164 ANNUAL REPORT SMITHSONIAN INSTITUTION, 1911. 
but the experimental difficulties were so great that very little was 
accomplished. A small box was made of brass and square plates of 
quartz. ‘The inside was varnished and blackened with soot, a drop 
of mercury introduced, and the box exhausted. The camera with its 
quartz objective was now trained on the box, and a beam of light 
from a mercury lamp (quartz) focused at the center of the box. 
Though the eye could see no trace of the cone of rays, the photograph 
brought it out as distinctly as if the box was full of smoke. An ex- 
posure of only 1 second was necessary, and with a 10-second expo- 
sure the spectrum of the light scattered by the vapor was secured. 
It was found to consist of a single line only (the 2536 line), though the 
light entering the box was the total radiation of the mercury are, the 
spectrum of which contained hundreds of lines. The pressure of the 
mercury vapor was about 0.001 millimeter, in other words, zeso00 
of the pressure of the air in the room. It seems most extraordinary 
that a vapor at such a very low pressure and at the temperature of 
the room should glow so brilliantly with invisible ight. A little fur- 
ther experimenting resulted in a further discovery. It was found 
that if the box was filled with air at atmospheric pressure, the cone 
of rays glowed feebly in the mercury vapor with which the air was 
saturated. As the pressure wasreduced the glowincreased in brilliancy, 
reaching its maximum at a pressure of about 5 millimeters. As the 
exhaustion was pushed further the mercury vapor outside of the 
cone became luminous, and at the highest vacuum attainable the glow 
filled the entire box. This is secondary resonance radiation excited 
by the primary radiation of the mercury vapor, which is excited by 
the cone of focused rays. The brilliancy of the cone remained about 
the same, so that we can not attribute the bursting out of this sec- 
ondary fluorescence to a mere increase in the brilliancy of the directly 
excited vapor. 
Experiments are now in progress to determine why the presence of 
a few millimeters of air destroys all trace of the secondary radiation. 
Photographs of the glowing vapor in air at pressures of 5 millimeters, 
1 millimeter, and 0 are reproduced in plate 6, d, e, f. 
If we put the drop of mercury in a small flask with very thick 
walls, exhaust the air, and seal the neck of the flask with the oxy- 
hydrogen flame, we are in a position to study this interesting type 
of radiation in mercury vapor at high pressures. I found that as 
the temperature of the flask was raised the radiation came from a 
region nearer and nearer the front surface, which was illuminated by 
tht rays from the lamp, and that when the pressure was about 10 
atmospheres the ray from the lamp, which had a wave length of 
2536, was selectively reflected from the surface of the vapor, pre- 
cisely as if the inner surface of the bulb were plated with silver. The 
other rays passed through the bulb with their usual facility. I am 
