622 
have been manufacturing carbohydrate material within their 
chloroplasts at least three-and-a-half times faster than those in 
normal air, and, although this rate of photosynthesis would per- 
haps not be maintained for very long, yet there would always bea 
general tendency for the carbohydrate supply in the leaves to be 
kept up to a higher point than in the controls grown in ordinary 
air, a fact which was shown by the leaves of set B always being 
gorged with starch. 
Since it is quite certain that this increased photosynthesis does 
not to any material extent contribute to the increase of dry 
weight of the plants, we can only conclude that the transforma- 
tion, translocation and general metabolism of the leaf-reserves 
under these conditions cannot keep pace with the increased 
tendency to produce an extra amount of plastic material from 
the atmosphere. Moreover, it is clear that the whole mechan- 
ism of the plant on which normal nutrition depends has its parts 
so completely and accurately correlated that any slight increase 
in the composition of the surrounding air which favours in- 
creased photosynthesis destroys the adjustment of the various 
parts and results in a more or less abnormal development of the 
plant. That any such disturbance of the economy of the plant 
should profoundly modify the reproductive functions might 
perhaps have been expected. 
It is somewhat remarkable to find that all the species of 
flowering plants, without exception, which have been the sub- 
ject of experiment appear to be accurately ‘‘tuned” to an 
atmospheric environment of 3 parts of CO, per 10,000, and 
that the response which they make to slight increases in this 
amount are in a direction altogether unfavourable to their 
growth and reproduction. It is not too much to say that a com- 
paratively sudden increase of carbon dioxide in the air to an 
extent of about two or three times the present amount would 
result in the speedy destruction of nearly all our flowering 
plants. 
To a certain extent, we may regard the facts recorded in this 
paper as indicating that the composition of our atmosphere as 
regards its carbon dioxide has remained constant, or practically 
constant, fora long period of time, but the authors leave altogether 
untouched the question of any variations of a secular kind. All 
we are justified in concluding is that if such atmospheric 
variations have occurred since the advent of flowering plants, 
they must have taken place so slowly as never to outrun the 
possible adaptation of the plants to their changing conditions. 
MAGNETO-OPTICAL ROTATION IN THE 
INTERIOR OF ABSORPTION BANDS. 
AN interesting confirmation of Voigt’s theory of absorption 
- has been afforded by an experiment by Prof. Zeeman, de- 
scribed in the Proceedings of the Amsterdam Academy of Sciences 
of June 25. In Voigt’s theory; the separation of a spectral 
line by the action of a magnetic field is found as the separation 
of an absorption line, and the theory requires a negative rotation 
of the plane of polarisation in the interior of the absorption 
band. Now in previous experiments, such as those of Corbino, 
the only observed result has been a very small positive rotation. 
The new experiment described by Prof. Zeeman is interesting, 
not only as showing the existence of a negative rotation in the 
interior of an absorption band, but also as being in perfect 
quantitative agreement with Voigt’s theory. 
By means of a system of quartz prisms such as have been used 
by Fresnel in his experiment on the division of a plane polarised 
ray into two circularly polarised rays, a number of interference 
fringes were formed at right angles to the bands of a spectrum, 
NO. 1720, VOL. 66] 
NATURE 
[OcroBER 16, 1902 
and when light from a bright source was allowed to pass through 
sodium vapour and analysed with a Nicol’s prism, the effect 
produced by the sodium absorption lines and the interference 
lines combined was as shown in Fig. 1. If, however, the 
sodium vapour was subjected to the action of a powerful mag- 
netic field, of from fifteen to twenty thousand units, the effects 
shown in Fig. 2 were observed. It will be noticed that 
the fringes moved upwards along the components of the 
doublets, whereas the parts of the fringes de/ween the com- 
ponents became disconnected from the exterior parts and moved 
downwards. As the density of the sodium was increased, the 
interior portions slid downwards with increasing velocity, and 
at a certain stage those in the interior, more particularly of the 
D, line, resembled inverted arrows. At last with increas- 
ing proportions of sodium these arrows entirely disappeared, 
and it was observed that this disappearance was more rapid 
with the D, lines than with the D, lines. Among subsidiary 
features it was noticed that the slope of the interior interference 
fringes is greater towards the side of the greater wave-lengths 
than towards the violet. The interior fringes also show a slight 
asymmetry, so that, eg., the points of the arrows in Fig. 2 
ought to be asymmetrical. 
Using very much denser vapours, however, results were 
obtained agreeing more with the experiments of Macaluso and 
Corbino. Figs. 3 and 4 show the effects with field intensities of 
about 4500 and 10,700 units respectively, and the absorption 
bands appear to contain horizontal parts of an interference 
fringe which have undergone a very small displacement wfpward’s 
Ds Dy 
Fic. 4. 
f Fic. 3. 
by the action of the field. These horizontal parts are, however, 
broader and more ill defined than the markings in the circum- 
stances previously described. _It is possible that the conditions 
assumed in these later experiments are different from those 
required by the theory, and that some explanation of the 
difference in the two kinds of phenomena may be found. 
In a paper communicated to the Reale Accademia dei Lincei 
of Rome, also on May 31, Prof. W. Voigt discusses the same 
phenomenon on a theoretical basis, and quotes the formula 
(A? — P?— 1) : 
(A+ P®+1)?- 4A7P? 
where x denotes the angular rotation of the plane of polarisa- 
tion, 7 the geometrical mean of the indices of refraction of the 
two waves propagated in the vapour, P is proportional to the 
magnetic field, and A is proportional to the number of wave- 
lengths in the distance of the point considered from the primitive 
position of the absorption band. From this formula are ob- 
tained the curves shown in Fig. 5, which correspond to the 
nx =KP 
Fic. 5. 
values P=0'5, P=1°5, P=3'0, and the resemblance between 
these curves and Prof, Zeeman’s photographs will be readily 
noticed. 
