and Diathermancy of Air and Hydrogen. 403 



mometer fixed in a horizontal position. The upper vessel is 

 filled with boiling water, which, during the experiment, is kept 

 in ebullition by the introduction of steam. To keep the sur- 

 rounding medium at as constant a temperature as possible, 

 the apparatus is placed in a beaker, and this again in a second 

 beaker tilled with water of a constant temperature to the height 

 of the boiling water in the inner vessel. 



In order to diminish radiation as far as possible, a cork disk, 

 or a thin silvered copper plate, was placed as a screen above 

 the bulb of the thermometer. But the desired object was very 

 imperfectly attained in this way ; for the thermometer stood 

 8 degrees higher than that indicating the temperature of the 

 water in the outer beaker, even when the air was almost en- 

 tirely exhausted. On allowing air to enter, this rise of tem- 

 perature diminished more and more with the increase of pres- 

 sure, thus proving that, as in solid and liquid bodies, the heat- 

 rays are partially absorbed by their passage through air. 

 Conductive motion, if such existed, was masked by the stronger 

 effect of radiation. 



Similar results were obtained by Magnus with a great 

 number of other gases investigated by him. With hydrogen, 

 on the contrar}', he found that the temperature increases with 

 the increase of density ; and even at ordinary atmospheric 

 pressure it has a greater diathermancy than a vacuum. 



It was more especially on account of this behaviour that 

 Magnus inferred the similarity of hydrogen to the metals as 

 regards the conducting-power. 



A more minute and careful investigation of the above method 

 led me, however, to a different conclusion, and eventually to 

 a renewed experimental investigation. 



My apparatus is similar to that used by Magnus, with such 

 alterations as seemed to me appropriate. My object being to 

 study the conductive power of gases, other influences being 

 excluded (such as the effect of radiating heat &c.), a brass 

 cylinder, instead of that of glass in Magnus's apparatus, was 

 cemented air-tight upon the glass vessel. The bottom of the 

 cylinder facing the thermometer is a polished clean metallic 

 surface ; a double w^all surrounds it, the space between being 

 filled with cotton-wool to prevent too rapid cooling. The 

 glass cylinder is 20 centims. high and 7*5 centims. in dia- 

 meter ; its lower edge is ground so as to fit air-tight onto the 

 plate of an air-pump. Through the plate (of the air-pump) 

 pass two wires, one of iron and the other of German silver, and 

 extend vertically almost to the bottom of the brass cylinder, 

 where they are soldered together. At the junction both wires 

 are flattened to 12*5 millims. width, and thus form a flat 

 arc whose centre was, in the flrst experiments, only 23 mil- 



2D2 



