JUNE 22, 1899] 
It is clear that if the broadened line is really a triplet, 
then the components of this triplet must be so close 
together that they overlap each other, and so appear to 
the eye merely as one broad line, as illustrated by the 
model which is here before you. [Model illustrating 
the overlapping shown here.] We know that the 
spectral lines are not infinitely narrow lines, but are 
really narrow bands of light of finite width, and conse- 
quently we are quite prepared to regard the magnetically 
broadened line as an overlapping triplet, but we cannot 
remain satisfied until we have proved beyond all doubt 
that it really is a triplet, and not merely a single broad 
line. Todo this, Dr. Zeeman made use of the second 
prediction of the theory—namely, that the constituents of 
the triplet must be plane polarised. If this is so, then the 
outer edges of the broadened line must be plane 
polarised, and therefore by introducing a Nicol’s prism 
into the path of the light it must be possible to turn the 
Nicol so that the plane polarised edges shall be cut off, 
and the breadth of the line shall be reduced to its normal 
amount. In fact in this position of the Nicol the out- 
side lines of the triplet are extinguished, and the central 
component alone remains. This component is, of course, 
the same in width as the original line, and consequently 
when the outer members of the triplet are extinguished 
all the magnetic broadening of the line is removed. 
When the Nicol is turned through a right angle the 
central component of the triplet is extinguished, while 
row shows the lines uninfluenced by the magnetic 
The row next above shows the same lines broadened by the 
Fic. 2.—The lower 
field. 
magnetic field. 
Nicol’s prism. 
The top double row shows the analysation by a 
(Reproduced natural size.) 
the side lines remain, and, if these side lines are suffi- 
ciently separated, so that they do not overlap, then, 
when the central line is removed, a narrow dark space 
will exist between the side components, which represents 
the space intervening between the outer members of the 
triplet, as illustrated by Fig. 2. 
But even though we may be able to so increase the 
strength of the magnetic field that when the central com- 
ponent of the triplet is removed by a Nicol the side lines 
stand apart witha clearly defined interval between them, 
yet this in itself does not absolutely satisfy us that the 
broadened line is a triplet. It might be contended that 
the broadened line is not really a triplet, but is merely a 
band of light polarised in one plane along its edges, and 
in the perpendicular plane along its centre, and that in- 
crease of the magnetic field might never separate it into 
distinct constituents, but merely continue to broaden it. 
This contention, however, might be disposed of by a 
careful study of the facts, even though we might not be 
able to produce a magnetic field strong enough to com- 
pletely separate the constituent lines of the triplet. 
Actual Triplets Obtained. 
But clearly the thing to be arrived at is to so arrange 
matters, in fact to so design our electro-magnet and to 
plan the conditions of our experiment, that the magnetic 
NO. 1547, VOL. 60] 
NATURE 
177 
field acting on the source of light shall be strong enough 
to completely separate the members of the triplet if such 
exist. You will understand that this is no easy thing to 
do when you remember that it was only after repeated 
efforts and many failures that even a slight broadening 
of the spectral lines was obtained. Nevertheless, in spite 
of the great difficulty which besets this investigation, and 
which arises from our inability to obtain a magnetic field 
of unlimited strength, yet, with a properly designed 
magnet and other properly arranged conditions, it is 
possible to obtain a magnetic field strong enough to com- 
pletely separate the constituents of the magnetic triplet, 
and thus to prove that the prediction of theory is verified 
by the actual facts. 
Other Perturbations, Complex Types. 
But with a magnetic field of great strength the facts as 
shown by these slides [photographs shown here] turn out 
to be more complicated and more interesting than the 
simple theory led us to expect. For while some of the 
spectral lines are split up into triplets as indicated by 
The top lines are not subject to the influence of the magnetic 
FIG. 3. 
field. Underneath the same lines are shown affected by a magnetic 
ee 
int 
field of increasing strength. The line on the right is resolved 
a pure triplet, while that on the right appears at first as a quartet, 
and finally in a very strong field as a sextet (easily seen on the 
negative). (Reproduced natural size.) 
theory, some, on the other hand, become resolved into 
sextets, or octets, or other complex types (see Fig. 3). 
Thus when the magnetic field becomes sufficiently in- 
tense, we realise to the full all the theoretical predictions 
and more. The reason of this surplus of realisation 
over expectation lies in the fact that the theory in its 
simplest form deals only with the simplest types of motion 
under the simplest conditions, and the conclusions 
arrived at are of course of corresponding simplicity. 
When more complicated types of motion are contem- 
plated, the theory furnishes us with the dynamical 
explanation of the more complicated types of effect 
produced by the magnetic field. That tripling pure and 
simple should occur in the case of every spectral line (as 
predicted by the simplest form of theory) is not a result 
which we should expect from a broader consideration of 
the problem. In fact, if we reflect on the subject, we are 
forced to the conclusion that deviations from the pure 
triplet type should be expected, and, as we have seen, 
such deviations actually do occur. In this respect, 
