by Cathode and Canal Rays to Exciting Current. 185 
a cathode in the case of canal rays. Hence it was only 
necessary to reverse the current before it entered the tube in 
order to have one or other of the rays impinging on the 
collecting electrode. The reversal was best accomplished 
by the aid of a large paraffin-wax four-cup commutator. 
Tke convection-current of the rays was measured by 
connecting the collecting electrode to earth through the 
galvanometer. The neighbouring net-electrode was also 
always connected to earth. In measuring the main current, 
the galvanometer was at first placed in front of the tube 
across a shunt which was connected between the influence- 
machine and the tube; iater it was connected beyond the 
tube (across the earth-connexion), as with the first arrange- 
ment the leakage was excessive. The collecting electrode 
AE was in all the experiments at a distance of 3°8 cms. from 
the neighbouring net-electrode. Thus the ratio of the ray 
convection-current to the main current could easily be deter- 
mined by the direct measurement of each. In these experi- 
ments the tube shown in fig. 1 was exclusively used. Since, 
however, on account of the unsteadiness of the current it is 
important to take the two readings almost simultaneously, 
I employed two galvanometers, Wiedemann’s, whose front 
coil without any shunt was used directly for the measurement 
of the main current, and which in this case had a sensitive- 
ness of 3°2 x 10-* ampere ; and Kohlrausch’s, for measuring 
the ray current. In the following tables are given the values 
of the ratios of the cathode and canal ray currents to the 
main current, in the form of percentages of the main current, 
the figures being rounded off with regard to the second 
decimal place. It must also be mentioned that the results 
in the case of high potentials can lay no claim to any great 
degree of accuracy, as with a very high vacuum _point- 
discharges and other disturbances occur. 
Metuop [I. 
Deternunation of Ratio by Differential Method. 
Before this method was actually applied, certain prelimi- 
nary experiments were carried out. 
By means of a secondary battery current the galvanometer 
was adjusted so that with equal currents no deflexion was 
obtained. Then by weakening the current in one branch the 
direction of deflexion was noted when, e. g., the negative 
current (corresponding to cathode rays) was weakened. The 
experiments then confirmed my surmise that by the with- 
drawal of the cathode rays the negative current was weakened, 
