714 Dr. S. B. Milner on the Current-Potential 



p.d. is apparently in the opposite direction to the current. 

 At first I thought that it was an effect of the inertia of the 

 ions in the spark-gap making itself evident when the 

 frequency becomes sufficiently great, since on certain sup- 

 positions the inertia can be shown to give rise to this form 

 of curve. However, closer examination showed that this was 

 not the case, and that the effect was due to the magnetic 

 field of other parts of the spark circuit becoming comparable 

 with the field of the deviating coil (which at the highest 

 frequencies was reduced to only two or three turns of wire), 

 and thus making the resultant line of magnetic deviation not 

 at right angles to the electrostatic one. When this effect was 

 eliminated, no difference in character in the current-potential 

 figures was observable down to periods of 5xl0~ 7 sec. 

 It seems to me very possible that Roschansky's results at 

 high frequencies are capable of being explained in the 

 same way. 



Observations on the figures were made with the spark 

 situated between the poles of an electromagnet, but no 

 effect could be detected on suddenly switching on a strong 

 transverse magnetic field. 



A striking difference was observed when the sparks, 

 instead of being free in air, were confined in a capillary 

 tube. The electrodes were copper wires about 1 mm. diameter 

 fitted tightly in the tube. There is now no clear distinction 

 between the glow and the arc p.d's., and the average p.d. 

 throughout the discharge is much greater than with sparks 

 in free air (see figs. 6 and 7, PI. XX.). The average 

 p.d's. increase with tlie spark-length (compare the two 

 figures — fig. 7 is taken under the same conditions as 6 except 

 with a longer spark); but alteration of the inductance and 

 capacity produces but little effect. When one electrode is 

 enclosed in the tube and the other is free in air, there is no 

 appreciable difference between the top and bottom halves of 

 the curve: this shows that the p.d. is not concentrated at the 

 electrode, but is a uniform fall over the whole length f t i ie 

 spark enclosed in the tube. This type of figure is probably 

 due to the rarefaction of the air in the tube and the consequent 

 reduction in the number of ions which carry the current ; 

 for at atmospheric pressure the effect does not show until a 

 sufficient number of sparks have occurred to heat the tube 

 and thereby rarify the air in the spark-gap — when the tube is 

 partly exhausted of air by a pump the effect is shown at the 

 first spark. 



Fig. 8 shows an interesting effect which is usually observed 

 when the plate Y is connected to I, the junction between the 



