
738 PROFESSOR TAIT ON THERMAL AND ELECTRIC CONDUCTIVITY. 
for a few seconds. Nothing but absolutely perfect adjustment could have 
made this possible when using the other methods. 
In my experiments the most unfavourable case gave 
g > 30,000 g, 
so that g and p are practically equal when there is no deflection. 
§ 21. I employed the bar C of inferior copper in all these comparative 
experiments. But the conductivity of the German silver bar is so much less 
that I could employ only 10 inches of it, as against 7 feet of the inferior copper. 
I therefore endeavoured, by experiments on short lengths of the two copper 
bars, to find approximately the correction required, in consequence mainly of 
the breadth of my contact pieces, very slightly, perhaps, in consequence of the 
great section of the bars. Here are the results in inches,— 
C. Crown Ratio. Mean of 
A'B. AB. Uncorrected. Corrected. Corrected 
49°66 85°7 1:726 1°732 
48°47 *83'°5 1:723 1729 
A147 71:25 L718 1-725 
18°8 32:'2 1713 1:728 se 
16°65 *28'46 1°709 1731 
9:25 a7 1697 1729 
[Note——In the experiments marked with an asterisk the arrangement was altered by 
shifting the crown bar to the other coil of the galvanometer. The agreement of these with the 
others is a good guarantee of the accuracy of the adjustments, and the goodness of the method 
is seen in the fact, that no observation deviates so much as } per cent. from the mean. This isa 
striking verification of what was said above about the small effect of the holes bored in the 
bars, for the nippers were placed quite at random in the various experiments. ] 
The contact pieces were nippers of polished copper, 0:42 inch broad, which 
were easily slipped along the bars, and were tightened on them by screw clamps 
when the final adjustment was nearly arrived at. 
It appears from the column of corrected ratios above, that it is only neces- 
sary to subtract 0°4 inch (the sum of the half breadths of the nippers, the wires 
being soldered to them symmetrically) from each of the measured distances to 
secure almost perfect uniformity. Thus I was led to see that the influence of 
the section of the copper bars is almost undiscoverable by such experiments. 
§ 22. For the Forbes iron bar the following results were obtained (but with 
the correction 0°2 inch) :— 
Ratio. 
Fe, C. Uncorrected. Corrected. 
20 74:3 3°715 3°74 
10 ones 3°73 3°79 
5 18°4 3°68 3°79 

