128 
Proceedings of Royal Society of Edinburgh, [jan. 16 , 
its thermoelectric properties for me by the indirect method. The 
following is their account of the work : — 
“ The experiment was carried out by heating two junctions, of 
PdCo and FeCu, joined together at the hot end, to a red heat by 
means of a cylinder of red hot iron ; the other ends being immersed 
in separate quantities of cold water (same temperature as the atmo- 
sphere). The rise in temperature of the cold junctions as the expe- 
riment went on was extremely small, A correction for this made 
no difference on points on the curve, and was therefore neglected. 
The currents from the two junctions were made to pass through the 
galvanometer in opposite directions ; and the deflections were alter- 
nately observed, from the PdCo and FeCu circuits, at equal intervals 
of time. At the higher temperatures readings were noted down as 
soon as the spot of light on the scale came to rest ; and observations 
were corrected by taking the mean of two consecutive readings of the 
PdCo as corresponding to the deflection of the FeCu between them. 
“ The next experiment was to find the thermometric value of the 
PdCo deflections. The same junctions were now heated, by means 
of an oil-bath, and the temperature observed by a thermometer, up to 
230° C., in the case of Battelli’s iron up to 250° C. The deflections 
were observed every 10° from 30° upwards j observations being 
also taken as the temperature fell (the duplicate readings coincided 
very closely). In order that the junction and thermometer should 
have the same temperature, when the readings were taken, heat was 
applied by means of a Bunsen burner until the mercury rose 5, and 
in some cases 6, 7, 8 and 9 degrees (according as the temperature 
became higher), above the temperature of last reading ; it was then 
removed, the oil during the whole time being vigorously stirred, the 
mercury continued to rise slowly, and finally reached the proper 
degree for the observation. The curve for PdCo was drawn, 
and constants were calculated, by means of which points on the. 
curve at higher temperatures were found, the curve in this way 
being extended to about 600° C. By the equation to the PdCo 
parabola points were verified, and found to coincide, very closely, 
in some cases exactly, with those observed. 
“ The PdCo curve was tested by lessening the resistance in the 
PdCo circuit, and the new curve thus found was compared in detail 
with the previous one. 
