AuGusT 26, 1915] 
NATURE 
701 

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.] 
The Analogy between Radicles and Elements. 
Tue remarkable chemical analogy of the ammonium 
radicle and the alkali metals may be explained with 
the help of Bohr’s theory of atomic structure. Accord- 
ing to Bohr, the atom of nitrogen consists of a nucleus 
with a positive charge of 7e, surrounded by two rings 
of electrons, the inner containing 4, the outer 3 elec- 
trons. N is, therefore, represented by 7 (4, 3). Ordinary 
chemical compounds are supposed to be held together 
by rings of electrons rotating in planes perpendicular 
to the lines connecting the nuclei of the composing 
atoms; so that NH, may be represented by (4, 3—3). 
This principle leads, however, to difficulties, if applied 
to NH,; a configuration (4, 3—4) seems impossible. 
It is better to suppose the nuclei and electrons to re- 
arrange themselves within the atom. I assume, there- 
fore, that the nuclei unite, or at least that they get 
quite close to each other, in the centre of the system, 
while the electrons arrange themselves in four rings 
(4,4, 2,1), rotating round the joined nuclei. It may 
seem difficult to bring the nuclei together against the 
repelling forces, but it must be remembered that the 
radicle NH, can only be formed by indirect methods; 
in no circumstances will ammonia and hydrogen unite 
to form NH,. The co-operation of another molecule 
containing hydrogen, like H,O, HCl, is absolutely 
necessary, an intermediate compound first being 
formed, from which the ammonium-ion is produced by 
electrolytic dissociation. Once liberated, the ammo- 
nium decomposes immediately; it is extremely un- 
stable, exactly conforming to our expectations. It is 
also important to remark in this connection, that, 
according to Coehn, electric charges are given off 
during the decomposition. 
The arrangement proposed at once shows analogies 
with Bohr’s representation of Na (8,2,1) and 
K (8, 8,2,1). These two configurations contain the 
same outer rings as NH, and differ only with regard 
to the inner rings. The number of electrons in Na 
and NH, is even the same; in K and NH, we find 
the same number of rings. The second similarity 
seems to be of primary importance, ammonium re- 
sembling potassium more closely than sodium. The 
fact that the analogy does not go further than the 
chemical properties follows at once from the general 
principle that the chemical character of a substance 
depends only on the outer rings. 
The same considerations apply, of course, to PH,, 
which, however, is only known as an ion, and AsH,, 
and perhaps to hydrazine, which shows some re- 
semblance to calcium. Whether the substituted ammo- 
nium bases may also be drawn within their scope 
is not yet settled. 
There is still another line of thought to which I 
wish to direct attention. In the periodic table nitrogen 
is placed in the fifth group. When it combines with 
four hydrogen atoms, according to the. above hypo- 
thesis, the nucleus charge increases by 4e; the result- 
ing radicle is analogous to a metal of the first group, 
i.e. to an element which occupies the fourth place on 
the right of the original element. This is exactly the 
reverse of the radio-active transformations in which 
by the loss of an a particle (2e) the atom removes two 
places to the left. 
This idea may also be applied to the analogous sulphur 
and iodine bases. 
NO. 2391, VOL. 95] 

in compounds, in which the hydrogen atoms are sub- 
stituted by aliphatic or aromatic radicles, we shall take 
for discussion the fundamental types; instead of 
SR,OH and IR,OH, we consider the radicles SH, and 
IH.. SH, also has the chemical properties of an 
allxali metal. Now, sulphur is placed in the sixth 
group; if the nucleus takes up three elementary 
charges, 3e, the atom shifts its position by three places 
to the right, and arrives in the first group, thus con- 
firming the general law. Also, the change of the 
iodine atom (seventh group) by taking up two 
hydrogen atoms, 2e, is an agreement with the theory, 
for the iodonium radicle also exhibits the properties 
of a strong base. JTodonium, it is true, especially re- 
sembles thallium, but it is known that thallium and 
silver are closely similar; so iodonium also fits in the 
first group. 
The above considerations may be extended to 
cyanogen and the halogens, which also show a striking 
resemblance. Though cyanogen is not unstable, it is 
not easily obtained. Its formation is only possible at 
high temperatures, and in the presence of alkalis, or 
by the use of electric discharges. It is a strongly 
endothermic compound. The structure of the CN 
radicle may be represented by 13 (4, 4,4,1), whilst 
Bohr writes fluorine 9(4,4,1), and chlorine 17 
(8,4; 4, 1). E. H. BucHNer. 
Chemical Laboratory, University of Amsterdam, 
IGN aao pus 

The Density of Molecules in Interstellar Space. 
In recent years evidence has been brought for- 
ward by several investigators! indicating that light 
from distant stars suffers a slight attenuation in 
travelling through interstellar space. In particular 
a recent investigation by Jones? assigns fairly definite 
numerical values to coefficients of attenuation corre- 
sponding to ‘“ photographic”’ and ‘‘visual”’ light from 
stars of known proper motions and spectral types 
the magnitudes of which had been carefully measured 
by Parkhurst * for light of these wave-lengths. If, as 
seems reasonable, this extinction is assumed to be 
due to attenuation by scattering in travelling through 
a ‘“‘residual’’ gas occupying interstellar space, we 
are enabled to estimate the average density of mole- 
cules in the intervening regions, following a method 
due originally to Larmor? for assigning an upper 
limit to the density of matter in comets’ tails. 
If we denote by x the coefficient of attenuation cor- 
responding to wave-length A, radiation of this 
wave-length originally of intensity E, is reduced, 
after travelling a distance x, to the value given by 
E=E,e-**. According to Rayleigh’s law of mole- 
cular scattering, « is given in terms of the refractive 
index », and the molecular density of the medium n 
by the relation «=§7°(u?—1)?\-4/n. The ability of 
Rayleigh’s law to account almost completely for the 
attenuation of solar radiation in travelling through 
the earth’s atmosphere was first pointed out by 
Schuster 5; a later investigation by the writer,® based 
on the results of the Smithsonian Astrophysical 
Observatory, indicated that formula based on this 
law were competent to explain atmospheric extinction 
as well as to account quantitatively and qualitatively 
ay? 0 . xxix. (1909), pp. 46-54% Xxx. 
(Bi SIN EAS ey aight pias pes 1H => Monthy Notices 
Roy. Ast. Soc., Ixix. (1908), p. 61. King, Pasko Harvard Annals, \ix., 
No. 6, p. 179, April, 1o11: Harvard Annals, Ixxyi., No. 1, pp. 1-10. 1913- 
Brown, F. G., Monthly Notices, \xxii. (1912), P- 195, also p. 718. 
2 Tones, H. S., Monthly Notices Rov. Ast. Soc , Ixxv (1914), pp. 4-16. 
3 Parkhuret, J. A., “Yerkes Actinometry,” Astrophysical Journal, 
xxxvi. (1912), p 160 
4 Larmor, Sir J.. Lectures, Cambridge, 1908. ia Wi ae 
5 Schuster, A., NaTurRE, July 22, 1909; ‘Optics,’ 2nd Edition, 1909, 
P. 329- : - 
Although these are only known | “ © King, L. V., Phil. Trans. Roy. Soc., ccxii.a (1912), Pp. 375-433. 
