Measurement of High Temperatures. 339 



denly close the cock which conducts the hydrogen into the an- 

 nular space, and immediately plunge the tube which lets out the 

 gas into a mercury trough, the mercury is seen to rise gradually 

 in this tube to a height of 602 millims. (the barometer being at 

 753). The hydrogen is therefore still forcing its way into the 

 platinum tube, and an almost complete vacuum is produced in 

 the annular space. And yet we also ascertained that our appa- 

 ratus, difficult as it was to construct, did not close perfectly. 

 This is perhaps the only reason which can explain this slight 

 difference of 15 centims. between the height of the barometer 

 and the height of the mercury in the delivery-tube, a difference 

 which measures the pressure of the gas remaining, or introduced 

 by leaks in the porcelain tube. 



The result of the foregoing is, that platinum behaves at high 

 temperatures like the porous vessels with which the beautiful 

 experiments by M. Jamin on the diffusion of gases and those 

 already published by one of us are obtained. 



A sufficiently convincing proof of this porosity may also be 

 had by substituting, in the preceding experiment, pure and dry 

 carbonic acid for air, and maintaining the current of hydrogen 

 in the annular space. At the end of the platinum tube a great 

 quantity of carbonic oxide and hydrogen may be collected at the 

 same time as carbonic acid*. 



These results show that it is impossible to construct gas-pyro- 

 meters of platinum when they have to come into contact with 

 the reducing gases, or with the hydrogen of a furnace f. It is 



* Working with the same tubes at high and increasing temperatures, 

 gases are obtained containing — 



Hydrogen 127 7'3 



Carbonic oxide 1 77 



Carbonic acid , 87*3 75'0 



UXK) 100-0 



The last analysis relates to the gas which issues from the platinum tube 

 and is collected at the highest temperature : the result establishes the pre- 

 sence of carbonic oxide and of moisture in the gases collected. 



t M. Pouillet, in his experiments on the pyrometer which he first em- 

 ployed, avoided to a great extent this source of error, as he heated his 

 apparatus in an iron muffle very nearly closed. His temperatures are also 

 much higher than those of M. Ed. Becquerel, especially if M. Pouillet's 

 results are modified by adopting, as he himself recommends, the new coeffi- 

 cient of expansion -^3 instead of Gay-Lussac's coefficients, with which they 

 were originally calculated (Traite de Physique, par M. Pouillet, 6 me edit. 

 1853, t. l er , pp. 237, 238, 239, 269, plate 9. fig. 8). M. Ed. Becquerel 

 places the platinum reservoir of his pyrometer inside an earthenware tube 

 — that is to say, in a porous material through which the hydrogen of the 

 furnace can penetrate with the greatest ease. 



