April 7, 







NATURE 



535 



It is not a little singular that the same lament should 

 have been uttered, with so long an interval between, by 

 two prominent technical men, and it suggests that but 

 little experimental work had been done in the meantime 

 with a view to the measurement of high temperatures. 

 This is, however, far from being the case. A vast amount 

 of work was done by physicists and metallurgists whose 

 chief masters were " indefatigable labour, the closest 

 inspection, and hands that were not afraid of the black- 

 ness of charcoal" ; and their more noteworthy efforts were 

 based on the employment of the air thermometer, in 

 which the expansion of air replaces the expansion of the 

 mercury in the ordinary thermometer, the bulb being of 

 some fire-resisting material. ^ For this purpose, Princep 

 (1827) used a bulb of gold, Pouillet (1836) one of platinum, 

 and finally, Deville and Troost, in a truly splendid series 

 of investigations, adopted bulbs of porcelain, with iodine 

 vapour as the elastic fluid. They ultimately reverted to 

 the use of air. 



You will remember that old mercurial thermometers 

 had much information, supposed to be useful, engraven 

 on their scales, and such statements as " water freezes," 

 "water boils," "blood heat," "fever heat," "summer 

 heat," were considered indispensable. It is by exposure 

 to known temperatures that a thermoscope can be con- 

 verted into a pyrometer for measuring intense heat ; and 

 the air or gas thermometer has, in the hands of Deville 

 and Troost, rendered excellent service by enabling such 

 gradations to be effected. The gas thermometer is not, 

 in itself, a handy appliance, for it requires much sub- 

 sidiary apparatus, and elaborate corrections of various 

 kinds have to be introduced into the numerical data it 

 <iffords ; but it has given many fixed temperatures — such 

 as melting-points and boiling-points of elements, and of 

 ■compounds — which may safely be made use of in graduating 

 pyrometers. For very high temperatures, 900° C. and 

 over, we rely on the excellent work of M. VioUe- on the 

 specific heats of platinum, silver, gold, palladium, and 

 iridium, which have enabled the melting-points of the 

 respective metals to be calculated. 



The determinations of temperatures between 300^ and 

 1000% which are now generally accepted, also rest upon 

 data accumulated by the aid of the air thermometer, 

 which has thus enabled the graduation to be effected of 

 •instruments widely differmg from it, that can be trusted 

 to give rapid and accurate indications in daily use. I 

 can only bring before you two of the many kinds which 

 have been devised ; they are, however, by far the best 

 that are available, and for the determination of tem- 

 peratures up to the melting-point of platinum, leave little 

 to be desired — 



(i) A pyrometer which depends on the increase in the 

 resistance of a heated conductor through which a divided 

 electrical current is passing ; and 



(2) One in which the strength of an electric current, 

 generated by the heating of a thermo-junction, is used as 

 a measure of the heat applied to the thermo-junction. 



The principle of the electrical resistance pyrometer 

 was indicated by Sir William Siemens (" Collected 

 Papers," vol. ii. "Electricity," p. 84, 1889) in a letter 

 addressed to Dr. Tyndall, dated December i860, and the 

 nature of the instrument may be made clear by the ac- 

 companying diagram, Fig. I. A divided current passes 

 from the battery B, to a platinum wire, c, coiled round a 

 clay cylinder, and to a resistance coil, R. At the ordinary 

 temperature the resistance of the platinum coil is balanced 

 by the standard resistance R. If, however, the platinum 

 coil be heated, its resistance will be increased, and, this 

 increase of resistance, which can be measured in various 

 ways, indicates the temperature of the coil c. The coil 

 itself may be adequately protected and exposed to tem- 



' See the excellent bibliography given by C. Barus, Bull. Geological Survey, 

 U.S.A., No. 54, 1889. 



^ Comptcs rendus, vol. Ixxxix. p. 702, 1879; vol. xcii. p. 866, 1881. 



NO. II 7 I, VOL. 45] 



peratures which have been determined by the air thermo- 

 meter ; the deflection of a suitable (differential) galvano- 

 meter, G, will then indicate temperatures directly. For 

 instance, the temperature at which zinc boils has been 

 accurately fixed at 940^ C, and if the coil is heated in the 

 vapour of boiling zinc, the angle through which the 

 galvanometer mirror is deflected marks the temperature 

 of 940" C. 



The Report of a British Association Committee showed 

 in 1874 that the instrument is liable to changes of zero, 

 but Mr. H. L. Callendar has recently (1887) restored 

 confidence in the method which had been shaken by the 

 Committee. He has proved that if sufficiently pure pla- 

 tinum wire be used, and if the wire be carefully annealed 

 and protected from strain and contamination, ^ resistance 

 pyrometers may be made practically free from changes of 

 zero even when used at temperatures as high as 1000'' C. 

 He attributes the changes of zero to which the Siemens 

 pyrometers are liable to the action on the wire of the clay 

 cylinder on which it is wound, and of the iron tube in which 

 it is inclosed. As the result of his experiments he has intro- 

 duced certain modifications, which render the instrument 

 not only trustworthy but very sensitive. He winds the 

 platinum wire on a thin plate of mica, and incloses it in 

 a doubly glazed tube of hard porcelain. He uses the zero 

 method of measuring the resistance ; but for these and 

 other details of manipulation his own very interesting 

 papers must be consulted. I will only add that I have 

 had the pleasure of working with him in the Mint Labora- 

 tory, and I am satisfied that at temperatures about 1000^ 

 the comparative results afforded by his method are accu- 

 rate to the tenth of a degree, a result which would certainly 

 have been deemed impossible a year or two ago.- 



The necessity for working with small volumes of fused 

 metals, into which the tube of Callendar's pyrometer could 

 not be plunged, has led me to prefer to adopt a method 

 that would be classified under the second heading I have 

 given. A very small thermo-junction may, in fact, be 

 employed in such cases. The use of thermo-junctions for 

 measuring high temperatures appears to have been sug- 



' Phil. Trans. Roy. Soc. vol.178 (1887). A. pp. 161-230. and vol. 182 (1891J, 

 A, pp. 119-157; Phil. Maf;. vol. xxxii. July 1891, p. 104, and vol. xxxiii. 

 Feb. 1892, p. 220. 



^ As this statement has been received with some surprise, it may be as 

 well to state briefly how this degree of accuracy and sensitiveness is attained. 

 The resistance-box is compensated for changes of temperature, and changes 

 of resistance in the wires leading to the pyrometer are automatically elimi- 

 nated. The resistance itself is measured by a modification of the welUknoivn 

 Carey-Foster method. The balancing resistance of the Wheatstone bridge 

 employed, is composed partly of resistance coils and partly of a bridge-wire 

 along which a contact key slides. The resistance of a centimetre of this 

 wire is made to correspond to the increase of resistance of the pyrometer 

 produced by a rise of 1' C. The galvanometer can easily be made sen>itive 

 to one-hundreth of a centimetre of this bridge-wire, so that one-tenth of a 

 centimetre, which corresponds to one-lenth of a degree, can, of course, be 

 measured with certainty. 



