14 A PHYSICAIv STUDY OP THE! IflRBlflvY. 



adult state. It would also be interesting to examine sections of Photuris 

 after the insect has flashed for an hour and a half or longer, as described 

 elsewhere, and of a similar specimen which has not undergone such a period 

 of flashing, to see whether the "reflecting layer" is markedly different in 

 the two samples. Of course numerous examples of each (of the same age) 

 would have to be examined to obtain an average value. 



VI. METHODS OF INVESTIGATION OF THE LIGHT EMITTED. 



As with other sources of radiation, the best method of investigating 

 the composition of the light of the firefly appeared to be one involving the 

 determination of the spectral energy curve. Because of the weakness and 

 intermittence of the radiation it is not possible to use a bolometer or a spec- 

 trophotometer, and the only satisfactory method is the one involving pho- 

 tography. While this seems to be an indirect method to obtain a spectral 

 energy curve, it will be noticed presently that it is not so cumbersome as it 

 appears. The photographic plate, being integrative in its action, is well 

 adapted to the investigation of weak and intermittent sources, but it must 

 be of a special kind which is sensitive to all parts of the visible spectrum. 

 The nearest approach to this condition is the Wratten and Wainwright 

 "Panchromatic" plate, which is sensitive, in a variable degree, to all the 

 frequencies from the ultra-violet far into the red. The method is really a 

 species of spectrophotographic photometry in which the light of the firefly 

 and that of the standard source are photographed, after which the "densi- 

 ties" of the negatives are compared. 



In ordinary spectrophotometry one compares intensities. In the photo- 

 graphic method the effects of these intensities on the photographic plate are 

 compared, the element of time being the unit of measure and the effect of a 

 given intensity upon the photographic plate being taken proportional to 

 the time of exposure. This, of course, is not true for long exposures. For 

 a limited range the density of the plate is proportional to the time of expos- 

 ure; beyond this range saturation begins and, still farther away, reversal 

 takes place. It is therefore necessary to determine the exposure equivalent 

 of each density measured; that is to say, the "characteristic curve" of the 

 plate must be determined. This is accomplished by making a series of 

 exposures of known relationship of a standard source in which the densities 

 cover the same range as those obtained on the unknown source, e. g., the fire- 

 flies. Such a series of exposures are shown in Plate i, A, the standard 

 source being a carbon glow-lamp operated on 4 watts per candle. The lamp 

 was placed at a fixed distance from the spectrograph slit, which was covered 

 with a ground-glass plate, the voltage was kept constant, and a series of 

 exposures was made corresponding to 2, 4, 6, 8, 10, 12, 20, 30, 60, 120, and 

 240 seconds. For densities on the firefly negatives, in which there were no 

 corresponding density measurements on the glow-lamp, the characteristic 

 curve of the latter was drawn for the particular wave-length in question, 

 and the firefly density value was determined by interpolation. 



Since the energy value, "mechanical equivalent," is not the same for all 

 rays, it is necessary to know the spectral energy distribution of the carbon 

 glow-lamp. This was determined anew for the present research by means of 

 a large vacuum spectrobolometer, fluorite prism, and mirror spectrometer. 



