241 
the two ends connected to the needles, not always at the points. When 
the magnets were excited there was no deflection observed in a field of 
1,500 gausses. The photographs taken show nothing between the points. 
Before putting on the cap containing the plate to take the photographs 
in series D a pencil drawing was made of the general form of each spark 
as seen from the end of the tube. Fig. 10 is a blue-print taken from these 
drawings, which shows that the directicn of the spark as it leaves the 
electrode has the same direction as the photographed stream. 
The width of the streams was measured in the proximity of the point 
with a micrometer microscope, and it was found that the width was inde- 
pendent of the potential between the points. The measurement was made 
between the outer edges of the central dark band. It will also be noticed 
in series D that the negative stream is almost as plain and almost as long 
as the positive stream. 
The photographs of series E show plainly the interference fringes as 
described by Professor Foley. Although no special pains were taken to 
show these fringes in any of the work, one or two can be seen on each 
photograph. 
SUMMARY OF RESULTS. 
1. The positive stream between the points for a spark or brush dis- 
charge was deflected by a magnetic field as low as 1,000 gausses, the direc- 
tion of deflection being in accordance with electro-dynamic¢ laws. 
2. The stream for glow discharge and the negative stream in any case 
were not deflected by a field of 1.500 gausses. 
3. The direction of the photographed stream for a spark discharge 
as it leaves the point is the same as the visible direction of the spark. 
4. The size of the stream at the points is independent of the potential 
between the points. 
5. The stream was deflected by an air current, the negative being de- 
flected more than the positive. 
6. The stream for the high potential spark increased in width as the 
distance from the point increased, while the stream for the glow discharge 
retained its original size as far as it could be traced. 
7. Some of the data indicate tlfat the stream photographed is one 
of the ionized air particles, the stream as a whole having an index of re- 
fraction different from that of the surrounding air, due to its pressure. 
[16—23003 ] 
