206 Holborn and Day — Gas Thermometer, etc. 



in perfect order. The best arrangement is always to measure 

 with two elements having a common hot junction and entering 

 the space in question from opposite sides, as has already been 

 described in our former paper. 



Furthermore, inasmuch as thermo-elements have come to be 

 employed very generally for technical purposes during the past 

 few years, we take this opportunity to say that it is impossible 

 to give the limits to which the electromotive force can be 

 affected by the breaking of the protecting tube or other acci- 

 dent through which the wires become exposed to combustion 

 gases in the manner above described — the higher the tempera- 

 ture the greater the danger. In case of such an accident the 

 element should at once be tested and restored as elsewhere 

 indicated. 



7. Conclusion. 



By way of conclusion we will bring together the tempera- 

 tures which we have obtained for the melting points of the 

 metals based upon observations with the gas thermometer and 

 some recent results of other observers employing other methods. 



These include the measurements with platinum resistances 

 as carried out by Callendar and by Heycock and Neville* 

 which are also based upon the gas thermometer up to 445°, 

 the boiling point of sulphur, and the higher temperatures 

 exterpolated from a formula of the second degree. 



The melting points of gold and silver have also been meas- 

 ured with thermo-elements by Berthelot,f using the wire 

 method and determining the temperature optically from the 

 •decrease in density of an air column in an electrically heated, 

 open, porcelain tube. 



Gas Platinum Optical 



Thermometer. Resistance. Method. 



Cadmium 321-7° 320V 



Lead.. 326-9 327-7 



Zinc. 419-0 419-0 



Antimony . 630-6 629-5 



Aluminium 657 - 654*5 



Silver, in air 955- 955- 962 



Silver, pure 961-5 960-7 



Gold 1064-0 1061-7 1064 



Copper, in air 1064*9 



Copper, pure 1084-1 1080*5 



Charlottenburg, May, 1900. 



* Most of the figures are taken from Heycock and Neville's paper — loc. cit.; 

 Callendar's values — Phil. Mag. V, xlviii, 619, 1899 — differ in several cases 1°. 

 fD. Berthelot, C. E., cxxvi, p. 473, 1898. 



