“ 
496 
NGTURE 
| Oct, 9, 1873 
as to the heat of the sun ;” and in a foot-note it is remarked : 
«‘ These inconveniences, however, are fully counterbalanced, if it 
be true that the cessation of hurricanes since 1789 has been caused 
by the great diminution of the woods.” As the history was 
published in or after 1801, it would appear that during the twelve 
years 1789-1801 no hurricanes occurred. Now, since according 
to the Table of Sunspots the years 1788 and 1804 were maxima 
years, and the intervening minimum occurred in 1798, our theory 
would lead us to expect a comparative cessation of hurricanes 
during the period mentioned. ‘ ; 
If time permitted, I would adduce similar evidence respecting 
the hurricanes of Bourbon and other parts of the world. 
The hurricanes of the Indian Ocean are well known to be 
attended with torrential rains. So much is this the case that the 
popular belief at Mauritius is that cyclones are the cause of our 
rains. Heavy rains over extensive areas are certainly concomi- 
tant with cyclones in the Indian Ocean, and it was therefore re- 
solved to examine whether there was a rainfall periodicity. As 
far as the Mauritius observations went, the matter was clear ; but 
it was desirable to extend the investigation to other localities. 
The Queensland and South Australia observations, which were 
the only ones available at the time, gave a similar result, and as 
Adelaide is far beyond the limits of the region of cyclones, it was 
surmised that there was a rainfall periodicity generally, The 
Cape observations afterwards gave additional support to this 
view. The rainfall of England was next examined, and also 
found to bear out the hypothesis. 
It would occupy much more time than I can at present spare 
to enter fully into this question of rainfall periodicity. With the 
help of the researches of Mr, Lockyer, Mr, Symons, and Dr. 
Jelinek of Vienna, I haye now examined 93 tables of rainfall for 
various parts of the world, and I find that, scarcely without ex- 
ception, more rain falls in the maxima than in the minima sun- 
spot years. I beg to append a table showing the general results 
for the quarters of the globe. It will be seen that, so far as the 
observations go, Europe, Africa, America, and Australia give 
very favourable results. Asia is only represented by three sta- 
tions, one of which is Jerusalem, where the excess of rain in one 
minimum period exceeds the excess in the maxima periods for 
two stations in India. France is the only European country (of 
which the rainfall has been examined) that gives an unfavourable 
‘return, but it must be remarked that we have as yet got only five 
stations there, which are all inland, and probably do not fairly 
represent the rainfall of the whole country. 
By taking the longest possible series of observations for several 
stations, the periodicity comes out, and there is, I think, strong 
evidence that the rainfall for the whole globe is subject to an 
annual variation. 
Having given the facts, as far as I have been enabled todo so, 
I abstain from offering any theoretical remarks. If cyclone and 
rainfall periodicities be fully established, a corresponding (di- 
rect) temperature periodicity should exist, and this presumable 
variation of solar heat may be the indirect cause of the periodi- 
city of auroras and magnetic disturbances. 
(The catalogue of cyclones was appended. ) 
On the Effect of Pressure and Temperature on the Widening 
of the Lines in the Spectra of Gases, by Arthur Schuster, Ph. D. 
One of the questions in Spectrum Analysis yet open to discus- 
sion, is what influence pressure and temperature exert on the 
widening of the lines, which is sometimes observed when an 
electric current passes through certain gases. The subject of 
this communication is to point out a little ambiguity which has 
crept into the very statement of the question at issue, and to 
show the only way by which a decisive answer can be arrived 
at, and, in my opinion, has already been arrived at. I shall 
begin by assuming that the convection of electricity has no direct 
influence on the character of the spectrum ; that is to say that 
under the same pressure, and at the same temperature, the gas 
will always show the same spectrum, whether the temperature 
has been produced by the passage of an electric current or by 
any other means. In the present state of science this is the only 
reasonable assumption that can be made, and it has been tacitly 
made, I think, by every one who kas written on the subject. 
Let us imagine a vessel filled with hydrogen, and let the tem- 
perature of the gas be brought up to incandescence. The heat 
communicated to the vessel is partly used to increase the trans- 
latory motion of the gas, and thereby to increase its pressure, 
and the other part of the heat has increased the periodical 
motion in the molecules of the gas, which is generally admitted 
to be the cause of its incandescence. If the temperature is 
such that the lines are widened we can account for this fact in 
two different ways. We may think that the forces which 
maintain the molecule in vibration, and which are such that 
at a lower temperature only perfectly isochronous vibrations 
can take place are somewhat altered, so that the bonds which 
keep the molecules together are loosened and now allow vibra- 
tions to take place, the period of which is somewhat altered and 
varying. We might secondly explain the widening of the lines 
by saying that they are caused by the disturbances caused by the 
frequent shocks of other molecules. If we increase the number 
or the force of these shocks by increasing either the number of 
molecules or their velocity we might well obtain disturbances 
large enough to change a little the period of vibration. These 
are the only two explanations that can be given, and if we say, 
therefore, that temperature is the cause of the widening of the 
lines we can only mean that part of temperature which has its 
equivalent in the vibrating energy within one molecule. If we 
say that pressure is the disturbing cause we include that part of 
heat which increases the pressure with increasing temperature. 
Let us now see whether we can obtain a clear answer to the 
question which has now been clearly put. 
It is evident that no result can be arrived at by subjecting the 
same quantity of gas in the same vessel to different tempera- 
tures, for we cannot increase the vibrating energy of the mole- 
cules without increasing at the same time and in the same 
proportion (as Clausius has shown), their translatory velocity. 
By varying in the same ratio the two possible causes we shall 
never be able to say which is the right view to be taken, 
There are two ways open to us to mend this difficulty, We — 
might increase the temperature of the gas under the same pressure. 
Ifthe perturbation caused by the shocks of other molecules cause 
the widening of the lines this widening ought not to take place 
as we have reduced the number of these shocks in the same ratio 
as we have increased their force. If on the contrary the disturb- 
ance in the period of vibration has its cause within the individual 
molecules it ought to remain. 
We might, secondly, decide the question in subjecting the gas 
at the same temperature to different pressures. If perturbations 
are the cause the lines would be widened. Which of these two 
ways is most easily pursued in experimenting? Can we easily 
heat up a gas to incandescence under constant pressure? I 
think not. If an electric discharge takes place in a gas only 
comparatively few particles of the gas are heated up, and at 
a very small distance from the points through which the dis- 
charge takes place the gasis hardly heated up at all. But if the 
heat is not diffused through the whole mass of gas, the increase 
of pressure caused by this gas will also be merely confined to 
the luminous streak, and we can therefore obtain no answer to - 
our question (as this has been attempted) by filling a tube with 
a certain quantity of gas, and altering the strength of the pass- 
ing current or the mode of discharge. ; 
We are, therefore, compelled to abandon this route and to 
turn our eyes to the second way which I have indicated ; but 
here we meet another difficulty, and even one over which we 
cannot easily get. We cannot alter the pressure of a gas with: 
out altering its electric resistance, and, therefore, also the 
strength of the electric current and the heat developed. We 
can only decide the question by subjecting the gas at the same 
temperature to different pressure. Now have there ever been 
any such experiments made? I think there have, and even 
very decisive ones. Frankland and Lockyer have found that 
if we increase the pressure of hydrogen while an electric cur- 
rent is passing through it the lines begin to expand till the spec- 
trum becomes continuous, and finally the resistance becomes so 
large that the electric current will not pass at all. On the other 
hand Gassiot and Plucker have observed that if we diminish the 
pressure of hydrogen its electric resistance force diminishes, 
attains a minimum, then increases again, and if we keep up ex- 
hausting the tube it becomes again so great that the current can- 
not pass. Plucker says that a tube exhausted to its utmost 
limits shows the lines of hydrogen and silica. He mentions 
at one place, I think, that the lines are very fine and distinct. 
If there would have been any widening he would have been 
sure to mention it. Now it is not too much to assume that 
the resistance of the gas at the moment when the discharge just 
ceases to take place is the same whether the increase of resist- 
ance is produced by too great a pressure or too great an ex- 
haustion. At this moment, therefore, the current is the same 
and the same energy must be converted into heat by resistance. 
But in the case in which the current does not pass on account ot 
the excessive diminution of pressure, only a much smaller 
