June 4, 1914] 
NATURE 
349 
LETTERS. LO 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. ] 
Efficiency of Damped Seismographs. 
In Nature (April 2, 1914, p. 119) I found reprinted 
a statement of Dr. Cavasino, concerning the influence 
of damping on recording seismographs, which I think 
is based upon a misunderstanding. 
There is no doubt that when damping is introduced, 
until the limit of aperiodicity is attained, the proper 
period of the instrument exists as such. But it is not 
that point at all which is important, but the fact that 
more or less heavy damping diminishes the influence 
of the proper motion of the instrument on the record 
and enables one to obtain a much more trustworthy 
picture of the true motion of the ground. 
Further, if you have an instrument of low sensitive- 
ness and simply introduce a heavy damping you will 
make a very bad seismograph; but nobody proceeds in 
this way. When damping is introduced, one must 
simultaneously provide to augment the magnification, 
be it by introducing magnifying levers like those in 
Wiechert’s instruments, or by using galvanometric 
registration as in the aperiodic seismographs. The 
lengthening of the proper period of the instrument also 
gives good results for the longer seismic waves. 
If these precautions are taken, heavily damped in- 
struments give absolutely the same moments for the 
commencement of both first preliminary phases of an 
earthquake as undamped ones; moreover, the three 
different components give absolutely the same results. 
There can be no question of a difference of several 
minutes, not even of seconds, as Cavasino states. The 
difference in the times of arrival of the first longi- 
tudinal waves for the aperiodic instruments of the 
Pulkovo seismological station for the three components 
of the movement of the ground differ from another 
only by a fraction of a second. 
Cavasino’s assertion, that damped pendulums give 
fewer records of earthquakes than undamped ones only 
depends upon the way in which the damping is intro- 
duced, and as a general statement does not hold good. 
In fact, the aperiodic instruments in Pulkovo regis- 
tered in 1912, 671, and in 1913, 576 earthquakes; the 
number of azimuths of the epicentre determined at 
Pulkovo were, in 1912, 137, and in 1913, 103. 
In 1913 the number of earthquakes registered by 
aperiodic instruments were :—In Tiflis, 456; Irkutsk, 
738; and Taschkent, 954. 
I doubt whether any other. seismological station 
using undamped seismographs has approached these 
figures. B. GatitTzin. 
Laboratoire de Physique de l’Académie 
Impériale des Sciences, St. Pétersbourg, 
May II. 
Spectra of Secondary X-Rays. 
HirHerto all X-ray spectra have been obtained by 
using the body, the spectrum of which is to be 
examined, as an antikathode inside the tube. - All the 
trouble of exhausting the tube, etc., can be avoided by 
illuminating the substance of which the spectrum is 
to be determined with the primary rays from a tube 
of any of the usual types, and only allowing the 
secondary rays to enter the spectrograph. I have 
photographed the secondary spectra of copper, zinc, 
mercury (amalgam of zinc), ete., by this method with- 
NO. 2327, VOL. 93] 
out any trouble. Zinc amalgam, for instance, shows 
five rays, two due to zinc, one due to some impurity— 
probably iron—and two due to mercury. 
If a quantity of the substance to be examined is 
placed in a thin paper bag, the experiment is par- 
ticularly striking. Using zinc oxide, for instance, the 
presence of zinc was evident immediately, so that the 
chemical analysis of a substance inside a _ closed 
envelope without in any way touching or manipulating 
it is no longer beyond the bounds of possibility. 
The method promises to be particularly useful for 
any experiment in which the rays emitted under 
various conditions are to be examined, such as those 
on the effect of an electric or magnetic field upon the 
spectra, with which I am now engaged. 
MaurRIcE DE BROGLIE. 
29 Rue Chateaubriand, Paris, May 30. 
Weather Forecasts. 
At the conclusion of my former note on this subject 
(NaturE, February 26, 1914, vol. xcii., p. 711) I said 
that it seemed improbable that trustworthy forecasts 
of the weather for twenty-four hours in advance ever 
would, or could, be made for latitudes far removed 
from the equator, and in the present communication 
I give the reasons for that opinion. 
Enough is now known concerning the average weather 
conditions on the globe to show that were it possible 
to make the surface wind currents visible and to 
observe their distribution from a distance, the appear- 
ance would be 
very much like 
that given in 
the accompanying 
figure, provided 
that the surface 
was level and 
uniform in 
quality; - 7.¢./.. ail 
land or all sea. 
The wind cur- 
rents on the actual 
earth if viewed 
in this way would 
no doubt be seen 
to be considerably 
affected by differ- 
ences in the 
nature of the 
surface over 
which they passed, more particularly where that 
surface was mountainous, but the general character 
of the flow would still be that given in the figure, 
namely, that round. the equator (leaving seasonal 
variations out of account), there would be a region of 
calms bordered on each side by fairly regular trade 
winds extending roughly to latitude 30° N. and S.; 
whilst outside this region again, the whole surface 
would be covered by eddies of various sizes in which 
the direction of circulation was left-handed in the 
northern, and right-handed in the southern hemi- 
sphere. 
If pressure could be observed, as well as the direc- 
tion of the wind, the central parts of the eddies would 
always appear as regions of low barometer. Were 
the observations extended over a few hours, it would 
be seen that the eddies themselves (i.e. their centres) 
were in motion travelling on the whole to the N.E. 
and S.E. in the northern and southern hemispheres 
respectively. 
Continuing the observations for days or weeks, it 
would be found that the eddies were mostly short- 
lived, few lasting more than two or three days, and 
that although their average course was N.E. or S.E., 
