TROWBRIDGE. — HIGH ELECTRO MOTTVI fORCE. 201 
in air with disruptive discbarges at atmospheric pre* ure the Faraday column extends 
to the cathode, leaving no interval for a cathode space. 
In air, however, at atmospheric pressure, as I have said, (he negative glow at ilie 
cathode is far more noticeable than the positive glow when the terminals are far apart. 
For instance, the negative glow can be detected several hundred Eeet awaj from the 
positive terminal of the battery of 20,000 cells; while the positive glow only appears 
when the positive terminal is brought within two f< *t of the negative terminal. One 
marvels at the extent of invisible ionization extending from the cathode. 
The characteristic brush discharge at the positive pole, a characteristic discharge 
which we see from Figure 8 is reproduced at every fork or bifurcation of the disrup- 
tive discharge, is shown in Figure 6. It is wen to consist of bifurcating discharges 
together with a flaming discharge. Turning to the negative pole (Figure 4) we find 
the brush discharges are straight, as if drawn in pencil with a ruler, except at their 
termini, where they end in a diffuse brush. 
If the two discharge terminals are in the form of Neptune's tridents, the discharge 
often separates near the positive terminal, and the two portions appear at first to be 
undecided which of the points on the negative pole they will select. Finally they 
decide to unite. This phenomenon is evidently an evidence of the oscillatory nature 
of the discharge, for it shows that there are points of varying potential along the 
line of discharge. 
Some years ago I showed that an explosion occurs whenever powerful sparks 
change their direction in zigzags, such as are evident in Figure 8. The spark passed 
between a plate of glass and a sheet of paraffined paper, and it was found that the 
paper was perforated at each forking of the discharge. This is shown in igure 6, in 
which the spark negative is placed alongside of the negative of the perforations. 
Possibly these explosions occurring alon an extended lightning discharge may be an 
important element in the phenomenon of the rolling of thunder, for the sound of such 
explosions would arrive at considerable intervals apart. 
An interesting account of the explosive effect at each turning point of a lightning 
discharge has been given me by Mr. Harvey N. Davis, an instructor in tli Jefferson 
Physical Laboratory, and I give it here, since it is an account by a skilled observer 
of both the above explosive effects and ball lightning. 
"During the 27th of August, 1906, a large boarding-house on the side of Mount 
Moosilauke, in the town of Warren, N. H., was struck by lightning in an unusu- 
ally sudden and severe thunder storm. The path of at least three independent dis- 
charges could be traced, but they nn: t have been practically simultaneous, for those 
