September 30, 1897] 



NATURE 



531 



metals with particular reference both to iridium and to rhodium. 

 Recently Mylius and Forster at the Reichsanstalt further 

 contributed to platinum metallurgy. But in view of the toils 

 in which the whole subject of high temperature measurement 

 languished in Becquerel's day, his results were not sufficient to 

 remove the discredit which Regnault had thrown upon thermo- 

 electric pyrometry. And so it happened that the return to the 

 method in recent date was of the nature of a resuscitation. 



It is amusing to note, as we pass on, the pranks of custom as 

 it bore down upon pyrometry. Following Deville and Troost, 

 every worker (I might mention at least five) felt in duty bound 

 to redetermine the boiling point of zinc— rather a difficult feat 

 in its way. Thus we find boiling zinc inseparably associated 

 with the destiny of the calibrated thermo-couple. Le Chatelier 

 broke this law of fateful sequence by ignoring the need of 

 calibration at the outset, and then using the couple so dignified 

 to determine the melting points of silver, gold, palladium and 

 platinum. But these are Violle's melting points. Hence the 

 pyrometric feature of Le Chatelier's platinum-rhodium couple 

 was in its inception due to Violle. 



Meanwhile, accompanying the geologic inquiries of Clarence 

 King, an extensive series of pyrometric investigations which had 

 been in progress in the United States since 1882 were completed 

 (1887). These contained a full examination of divers efficient 

 methods of pyrometry and a study of the porcelain air-thermo- 

 meter with particular reference to the calibration of thermo- 

 couples. In the course of this work the admirable pyrometric 

 qualities of the platinum-iridium alloy were exhibited by detailed 

 and direct comparisons with the air thermometer. It was shown 

 that the calibration could be made permanent by referring the ' 

 thermo-electromotive forces to a Clark's cell ; that the character | 

 of their variation with temperature is uniformly regular, and that i 

 the thermal sensitiveness of the couples increases as the higher 

 degrees of red and white heat are approached. Finally it was 

 pointed out that couples destroyed by silicate corrosion, or in 1 

 similar ways, could be restored by fusing over again on the lime i 

 hearth with merely negligible changes of constants. Elsewhere, j 

 LeChatelier's clever combination of the platinum-rhodium couple ' 

 with the D'Arsonval galvanometer, then a comparatively new \ 

 instrument in the laboratory, secured immediate favour. Prof j 

 Roberts-Austen, ever on the watch to waft good things across i 

 the channel from Gaul into Albion, hailed the new-comer with i 

 no uncertain sound. Some time after, the platinum-rhodium j 

 couple entered Germany, and was there definitely calibrated 

 (1892) for the first time, as already stated, at the Reichsanstalt. 1 



Of the three available couples, palladium, platinum-rhodium 

 and platinum-iridium, the former is excluded from competition I 

 by reason of its low fusibility. Between platinum-iridium and i 

 platinum-rhodium, the latter has been more extensively adver- ! 

 tised but is otherwise inferior to the older platinum-iridium alloy. | 

 In other words, platinum-iridium, when suitably alloyed, can be [ 

 made more sensitive than platinum-rhodium in the ratio icx) to | 

 76. Beyond this the alloys are much alike ; both are tenacious, j 

 resilient, refractory metals, and their thermo-electric forces under I 

 like conditions of temperature show a constant ratio even at j 

 extreme white heats. The thermo-electric activity of these two j 

 alloys is exceedingly remarkable. Among over fifty different 

 platinum alloys examined no similarly sensitive combinations | 

 were found. Moreover, whereas platinum alloys of extremely 

 large electrical resistance are not unusual, such metals are not 

 to be distinguished thermo-electrically. 



To conclude : the small dimensions of the sensitive point of 

 the thermo-couple, the independence of the intermediate temper- 

 atures between the junctions (apart from small corrections due 

 to the Thomson effect), and therefore the removal of the ter- 

 minal difficulty, the high upper limit of the measurable temper- 

 atures, the permanence of its constants in relation to Clark's 

 cell in the lapse of time, the instantaneity of its indications, the 

 easy reproduction of destroyed couples, their relative insensitive- 

 ness to furnace gases, the regular and simple character of the 

 temperature function, the sustained sensitiveness throughout all 

 temperature ranges even quite into the fusion of platinum — all 

 these facts are a sufficient if not an overwhelming recommenda- 

 tion of the method. 



In speaking of long range temperature variables, one is hardly 

 permitted to overlook the remarkable work which has recently 

 been done in the direction of low temperature ; but with these 

 subjects I am less familiar, and can therefore only refer to in 

 passing. The progress made in the subject is sufficiently evi- 

 denced by the growth of large low temperature laboratories 



NO. 1457, VOL. 56] 



throughout the world, laboratories which undertake " the cold 

 storage" of " cold storage," as it were, like those of Pictet in 

 Berlin and Paris, of Dewar in London, of Kamerlingh Onnes 



' in Leyden, of Olszewski in Krakau, and others. Dewar and 

 Fleming have added to our knowledge of the probable consti- 

 tution of bodies at the absolute temperature. Olszewski has 

 found the critical temperature of hydrogen 'at - 230° and its 

 atmospheric boiling point at - 243°. Dewar and Moissan have 



I liquefied fluorine. There is much here which I must reluc- 

 tantly forego. The hydrogen thermometer.'the platinum balance 



j (Callendar), and the thermo-couple are again doing excellent 



I work in thermometry. 



I Applications of Pyrometry. 



I Turning now to the applications of recent pyrometry, we 



i meet first many series of valuable data on melting points and 



j similarly valuable data on the dissociation temperatures of 



t chemical compounds. To these I merely refer, not being quali- 



j fied to enter into chemical interpretations. High temperature 



boiling points have also been treated, and I will especially con- 



I sider the case of the variation of metallic ebullition with pressure. 



■ The relation of vapour pressure to temperature has thus far 



defied the counsels of the wise, even though such men as 



Bertrand and Dupre have given the matter close scrutiny. One 



would suspect the simplest relation to hold for metallic boiling 



points, and investigations have therefore been undertaken in 



which the temperature of ebullition of Hg, Cd, Zn, Bi, were 



studied for pressures decreasing from one atmosphere down 



indefinitely. 



The results so obtained show an effect of pressure regularly 

 more marked as the normal boiling point is higher, so that the 

 attempt to express the phenomenon for all these bodies by a 

 common equation is roughly successful. By far the most rapid 

 reduction of boiling point occurs when the pressure decreases 

 from ^ atmosphere indefinitely. For the case in which the 

 normal boiling point is to be predicted from a low pressure 

 value in case of a metal which, like bismuth, boils with great 

 difficulty, very high exhaustion is essential. 



Igneous fusion, by which I mean the fusion of rock-forming 

 magmas, is particularly interesting in its relation to pressure. 

 This has been again recently pointed out by Clarence King in 

 his discussion of the age of the earth. If the earth is solid 

 within, as is now generally admitted, such solidity can only result 

 from superincumbent pressure withholding fusion. To study 

 the relation of melting point to pressure directly is out of the 

 question when white heat is the condition of fusion. In this 

 respect the laboratory in the interior of the sun or even of the 

 planets has some salient advantages ; but we cannot comfortably 

 put such a laboratory under strict surveillance of protoplasm. 



Fortunately the Clapeyron equation, successively improved by 

 James Thomson and by Clausius, is here usefully available. To 

 measure the melting point, the difference of specific volumes of 

 the solid and the liquid body and the latent heat of fusion at 

 this temperature, with the aid of Joule's equivalent, is to 

 measure also the relation of melting point to pressure im- 

 plicitly. Based on the first and second laws of thermo- 

 dynamics, this equation is generally true, no matter what 

 specific properties may characterise the body. The process has 

 thus far been completely pushed through for diabase only. 

 Thermal change of volume may be measured by enclosing the 

 rock in a platinum tube of known expansion, and the contraction 

 of the contents from liquid to solid found by an electric micro- 

 meter probe reaching within the tube. Given a furnace fully 

 under control, then, as experiment has shown, the cooling can be 

 made to take place so slowly that platinum remains rigid rela- 

 tively to its charge of red-hot magma, and under these con- 

 ditions the contraction can actually be followed into the solid 

 state. At the same time, the temperature at which marked 

 change of volume occurs is the melting point. Apart from 

 difficulties of manipulation, the latent heat may be found from 

 measurement of thermal capacity on either side of the tempera- 

 ture of fusion, by a modification of known methods. 



The rate at which fusion is retarded by pressure, computed 

 from these data in the manner specified, showed an increase of 

 the melting-point of a silicate of about 0-025° C. per super- 

 incumbent atmosphere. But this datum falls within the margin 

 (o*02 . . . 004) of corresponding data much more easily and 

 directly derived for organic bodies. One may therefore argue 

 that if the melting-point pressure rate is so nearly constant on 

 passing from the class of siliceous to the thoroughly different and 



