CONTINUOUS ELECTRIC CALORIMETRY. 
79 
current-standards at McGill College are, therefore, deserving of greater weight than 
might be supposed at first sight. 
Taking the mean of the observations of Barnes and F baser as fairly applying to 
the mean temperature of the Heat Laboratory at the time, we obtain 
No. 4086 at 18 0, 4 C. = 1'00042 ohms in terms of ten McGill standards . (10). 
Taking the long-period coefficient '000047 as appropriate for deducing the value 
of No. 4086 at the same temperature, from Glazebrook’s observations we find 
No. 4086 at 18°'4 C. = 1'00046 ohms in terms of original B.A. standards (11). 
This agreement is as close as could reasonably be expected, and we may conclude 
that the value of No. 4086 is known from the comparisons to at least 1 part 
in 20,000. Adopting the latter value for 4086, we obtain finally for the value of the 
mica current-standards, 
Mica current-standards in parallel at 18°'36 C. = '50037 ohm . . (12). 
Assuming the temperature coefficient '000248 for the mica current-standards, we 
find for the values at 5°, 10°, 15°, and 20°, respectively, 
Values of Mica Current-standards in Parallel. 
Temp. Cent. 5° 10° 15° 20° 
Value. '49871 '49933 '49995 '50057 
These values are in practical agreement with those obtained by Barnes from the 
McGill comparisons alone, but he does not make any attempt to explain the apparent 
discrepancies. They are probably correct in terms of the International Standard 
Ohm to 1 part in 20,000, but the uncertainty of the absolute value of the 
International Ohm itself may amount to 2 or 3 parts in 10,000. If the Board 
of Trade Ohm really exceeds the C.G.S. Ohm by nearly 3 parts in 10,000, according 
to the result of the measurements of Professors Jones and Ayrton with the McGill 
College apparatus, the electrical watts supplied in the calorimeter have been over¬ 
estimated to that extent, so that the absolute value of the equivalent of the calorie 
in terms of electrical energy as deduced from this investigation would require to 
be reduced by about 1 part in 4,000. 
It should be observed that the absolute values of the current-standards are of 
comparatively little importance for the present investigation, as they do not at all 
affect the question of the variation of the specific heat. I have thought it worth 
while, however, to discuss the question somewhat fully for two reasons. In the first 
place, the comparisons illustrate a possible objection to the use of manganin coils for 
