States and other universities should be framed by joint 
representatives of all our universities. 
It would be a misfortune if anything in the nature 
of competitive schemes for attracting students were 
to be evolved independently by different university 
authorities. J. B. Couen. 
Leeds, December 23. 
The Deterioration of the Atmosphere in the Swiss Alps. 
Dr. Maurer, Director of the Swiss Federal Meteoro- 
logical Service, sends a further communication about 
the deterioration of the atmosphere of the Swiss Alps, 
referring to Prof, Riccé’s letter in Nature of Novem- 
ber g on the occurrence of an eruption at Stromboli. 
I am forwarding a translation for your information. 
Napier SHaw. 
Meteorological Office, South Kensington, 
London, $.W., December. 11. 
This remarkable optical deterioration of the atmo- 
sphere was visible here (Zurich) until about the middle 
of November. The thin, cirrus-like layer could ‘be seen 
on clear mornings just before sunrise at a height of 
about 14-15 km. above the earth’s surface, according 
to our reckoning—that is to say, it was situated con- 
siderably above the usual cirrus region. It consisted 
of thin horizontal bands, extremely delicate and soft, 
which soon disappeared after sunrise. A curious fact 
was that no appreciable effect, either actinometric or 
photometric, was produced by this thin, mist-like layer. 
The impression made was that of a most delicate, 
comet-like veil of mist, yet not dimming the starlight. 
After sunrise absolutely nothing was to be seen of 
the phenomenon, in spite of the keenest observation 
through field-glasses of a weak magnifying power. 
Synchronising with this remarkable phenomenon, the 
sun had a large aureole with a diameter of 100°. Here 
and there the extreme outer edge of this ring was of 
a pale brown colour. So far we have not been able 
to offer any explanation as to the cause. 
During the period of maximum visibility of the thin 
veil (twilight cirrus), such a conspicuous layer was to 
be seen in the eastern sky, shortly before sunrise and 
at a height of 40°, that even an unskilled observer 
would have noticed it at once.° The structure of the 
layer was often so regular and definite in its remark- 
able horizontal stratification that it looked as though 
an artist with a coarse brush had coloured the whole 
eastern sky with long horizontal strokes not too neatly 
laid on. 
Each time this peculiar veil of “twilight cirrus” 
reached a maximum of intensity we had a colourless 
morning twilight with interrupted “purple light.” 
What can be the cause of all these remarkable pheno- 
mena? 
Winter Thunderstorms. 
_A YEAR ago (December 16, 1915) I asked readers of 
Nature if they would let me know when they observed 
thunder or lightning during the first three months of 
the year. I made a similar request to the observers 
of the British Rainfall Organisation. The number of 
replies I received amounted to nearly one thousand, 
and the rather remarkable fact came to light that 
during the period in question thunder or lightning 
occurred somewhere in the British Isles on sixty-four 
out of the ninety-one days. So numerous were the 
replies received that I was unable to answer each one 
personally, but I wish to thank those correspondents 
who kindly -sent information, and to assure them that 
every report was of value. 
The information obtained last winter was so re- 
markable, and the number of days on which thunder 
NO. 2461, VOL. 98] 
NATURE 
[DECEMBER 28, 1916 
or lightning occurred so unexpected, that I am 
anxious to collect information again. Readers of 
Nature could assist if they wouldgsend me a note by 
postcard or letter if they observe thunder or lightning 
between January 1 and March 31, 1917. The follow- 
ing points are of interest :—(1) Time when storm was 
nearest or overhead; (2) direction of storm when first 
observed, and time; (3) direction when last observed, 
and time; (4) note if there was a change of wind 
during the storm and if there was a drop in tem- 
perature; (5) any other information as to heavy rain, 
hail, snow, or any remarkable feature; (6) if an 
observer has accurate time, a list of the times of 
occurrence of flashes would be useful. 
Many observers may not have the time or oppor- 
tunity to record all these points, but I should be grate- 
ful for information on even one of them; No. 1 is the 
most important. The information is of real value, 
and every record, however short, is of use. 
C. J. P. Cave. 
Meteorological Office, South Farnborough, 
December 20. 
GRAVITATION AND THE PRINCIPLE OF , 
RELATIVITY. 
pa Geass gia to the principle of relativity in 
its most extended sense, the space and time 
of physics are merely a mental scaffolding in which 
for our own convenience we locate the observable 
phenomena of Nature. Phenomena are conditioned 
by other phenomena according to certain laws, 
but not by the space-time scaffolding, which does’ 
not exist outside our brains. As usually expressed, 
the laws of motion and of electrodynamics pre- 
suppose some particular measurement of space and 
time; but, if the principle is true, the real laws 
connecting phenomena must be independent of our 
framework of reference—the same for all systems 
of co-ordinates. Of course, it may be that pheno- 
mena are conditioned by something outside ob- 
servation—-a substantial azther which plays the part 
of an absolute frame of reference. But the fol- 
lowing considerations may show that the ideal 
of relativity is not unreasonable. Every observa- 
tion consists of a determination of coincidence in’ 
This is sufficently obvious in 
laboratory experiments; and even the crudest 
visual observation resolves itself into the co- 
incidence of a light-wave with an element of the 
human retina. If, then, we trace the path of 
adventure of a material particle, it intersects in 
succession the paths of other particles or light- 
waves, and these intersections or coincidences con- 
stitute the observable phenomena. We can repre- 
space or time. 
sent the course of Nature by drawing the paths — 
of the different particles—on a sheet of paper in 
a two-dimensional case. The essential part of the 
diagram is the order of the intersections; the 
paths between the intersections are outside ob-. 
servation altogether, and are merely interpolated. 
The sequence of phenomena will not be altered if 
the paper is made elastic and deformed in any 
way, because the serial order of the intersections 
is preserved. This deformation of the paper cor- 
responds to a mathematical transformation of the. 
space in which for corvenience we have located 
the phenomena. 
