70 
the surfaces of the glowing electrodes. The velocities of the 
ions in the flame were estimated by finding the electric intensity 
required to cause them to move down the flame against the 
upward stream of gases. The positive ions of all the alkali 
metal salts had a velocity of about 60 emis. 
sec. 
The corresponding velocity of the negative ions was about 
cms. 
for one volt per 
cm. 
— In a current of hot air the corresponding velocities 
were Baral — 
(1) Negative ions of salts of Li, Na, K, Rb, Cs, Ca, Sr, and 
cms. 
Ba, 26:0 —— 
sec. 
(2) Positive ions of salts of Li, Na, K, Rb, and Cs, 7°2 cms: 
cms. 
sec. 
The greater velocity of the negative ions enables the phe- 
nomena of unipolar conduction &c., to be easily explained. 
Physical Society, May 12.—Prof. Perry, Vice-President, 
in the chair.—Dr. Lehfeldt read a note on the vapour pressure 
of solutions of volatile substances. The change in vapour 
pressure of a solvent due to the solution in it of a small quantity 
of volatile material has been calculated on the basis of Raoult’s 
rule for the corresponding case of a non-volatile dissolved body. 
The author has interpreted the formula of Nernst in the follow- 
ing words :—When asmall quantity of volatile substance is dis- 
solved in a liquid the vapour pressure of the liquid is altered 
in the ratio of the molecular fractional amount of solvent in the 
liquid to that in the vapour. In order to test’ this formula, it 
has been applied to the results of experiments made on four 
series of liquids, viz. alcohol with benzene and toluene, and 
carbon tetrachloride with benzene and toluene. _ In the case of 
normal solutions, such as carbon , tetrachloride in toluene, 
carbon tetrachloride in benzene, and benzene in carbon tetra- 
chloride, the agreement between the observed and calculated 
values of the percentage composition of the liquid was remark- 
ably good. | In the case of toluene in carbon tetrachloride the 
solution contained about 29 per cent. of the dissolved body ; 
and as the range of applicability of the formula had probably 
been exceeded, the agreement was not so good as in the 
previous examples. The mixtures containing alcohol show 
maxima of vapour pressure, and on this account the departure 
from the formula is so much more marked that it is impossible 
to apply it except in the case of very dilute solutions. The 
temperature used throughout the experiments was 50° C.—The 
Secretary read a note by Prof. W. B. Morton and Dr. Barton on 
the discussion of their paper, on the criterion for an oscillatory 
discharge of a condenser. In the discussion which followed the 
reading of the paper, it was pointed out that the result obtained, 
viz. that on taking into account the distribution of the current in 
the wire—a condenser having the critical capacity on the simple 
theory gives an oscillatory discharge—seems to be contradicted 
by the well-known fact that the resistance of the wire Is greater 
and the inductance less for oscillatory than ‘for steady currents. 
The explanation of the apparent paradox is to be found in the 
effect of the damping on the inductance. When the damping is 
great and the frequency small, as in the neighbourhood of the 
critical case, what may be termed the equivalent inductance 
becomes greater than the steady current value. It is shown 
that this increase in ‘‘ L.” outweighs the increase of“ R” in its 
effect upon the criterion for oscillatory discharge. An examin- 
ation of the expression for the equivalent inductance in the case 
of iron shows that it is greater than the steady current value if 
the ratio of one amplitude to the next is greater than e”!?0° where 
n is the frequency of the oscillation. Since the decrease of 
«],” with maintained oscillations is due to a surface con- 
centration of current, it is suggested that there must be an axial 
concentration in the case of damped vibrations. Following the 
method of Maxwell for determining the current density at a dis- 
tance from the axis of a wire, an expression for the current was in- 
troduced containing a damping coefficient. The ‘ quasi-ampli- 
tude” of the disturbance at any point in the wire was thus ob- 
tained. An examination of the result shows that making the 
damping zero indicates a surface concentration. If, on the 
other hand, the damping is great, the expression for the ampli- 
tude increases as the distance from the axis decreases, and we 
get an axial concentration. Assuming sufficient damping to 
produce this effect, it is shown that as we go through the point 
NO. 1542, VOL. 60] 
(3) Positive ions of salts of Ca, Sr, and Ba, 38 
NATURE 
[May 18, 1 899 
y = _, where a is the radius of the wire, we pass from a 
2 
greater value of current density in the inner parts to a less in 
the outer than would correspond to a uniform distribution 
throughout the wire. From general reasoning the authors 
think that if a rapidly damped disturbance is propagated into 
a wire from its boundary, and if the oscillations are slow enough 
to allow the current to penetrate to the core, we should expect 
to find an axial concentration in the latter stages of the 
phenomenon. Dr. Lehfeldt said that Prof. Lodge had pointed 
out, at the reading of the paper, that the solution the authors 
obtained changed character at the critical resistance. As this 
point had not been considered in the note, he supposed that 
the change in character made no difference to the results ob- 
tained, The Chairman expressed his interest in the proof of 
the existence of an axial concentration. —Mr. Addenbrooke ex- 
hibited and described a quadrant electrometer for application 
to alternating current measurements. The author has substituted 
for cylindrical quadrants two sets of flat plates, the top set being 
adjustable. In this way the range of the instrument is con- 
siderably increased. The ability to remove one or more of the 
top quadrants makes the needle very accessible. By lowering 
the needle on to the bottom quadrants, and then bringing down 
one of the top plates, the instrument can be carried with safety. 
One of the top quadrants can be worked up and down by a 
worm gear, and by this motion the ‘electrical zero” of the 
electrometer is obtained. The suspension consists of a flat 
phosphor bronze strip, the torsion of which is found to be per- 
fectly uniform, there being no fatigue effect. The case of the 
instrument contains windows, so that the needle can be viewed 
from two directions at right angles, and there are screw motions 
to centre the needle with respect to the quadrants. To reduce 
the effect of air convection currents upon the needle, the inside 
of the case is lined with cotton velvet. The quadrants are 
supported on brass bars passing through long ebonite sleeves 
inthe bottom of the instrument. This gives good insulation 
without the use of sulphuric acid, and there is no Leyden jar or 
condenser in connection with the needle. When using the electro- 
meter idiostatically with the finest strip, a light needle, and the 
quadrants one-tenth of an inch apart, a difference of potential 
of one volt will produce a deflection of about 5 mms. upon 
a screen two metres distant. Using the instrument hetero- 
statically with 100 volts on the needle one-fifth of an 
inch between the quadrants and half a volt acting across them 
a deflection of 200 mms. can be obtained. This sensitiveness 
is about twelve times as great as that got from instruments 
designed by Kelvin, Mascart, and Haga. Mr. Addenbrooke 
then showed how, in conjunction with a voltmeter and an 
ammeter, it was possible with his instrument to determine all 
the factors of an alternating current system. The increased 
sensitiveness of the electrometer renders it possible to measure 
currents of any magnitude with a very small waste of energy. 
Mr. Gaster pointed out that the measurement of self-induction 
with an electrometer could only be carried out practically if the 
current curve was a sine curve. He said that in curves ob- 
tained from a Ganz motor a correction amounting to 7 per cent. 
had to be applied. The Chairman said that even if the curve 
obtained was a sine curve, the electrometer was never used in 
this country for measuring Self-induction. Prof. Herschelasked 
if it were possible to adjust the quadrants after the needle had 
been charged Mr. Addenbrooke then purposely disturbed the 
position of the adjustable plate, and, after charging the needle, 
reduced the deflection to zero by the worm gear. The author 
said that for high voltages the curve of calibration was different 
to that obtained from the ordinary formula. The Chairman 
said that this discrepancy was probably due to want of perfect 
symmetry. In a paper read before the Royal Society by 
Perry, Ayrton and Mather, it was shown that the presence of 
the guard around the mirror of an ordinary electrometer was 
sufficient to affect the needle when working with high voltages. 
In working with the plates very close together he was afraid 
the symmetry would be liable to be disturbed by a slight tilting 
of the needle due to electrostatic attraction. The author ob- 
served that the plates were only very close together when 
working with low voltages. 
Chemical Society, May 4.—Prof. Thorpe, President, in 
the chair,—The following papers were read :—On the com- 
bustion of carbon disulphide, by H. B. Dixon and E. J. Russell. 
Carbon disulphide undergoes a phosphorescent combustion in 
air at temperatures below its ignition point, the lowest observed 
