May 7, 1914] 
NATURE 
241 
work-like and flocculent appearances so often observed 
in the froth which is formed when the tide breaks on 
the seashore may be explained in a similar manner. 
HaRo_D WaGER. 
West Park, Leeds, April 14. 
An Extension of the Spectrum in the Extreme 
Ultra-Violet. 
Tue researches of Schumann led him to extend 
the spectrum to the neighbourhood of wave-length 
1250, His limiting wave-length was determined by 
the absorption of the fluorite which formed a neces- 
sary part of his apparatus. In 1904 I succeeded in 
pushing the limit to wave-length 1030 by the use of a 
concave diffraction grating. — 
Recently I have renewed the attack on the problem, 
with the result that I have succeeded in photograph- 
ing the spectrum of hydrogen to wave-length 905. 
The extension is due, not so much to any fundamental 
change in the nature of the apparatus as to an im- 
provement in technique consequent on an experience 
of ten years. ; 
It is a characteristic of the region investigated by 
Schumann between wave-lengths 1850 and 1250 that, 
while hydrogen yieids a rich secondary spectrum, with 
the possible exception of one line, no radiation has 
been discovered belonging to the primary spectrum. 
On the other hand, in the new region between the 
limit set by fluorite and wave-length 905, a disruptive 
discharge in hydrogen produces a primary spectrum 
of great interest made up of perhaps a dozen lines. 
These lines are always accompanied in pure hydrogen 
by members of the secondary spectrum, but they may 
be obtained alone if helium containing a trace of 
hydrogen is employed. — 
Results obtained from vacuum tubes when a strong 
disruptive discharge is used, must always be inter- 
preted with caution since the material torn from the 
tube itself sometimes furnishes impurities. In the 
present case, it will be some time before the effect of 
such impurities can be estimated. However, it mav 
be stated with some degree of certainty that the 
diffuse series predicted in this region by Ritz has 
been discovered. The first member at 1216 is found 
to be greatly intensified by the disruptive discharge, 
and the next line at 1026 appears also, though very 
faintly. This diffuse series bears a simple relation to 
Balmer’s formula. Following the same kind of argu- 
ment, a sharp series corresponding to the Pickering 
series might be expected. The new region appears 
to yield two lines belonging to such a relation at the 
positions demanded by calculation. 
THEODORE LyMAN. 
Harvard University, April 20. 
The Structure of Atoms and Molecules. 
SINCE in an elaborate criticism of Bohr’s theory on 
the constitution of atoms and molecules, Prof. J. W. 
Nicholson, as in his letter to Nature (February 5, 
p 630), comes to the conclusion (Phil. Mag., xxvii., 
p- 560, 1914) that the valencies of lithium, beryllium, 
boron, ete., on Bohr’s theory are not in accord with 
experience, and if the electrons in the atoms are to be 
in one plane, we must either abandon Bohr’s method 
of calculating valency—and - (generally) | Bohr’s 
theory of the atoms more complex than hydrogen and 
helium—or give up van den Broek’s hypothesis, that 
the charge of the nucleus of Rutherford’s atom is 
equal to the atomic number (which hypothesis was 
accepted by Bohr as one of his fundamental assump- 
tions), I may be allowed to add some remarks to my 
previous letter on this subject (Nature, ‘March 5, 
1914). 
NO 423, VOLi.93) 
For these atoms at least this hypothesis is a mere 
expression of experimental facts. The hydrogen atom 
is known to lose never more than one electron, and 
the helium atom never more tnan two, and, of course, 
never one to form an electrolytic ion, while lithium, 
beryllium, boron, and carbon can lose, or, in chemical 
combination, dispose of 1, 2, 3, 4 electrons respectively. 
Further, the number of electrons per atom has been 
proved to be nearly equal to half the atomic weight 
(Rutherford, Barkle), and in the case of carbon to be 
six (Rutherford, Phil. Mag., vol. xxvi., p. 711, 1913). 
Since the number of electrons per atom must be 
an integer, here, at least, no other solution seems to 
be possible than that the number of electrons per 
atom surrounding the nucleus, and hence the nuclear 
charge, is equal to the atomic number. 
Further mentioning Moseley’s previous experiments 
on high-frequency spectra (undertaken for the express 
purpose of testing the atomic number hypothesis), and 
criticising the theoretical deductions, derived by Mose- 
ley from these experiments, Nicholson concludes that 
they have shown no relation to Bohr’s theory (loc. cit., 
p- 564). Now in another paper Moseley, from further 
experiments on high-frequency spectra, proves (Phil. 
Mag., vol. xxvii., p. 703, 1914) that the frequency of 
any line in the X-ray spectra is approximately propor- 
tional to A(M—b)*?, where A and 6b are constants for 
each series, and M, the atomic number (called by 
Moseley N) of the element, is identified with the num- 
ber of positive units of electricity contained in the 
atomic nucleus, so that these experiments “‘ give the 
strongest possible support’’ to this atomic number 
hypothesis (loc. cit., p. 712). The number of rare- 
earth elements as given by Moseley is the only excep- 
tion. 
That b is much larger for -he ‘‘L” lines than for 
the “K” lines suggests, according to Moseley (in 
agreement with my own views, Nature, December 
25, 1913) that the ‘‘L”’ system is situated the further 
from the nucleus. If so, b=the number of electrons 
nearest the nucleus, and not=oc,, the term arising 
from the influence of the electrons in a ring on each 
other, and, for the ‘“‘K ’ lines, n, like b, must be 
unity, as calculated by Nicholson on Bohr’s theory. 
For the ““L” lines, according to Moseley, b=7-4, but 
it can easily be seen from the tables that if (M—b) 
be here augmented by o-8 per cent., all values are 
integers (+0-2), and b=7 and n=1 again, but perhaps 
the factor 5/36 in Moseley’s interpretation cannot be 
retained. 
Hence, though this number 7 requires confirmation, 
principally, for the ‘‘K” line at least, Bohr’s 
theory is here in agreement with Moseley’s experi- 
ments, and with the atomic number hypothesis. Not 
only the frequencies, but also the minimum velocity 
of electrons requited to excite this radiation, and the 
absorption of it (in aluminium) have been proved (loc. 
cit.) to depend on the atomic number very nearly, and 
Nicholson’s conclusion that the atomic numbers are 
not correct does not hold, for (M—b), not M, is one 
unit less for the K radiation than the corresponding 
atomic number. But, from analogy, Bohr’s lithium 
atom, as well as Nicholson’s ring of three electrons, 
must be given up, for of three, one electron (b) must 
be very near the nucleus, one (n) near but outside this 
first one, and one as electron of valency must be 
peripheric. 
Further, the velocity of electrons, required to excite 
this radiation, according to Widdington equal to 
10° x atomic weight cm./sec., is more accurately equal 
to 2-24 x 10°(M—1:) cm./sec., than for Cr, Fe, Ni, Cu, 
Zn, and Se; the last formula gives for the constant 
reduced to unity 0-99, 1°04, 1-02, 1-00, 0°97, 1-00, while 
the first gives 0-99, 1-05, 1°06, 0:99, 0:98, 0°94 respec- 
tively. Since the absorbability of the excited radia- 
