572 
MESSRS. R. T. GLAZE BROOK AND S. SKINNER ON THE 
It was, therefore, constructed of a strip of platinoid about 1 cm. wide and '05 cm. 
thick. This was wound on an open frame about 50 cms. long and 50 cms. wide 
by 6 cms. deep. The frame was of wood covered with strips of ebonite which 
insulated the metal from the wood. 
The whole was immersed horizontally in a bath of paraffin, containing about 60 litres 
of oil. A stirrer, in the form of a T-shaped piece of wood, passed under the coil 
between it and the bottom of the bath, and by means of this the liquid could be kept 
in a state of agitation. This stirrer was worked frequently during an experiment. A 
delicate thermometer, placed with its bulb very close to the strip, showed an immediate 
slight rise in temperature, one or two tenths of a degree, if the oil was not stirred. If 
the stirring w 7 as maintained the rise during an experiment was usually only slight. 
The ends of the platinoid strip were soldered to thick copper rods, which dipped 
into two mercury cups insulated on a paraffin block. 
By means of copper rods, these cups could be put into communication with the 
B.A. bridge, and the resistance of the strip measured in situ in terms of the standard 
coil “ Flat.” This was done without shifting the coil both before and after each experi¬ 
ment ; the two results never differed by more than '0007 B.A. unit, and the mean has 
been taken in calculating the result. The resistance of the strip at 17° C. was 
1-0011 B.A. unit. 
For some of the experiments this resistance might, with advantage, have been 
higher, though the fact that it was nearly equal to the standard secured accuracy in 
the comparison with it, and it was thought desirable, from this point of view, to 
use the same resistance throughout. 
In order to reduce the effects due to variations in the resistance of the voltameters, 
a resistance of thick platinoid wire, usually some twelve or fifteen ohms in amount, 
was included in the circuit. This was varied until the current was of the right 
amount. 
§ 6. The Balance. 
A short beam balance by Oertling was used for the weighings. This and the 
weights were the same as those employed in the determination of the specific 
resistance of mercury by Glazebrook and Fitzpatrick (‘Phil. Trans.,’ A., 1888). 
The weights had been compared with the standard, and the correction found to be 
inappreciable. Some of the weighings were done in both pans in order to determine 
the ratio of the arms of the beam. This was found to be 1'00002. The correction 
to the weight in vacuo varied between P000030 and 1 — '000013, according to the 
number of platinum weights used ; both these are too small to affect our results, and 
have not been introduced into the calculations. 
