Royal Society, 151 



In measuring temperatures not exceeding 100° Cent., the instru- 

 ment is so arranged that two similar coils are connected by a light 

 cable containing three insulated wires. One .of these coils, "the 

 thermometer-coil," being carefully protected against moisture, may 

 be lowered into the sea, or buried in the ground, or fixed at any ele- 

 vated or inaccessible place whose temperature has to be recorded 

 from time to time ; while the other, or " comparison-coil," is plunged 

 into a test-bath, whose temperature is raised or lowered by the ad- 

 dition of hot or cold water, or of refrigerated solutions,* until an 

 electrical balance is established between the resistances of the two 

 coils, as indicated by a galvanoscope, or by a differential voltameter, 

 described in the third part of the paper, which balance implies an 

 identity of temperature at the two coils. The temperature of the 

 test-solution is thereupon measured by means of a] delicate mercury 

 thermometer, which at the same time tells the temperature at the 

 distant place. 



By another arrangement the comparison-coil is dispensed with, and 

 the resistance of the thermometer-coil, which is a known quantity at 

 zero temperature, is measured by a differential voltameter, which 

 forms the subject of the third part of the paper ; and the temperature 

 corresponding to the indications of the instrument is found in a table, 

 prepared for this purpose, in order to save all calculation. 



In measuring furnace-temperatures the platinum-wire constituting 

 the pyrometer is wound upon a small cylinder of porcelain con- 

 tained in a closed tube of iron or platinum, which is exposed to the 

 heat to be measured. If the heat does not exceed a full red heat, 

 or, say, 1000° Cent., the protected wire may be left permanently in 

 the stove or furnace whose temperature has to be recorded from 

 time to time ; but in measuring temperatures exceeding 1000° Cent., 

 the tube is only exposed during a measured interval of, say, three 

 minutes, to the heat, which time suffices for the thin protecting casing 

 and the wire immediately exposed to its heated sides to acquire within 

 a determinable limit the temperature to be measured, but is not suffi- 

 cient to soften the porcelain cylinder upon which the wire is wound. 

 In this way temperatures exceeding the welding-point of iron, and 

 approaching the melting-point of platinum, can be measured by the 

 same instrument by which slight variations at ordinary temperatures 

 are told. A thermometric scale is thus obtained embracing without 

 a break the entire range. 



The leading wires between the thermometric coil and the measuring 

 instrument (which may be, under certain circumstances, several miles 

 in length) would exercise a considerable disturbing influence if this 

 were not eliminated by means of the third leading wire before men- 

 tioned, which is common to both branches of the measuring instru- 

 ment. 



Another source of error in the electrical pyrometer would arise 

 through the porcelain cylinder upon which the wire is wound beco- 

 ming conductive at very elevated temperatures ; but it is shown that 

 the error arising from this source is not of serious import. 



The third part of the paper is descriptive of an instrument fqr 



