CONTINUOUS ELECTRIC CALORIMETRY. 
89 
be accurately at the same mean temperature. The necessary changes of connection 
were effected by means of mercury cups and thick copper connectors, similar to those 
employed for standard resistance coils. The mercury cups were made by boring 
suitable holes in a flat plate of ebonite, to the under side of which thick copper plates 
were screwed, the joint being made mercury tight with a thin sheet of rubber. 
The apparatus above referred to was first used for some determinations of the 
linear expansion of standard yards at the Standards’ Office in 1887. It was subse¬ 
quently employed by Dr. A. S. Lea, F.R.S., and Dr. W. H. Gaskell, F.R.S., and 
later by Dr. Rolleston, in some physiological experiments on the heat produced in 
muscle and nerve by electrical and other stimuli. For this purpose a very delicate 
pair of differential thermometers were constructed of ‘001-inch wire wound on mica, 
weighing a few milligrammes each, and sensitive to the ten-thousandth of a degree C. 
One of these thermometers was described and figured in the specification (No. 14,509, 
1887). This apparatus is still in the possession of the Physiological Laboratory, 
Cambridge. 
The chief defect of the original form of apparatus was the uncertainty of the 
temperature correction of the platinum-silver bridge-wire owing to its length. In 
making a new form of apparatus in October, 1890, bridge-wires of manganin 
were employed, annealed at a red heat in coal gas. The pair of bridge-wires could be 
very accurately calibrated throughout their length by the Carey-Foster method, but 
owing to the trouble of determining and applying the bridge-wire correction, it was 
eventually decided to use a bridge-wire of low resistance in conjunction with a larger 
number of resistance coils. It also proved to be unnecessary in practice to make all 
the coils interchangeable in pairs, provided that the ratio coils were tested for equality 
of temperature-coefficient. In this case, it was sufficient to calibrate the bridge-wire 
and resistances by a method of substitution, which was much simpler than the 
Carey-Foster method. Apparatus constructed on this principle was described in the 
‘ Phil. Mag.,’ July, 1891, and figured in the patent specification (No. 5342, 1891). As 
the resistance coils were no longer required to be interchanged, they were permanently 
connected to the copper plates in a single box instead of being connected to copper 
rods in separate boxes like standard coils. The mercury cups, however, were still 
retained, in preference to plugs for short circuiting the resistance coils in accurate 
work, and were constructed precisely as originally described. The simplification 
consisted in connecting the resistance coils permanently in series, and using simple 
bridges of thick copper connected in pairs for short-circuiting each resistance coil 
and its compensator simultaneously. 
The particular resistance box employed in this investigation is shown in the 
accompanying fig. 2. It was made to my designs by the Instrument Company, 
Cambridge, at the beginning of 1893, but I had personally to undertake the delicate 
work of compensating and adjusting the resistance coils. It contained 9 resistance 
coils, A, B, C, D, &c., on the binary scale, ranging from 25GO to 10 units, constructed 
VOL. CXCIX.—A. 
N 
