74 
PROFESSOR HUGH L. CALLENDAR ON 
observation with stirrer and cooler (Barnes, p. 174), by comparison with a 
manganin standard ohm, No. 4086, which was kept at a steady temperature. The 
value of the temperature coefficient of the mica current-standards is, therefore, 
independent of any assumption with regard to the temperature coefficients of the 
other standards employed. The low value '000248 of the coefficient found is explained 
by the perfect annealing of the wire. 
The absolute values of the mica current-standards were referred to the manganin 
standard No. 4086. The comparisons of this coil with the mica standards, which 
extended over two years, are given in Table IX. (Barnes), and do not show any 
relative changes of importance. The observations were taken on twelve different 
dates at temperatures varying from 15° to 21° C, the mean temperature of all the 
comparisons being 18°'3 6 C. 
The temperature coefficient of the manganin standard No. 4086 was assumed to be 
•000018 in the reduction of these observations. This value was determined by 
observing the change of resistance due to changes of temperature of the oil-bath in 
which the coil was immersed, the changes of temperature being effected at the rate of 
L0° C. in two or three hours. The results in Table IX. (Barnes) are reduced to a 
temperature of 20° C. If, however, we reduce the results to 18 0, 4 C., the mean 
temperature of the comparisons, we shall be independent of any assumption with 
regard to the temperature coefficients. We thus obtain the values :— 
Mica current-standard, No. I. at 18°*27 = No. 4086 -f- ’00091 at 18°’4C. 
„ II. „ 18 0, 46 = „ „ - -00035 „ „ 
No. I. and No. II. in parallel at 18 0, 36 C. = Half No. 4086 -p ‘00014 . . (1). 
The manganin standard No. 4086 had a Cambridge certificate No. 378 dated 1893, 
which stated its value to be ’99978 of a true ohm at 1 5’9° C. The preliminary results 
for the value of the calorie in terms of the electrical units, communicated to the 
British Association in September, 1899, were expressed in terms of this certificated 
value as correct. A comparison was subsequently made at McGill College of No. 4086 
with a set of 10 platinum-silver standard ohms of the old pattern, the coils of which 
were embedded in paraffin wax. These 10 coils all possessed Cambridge certificates, 
and their relative values when tested, agreed fairly with those obtained several years 
previously, showing no certain indication of change. The mean of the certificates of 
these 10 coils gives the value :— 
Mean of 10 Pt-Ag standard ohms = ’99962 at 16°’5 C. ... (2). 
The details of the comparison of No. 4086 by Dr. Barnes # with 9 of these 
* The standard No. 3566 omitted by Barnes happens to be so near the mean of the 10 as to make no 
material difference in the result. 
