January 6, 1923] 
NATURE 13 

Letters to the Editor. 
[The Editor does not hold himself responsible for 
opinions expressed by his correspondents. Neither 
can he undertake to return, or to correspond with 
the writers of, rejected manuscripts intended for 
this or any other part of NATURE. No notice is 
taken of anonymous communications. | 
Broadcasting Transmitter. 
I expect that broadcasters would find that a 
definite generatorof E.M.F. would givecleaner articula- 
tion than is probable with a transmitter depending on 
random variation of resistance. In other words that 
one of those telegraphic devices which I described 
long ago (Proc. Inst. E.E., vol. 27, p. 838, Dec. 1898), 
consisting of a light stiff coil of fine wire suspended 
elastically in an annular magnetic field, would repro- 
duce speech and music better than a microphone. 
The fluctuations of the induced current in such an 
instrument are not capricious, and apart from stimu- 
lation it is silent ; always provided that the exciting 
magnetic field is kept steady—a condition likely to 
be assisted by saturation of the iron magnet core, or 
by use of a sufficiently strong permanent magnet. 
If an electromagnet is used, some contrivance is 
advisable whereby the coil circuit is automatically 
opened whenever the rousing current is put on or off. 
Otherwise, the response may be too violent for the 
valve and rest of the apparatus, not to mention a 
receiving ear. OLIVER LODGE. 
Normanton House, Lake, Salisbury. 

The Green Flash at Sunset. 
Srr ARTHUR SCHUSTER in his réview of Dr. Mulder’s 
book on this subject states that ‘‘ there seems no 
reason to doubt that dispersion combined with ab- 
sorption of light completely accounts for the effect ”’ 
(NaturRE, September 16, p. 370). Yet Dr. Mulder’s 
own view is that a complete explanation is still want- 
ing. 
Sty apology for again raising this question is that I 
believe I can supply from some recent observations 
what seems to be lacking in the dispersion theory, 
which fails to account for the remarkable variations 
in visibility of the green flash under apparently 
favourable conditions. 
On the outward voyage to Australia to observe the 
solar eclipse of September last, I was struck with the 
faintness of the green flash at sunset, although the 
sky was clear down to the horizon. It was visible in 
binoculars ( x8), but scarcely, if at all, to unaided 
vision. On the return journey, on the other hand, 
the phenomenon was brilliant every evening on the 
tun between the north-west coast of Australia and 
Java, and I was able to observe also what happened 
when Venus set in the sea. On this voyage the 
ordinary mirage effect was conspicuous, that is, dis- 
tant land appeared raised above the sea horizon by 
a small interval, due to the total reflection of sky 
and land at the surface of a thin layer of air of low 
density in contact with the sea. At sunset the last 
segment of the disappearing limb was similarly re- 
flected and reversed, causing a lenticular shape with 
the cusps raised about a minute of arc above the 
horizon. The green flash occurred when the green- 
edged cusps coalesced into a single bright patch, and 
this on one occasion turned to violet at the last 
moment. 
The striking thing about the setting of Venus was 
the sudden appearance of a reflected image moving 
upwards to meet the descending image, and the in- 
stantaneous and conspicuous change of colour from 
NO. 2775, VOL. 111] 
dull red to green at the moment of meeting of the 
two images. The vertical spectrum of the planet 
caused by atmospheric dispersion was at no time 
visible in the binoculars, but the change of colour 
was probably due to the setting of the lower red of 
the spectrum. 
It seems to me evident from these observations that 
the mirage layer greatly intensifies the ordinary dis- 
persion effect, by adding the light from the reflected 
image to the direct image at the moment of setting. 
The normal dispersion effect at sunset under condi- 
tions when there is no mirage is scarcely visible to 
unaided vision, although easily seen in a telescope of 
low power. J. EVERSHED. 
Kodaikanal Observatory, September 26. 

Thermal Opalescence in Crystals and the Colour 
of Ice in Glaciers. 
In a previous communication to NATURE (vol. 109, 
page 42) it was pointed out that the thermal agitation 
of the atoms in crystals causes optical heterogeneity 
which should give rise to a noticeable scattering when 
a beam of light is sent through the substance, and that 
this effect may actually be observed with suitable 
arrangements in clear quartz or rock-salt. I have 
recently found that the same phenomenon is con- 
spicuously exhibited by ice. If a block of clear ice, 
free from air-bubbles, stria, or other obvious inclusions, 
and having flat sides, be held squarely and a narrow 
pencil of sunlight concentrated by a lens be passed 
through it, the track of the pencil shows a beautiful 
blue opalescence. It is advisable not to use a very 
highly-condensed cone of rays, as this would cause 
internal melting of the ice with formation of cavities 
which reflect white light and distract the eye. A dark 
background should be provided against which the 
track may be viewed. With small or irregular lumps 
of ice, the observation may easily be made by im- 
mersing the ice in clear distilled water contained in 
a glass flask which is painted black outside, windows 
being provided for ingress and egress of light and for 
observation of the opalescent track. Even with ice 
which at first looks unpromising owing to internal 
flaws or inclusions, portions in which the blue opal- 
escence is not overpowered by disturbing effects may 
be picked out. A suitable orientation of the block 
with reference to the direction of the incident rays is 
often useful in avoiding reflections from cavities in 
the ice. 
A comparison of the relative scattering powers of 
clear water and of ice at o° C. is instructive. Accord- 
ing to the measurements of Bridgeman, the com- 
pressibility of ice is 35 x 10-* per atmosphere, and its 
refractive index is 1:310, while the corresponding 
figures for water are 52 x1o0-® per atm., and 1°334. 
The Einstein-Smoluchowski formula gives the scatter- 
ing power of water at 0° C. as 144 times that of dust- 
free air, and if it could be applied in the case of solids, 
the scattering power of ice should be 79 times that 
of air. As has already been pointed out, however, 
the formula has to be modified in the case of crystal- 
line solids, and a revised calculation indicates the 
scattering power of ice as about 30 times that of air, 
which is of the order actually observed in experiment. 
The atomic scattering of light in block-ice demon- 
strated and measured in these experiments should 
certainly be capable of being observed on a large scale 
under suitable natural conditions. Indeed, it is well 
known that masses of ice in glaciers and icebergs often 
exhibit a blue colour, and it appears to the writer very 
significant that the circumstances in which natural ice 
shows a blue colour are precisely those found to be 
A2 
