 
gives the values of the square root of 5- corrected for change in 
\jt 
specific heat, by T2 % per degree C. [Assumption (1st).] This is 
not strictly accurate. 
Table XII. 
Temperature. 
e. 
Thermopiles 
Deflection. 
D. 
D 
C t 
/- 
V c t 
(corrected). 
p 
7°'l C. 
3191 
1635 
1283 
Q 
7° -3 
3204 
1606 
1272 
P 
70.7 
2504 
1568 
1259 
Q 
7° *8 
2447 
1568 
1257 
Q 
91° -7 
3811 
1585 
1326 
P 
93° -8 
3972 
1571 
1322 
Q 
94°-8 
3788 
1564 
1320 
P 
98°T 
4846 
1555 
1318 
Q 
100°-5 
6098 
1573 
1327 
P 
151°*3 
4932 
1601 
1377 
Q 
153° -0 
5063 
1581 
1369 
P 
153° -9 
5120 
1603 
1378 
It will be found that the values of ^/5L( C orrected) lie very nearly 
on a straight line ; we may take the coincidence as exact (assump- 
tion 2nd), i.e ., that 
= ( a + Ptf ( 3 )" 
Xow, supposing (assumption 3rd) that the four sets of thermopile 
junctions have, all of them, the same heat capacity (which is true 
to within about 1 %), we have, following the notation used above — 
