BETWEEN THE FREEZING AND BOILING-POINTS. 
177 
From 3565 to 3406 each ohm was of the older form, with the wire embedded in 
paraffin wax and made to insert in a water-bath, with long heavy wire connectors for 
the terminals of the Carey-Foster bridge. No. 4086 was the best form to use with 
the bridge, as the method of having the other coils always in paraffin wax is bad, and 
it is never jmssihle to know exactly the true temperature of the coils. The German 
standard was evidently made of manganin wire on account of its very small 
temperature coefficient. Unfortunately coil 4086, to which all of my results were 
referred (‘ B.A. Report,’ 1899), seems to be different to the others by as much as 
6 parts in 10,000. It is difficult to see how it could have been injured in any way 
since it came into the possession of this laboratory, and, as will be seen presently, the 
comparisons of this ohm with both of the specially constructed platinum-silver ohms 
does not indicate any possible change since May, 1898. The cause that has been at 
work to alter its resistance has left it entirely unaffected during the last two years. 
Two tables of comparisons are now given of all the 1-ohm coils. The first set in 
Table VII. was taken by myself and expresses all the ohms, except 3566 and 1214, 
in terms of 4086. For the second set in Table VIII. I am indebted to Mr. Fraser. 
For reducing the values of the ohms to one temperature, temperature coefficients 
were used which were obtained either by myself or Mr. Fraser in duplicate by 
special experiment, and verified repeatedly by the students. All the platinum-silver 
standard ohms were found to have the coefficient + '000254 t°. 
The manganin ohms 40S6 and 1214 were found to have the coefficients '000018 
and -f- '00M022 respectively. 
The different columns of Table VII. are arranged so that the first gives the 
number of the ohm, the second the certified resistance at temperature given in the 
third column, and the fourth column gives the length of bridge-wire multiplied 
by '001020 to reduce to ohms, which represents the difference in resistance between 
each ohm and No. 4086. In the fifth column is given the temperature of the 
different ohms during the comparison, and in the sixth the temperature of 4086. 
The seventh column contains the value of 4086 at the temperature of comparison 
found from the certified value by the temperature coefficient. As all the platinum- 
silver standard ohms were larger than 4086, the eighth column is obtained by adding 
columns 4 and 7. This gives the resistance of each ohm in terms of 4086. In the 
last column, for comparison, I give the value of each ohm in terms of its own 
certificate, and corrected to the temperature of comparison in column 5 by the 
temperature coefficient. The values in the eighth column in terms of 4086 are all 
systematically smaller than the values in the ninth column, whereas they should be 
equal. The observations differ amongst themselves somewhat, but they are as good 
as can be expected from the difficulty of knowing the true temperature of the coils 
embedded in the wax. 
In all these tests the standard ohms were left for several hours near the place 
of test, so that they could assume, as nearly as possible, the temperature of the air. 
VOL. CXCIX.-A. 2 A 
