314 
NATURE 
[ NovEMBER 13, 1913 
tion of a promising career in science, viz., that 
in each spectral series the magnetic separations 
measured in frequency are the same for all lines, 
and that there is close parallelism for elements 
of the same chemical group, remains the chief 
generalisation in this branch of the subject. It 
was fully confirmed by the much more extensive 
investigations of Runge and Paschen published 
three years later. But in fact the narrower 
foundation on which Preston built may well have 
appeared at the time to be sufficient, in view of 
the pertinent theoretical considerations. 
The fundamental puzzle, why there should be 
definite resolution at all, instead of hazy broaden- 
ing, has already been referred to. The most 
general theoretical system for which definite 
resolution can be predicted remains now, as then, 
one composed of any number of negative elec- 
trons describing orbits, however entangled, under 
their mutual repulsions in a field of force steady 
(or nearly steady), thus due to positive charges 
fixed (or nearly fixed, as they may well be, even 
though free, on account of attached inertia), and 
symmetrical with respect to the axis of the im- 
pressed magnetic field. In such a case the effect 
of an impressed magnetic field H on the system 
is the same as that of an impressed rotation round 
the axis of the field with velocity w=eH/2m; and 
in the analysis of the radiation which the system 
sends out, all its spectral lines are therefore 
divided into normal triplets, i.e. according to the 
elementary Lorentz rule, with the common interval 
w/2m in their frequencies. If a natural spectral 
series had been found to behave differently from 
this theoretical system, it would at that time have 
been a matter for surprise: yet in Runge and 
Paschen’s work, though Preston’s rule is obeyed, 
the resolution proved often to be very different 
from the normal triplet type which is characteristic 
in the proposition above quoted. Instead, how- 
ever, these experimenters found order of a more 
general kind, the components, often more than 
three, being usually symmetrically spaced at inter- 
vals which are equal to or exact sub-multiples 
of the standard Lorentz amount. 
Not a few attempts have been made for the 
theoretical elucidation of this remarkable rule; 
but it probably still remains as a touchstone for 
the next substantial advance in the dynamics of 
molecular structure. Prof. Zeeman rather hints 
his opinion that its range of approximate applica- 
tion may be limited, just as the original standard 
triplet resolution proved to be exact only in 
special systems. Large accumulations of material 
exist for detailed comparative study: the subject 
has in fact- now definitely entered the chemical 
laboratory, and attention is specially directed by 
NO. 2298, VOL. 92) 
| the simplification, in fact fusion, which has been 
our author to the work of J. E. Purvis with Dr. — 
Liveing’s spectroscopic equipment at Cambridge, — 
revealing identical types Of resolution in the 
spectra of numerous elements in which series are 
not as yet known. . 
For further progress on the physical side, much — 
hivher resolving power is a desideratum, which, — 
indeed, is now rapidly being applied. A beginning” 
has been made (by Nagaoka in a recent letter in 
Nature, August 25) in the mapping of the re- 
markable changes of type of resolution of the 
definite satellites attached to certain lines, as the 
magnetic field is increased: this phenomenon, and — 
found by Paschen and Back to ensue in the resolu- 
tion of close multiple lines, when the field becomes — 
very great, and more recently by Fortrat, follow- 
ing early isolated observations by Michelson and 
others, lend weight to Voigt’s hypothesis of some 
kind of vibrational linkage between adjacent lines, 
even when their own modes of resolution are of 
different types. 
Such difficulties as these have obstructed the 
general theory, as approached from the side of the — 
radiation from magnetised flames. But at an 
early stage Voigt had formulated the problem— 
and has since developed it in many directions, 
analytically and experimentally, with his usual 
mastery—from the point of view of propagation 
of incident radiation through a magnetised 
medium, a subject already discussed for trans- 
parent media in theories of Faraday’s rotation of 
the plane of polarisation and of the related Kerr 
effect of reflection. If that type of theory is ex- 
pressed so as to exhibit the mechanism of selec- — 
tive absorption, by the explicit introduction of — 
terms appropriate to molecules vibrating by reson- 
ance and attached to the medium, and also of 
general damping terms when expedient, a dark 
narrow band which would be single in the 
absence of an impressed magnetic field should 
become resolved into Zeeman components 
when such a field is included; or at any rate 
this fact will be a guide to the form of the 
equations. 
Almost simultaneously with this theoretical dis- 
cussion, the Italian physicists Macaluso and 
Corbino broke the cognate experimental ground, 
by the detailed observation of an absorption line 
under very high dispersion, showing that the 
known excessive and anomalous refraction at its 
borders was accompanied by excessive and anoma-— 
lous magnetic rotation, superposed on the mag- 
netic resolution of the line. Indeed, very soon 
after Zeeman’s first discovery, Righi had put the 
resolution of the line in evidence in a most effective 
and beautiful manner, in an absorption experi- 
