118 
MESSRS. W. DE LA RUE AND H. W. MULLER ON THE 
Lengths of Spark. 
No. of cell?. 
Two points. 
Point + 
and 
disc — 
Concentric 
cylinders. 
Plane surfaces. 
Spherical 
surfaces. 
inch. 
inch. 
inch. 
inch. 
inch. 
8,000 
0-4000 
0-3400 
0-0990 
0-0882 
0-0818 
9,000 
0-4760 
0-3805 
0-1142 
0-1020 
0-0938 
10,000 
0-5515 
0-4248 
0-1310 
OT157 
0-1050 
11,000 
0-6200 
0-4730 
0-1496 
0-1280 
0-1170 
The foregoing numbers were obtained from curves laid down from the means of the 
actual observations of differences in the length of spark between 8040 and 9540, and 
11,000 cells respectively; the diagram, fig. 31, shows the curves on a reduced scale. 
It is presumable that if the whole battery had the potential of cells newly charged 
the striking distances for 11,000 cells would be for — 
Cells. Two points. Points and disc. Concentric p , surfaces . Spherical 
cylinders. surfaces. 
i nc h. inch. inch. inch. inch. 
11,000 0-6412 0-4889 0-1546 0-1323 0-1210 
It will be seen that the curve for two points runs on 
pretty smoothly with that obtained up to 8040 cells, 
whilst that for a point and disc bends inwards and no 
longer conforms to a curve, laid down on the hypothesis 
of the length of spark being in the ratio of the square of 
the number of elements. 
We must state that our insulation is put to a severe 
test when 11,000 cells are connected in series, and that 
there is an evident leakage when one pole is connected 
to earth. This may have some but not a great effect on 
the length of spark. 
The 11,000 elements made manifest an interesting 
phenomenon in the streamer-discharge, which precedes 
the jump of the spark and formation of the arc, con- 
firming what we have previously stated at pages 88-100, 
namely, that the discharge is more continuous at the 
negative pole than at the positive. It was seen that 
when two points were used as terminals in the discharger 
(fig. 1, page 58) with this number of elements there was 
a continuous brush-discharge at the negative, while 
from the positive an intermittent streamer-discharge took 
place, and that the streamers emanating from it enve- 
loped the negative brush-discharge without in the least 
disturbing its form. This phenomenon is well represented 
on a scale of 4 to 1 in fig. 32. 
Fig. 32. 
