DOL Mr. 8. ©. Laws on the Thomson Hffect 
deflexion of 1 division of a scale reflected in a telescope for a 
current of 10-7 amperes, whilst a difference of temperature 
between the junctions of 1° C. produced a current of 
2°2x 10-5 amperes in the galvanometer circuit, so that a 
deflexion of 1 division corresponded to a change of tempera- 
ture of =}, degree ©. To measure the current in the 
galvanometer a null method was employed, the current 
produced by the E.M.F’. of the couple bemg compensated by 
adjusting a current in the other set of coils of the galvano- 
meter. This compensating current was obtained from a single 
secondary cell joined up with a set of known resistances, and 
by means of a reversing key could be sent in either direction 
through the galvanometer. rom a point R, (fig. 1) of the 
main circuit a current was taken off to pass through the 
galvanometer and high resistance R;. The required magni- 
tude of this compensating current was then obtained by first 
choosing a suitable value for R, (106 ohms say), then 
adjusting R, (5000 ohms), and finally R; (1750 ohms) ; the 
current is calculated from the E.M.F. of the cell and 
known resistances. 
In order to decide accurately on the value of the compen- 
sating current necessary to annul the effect of the current in 
the junction circuit, it is necessary that the two circuits 
should be completed at the same instant. This was effected 
by placing in each circuit a key consisting of a vertical 
copper rod dropping into a mercury cup. The two rods, 
which were insulated from each other, were connected to the 
same movable upright, worked by a string over a pulley, and 
their lengths were adjusted so that they touched the surface 
of the mercury in the cups at the same instant. In this way 
the compensating current could be adjusted so that the 
galvanometer gave no deflexion when the upright was 
released and the circuit completed. 
This value of the compensating current is then proportional 
to the difference of temperature between the points of the 
rods to which the junctions are attached. 
IV. Practice of Method. 
The method of carrying out an experiment is then as 
follows :—The hot bath is filled with boiling water, which is 
kept boiling by means of a small gas-flame underneath ; the 
cold bath is filled with melting ice, the ice being prevented 
from coming into contact with the rods by means of wire 
gauze caps surrounding their free ends. 
When the temperature of the rods has become steady 
throughout, a constant current—3 or 4 ‘amperes—is passed 
jlo 
