DISTRIBUTION OF ENERGY IN FLUORESCENCE SPECTRA. 



179 



20 



For this purpose we made a careful spectrophotometric comparison between 

 our acetylene standard and a Hefner flame (see Table 24 and Fig. 174). From 

 Angstrom's curve of the distribution of energy in the spectrum of the Hefner 

 standard we then computed the relative intensities of our comparison flame 

 for several regions lying between 0.656 /1 and 0.483 ju. The points so deter- 

 mined are shown in Fig. 173 by crosses marked A. 



Several points determined by Stewart by direct measurement with the 

 radiometer are also shown in Fig. 173, being marked S. 



It is a matter of some interest to determine to what extent the visible 

 radiation from the acetylene flame corresponds to the visible radiation 

 from a black body. We find that the curve computed from Wien's equa- 

 tion is practically identical with our experimental curve from 0.56 n to 

 0.65 i*. But for wave-lengths less than 0.56 n the curve based upon Wien's 

 equation (shown by the broken line in Fig. 173) deviates considerably from 

 that determined by experiment. The acetylene flame appears to possess a 

 band of abnormally high radiating power in the region lying between 0.55 n 

 and the violet end of the 

 spectrum. 



Since the first account 

 of our own work appeared 

 Coblentz 1 has published 

 the results of direct 

 measurements of the 

 spectrum of the acetylene 

 flame with the vacuum 

 bolometer. While Cob- 

 lentz's results are in fair 

 agreement with ours in 

 the region from 0.48 /j. to 

 0.56 jit,they differ very appreciably for the longer waves, lying between our 

 own results and those of Angstrom. 



COMPARISON OF THE FLUORESCENCE SPECTRA WITH THE SPECTRUM OF THE 



STANDARD ACETYLENE FLAME. 



Before making the final spectrophotometric determinations from which 

 the distribution of energy in the spectrum of various fluorescence spectra 

 were to be computed, a careful study was made of the effect of slit-width 

 upon the form of the observed curves. 



In the ordinary use of the Lummer-Brodhun spectrophotometer slit a 

 (Fig. 172) would be of constant width and slit b would be varied. The 

 accuracy of the screw of the latter was therefore tested, as already described 

 at length in the opening paragraphs of Chapter XI, by mounting the 

 slit, which had been removed from the instrument, in the field of a pro- 

 jecting lantern and measuring the width of the slit-image, focused upon a 

 horizontal millimeter scale at a distance of about 8 meters. Variations 

 from constancy of the ratio of widths to micrometer readings were found 

 negligible for a range of two turns of the screw. Studies of the brightness 

 of the spectrum obtained when this slit was used showed, however, marked 



10 



p 



0.50 



0.54 0.58 



Fig. 174- 



0.62 



0.66//, 



'Bulletin of the Bureau of Standards, vol. 7, p. 243, 191 1. 



