1890-91.] Prof. C. G. Knott on Cobalt and Bismuth. 313 
we obtain for the coefficients A and B the following values : — 
A 
B.10 2 . 
Lead, . . 
0 
0 
Palladium, . 
- 6-18 
- 3-55 
Cobalt, 
-13-18 
-13-9 
Bismuth, 
-92-2 
- 6-4 
According to the numbers deduced by Fleeming Jenkin from 
Matthiesen’s experiments, bismuth lies four times further from 
lead than does cobalt. Here we have it seven times. Professor 
Tait’s electrolytically-deposited cobalt lies four and a half times 
further from lead than does palladium. Here we have it a little 
over two times. According to Becquerel’s numbers, given at the 
end of the English translation of Mascart and Joubert’s Electricity 
and Magnetism , the ratio at 50° C. of the thermoelectric powers of 
palladium and bismuth relatively to lead is as 7 to 40. Here we 
have it 1 to 16. 
These discrepancies are not surprising. We know* how variable 
are the thermoelectric properties of stable alloys intended to have 
the same composition, and how a very slight change in composition 
may be accompanied by a very large change in thermoelectric 
quality. The present experiments must therefore be judged of 
altogether on their own merits. Now, a simple comparison shows 
that Professor Tait’s electrolytically deposited cobalt will fit in to 
the region between lead and bismuth very much as Matthiesen’s 
cobalt fits in to his own series. Thus the cobalt investigated here 
seems to differ from the other specimens in much the same way. 
The new cobalt, indeed, has its diagram line at ordinary atmos- 
pheric temperatures above the line of Tait’s nickel, for which 
A=— 21 ‘8. This unexpected result was at once tested. A rough 
experiment was made with a nickel cobalt couple, and a neutral 
point was obtained at a moderately low temperature. The cobalt 
line, therefore, begins above the nickel line, but because of its 
greater downward inclination gets below it at temperatures above 
100° C. 
* See the paper by MacGregor and myself, already referred to ; also my paper 
on “ The Electrical Properties of Hydrogenised Palladium” {Trans. Roy. Soc. 
Edin vol. xxxiii., 1886). 
