in their Relations to Composition and Luminosity. 291 



In fact, our studies of the subject have led us in the di- 

 rection of the general conclusion that, all other conditions 

 being equal, the temperature in a given flame is the main factor 

 of luminosity. This, however, may as yet be regarded merely 

 as an hypothesis, in consequence of the imperfection of our pre- 

 sent means of actual experimental demonstration of the tempe- 

 rature of flames ; it is an hypothesis, nevertheless, which is in 

 general accordance with known facts. By the spectroscope, for 

 example, which can recognize only luminous rays, we find that 

 the higher the temperature the greater the number of these lumi- 

 nous rays. The recent results of Frankland upon the develop- 

 ment of luminosity by increased pressure in flames which are 

 non-luminous under atmospheric pressure, are in accordance 

 with this view, increase of temperature necessarily following in- 

 crease of pressure. 



Very vague views have been rife, even among chemists, with 

 regard to the temperatures of luminiferous flames. Some have been 

 satisfied with believing crude hypotheses, such as that the heat- 

 power of a flame is always proportional to the density of the gas 

 or vapour undergoing combustion, or that it is proportional to 

 the amount of oxygen consumed by a given volume of the gas, 

 and so on. This latter hypothesis has been one of very common 

 acceptation. A view which is even now entertained by some 

 skilful chemists (than which, however, nothing, as will be shown 

 below, could be more fallacious) is, that those individual gaseous 

 compounds which impart the highest luminosity under ordinary 

 conditions, are also the most productive of heat. 



The admirable researches of the great gas- chemist, Bunsen, 

 of Heidelberg, placed in our possession some years ago the 

 means of computing, at least with approximate accuracy, the 

 heat of flames of gases of known composition. Few, however, 

 have properly and successfully applied Bunsen's methods in 

 practice. We consider it quite time that these methods should 

 be introduced to the knowledge of gas-engineers in forms avail- 

 able to them. 



Bunsen's formulae for these computations are based upon the 

 actual experimental determinations of the total amounts of heat 

 developed by the combustion of different pure combustible gases 

 with pure oxygen made by Favre and Silbermann, and upon 

 Regnault's determinations of the specific heats of gaseous pro- 

 ducts of combustion. 



It is not to be maintained that Favre and Silbermann's num- 

 bers are strictly correct ; but they are doubtless approximate, and 

 at least proportionally correct among themselves ; at any rate 

 they are the best data we have. Those employed here are in- 

 cluded in the following Table. They are usually given in the 



U2 



