Measurement of High Temperatures. 341 



which we wish to measure, and enclosed in impermeable vessels 

 of Bayeux porcelain absolutely rigid at this temperature, at which 

 they hold a vacuum perfectly, and are still very far from their 

 point of fusion or softening. These vessels, through which a 

 current of dry air is sent until incipient red heat, are closed at 

 the end of the experiment by an oxyhydrogen blowpipe, just as 

 we close a glass air thermometer with a mouth-blowpipe. The 

 only sources of error consist in the observation of barometric 

 pressure, of the temperature of the surrounding atmosphere, and 

 lastly in the reading of the volume of the gases in a graduated 

 tube. This latter being all that we have to consider, we should 

 say that in working with flasks of 300 to 330 cubic centims. 

 capacity, a very considerable and inadmissible error of half a 

 cubic centim. would not sensibly affect our results. The tem- 

 peratures found by M. Ed. Becquerel would ascribe to us an 

 error of about 10 cubic centims. upon 80, which is out of the 

 question. 



The only causes of uncertainty are such as may arise from 

 the apparatus in which the metals are heated to ascertain their 

 boiling-point. To arrive at a degree of exactitude which does 

 not affect our vapour-densities, we are now repeating our mea- 

 surements of temperature by aid of the most delicate methods. 

 But as M. Ed. Becquerel seems to have placed himself very 

 nearly in the same conditions as ourselves, it was needful to look 

 elsewhere for the cause of our difference. It is this considera- 

 tion that led us to seek in platinum for a property unknown 

 hitherto, though in perfect harmony with its catalytic action, and 

 the feebleness of its conducting-power for heat and electricity*. 



* While this paper was printing, we received from Mr. Matthey, the 

 great London manufacturer, a cast platinum tube weighing 1070 grammes, 

 60 centims. long and about 2 milJims. thick. All the experiments which we 

 have already described have been repeated with this tube, and gave exactly 

 the same result, causing the formation of water by means of hydrogen and 

 at the expense of the air, and producing a vacuum in the annular space, &c. 

 The immense thickness of this tube and its quality have in no respect 

 altered the phenomena, either in their direction, in their intensity, or in 

 their activity, Mr. Matthey, knowing what we were in need of, has had 

 this tube manufactured with part of the ingot of 100 kilogs. of cast plati- 

 num which he exhibited in 1862, and, anticipating our wishes, has obligingly 

 sent it to us for the benefit of science. We publicly express our obliga- 

 tions to him. 



