416 TRANSACTIONS OF SECTION A. 
change of energy only when an electron enters or leaves the atom. Such systems 
constitute unstable elements of a simpler nature than those on the earth. The 
spectra of such elements do not exhibit the usual series, but a series in which 
the cube roots of the wave-lengths differ by a constant amount. This is in 
accordance with a radiation of energy in discrete amounts proportional to the 
frequency. 
The main difficulty in supposing that the dynamical vibrations of systems of 
electrons can give spectrum series such as Balmer’s lies in the fact that the 
square of the frequency is then a rational function of integers, and not the 
frequency itself. This difficulty is absent from the model of Ritz. But it is 
more probable that Lord Rayleigh’s suggestion is correct—that the origin of 
spectral series is kinematical rather than dynamical. 
A process was sketched by which a series of lines 
n* 
om? — at 
(a = constant) can be obtained for an atom with two rings of electrons by 
simple kinematical principles. If the outer ring contains only one electron, the 
lines are doublets. Hydrogen is perhaps built up in this way, which is in accord 
with its unit valency. The infinite number of lines is due to the infinite number 
of degrees of freedom of the ether. 
a 
The following Report and Papers were then read :— 
1. Report of the Seismological Committee.—See Reports, p. 69. 
2. Some Notes on Periodograms. By Professor H. H. Turner, F.R.S. 
An analysis of the catalogue of large earthquakes indicates a periodicity of 
nearly fifteen months, which is dealt with in the report of the Seismological 
Committee. Further examination suggests that some of the deviations there 
shown may be due to the existence of neighbouring periodicities which have not 
yet been fully examined. But attention has been concentrated on the existence 
of pairs of periodicities, or groups of periodicities (analogous to double lines or 
groups of lines in a spectrum), by related work on the variations of level and 
azimuth of the Greenwich and Cape transit circles. As an example the following 
results may be tabulated for periods near twelve months, hitherto perhaps 
obscured by the well-known twelve-month period :— 
Azimuthal Variations near Twelve Months. 
Shee: Coefficient Phase 
Period in Se 7 et 
Months | Greenwich Cape G0. HE Gia 
11-76 0-20 2'"-6 294° 106° =| ~—«188° 
11:60 0-19 1-0 315° 145° =| 1702 
12:40 0-16 1-0 2150 48° «=| ~—(167° 
The 11°76 month period occurs in the Greenwich rainfall (eighty years’ 
observations). Such independent confirmation seems to render the real existence 
of these close periodicities very probable. The examination of such points is 
being continued. 
2 
3. The Periodogram and the Method of Correlation. 
By J. I. Crate, F.R.S.E. 
1. At Portsmouth last year, Professor Turner gave an explanation of Professor 
Schuster’s method of analysing a series of figures for suspected periodicities, and 
applied the method to examine seismic records for possible periods. Professor Karl 
