THE FLUORESCENT SUBSTANCE IN FIREFLIES. 37 
serious matter. To measure such a temperature difference is no small 
undertaking. To go one step farther and determine how this temperature is 
modified, as for example by the flash, is a different problem requiring further 
investigation. 
In Fig. 13, C, is shown a type of (vacuum) bismuth-silver thermopile with 
which it is purposed to make further radiation measurements. This pile 
consists of 10 junctions of silver wire 0.019 mm. in diameter and bismuth 
wire 0.04 mm. in diameter, blackened with a composition of platinum black 
and lamp black. The central junctions are insulated with shellac, placed 
in contact and backed with a receiver of pure tin, 0.01 mm. thick and 1 mm. 
indiameter. In air it is one-half as sensitive as a 20-element Rubens linear 
thermopile. It is about 4 times as sensitive as the radiometer used to 
measure heat from stars, when used with a galvanometer of the same period 
as the radiometer. Its resistance is 19.3 ohms. 
X. THE FLUORESCENT SUBSTANCE IN FIREFLIES. 
During the photographic work two years ago, in cleaning the test tubes 
which had contained fireflies, the writer noticed that the wash water had a 
peculiar faint bluish tinge, which he considered due to some fluorescent 
material. In order to demonstrate to others that it was not “blue sky,” 
““decomposition bacteria,” etc., an attempt was made to obtain a pure solu- 
tion and photographs were made of the fluorescent material.* 
The usual procedure to obtain a pure solution was to place the insects in a 
strong solution of alcoholand water. Thealbuminous material was precip- 
itated with lead acetate and the solution was concentrated by boiling. 
The solution was then neutralized with potassium oxalate to prevent tur- 
bidity while photographing the fluorescence spectrum, which was obtained 
with the large spectrograph employed in photographing the light emitted by 
the firefly. Such a photograph of the fluorescent spectrum is shown in 
Plate1,C. The cadmium spark produced a greater fluorescence than either 
zinc or aluminum, and was therefore used as an exciting source. The ultra- 
violet light tends to produce a cloudiness in the solution which must then be 
filtered. The fluorescence spectrum shows no structure and seems to be of 
the same nature as is commonly observed in many substances. The main 
point of interest is the extent (from 0.38 to 0.51) of the fluorescence spec- 
trum, which is complementary with the light emitted by the firefly. (See 
the right-hand side of Plate 1, C, which givesthe spectrum of the lightemitted 
by Photinus.) However, no special significance is attached to this fact. 
The maximum emission of the fluorescence spectrum lies at about 0.414. 
The absorption in the ultra-violet appears to be very strong, as indicated by 
the fact that, unless the solution is very dilute, the fluorescence extends but a 
few millimeters into the solution. When concentrated the solution becomes 
yellowish and but little fluorescence is noticeable. On dilution, however, 
the fluorescence reappears. The yellowish color is attributed to the exten- 
sion of the ultra-violet absorption, of the fluorescent substance, into the 
visible spectrum on concentrating the solution. The material decomposes 
*Coblentz, Phys. Zeit., 10, p. 955, 1909; Bull. Bur. Standards, 6, p. 321, 1909, 
