420 Prof. H. Buff on the TJiermal Conductivity 



To fill the glass cylinder with aqueous vapour, a small vessel 

 of water was placed within it ; near it a thermometer was 

 hung, the readings of which are given in brackets in the last 

 column but one. 



The first two experiments with moist air were made in a 

 space as highly rarefied as was attainable by means of the air- 

 pump. The manometric numbers indicate therefore the pres- 

 sure of a mixture of air and water-vapour, in which that of 

 the former did not exceed 1*5 millim. After the termination 

 of these experiments, air of 75*7 millims. barometric pressure 

 was allowed to enter the space already saturated with vapour. 

 The amount of vapour had to be determined from the tempe- 

 rature indicated by the thermometer. 



It appeared of particular interest to investigate defiant gas 

 according to this method — partly because its specific gravity 

 so closely approximates to that of air, and partly because, 

 notwithstanding the large proportion of hydrogen it contains, 

 it was found both by Magnus and Tyndall to absorb more of 

 the heat-rays than air. 



In has been pointed out that the basis of Tyndall's compa- 

 risons was not reliable ; and it is probable that in Magnus's 

 experiments conditions obtained such as led him to find the 

 diathermancy of hydrogen smaller than that of the air. 



On the other hand, Dulong and Petit, in their extensive 

 researches on the laws of cooling *, compared the rapidity 

 with which a thermometer cooled in contact with air, hy- 

 drogen, carbonic acid, and defiant gas successively (all four 

 gases taken at the same pressure), and found that the heated 

 body cools most rapidly in hydrogen, and next in olefiant gas. 

 Considering that equal volumes of air and of olefiant gas have 

 almost the same weight, the idea naturally presents itself of 

 supposing that the above-mentioned behaviour is due to the 

 fact of the diathermancy of olefiant gas being greater than 

 that of air and of carbonic acid. 



The gas used in my experiments was prepared from a mix- 

 ture of sulphuric acid and alcohol by Messrs. H. W. Will and 

 A. Winther, assistants in the chemical laboratory of this Uni- 

 versity. It contained a trace of ether, and might, whilst being 

 introduced into the apparatus, have taken up a small quantity 

 of air, though the cylinder was repeatedly filled Avith fresh 

 quantities of the gas before the experiment was begun. The 

 numbers of the following Table cannot be directly compared 

 with those of Y. and X., because in those two series the junc- 

 tion was covered with lampblack, and the bottom of the metal 

 cylinder was clean and bright, whilst in the present experi- 

 * Ann. Chim. Phys. t. vii. p. 351. 



