OP THE MANUFACTURE OF GAS. 



79 



separate retort, and the introduction of these gases, along 

 with an excess of watery vapour, into the retort in which the 

 illuminating gases are being generated, and in such a manner 

 that these latter gases shall be swept out of the retort as 

 rapidly as possible, and thus removed from the destructive 

 influence of a high temperature. 



The excess of steam accompanying the water gas into the 

 second retort performs there a remarkable office; it reacts 

 upon the tar and other fuliginous matter in a manner that 

 will be described below, and gives rise to the formation of a 

 large additional quantity of gas, a very large proportion of 

 which is pure hydrogen. That this reaction of steam should 

 be confined entirely to the tar and other refuse matters, and 

 should not affect the luminiferous gases generated in the 

 same retort, is scarcely conceivable, since the constitution of 

 tar and of gaseous hydrocarbons is so nearly alike ; but any 

 destruction of illuminating principles that may be thus caused, 

 is immensely overbalanced by the quantity of these principles 

 which are saved from decomposition, by their rapid removal 

 from the influence of a high temperature and by the vapours 

 of volatile hydrocarbons with which the water gases remain 

 more or less saturated. 



My first experiments were made upon the application of 

 the process to resin ; but as these are of less scientific interest 

 than those on its application to coals and cannels, on account 

 of there having been no comparative experiments on resin 

 gas produced by the old process, I will confine myself prin- 

 cipally to a summary of the results, entering into detail only 

 on such points as bear upon, and illustrate the principles 

 which I have laid down. 



white's process applied to resin. 



These experiments were conducted at the gas works 

 attached to the mill of Messrs. George Clarke and Co., 

 Ancoats, Manchester. These works consisted, at that time, 



