28 



done to gas by acids. Clean deal sawdust is well moistened with 

 pure sulphuric acid, diluted with five or six volumes of water, so that 

 the sawdust may not be discoloured, and gas is passed through it in 

 a slow stream. With rich gases, which give the light of from 20 to 

 25 sperm candles for a consumption of five feet an hour, the saw- 

 dust instantly changes to a most beautiful pink colour, and the tint 

 gradually deepens until the whole becomes of a dark mahogany. 

 With poor gases, which give the light of from 10 to 12 candles, this 

 coloration is exceedingly faint at first, and deepens very slowly. The 

 differences of coloration are so considerable and constant, that I have 

 no doubt of the possibility of thus determining the value of gas as an 

 illuminant. By using a standard acid, the same kind of sawdust, 

 a uniform volume of gas, and the same sized U-tubes, notation of time 

 and depth of colour would give a close approximation to the illumi- 

 nating value of the gas. At all events, the sources of error are not 

 greater than those of photometry in the hands of all but the most ex- 

 perienced, and the process is quite as close an approximation to truth 

 as an ultimate analysis of gas, containing, as it does, impurities which 

 render skill and precaution useless. A comparison of the analysis 

 of coal-gas given in ' Bunsen's Gasometry,' with the substances now 

 known to exist in gas, will convince us that at present we cannot 

 attach any value to such analyses. 



To determine the substances in gas which produce this coloration, 

 some of its chief illuminating constituents were prepared and passed 

 separately through the acid sawdust. 



Olefiant gas made in the usual manner, and carefully purified, 

 reddens the acid sawdust. Ether vapour does not affect it, and there- 

 fore need not be removed from the gas for this experiment. 



Propylene, produced by passing the vapour of fusel-oil through a 

 red-hot combustion- tube filled with cast-iron nails, but kept at so 

 low a temperature that a small portion of oil passed over without de- 

 composition, reddened the acid sawdust. 



Commercial benzole, with the exception of one specimen, reddened 

 the acid sawdust. 



I have not yet had leisure to prepare and test acetylene. 



The coloration of fir-wood, moistened by hydrochloric acid, has 

 been mentioned by Williams as characteristic of pyrrole. 



To show that the colour was produced by illuminating matter 



