Prof. Magnus on the Propagation of Heat in Gases. 95 



Deflection. Rays. 



Vacuum 15-8 



corresponding to . . 16 "2 = 100 



Atmospheric air . . . 144 88-88 



Oxygen 14-4 88-88 



Hydrogen 13-9 85-79 



Carbonic acid . . . 13-0 80-23 



Carbonic oxide . . . 12-8 79-01 



Protoxide of nitrogen . 12-0 74-06 



Marsh-gas .... 11-7 72-21 



Cyanogen 11-7 72-21 



Olefiant gas .... 7'5 46-29 



Ammonia ..... 6-3 38-88 



Although these values cannot be looked upon as quite reliable, 

 inasmuch as variations may occur from imperfect purity of the 

 gas, or from other almost unavoidable impurities, they yet show 

 how considerable are the differences which perfectly transparent 

 gases exhibit in reference to the property of transmitting heat. 

 This surprising deportment, which I had already established 

 before I laid the first part of this treatise " On the Conduction of 

 Heat " before the Academy, led me to make a separate investiga- 

 tion of the radiation through gases, and first of all to ascertain 

 whether similar differences prevailed when another source of heat 

 was used. 



A Gas-flame as source of Heat. 



I desired first of all to use a source of heat at a higher tempe- 

 rature, for which purpose the apparatus depicted in fig. 2, Plate 

 I. was unfitted. I was accordingly compelled to use the gases 

 in a tube closed at both ends by plates. In testing this method, 

 I had occasion to make some observations which have probably 

 also been made by others, but which I have nowhere found 

 mentioned. 



Influence of the side of the Tube. 



If the rays from any source of heat be allowed to act upon a 

 thermo-pile without having passed through any tube, a smaller 

 deflection is obtained than when the rays from the same source of 

 heat placed at the same distance from the thermo-pile are allowed 

 to pass through a tube open at both ends, that is, not closed by 

 any kind of plate. This increased action is obviously caused 

 by the rays reflected from the inner side of the tube, so that the 

 thermo-pile is acted on, not only by the rays which come directly 

 from the source of heat, but also by those which fall obliquely into 

 the tube and are again reflected. Even if the tube were black- 

 ened on the inside, or if, as was usually the case in the following 



