586 
NATURE 
[OcToBER 12, 1899 
earth’s atmosphere, and of water vapour in the atmosphere of 
Mars. The method of calculation is to determine the number 
of years which would be required for the planet to lose from 
its surface a layer of the gas one centimetre thick at various 
temperatures. The results show that the earth might retain 
helium, but would lose hydrogen appreciably at ordinary tem- 
peratures, and that Mars might retain water vapour at ordinary 
temperatures. If helium ever existed on the earth’s surface, it 
must have escaped when the surface was much hotter than at 
present, whereas a smaller elevation of temperature would cause 
water vapour to escape from the surface of Mars. 
Prof. W. F. Barrett described the thermo-electric properties 
of an alloy containing iron 68°8 per cent., nickel 25'0, man- 
ganese 5°0, and carbon 1*2._ When a thermo-electric couple is 
formed of this metal and iron, the electromotive force rises with 
temperature to 300° C, ; it then remains steady until 500° C. is 
reached, after which it falls slightly and rises again to 1100° C. ; 
the fluctuations of electromotive force do not exceed 4 per cent. 
of the total value. When the alloy forms a couple with nickel 
the results are similar, but the range of variation is slightly 
greater. 
The committee on the heat of combination of metals in the 
formation of alloys, appointed last year to assist Dr. A. Galt in 
his experiments on this subject, reported the completion of their 
work. Only alloys of zinc and copper have been examined, 
twenty-two in number and containing from 5 to 90 per cent. of 
copper ; the difference between the amounts of heat evolved by 
dissolving in nitric acid unit mass of the alloy and corresponding 
amounts of the mixed metals was taken as the heat of combin- 
ation of the metals. The results indicate a negative heat of 
combination for alloys rich in zinc, the numerical value of 
which is a maximum when the alloy contains 16 per cent. 
of copper. The formation of an alloy containing about 24 per 
cent. of copper takes places without absorption or evolution of 
heat, while for 38 per cent. of copper the heat of combination 
is a maximum and positive ; beyond this it diminishes to zero 
for pure copper. In the absence of Dr. Galt and other members 
of the committee no reply was given to a serious criticism by 
Prof. Vernon Harcourt, that in the experiments no account was 
apparently taken of the fact that the products arising from the 
solution of an alloy in nitric acid are not the same as would be 
obtained from the mixed metals. In his paper read last year 
at Bristol, Dr. Galt mentioned that he had made many pre- 
liminary experiments, and possibly he has examined this point ; 
if not, the results obtained by the committee will be somewhat 
vitiated. 
A preliminary report of the committee on radiation from a 
source of light in a magnetic field was communicated to the 
Section, the chief points in which were (1) the discovery that 
light passing through a magnetic field at right angles to the 
lines of force suffers absorption (see NATURE, vol. lix. pp 
228-9, January 5, 1899) ; (2) the various modified forms of triplet 
are true magnetic perturbations of the same kind as the normal 
triplet ; (3) the spectral lines of a substance may be divided into 
groups such that all members of one group suffer the same kind 
of perturbation (see NATURE, vol. lix. p. 248, January 12, 1899). 
The Zeeman effect is attributed to the action of a mag- 
netic field on the moving ions; recently Mr, C. E. S. Phillips 
has discovered an apparently cognate phenomenon, which 
he described in his paper on the production in rarefied gases 
of luminous rings in rotation about lines of magnetic force. 
An electric discharge is passed between soft iron electrodes 
in a Crookes’ vacuum tube; on stopping the discharge 
and setting up a magnetic field between the electrodes, a 
luminous ring forms with its plane at right angles to the lines 
of force and in rotation about the magnetic axis. The direction 
of rotation is that which would be communicated to negatively 
charged particles, and is reversed on reversing the magnetic 
field; the luminosity persists sometimes for a minute, and re- 
versal of the magnetic field causes it to brighten momentarily. 
Two explanations of the phenomenon have been given; one is 
that the rotating matter consists of ions or electrons, and the 
other that the matter consists of gas particles which have ac- 
quired a negative charge by contact with the walls of the tube. 
From experiments of Prof. J. J. Thomson, it appears that neg- 
ative ions move more quickly than positive, which would account 
for the greater luminosity of the negative ions when set in 
rotation, " 
In a note on deep-sea waves, Mr. V. Cornish endeavoured to 
trace relations between the amplitude, wave-length, and wind- 
NO. 1563, VOL. 60] 
velocity for waves on the surface of deep water. Sir George 
Stokes pointed out that the amplitude observed is not that of a 
simple wave, but is the resultant effect of a train of waves of 
different periods and lengths. 
At the meeting of the Section on Saturday the visitors fron 
the French Association at Boulogne were present, and the 
President extended to them a hearty welcome, which was 
acknowledged by M. Benoit, as president of the Physical 
Section of the French Association. A paper was then com- 
municated by Prof. J. J. Thomson, on the existence of masses 
smaller than the atoms. He stated that several lines of research 
lead to a determination of the ratio of the mass of ‘an atom (7) 
to the charge carried by the atom (e). Among these are electro- 
lysis, the velocity of charged particles in a magnetic field, and 
the magnetic deflexion of kathode rays. The two latter 
methods are comparatively simple, because they depend on the 
observation of luminous effects, but although they agree with 
each other fairly well, they furnish a value of %/e which is 
about 1/1000 of that calculated from electrolytic phenomena. 
It becomes, therefore, a matter for inquiry whether in the 
former experiments the atom carries a charge greater than that 
required by Faraday’s laws, or whether the charge is carried by 
a portion only of the atom—in other words, whether a smal} 
fraction of the mass of the atom is detachable which has associated 
with it a negative charge. The simplest crucial experiment is 
obtained by determining separately either 7 or e, and the author 
has devised a means of measuring the latter quantity. Hetakes 
a negatively charged metal plate supported horizontally ; below 
this and parallel to it is a very large perforated metal plate, the 
whole being in rarefied gas at a pressure of about 1/100 mm. 
mercury. When ultra-violet radiation is directed through the 
perforated plate to strike the upper plate the latter is discharged, 
the discharging particles moving along straight lines normal to 
the two plates. If a magnetic field be now excited with its 
lines of force parallel to the plates, the particles describe 
curved paths which are in fact portions of cycloids. When the 
plates are near together the particles which leave the upper one 
strike the lower one; if, however, the plates are separated 
further, the vertex of the cycloidal path comes between them, 
and the particles do not reach the lower plate, so that the dis- 
charge ceases. In the actual experiment there is a gradual, but 
not abrupt, change in the rate of discharge, possibly because all 
the particles do not start from the surface of the upper plate. 
From observations on the distance apart of the plates when the 
change in the rate of discharge commences, the form of the 
cycloidal path is determined, and the results show that the 
smaller value of m/e is applicable to this case and to that of 
illumination by kathode rays. Further, the amount of electricity 
discharged by the illuminated plate per second is proportional 
to the number of particles between the plates, to the charge 
carried by each (e), and to the velocity of the particles. The 
last-named quantity is measured by a method due to Prof. 
Rutherford, so that if the total number of particles in the space 
is known the value of e can be determined. To count the 
particles use is made of the fact that they serve as nuclei for the 
formation of drops out of a condensing vapour, each particle 
giving rise to one drop, Leta known amount of air of given 
humidity be suddenly and definitely expanded in the presence 
of the particles, and observe the rate at which the drops fall ; 
this rate gives the size of the drops, andhence their mass, and 
since the whole mass of water deposited is known, the number 
of drops is thus determined. For negative charges the ratio 
m/e is independent of the nature of the gas, whereas for positive 
charges its value varies from one gas to another, and corre- 
sponds generally with the values given by electrolytic 
phenomena. Prof. Thomson considers that electrification con- 
sists in the removal from the ‘‘atom”’ of a small corpuscle 
with which the negative charge is associated; the remaining 
large portion of the mass is positively charged. This view is 
supported by Prout’s hypothesis that the mass of an atom is not 
invariable, and by the evidence derived by Lockyer and others 
from spectroscopic observations. 
In the discussion which followed upon Prof. Thomson’s paper, 
M. Broca described spectroscopic observations of a spark 
obtained between two platinum electrodes 4 mm. apart in a 
Crookes’ vacuum tube; the spectra of the regions near the 
electrodes and the space between them were not alike. Prof. 
Riicker drew attention to Schuster’s experiments, in which the 
spectrum of a substance not present in the material examined 
sprung into being in the arc itself. He believed matter to be a 
EE 
