1895 - 96 .] Dr Beattie on Thermo-electricity in Bismuth. 153 
Mark off the values of — along the perpendicular axis and 
dt 
t along the horizontal axis, we get then a straight line cutting the 
d 
horizontal axis in the point t — — — 
1 2 b 
At this temperature we have the maximum transverse effect : 
thus for Plate I. the maximum is at - 23° C., and for Plate II. at 
- 73° C. — values which agree with the experimental results. 
Let such a line he ADEN", then at C temperature t v the total 
transverse effect will be represented by the area ADCB ; this we 
can resolve into two parts, one represented by BD, it is propor- 
f?E 
tional to — and the absolute temperature L, the other by ADM, 
dt 
which is proportional to the absolute temperature. Similarly, for 
the temperature t 2 we have the total transverse electromotive force 
represented by PL + ELA ; at N, on the other hand, we have no 
component proportional to the transverse power — , and the total 
dt 
transverse component is here a maximum ; at t 3 the component 
proportional to — has in the above figure a negative value, the 
dt 
total transverse effect has decreased and will continue to do so 
until a temperature is reached as far beyond PT as N is distant from 
the origin B, supposing, that is, that E can be expressed by a para- 
bolic curve up to that temperature. 
In analogy with the notation of thermo-electricity, we may call 
dE 
— the transverse power of the conductor ; and we can speak of 
W f/ 
