338 
countless dark flashes covering the sky around 
it, tells me that the appearance to the eye was 
a brilliant white discharge, with fainter rose- 
colored ramifications, the latter developing in 
the negative, or rather positive, as dark flashes. 
Some years ago it occurred to me that a dark 
flash might be produced by a preponderance of 
infra-red radiations, which, as Abney has shown, 
undo the work of ordinary light on the plate. 
If we had a form of discharge capable of giving 
off very little actinic light, and an abundance of 
infra-red light, it might come out dark on a 
feebly illuminated background. This is, of 
course, a very wild guess, with nothing to sub- 
stantiate it, but the dark flash appears to be a 
reality, and a poor hypothesis is perhaps better 
than none at all. 
I have recently thought that the phenomenon 
might perhaps be explained in another way. 
We have a flash which appears darker than the 
sky behind it. It is inconceivable that the dis- 
charge could render the air in its path opaque 
in the ordinary sense to white light. But the 
light which illuminates the sky in the case of 
these pictures is not daylight, but light coming 
from another flash, that is made up of wave- 
lengths corresponding to the periods of vibra- 
tion of the dissociated matter in the path of the 
discharge. Now, may it not be possible that 
in the dark flash we have a discharge, weak or 
nearly wanting in actinic light which, however, 
renders the air in its path capable of absorbing 
to some extent the radiations of the wave- 
lengths which come from the bright flash. 
Such a flash might possibly appear dark on a 
blackground feebly illuminated by light ex- 
clusively of these wave-lengths. 
In other words, may we not have in the path of 
the dark flash, dissociated molecules, radiating 
but feebly, and capable of taking up vibrations 
of periods similar to their own, coming orig- 
inally from a simultaneous brighter discharge? 
It might not be impossible to reproduce the 
phenomenon by photographinga sparkin front of 
a white background inan absolutely dark room. 
Sparks are almost always taken against a dark 
background, which would account for the ab- 
sence of dark flashes in pictures of artificial dis- 
charges. A heavy main spark with lateral branch- 
es would seem the most suitable kind to employ. 
SCIENCE. 
[N.S. Vou. X. No. 245. 
The best method, however, of attacking the 
problem experimentally, it seems to me, would 
be a search for selective absorption in a partially 
exhausted tube. If the source of light were 
continuous any absorption would be unnotice- 
able unless it persisted for some time after the 
discharge (which is unlikely), for the time be- 
tween successive discharges is very great in 
comparison to the actual duration of one of 
them. Even in the case of so-called continuous 
discharges produced by high potential storage 
batteries the discharge is often, and may al- 
ways be, intermittent in character. The source 
of light should then be of no longer duration 
than the discharge occurring in the gas, the ab- 
sorption of which is to be examined. 
I can think of no way of producing a white 
or continuous spectrum source of as short du- 
ration as, and contemporaneous with, the dis- 
charge in the tube, but by employing two tubes 
differently excited, the one as a light source, 
the other as an absorption tube, some results 
might be obtained. Professor Trowbridge found 
that an argon tube emitted a blue light or red 
light, according to whether it was illuminated by 
means of an oscillatory or non-oscillatory dis- 
charge. By using the blue tube as the source 
and the red tube as the absorption tube, the two 
being arranged so as to be illuminated simul- 
taneously, it might be found that the red tube 
had the power of absorbing to some extent the 
blue radiations from the other. I hardly think 
results would be obtained, but the experiment 
seems worth trying. 
A picture taken by Mr. H. B. Lefroy, of 
Toronto, sent to me by Mr. Lumsden, Secretary 
of the Astronomical and Physical Society of 
Toronto, has some very curious appearances. 
There is an exceedingly brilliant flash running 
down the center of the plate, illuminating the 
sky quite brilliantly in its neighborhood. In 
its immediate vicinity, though not joined to it 
in any way, are innumerable dark, thread-like 
markings, which in places seem to cross each 
other, forming meshes. 
Mr. Lumsden assures me that the testimony 
of all photographic experts who have seen 
the plate is to the effect that markings of that 
description could only be produced in the ex- 
posure ; that is, they are not due to faults in the 
