TRANSACTIONS OF SECTION A. 299 
been made to apply Planck’s theory of radiation to the explanation of the laws 
of spectra. The most ingenious and suggestive is that of Bohr. It is based on 
the Rutherford atom, but throws no further light on the structure of the atom 
itself, as the mechanism of radiation is totally unexplained, and it is this 
which we are in search of. The most remarkable result is the derivation of the 
value of Rydberg’s constant from known electric constants, and Planck’s con- 
stant. This result has certainly caught the scientific imagination, and one feels 
convinced, especially on a first reading of his paper, that there is some truth 
at the bottom of his theory. But Lindemann has pointed out, by consideration 
of dimensions, that a large number of theories would give values in which the 
various constants enter in the same way. In Bohr’s theory the exactness of the 
numerical relation depends on an apparently arbitrary assumption as to the 
frequency of the energy emitted when an electron is combined. It is true that 
later he attempts to justify this by making his formula conform to certain 
observed properties of series. But with the introduction of this his value of 
Rydberg’s constant ceases to be a direct deduction from his theory. Moreover, 
in doing so he assumes the frequency of an electron to be proportional to its 
angular velocity, which can only be the case for one electron—i.e., for an atom 
built on a planetary system, and not on a Saturnian, as is his. Nicholson has 
recently criticised the theory on other grounds, and as he is to take part in the 
discussion I will leave this point to him. From the spectral point of view the 
weightiest objection would seem to be that it is capable only of giving a 
formula of the Balmer type—which holds for hydrogen alone. In the best- 
known series types, the P, 8, and D depend on formule of the type pe ea) 
The wdepend on atomic constants, and are always considerable for elements of 
large atomic weight. As the atomic weight diminishes we get the following 
general changes:—In P, pm decreases to 1; in D it increases to 1; in § it 
approaches the value ‘5. In other words, for P and D the formule approach a 
Balmer type. For H it is indistinguishable from Balmer’s. For He, though 
approaching Balmer’s, it is decisively not 1. As a fact, Béhr’s theory does 
not represent any of the six known series of He, but he postulates that certain 
lines hitherto allotted to H belong really to He. Moreover, he supposes He to 
have 2 electrons, whereas, as I have attempted to show above, the number 
is more probably 4. Fowler has recently presented a paper to the Royal 
Society in which he supports Béhr’s allocation on observational grounds, but as 
it is not yet (August) published it is not possible to weigh the evidence. In 
concluding, I should like to say that although I have criticised certain parts of 
Bohr’s theory adversely, no one can admire more its ingenuity and great 
suggestiveness. 
Mr. H. G. J. Mosety explained the results of his classification of elements 
by their X-ray spectra. The frequency of the principal line in the X-ray 
spectrum is represented very closely by the formula 
yt=K (N—B) 
where K and B are constants, and N an integer increasing by a unit as we pass 
from element to element up the periodic table. If we take this atomic number 
N as ordinate, and the square root of the principal frequency as abscissa, the 
different elements will therefore give points lying approximately on a straight 
line. The secondary frequencies will at the same time give points on other 
straight lines. The order of the elements determined by N is nearly that of 
increasing atomic weight; there are one or. two exceptions, and in such cases the 
order given by N, and not the atomic weight, is evidently the correct order cor- 
responding to chemical properties. For example, the atomic weight gives the 
order Cl, K, A, whereas the X-ray frequency gives the order Cl, A, K. The 
latter is the order required by the periodic table. There are between aluminium 
and gold four missing elements, indicated by the double jump of N required to 
make the formula fit. These correspond generally to gaps indicated also by the 
periodic law. 
Professor Nricuotson : I prefer not to introduce new difficulties, which would 
ouly make the discussion too long, and will therefore confine my remarks to 
