BETWEEN THE FREEZING AND BOILING-POINTS. 
175 
coils, water at different temperatures was run through the circulating tube so as to 
change the temperature of the oil. 
Up to the time of writing, I have been unable to compare these ohms directly with 
official standards. It will, therefore, be necessary for me to describe a series of 
comparisons of these ohms with eleven certified ohms in the possession of this 
laboratory, together with a standard 1-ohm coil from the German Beichsanstalt, sent 
us for comparison by the Physical Department of the Massachusetts Institute of 
Technology. By means of these comparisons, we may possibly arrive at a result for 
the value of B somewhere near the truth. 
Throughout the present work I have used only one of the certified standards 
(No. 4086) to check the constancy of the two platinum-silver ohms, as it was of a 
better and more convenient form to use on the bridge than the others and had a 
much smaller temperature coefficient. 
Before describing the tests I will briefly review the method of comparing the ohms 
and the method of finding the value of the bridge-wire used on the Carey-Foster 
bridge throughout these tests. The 1-ohm pair of coils supplied by the makers of 
the bridge were used for the ratio coils, and a bridge-wire having a resistance of 
about ‘002 ohm per centim. was used. The bridge-wire was just 10 centims. long, 
with a scale graduated into half centimetres and millimetres. A lens was also 
supplied for reading the position of the balance point. The galvanometer for 
obtaining the balance point was a very sensitive 9-ohm Thomson reflecting galvano¬ 
meter, which was used for the thermometer work. It had a telescope and scale, and 
was sensitive to 50 scale-divisions for 1 millim. of bridge scale, which, of course, was 
far more sensitive than was required, or even quite convenient to work with. The 
current supplied to the coils was from one accumulator through 20 ohms external 
circuit. By simply lifting the commutator from the mercury cups on the bridge and 
revolving it through half a revolution, the connections could be made so as to reverse 
the position of the two resistances relative to the ratio coils. If P and Q are the 
ratio coils, B and S the resistances to be compared, then, when the current is reversed 
in B and S, but not in P and Q, we have, B/S — S/B = p(d l — d. 2 ), independent of 
P and Q. 
Here d l and d 2 are the readings of the balance points on the bridge-wire, and p a 
constant to reduce to ohms. 
Let B/S = (1 -j- r), where r is a small quantity ; then S/B = (1 — r) and 
B/S — S/B = 2 r — p [d Y — d. 2 ). 
To find p, S may be changed to S } by shunting with a known large resistance, say 
100 ohms. 
We have then, if B/Sj = 1 + r + dr, S x /B = 1 — r — dr, and reversing the 
current as before, 
