286 
ever, to justify a fresh effort, and within the last 
few days I have been able to get the effect as satis- 
factorily as could be wished. 
The chief essential is a source giving a sharp line 
of great intensity at 43303. This was found in a 
sodium vapour lampin quartz analogous to the mercury 
lamps in general use. Details of the construction and 
manipulation of these sodium lamps will be published 
later. The visible light from such a lamp was filtered 
out by means of a screen consisting of cobalt-blue 
uviol glass, combined with nitrosodimethylaniline. 
The light which came through was photographed with 
a quartz spectrograph, and was found to consist of 
A 3303 exclusively. 
This radiation was concentrated by means of a 
quartz lens on a quartz bulb containing some sodium. 
The bulb was made nearly red-hot with a bunsen 
burner, which was then extinguished. A patch of 
luminosity could be seen on the wall of the bulb when 
the ultra-violet beam fell upon it. As the bulb cooled 
and the vapour pressure of the sodium diminished, 
this patch of light gradually expanded, and filled the 
entire bulb; it then faded away, and had disappeared 
when the bulb was cold. This behaviour is exactly 
the same as is seen when D light is excited by the 
incidence of D light, and although in the present case 
the light is much fainter, the conditions of observa- 
tion are in some respects more favourable, for there 
is no disturbance from visible light scattered or re- 
flected by the walls of the vessel. 
Critics of this experiment will naturally concentrate 
their attention on two questions :— 
(1) Was the light observed really due to ultra-violet 
excitation ? 
(2) Was it of the same wave-length as the D line? 
As regards (1), a sheet of plate-glass 1-2 cm. thick 
Was interposed between the source and the bulb. The 
excited light was completely extinguished. 
As regards (2) the light was rather below the in- 
tensity which would easily allow of direct spectro- 
scopic examination, though with a little further im- 
provement of the conditions it might be made strong 
enough. I have, however, proved it to be of approxi- 
mately this wave-length by absorption methods. The 
luminosity was seen undiminished through a_ thick 
cell containing potassium bichromate solution, held 
before the eyes. It was absolutely invisible througha 
cell containing praseodymium nitrate. Thus the wave- 
length must fie in the region from A 5820 to 6020, 
for this is the only region transmitted by bichromate 
and absorbed by praseodymium. The D line at A 5890 
lies in this narrow region, and I think, therefore, that 
there is no reasonable doubt that the emission does 
consist of D light. Discussion of the theoretical 
bearing of this result is deferred. 
Ree sDRone 
Imperial College, South Kensington, May 8. 

The Green Flash. 
Many descriptions of the green flash have been pub- 
lished sin letters to NaTurE and elsewhere, but I do 
not remember to have seen a satisfactory explanation 
of this curious phenomenon. Atmospheric dispersion 
is invoked, but this does not explain the absence of 
the red end of the spectrum. My observations agree 
in every particular with those described by Mr. Whit- 
mell in Nature of March 11, p. 35. At sea I have 
observed a violet or blue tint occasionally, and on one 
occasion a red flash as the lower limb of the sun 
emerged from a cloud into a clear space very near the 
horizon. 
Normal atmospheric dispersion will, of course, pro- 
duce a red fringe to the sun's lower limb, and a blue 
fringe at the upper limb, 
NO. 2376, VOL. 95] 
NATURE 

[May 13, 1915 

time with a telescope free from secondary colour whem 
the sun is as high as ten or fifteen degrees above the 
horizon. When, however, a point of sunlight only is 
visible, the rest of the ‘disc being hidden beneath the 
horizon, atmospheric dispersion, if it could be per- 
ceived with unaided vision, should produce a complete 
vertical spectrum from blue to red, as in the case of 
stars when near the horizon. The red end of this. 
spectrum shculd be most evident, since these rays are 
least absorbed. In the flash, however, the red is 
completely suppressed, and the vivid green which is. 
obvious to the naked eye can only be seen at very 
low horizons. Moreover, it is not always seen, as. 
Mr. Whitmell remarks, when the conditions seem 
otherwise favourable. 
It seems to me very probable that the phenomenon 
is in some way connected with the abnormal condi- 
tions which at sea produce mirage effects. The layer 
of dense air in contact with the sea might produce 
total reflection for solar rays refracted from below 
the horizon, but the critical angle of reflection will 
depend on wave-length, and it is possible under cer-- 
tain conditions that the green rays may be totally 
reflected whilst the red dre refracted. 
I have one more observation to add to those de- 
scribed by Mr. Whitmell, and this will, I think, give 
the coup de grace to the theory of a subjective effect 
due to retinal fatigue. In May, 1900, I happened to: 
observe the setting of Venus in the sea from my 
eclipse camp on the Algerian coast. Observing with 
a 3-in. inverting telescope, I saw the planet when 
very near the horizon suddenly change in colour from 
dull red to vivid green, and as I lowered the telescope: 
to the point where-the sea horizon about bisected the 
field of view I was amazed to see two green images 
of Venus, one, the normal image, ascending from 
below, and the other sloping down from above. This 
was probably reflected from the sea itself. The setting 
took place at the moment of meeting of these two: 
images. The whole apparition, from the moment 
when the colour changed from red to green, to the 
instantaneous disappearance of the two images, cannot 
have lasted more than four or five seconds. The sea 
about this time was found to be excessively cold, 
although the air was hot during the daytime, and this 
state of things would doubtless favour the production 
of a relatively dense layer of air on the surface of the 
sea in calm weather. Joun EveRSHED. 
Kodaikanal, April 13. 

The Larger lons in the Air. 
In addition to the well-known small ions, which 
are of a type common to all gases, two classes of 
larger ions exist in the air under ordinary conditions. 
One of.these consists of the large ions of Langevin 
which have a mobility of about 1/3000, while the 
other contains ions with a mobility of about 1/50. 
As the latter value lies between those of the mobilities 
of the small and large ions, the members of this 
latter class may be called the ions of intermediate: 
mobility, or, shortly, the intermediate ions. 
The slow movement of these larger ions in an 
electric field clearly indicates that they are molecular 
clusters of more or less complexity. Ordinarily the 
value of the mobility is the only guide to the nature 
of the ionic structure, but in the case of the large: 
ion, at least, an important deduction is to be made 
from the outcome of experiments on the formation of 
clouds in closed vessels. 
It is well known, since Aitken’s notable work on 
the subject, that, in ordinary circumstances, the air 
is crowded with particles, in suspension, oa which 
the water vapour condenses into visible drops if the 
as may be seen at any | air becomes slightly supersaturated. These particles, 


