402 TRANSACTIONS OF SECTION A. 
between appled e.m.f. and current is mainly due to resistivity changes produced 
by Joulean heating. In the second class there is in general some thermoelectric 
action imposed on the resistivity-temperature phenomenon; and in the case of 
the crystalline substances mentioned the thermoelectric actions may be the 
more important. 
For the first class, theory yields a cubic equation connecting e.m.f. and 
current. Experiments are adduced by the author which show that the thermal 
explanation is in many cases sufficient. In the second class, thermeelectric 
theory yields a quantic equation. The curve connecting e.m.f. and current takes 
very various shapes according to the signs and the relative magnitudes of the 
Peltier and Thomson coefficients in the substances forming the contact. The 
author has measured these coefficients in some typical substances and has thus 
carried out a comparison between the theoretical curves and the experimental 
curves of contacts between pairs of these substances. The evidence gathered 
in this way tends in the main to support the theory. 
3. The Twisting of India-rubber. 
By Professor J. H. Poyntine, F.B.S. 
4. The Resistance of Air to Halling Spheres. 
By G. A. Swaxesprar, D.Sc. 
The experiments described fall into two classes :— 
(a) The determination of the time of fall of a hollow celluloid sphere of 3-70 cm. 
diameter through different distances, for a series of different weights of the sphere ; 
the resistance being assumed to be equal to the weight in air of the sphere when the 
limiting velocity is attained ; and ; 
(b) The relation between the resistances for a series of spheres of different 
diameters (from 2-01 to 7:52 cm.) for a velocity of 1,030 cm./sec. In the formula 
R = KSYV?, where R = resistance in dynes, S the area of diametric circle of sphere 
in square cm., and V the velocity in cm./sec., the value of K ranged from 2:50 to 
2-66 x 10-4 grms/cm.® for the sphere of 3-70 cm, diameter when the velocities 
ranged from 850 cm./sec. to 1,320 em./sec. For the different spheres at velocity of 
1,030 cm./sec., the value of K ranged from 2:85 x 10-4 for that of 7:52 cm. to 2:58 x 
10-‘for that of 2:0lem. The value of K diminishes gradually with the diameter 
and with the velocity. 
5. A New Method of Starting Mercury Vapour Apparatus.* 
By J. S. ANDERSON. 
In the best types of mercury vapour lamps and rectifiers at present on the 
market, the arc is started by tilting the lamp or rectifier, either by hand or 
automatically. Now this tilting arrangement is very often inconvenient. This 
is found to be especially the case when one is dealing with lamps used for 
scientific purposes. For example, in carrying out experiments on the Zeeman 
effect a mercury vapour arc lamp is extremely suitable. But the difficulty arises 
from the fact that the lamp must be placed between the poles of an electro- 
magnet, the distance between the poles being usually so small that any tilting 
apparatus that may be employed interferes with the proper mounting of the 
lamp. 
Mr. G. B. Burnside and the author have overcome this difficulty by con- 
structing a lamp which may be fixed in position between the poles of an 
electro-magnet, or in any other suitable position, and then started without having 
to be tilted. This is brought about by the employment of a heating arrangement 
near one of the electrodes, preferably the negative electrode. The lamp tube is 
provided with a small vessel near this electrode, the vessel having a re-entrant 
1 John 8. Anderson and George B. Burnside, Proc. Roy. Soc. Bdin., XXxill. 
(1918), p. 117. ; 
