306 F. J. Rogers — Magnesium as a Source of Light. 



in figure 4. The previously determined temperatures of the 

 other flames were plotted as abscissae and the corresponding- 

 galvanometer " swings " as ordinates. The ordinate on these 

 curves, corresponding to the " swing " due to the magnesium 

 flame was taken as indicating the position of that flame upon 

 the curve. The abscissa reading of this point gave the flame 

 temperature to be inserted in the table. In figure 4 are given 

 the curves for the third method. The galvanometer "swings" 

 for the same flame in the two cases were not exactly the same 

 on account of the bead of metal at the junction being larger 

 in one case than in the other. 



Whatever question may be raised touching the accuracy of 

 the absolute values of the above determinations of flame tem- 

 peratures ; there can be little doubt that they are relatively 

 nearly correct ; and we may be sure that the temperature of 

 burning magnesium is a little below that of the hottest part of 

 an air blast flame. 



When the quality of the light from sources of different 

 temperature is examined, it is almost uniformly found that the 

 light from the sources of higher temperature is richer in the 

 more refrangible rays. Prof. Pickering, in the course of the 

 series of investigations on the quality of the light from dif- 

 ferent sources already cited, has made use of this fact to calcu- 

 late flame temperatures. The following passage from ' his 

 paper will serve to indicate the basis of his estimates : " Then 

 if the temperatures we adopted were correct, this would give 

 us a very simple empirical law, viz : The temperature is always 

 proportional to some function of the ratio of any two assumed 

 Avave-lengths. For artificial sources for the wave-lengths *5S5 



and "455 it varies directly as this ratio Upon 



this principle the temperature of the magnesium-light, perhaps 

 the highest terrestrial temperature we have yet attained, would 

 be 4900° C." As direct measurement shows this estimate to 

 be altogether too' high, it might be inferred that the magnesium 

 light, unlike that of most sources, is not solely due to incan- 

 descence, but that it is in part due to something like fluores- 

 cence by heating. Further evidence in favor of this view will 

 be presented in the course of the following study of the 

 efficiency of the magnesium light. 



III. 



Radiant Efficiency of Magnesium light. 



By radiant efficiency of a source of light I mean the ratio 

 of luminous energy to total radiant energy. This ratio has 

 often been rather ambiguously called, net efficiency. 



