92 Prof. Magnus on the different Properties of the 



in which the other rays, especially the non-luminous rays, of the 

 two flames stand to each other, I have compared together their 

 spectra, and found them alike throughout their whole extent. If 

 there do exist any differences, they lie within the limits of the 

 errors of observation. It is of course understood that the proper 

 precautions must be taken in these experiments to prevent rays 

 falling upon the thermo-pile from the solid or liquid soda, or 

 from the platinum wire on which it is supported. Care must 

 likewise be taken that the extent of radiating surface is the 

 same in the luminous and non-luminous flames ; for the intro- 

 duction of the soda alters the size and form of the flame; and 

 hence, if we were to compare the entire flames by letting them 

 radiate into the conical reflector of the thermo-pile, the radiating 

 surface of the luminous flame being greater than that of the 

 non-luminous one, more rays will fall into the cone, although no 

 more heat is produced. 



The same thing applies to the comparison of a flame rendered 

 luminous by the presence of carbon, with the non-luminous flame 

 of a Bunsen's burner. In consequence of the gas being already 

 mixed with atmospheric air when it issues from the opening of 

 the burner, the flame is smaller than when the lower holes of the 

 burner are closed so that the air has access only to the outside 

 of the flame, and consequently the carbon which is separated at 

 the inside cannot burn till it gets to the edges of the flame. 

 If, however, the heat-spectra of equal-sized portions of the lumi- 

 nous and non-luminous flames are compared, the two are found 

 to be exactly alike. 



The greater part of the heat which the flames give out lies 

 beyond the red ; and it is not surprising that in this dark portion 

 of the spectrum the intensity should be the same for both flames ; 

 but it is astonishing that, with such a great difference in the illu- 

 minating power, the heating effect should remain the same. 



At the first glance this result may seem surprising ; for it is 

 well known that the solid particles of soda or carbon, which im- 

 part luminosity to the flame, radiate more heat than the gaseous 

 particles. But when we consider the small amount of these par- 

 ticles, the smallest trace of which suffices to give the greatest bril- 

 liancy to the flame — when we also consider that these particles 

 are ever being renewed, and that their heat must be imparted to 

 them by the flame itself, it no longer appears strange that the 

 difference between the heat radiated by the two flames should 

 be so small as to escape observation. The extraordinary increase 

 in the radiation of light, which this small amount of solid parti- 

 cles occasions, remains, however, on this account only so much 

 the more remarkable. 



