Marcu 24, 1904] 
NATURE 
487 
the deviations of which differ by only 1-5 or 2 2 degrees, and | 
by making an accurate geometrical diagram it will be seen 
that these beams never entirely separate out from each 
other, but continue to overlap no matter how far one passes 
away from the prism. Thus under the conditions of the 
experiment it would hardly be possible to detect the existence 
of separate beams at all. Blondlot does not mention the 
use of a lens to focus the rays, and if one were used it 
would be necessary to re-focus it separately for each beam, 
according to the different values of the indices of refraction. 
In measuring wave-lengths of light by a diffraction 
grating, everyone knows how enormously the intensity of 
the incident light is reduced in the different diffracted 
images, yet Blondlot was able, apparently with the greatest 
‘ease, to split up a divergent pencil of n-rays, coming 
through a slit 5 mm. wide, into eight divergent homo- 
geneous beams by passing it through a prism, then to take 
only as much of one single beam as would pass through 
a second slit 1-5 mm. wide, having perhaps 1/50 the in- 
tensity of the original beam, and after allowing this small 
fraction of the whole radiation to fall on a grating, to 
detect the existence of, and measure up accurately, a central 
image and no less than twenty diffracted images, the in- 
tensity of each of which must have been considerably less 
than 1/1000 of the original beam. All this was done with 
a radiation so feeble that no observer outside of France 
has been able to detect it at all. 
But it is questionable from another point of view whether 
the different diffracted images could be observed at all, at 
least in certain cases, under the conditions of the experi- 
ment, for the slit was quite broad, 1-5 mm., and apparently 
no lens at all was used to bring the spectra to a focus. 
The central beam and the various diffracted beams would 
thus continue to broaden out and become more and more 
diffuse. Now using the ordinary formula for a_ plane 
grating and calculating back from one of Blondlot’s wave- 
lengths, o-oo81m, it follows that for radiation of this wave- 
length the distance apart of adjacent spectral images at 
a distance, say, of 50 cm. from the grating would be only 
.o-8 mm., or considerably less than half the breadth of the 
central beam itself. This is with the grating mentioned 
as having 200 lines to the millimetre. With the grating 
containing 50 lines to the millimetre, the distance between 
adjacent spectral images would be only 0-2 mm., or less 
‘than 1/8 the width of the central beam. In other words, 
there would be no definition, and the broad central band, 
together with the broad diffracted bands, would hardly 
separate out at all, even using as large an angle of in- 
ccidence as 75 degrees. 
In measuring wave-lengths by means of Newton’s rings, 
it is well known that the rings produced by a fairly bright 
source of light, such as a sodium flame, are quite faint, and 
a dark background is necessary in order to see them at all. 
Yet if we accept one of Blondlot’s wave-lengths, 0 0085u, as 
correct, he must have succeeded in counting up no less than 
7o n-ray rings in the space between two adjacent sodium 
rings, and this by the use of a source of radiation only 
1/8 the intensity of the original source, as the latter must 
have been split up into homogeneous beams before the 
rings were formed. It would be interesting to know just 
where the phosphorescent screen was placed in this experi- 
ment, as the rings are formed in the thin air gap between 
the lenses, and the eye must be focused on that point to 
see them sharply. But of course, the screen could not be 
put between the lenses, as the latter could not then be 
brought into close contact, and if it were placed anywhere 
else the rings would be somewhat blurred. 
C. C. SCHENCK. 
McGill University, March 1o. 
Escape of Gases from Atmospheres. 
In a recent number of Nature (January 14) there appears 
an article on the above subject by Dr. G. Johnstone Stoney, 
in which he corrects a statement in the literary supple- 
ment of the Times of December 25, 1903, in regard to the 
escape of helium from the earth’s atmosphere. The 
permanence of planetary atmosphere is of so much import- 
ance to science that I trust I may be permitted through 
your columns to add a word to what Dr. Stoney stated in 
his letter of January 14. : 
The problem of the escape of gases from planetary atmo- 
NO. 1795, VOL. 69] 
spheres has, as Dr. Stoney remarked, been approached by 
two distinct methods :— 
(1) The inductive method, by taking the conditions as 
they appear in nature and arguing upward to results con- 
cerning our atmosphere which may then be applied to other 
planetary atmospheres. 
(2) The deductive method, by using the laws which are 
acknowledged to appertain to gases under known conditions, 
and by assuming conditions under which these laws are 
known to apply for the outer stratum of our atmosphere, 
and to apply these laws to the escape of molecules from the 
atmosphere. 
The first of these methods was made use of by Dr. 
Stoney in his memoir on ‘‘ Atmosphere upon Planets and 
Satellites’? in the Astrophysical Journal, 1898. In this 
paper Dr. Stoney argues that since helium is coming into 
the atmosphere at a greater rate than it is being removed 
from the atmosphere by natural carriers, and since it has 
not been proved to be increasing as a constituent of the 
atmosphere, it must. be escaping from the outer stratum 
of the atmosphere, and in doing so must attain a speed of 
9 27 times its mean velocity at a temperature of —66° C. 
—the velocity that would carry it beyond the earth’s 
attraction. 
In the Astrophysical Journal, January, 1900, I have shown 
by the Maxwell-Boltzmann distribution of velocity that if 
we assume the outer stratum of the atmosphere to be at 
a temperature of 5° C. with a density equal to that at the 
earth’s surface, and to be composed entirely of helium, only 
10:34 X 10-* c.c. of helium would be favourably situated, and 
would attain a velocity sufficient to escape in 10’ years—the 
computed age of the earth; and also that if we assume a 
temperature of 66° C., the number of c.c. that would attain 
to that velocity would be 22-10X 10-**, or less than a single 
molecule in the same length of time; while if we assume a 
temperature of —180° C., which I believe to be much more 
probable for the temperature of the ultimate stratum, only 
91-6 10-*° c.c. will escape, which, of course, means that 
an atmosphere of helium at normal pressure and at the 
average yearly temperature could not escape from the earth. 
If these results, deduced from the kinetic theory under 
conditions to which it is generally acknowledged that the 
kinetic theory does apply, have any value whatever, it seems 
to me that they completely refute the assumption made by 
Dr. Stoney that helium is escaping from our atmosphere. 
But these results do not stand alone as evidence of the 
permanency of our atmosphere. Prof. Bryan by an entirely 
different method (see Transactions of the Royal Society, 
London, 1901) reaches the same conclusion, both in regard 
to hydrogen and helium. 
In the Monthly Weather Review for August, 1902, I 
further discussed the probability of molecules, in a highly 
attenuated atmosphere, reaching velocities much greater 
than under normal conditions, and it is there shown that 
no conceivable effect could influence the results sufficiently 
to allow the escape of helium from the atmosphere. In 
further evidence of the fact that the latter view has been 
accepted by other writers, I may cite the work of M. E. 
Rogovsky (see Astrophysical Journal, November, 1901), 
who, after having published the above article, published a 
note in Nature (July 3, 1902) in which he stated that his 
results would have to be modified in accordance with the 
results obtained for the escape of gases according to the 
kinetic theory. 
In conclusion, permit me to say that although I fully 
recognise the imperfection of the kinetic theory in dealing 
with problems of attenuated atmospheres, yet I believe that 
the results arrived at under the special assumptions made 
will have to stand until it can be shown by other a priort 
reasoning that these conclusions are not within the limits 
of the probable results, i.e. that the escape of helium from 
our atmosphere is practically nil. S. R. Cook. 
Case School of Applied Science, Cleveland, O., 
February 22. 
Demonstration of Magnetostriction by Means of 
Capillary Ripples. 
In his experiments on the change of length by magnetisa- 
tion, Joule (‘‘ Papers,”’ vol. i. p. 50) mentions that ** the ex- 
pansion, though very minute, is indeed so very rapid that 
it may be felt by the touch.’’ If everybody were endowed 
