265 
which takes on more and more the form of an are as the speed of the 
machine increases, the air insulation is broken down, the air is more 
highly heated and more highly ionized along the spark path, and a greater 
number of ions will travel along this narrow path with great speed and due 
to the outer ones encountering the air molecules, the stream will follow 
more nearly the curvature of the spark. Farther from the point their 
speed becomes so small and they become so much scattered, they do not 
set up a stream so well defined. This same hypothesis applies to the expla- 
nation of the scattered stream when it was defiected by an electrostatic 
field. The stream retains practically its original diameter past the oppo- 
site terminal for the magnetic deflection in case of the glow and brush dis- 
charge, and although scattered may be traced nearly to the opposite point 
in the spark discharge. In case of the electrostatic deflection, the dis- 
charge without the transverse field is quite as well defined as those of the 
magnetic deflected series, while with the transverse electrostatic field the 
stream is short and not so well defined. If the ions moving with great 
speed start from the point, and soon by their bombardment start a current 
of air, at the same time lowering their own speed, they will certainly 
be scattered, part of them going to the oppositely charged plates, and part 
to the opposite point. 
If the majority of the negative ions are considered to be ordinary elec- 
trons and those from the positive point equal in mass to the hydrogen 
atom, the kinetic energy of the positive ions will be far greater than the 
negative. They will therefore carry with them a greater current of air. 
Perhaps it may be permissible to assume that the negative ions are not 
all single electrons, since it has been shown by J. J. Thomson‘ in case 
of discharge is rarefied gases, that negative ions exist nearly equal in 
mass to the positive ions, and have the same initial speed. The greater 
the per cent. of these large ions the greater will be the amount of air 
set in motion, the greater the velocity of the stream as a whole, and the 
more defined the stream. If, then, the assumption is made that the stream 
fs produced by the larger ions, it explains the equal deflection of the posi- 
tive and negative streams in case of the magnetic deflection. 
A few of the photographs show peculiar characteristics. In some 
there are two streams from the positive point. It was not learned whether 
«J. J. Thomson (Phil. Mag. Ser. 6, Vol. 16, pp. 657-691), 1908; also (Phil. 
Mag. Ser. 6, Vol. 18, pp. 821-844), 1909. 
