neighbourhood of strongly Illuminated Bodies. 229 



have been conducted in common air, but we considered it 

 highly important to observe the behaviour of various dry- 

 gases of very different density from air, and also to see whether 

 any phenomena of the same kind could be detected in liquids. 



With the apparatus shown in fig. 8, and already described, 

 we have ascertained that hydrogen gives a much thicker 

 coat, and carbonic anhydride a thinner coat, than is obtained 

 in air, under conditions which we believe to be fairly com- 

 parable. 



The figures in the following Table will serve to indicate 

 roughly the differences w r hich have been observed : — 





Temp, of 

 solid. 



Temp, of gas 

 in observing vessel. 



Thickness of coat 

 in centim. 



Hydrogen 



o 



41 

 65 



82 

 63 



o 



17 

 26 

 14 

 20 



•033 

 •013 (?) 

 •020 

 •011 



Ammonia 



Air . 



Carbonic Acid 



The first column gives the temperature indicated by the 

 thermometer upon the blackened bulb of which the coats 

 were measured. The third column gives the measurements 

 made at the middle of the bulb on the side near the light. 

 In the middle column the temperature of the gas in the ex- 

 perimental chamber, as indicated by the second thermometer, 

 is given. 



The quantitative portions of this paper are as yet incom- 

 plete, and we consider the above figures only as indicating the 

 differences provisionally. Some gases, moreover, appear to 

 vary considerably in their action towards the camphor-smoked 

 surface; and, in short, there are obscure conditions which affect 

 quantitative results in a way which we do not yet fully 

 understand. 



(b) Liquids. — The outline of the method employed for the 

 study of the appearances observed in liquids holding fine 

 particles in suspension has already been described (fig. 10). 

 The box is best filled with cold water from which the air has 

 been expelled by recent boiling, and to which a little ferric 

 oxide has been added, to render it turbid. When the beam 

 of the lamp is now converged on to the surface of the empty 

 brass steam-tube, there is usually to be seen a slightly dust- 



