28 Prof. Magnus on the Diathermancy 



by means of three screws, the rims of the brass mountings are 

 held firmly together. "When it is required to interpose a plate 

 of rock-salt, it is placed between two tubes the rims of which have 

 been very slightly covered with grease. It is of course also need- 

 ful that the plate should have been ground quite true and parallel. 

 In order to be able to apply a screen inside the tube, the four- 

 cornered brass box P P is interposed. It contains two flaps, 

 made of double tin-plate, and moveable about axles which pass air- 

 tight through the sides of the box and project on the outside. By 

 means of these axles, the flaps can be opened from without and 

 placed vertically so that they let all the rays of heat pass. They 

 can be closed in the same way — that is, brought to the horizontal 

 position, in which case no heat can pass through. 



The upper section, S S, was 0*15 metre high. When this sec- 

 tion was cut off from the rest of the tube by the interposition of 

 a plate of rock-salt and was used alternately full of air and vacu- 

 ous, the rest of the tube, B ¥, below the rock-salt plate being all 

 the time filled with dry air, the effect on the thermo-pile was 

 perfectly identical in the two cases. I have made this experi- 

 ment repeatedly. It refutes, as it seems to me, Prof. TyndalPs 

 assertion that the air to be examined must not be allowed to 

 come in contact with the source of heat. When the air is 

 below the heated surface, the latter can only be cooled by con- 

 duction; but, as I have shown, the conducting-power of all 

 the gases except hydrogen is so small that less heat traverses a 

 space filled with gas than would traverse a vacuum ; and hence 

 it follows that the effect of conduction in the case of the gases 

 is so small as to be inappreciable in comparison with the effect 

 of athermancy. The only possible exception to this would be in 

 the case of hydrogen. But even here, the cooling which results 

 from conduction is so trifling as to be of no consequence what- 

 ever when the experiments are made with a vessel kept constantly 

 at the temperature of boiling water by means of steam led into it. 



Prof. Tyndall* disputes the conduction of heat by hydrogen. 

 The grounds upon which he does so are not quite intelligible to 

 me ; but the fact which is alone conclusive as to the conducting- 

 power of this gas, he has left entirely unnoticed. This fact is, 

 that, as already stated, heat applied from above traverses a space 

 filled with any of the other gases less easily than it traverses a 

 vacuum : hydrogen alone transmits more heat than is transmitted 

 by a vacuum. Since, however, hydrogen does not allow more 

 rays of heat to pass than atmospheric air, and since, moreover, 

 even when its motion is hindered by means of eider-down, heat 

 is propagated through it more easily than through a vacuum or 



* Phil. Trans. 1862, p. 96. 



