234 
Proceedings of the Royal Society of Edinburgh. [Sess. 
The slider was also highly insulated, and the index line could be read to 
one-fifth of a millimetre ; consecutive and independent readings agreed to 
half a millimetre. 
The Resistance Coils. — It was found convenient to work with only 
one coil, and the coil selected as fitting the needs of the cell was one 
marked 400 ohms in a box of standard coils made by Groves. These had 
been recently standardised by Mr Manley, and the value of the coil used 
was 396*224 legal ohms at 15° C. The temperature of the coils was kept as 
constant as possible by placing the box in a larger box, the space between 
the two being packed with a non-conducting material. The temperature 
varied, however, by several degrees, and the necessary corrections were 
made. The temperature coefficient of the coils was ‘00015 per ohm per 
degree. 
The Leads and Switchboard. — The switchboard consisted of eight highly 
insulated mercury cups, and the connecting leads used for joining the coils 
with the bridge had a resistance of 0*006 ohms, while the leads to the cell 
had a resistance of 0*0669 ohms. The variation in the temperature of the 
room would not affect the resistance of these leads within the limits of 
experimental error. 
The Current and Telephone. — The current was derived from a single 
large bichromate cell, and the alternating current from a small Kohlers 
induction coil. The telephone was a French one. 
The Cell. — This was made from a U-tube whose diameter at the bend 
was about 6 mm. and much wider at the ends. The electrodes were 
about 10 mm. in diameter and were just above the constricted portions 
of the tube, coming close to the glass but not touching it. One was fixed 
in position and never moved ; the second was sealed into a piece of glass 
tubing which passed through a sound selected cork. The latter just fitted 
one limb of the cell. Into the cork three pins were firmly fixed, so that 
each of these rested on the rim of the cell. The cork entered the cell far 
enough and tightly enough for no play to be possible, and the cell was 
marked so that each pin came into exactly the same position when the cork 
was removed and replaced. This cell was found to answer admirably. 
Comparison tests were made with sea- water No. 1 at intervals during the 
measurements of the other waters, and the constant of the cell was found 
not to change at all. The latter was determined by means of a solution of 
potassium chloride of normal strength at 25° C. and was found to be 22*8714. 
The electrodes were platinised in the usual manner. 
Thermometers. — These read directly to 0*1, and each of those used were 
compared with a Kew standard and the correction applied. 
