394 TRANSACTIONS OF SECTION A. 
element to element in general. Comparison has been made in each case with the 
iron scale, as shown below, for titanium, lanthanum, and cerium :— 
Intensity . | Ti—Fe La—Fe | Ce—Fe 
1 |  —0-002 +-0-003 | +0-007 Ang. 
2 —0-003 -+0-003 
3 —0-002 +0-007 +0-007 
4 —0-004 
5 | 0-004 
Weighted Mean . | —0-0026 +-0-0033 --0-007 
On the assumption that the displacements may ke used as a criterion of level, 
the heavy elements lanthanum and cerium are low-level elements. By comparing 
all the elements with iron a distribution of the elements in the solar atmosphere 
is found that resembles that in the terrestrial atmosphere. As one penetrates 
the solar atmosphere from without the sun he would first encounter ‘that form of 
calcium vapour that produces the H and K lines’; to this is successively added 
hydrogen and the vapours of magnesium, sodium, iron, &c., each increasing in 
absolute density with the depth until in the lowest portion of the reversing layer 
occur also the vapours of all the elements whose lines appear in the solar 
spectrum. The indicated distribution confirms that shown by Adams’ measures 
of the solar rotation, and is in harmony with a wide range of solar observations 
that, taken separately, have shown indications of differences of level, such as the 
flash spectra, the weakening and strengthening of the lines in spots, the differ- 
ences between centre and limb spectra, and the variation of the magnetic field 
due to the spot vortex with level. From this point of view the displacements of 
the Fraunhofer lines in the penumbre of sun-spots give a means of sounding the 
solar atmosphere and of assigning relative levels to the sources of the lines, and 
open the way to further solar research. 
2. On the Fourier Sequence as a Substitute for the Periodogram. 
By Professor H. H. Turner, F.R.S. 
1. The use of the periodogram formed the main topic of Professor Schuster’s 
Address to the Sub-section of ‘Cosmical Physics in 1902. Hence it is appropriate 
that any alternative procedure should be brought before Section A. 
2. It is suggested that the Fourier coefficients should be calculated for exact 
submultiples of the whole available period of observation, and for these only. 
The advantages of this procedure are :— 
(a) That it is necessary. Since it is only a portion of the procedure advocated 
as necessary by Professor Schuster, this will probably be conceded. 
(6) That it is sufficient. It can readily be shown that the coefficients for 
any other period can be deduced from these Fourier constituents. 
(c) That it is convenient in practice. It directs attention to definitely 
specified periods, whereas the method of the periodogram proceeds by trial and 
error, or some equivalent. 
(d) It is instructive, especially in showing what will remain if any single 
periodicity or group of periodicities is eliminated from the observations. 
The first fifty-four terms of the series for sun-spots have been calculated 
and the calculation of other terms is being undertaken. 
3. Photographic and Spectrographic Arrangements of a Reflector. 
By J. H. Reynotps. 
4. Solar and Terrestrial Magnetic Disturbances. By Rey. A. L. Corti. 
Sun-spots are not only indices of the general state of solar activity, but are 
also most frequently the centres of disturbed regions covering considerable areas 
on the sun’s surface. The zones of sun-spot activity are also closely connected 
