in their Relations to Composition and Luminosity. 297 



Table III. — Analyses of Anthracite Hydrocarbon Gas, by Silli- 

 man and Wurtz. 





No. I. 



No. 2. 



Mean. 





60-43 



35-44 



413 



59-32 



3714 



3-54 



59-87 



3629 



3-84 



10000 



Carbonic oxide 



Marsh-gas 







10000 



100-00 



In Table IV., column I gives the results of the analysis of the 

 street gas served out at this period by the New Haven Gas-Light 

 Company, made from Westmoreland coal enriched with about 

 6 per cent, of Albertite ; column 2 the mean of four analyses 

 of the completed hydrocarbon gas made by us at Fair Haven 

 during the same time, by combining gas from the same West- 

 moreland coal (with 10 per cent, of Albertite) with half its vo- 

 lume of the anthracite gas. Columns 3 and 4 are obtained from 

 1 and 2 by centesimal reduction, after deduction of the illumi- 

 nant ingredients, being what we propose to designate as the non- 

 illuminating substrata of illuminating gases. 



Table IV. — Gas Analyses, by Silliman and Wurtz. 





1. 



New Haven 

 city gas. 



2. 



Fair Haven 



hydrocarbon 



gas. 



3. 



Substratum of 



New Haven 



gas. 



4. 



Substratum of 



Fair Haven 



gas. 





43-58 



2-14 



47-42 



6-86 



46-77 

 9-56 



36-71 

 6-96 



46-79 

 2-31 



50-90 



50-27 

 10-27 

 39-46 



Carbonic oxide ... 

 Marsh-gas 



Illuminants 





100-00 



10000 



10000 



10000 



Table V. gives the results of the computation, from our formulae, 

 of the calorific powers of these five gaseous mixtures, for com- 

 municating temperatures both above and below that of aqueous 

 ebullition. We should remark that we have here been obliged 

 to regard the volumes of illuminant hydrocarbons as representing 

 olefiant gas solely — both because we have no certain data as to 

 their real nature, and particularly because, if we actually knew 

 or should assume the nature of the hydrocarbon vapours present, 

 still we have no experimental calorific equivalents, as we have 

 for olefiant gas, from which to start in such a computation. We 

 have reason to believe, nevertheless; that the errors thufc intro- 

 duced are not important in amount. 



