706 PROFESSOR THOMSON ON THE ELECTRO-DYNAMIC QUALITIES OF METALS. 
Of the two results for the neutral point between silver and gold, only the last can 
be reconciled with the indications derived from the previous results as to the relative 
positions of these and the other metals tried along with them ; and accordingly — 5°7 
has been taken as the neutral point of gold and silver in the thermo-electric diagram 
given below (§ 101). The first result, 70°*8, was found as the mean of several deter- 
minations, from none of which it differed by more than 0°7, and the discrepance can 
scarcely be attributed to errors of observation, but is probably due to slight differ- 
ences in the specimens of gold and silver used in the different experiments. That 
very slight chemical differences in specimens of gold and silver wire may make great 
alterations in the temperature at which they become thermo-electrically neutral 
to one another, is readily understood by glancing at the diagram given below 101), 
and observing how close together the lines for gold and silver lie. 
99. The question. Does the difference between the specific heat of electricity in two 
metals vary with the temperature maybe answered by experiments showing the 
law according to which the means of widely different temperatures of the junctions 
giving no electro-motive force deviate from the true neutral point, which is the mean 
of any infinitely small difference of temperature giving no electro-motive force. 
I have not yet obtained indications of such a deviation in any case, having been 
prevented from prosecuting the inquiry by delays in the construction of a suitable 
air-thermometer. The examination I have been able to give the subject is only suffi- 
cient to show that the arithmetical mean of the temperatures of the two junctions 
giving no current, is probably in general within a degree of the true neutral point, 
when the difference between those temperatures does not exceed 100° Cent. 
The following summary of a series of experiments made on two consecutive days 
may serve as an example of the degree of consistence of the results obtained by 
the method which has been explained, in a case in which the two metals deviate 
rapidly from one another above and below their neutral point. 
Sheet-lead Electrodes ; Pi mean. 
Determinations by Mr. C. A. Smith, May 17 & 18, 1854. 
Difference 
Half sum of mercurial thermometer temperatures 
of temperatures. 
giving no current. 
m 
12f 
71 
7li 
7ii 
1211 
121f 
121f 
Mean 
^121°-5 
70 
122 
1851 
120f1 
1581 
143 
133 
121|- 
1211 
1211 
Mean 
^121°-4 
1251 
121L 
681 
122 ] 
Mean 
50 
123 J 
122°-5 
Mean of temperatures by mercm'ial thermometer 
giving no current. 
Differences from 50° to 77° 
122°* 15 
Differences from 125° to 185°... 
121°-4 
* See “Dynamical Theory of Heat,” § 115, equations (15) and (17). 
