162 Proceedings of Indiana Academy of Science 
are given. Results with other spacing of the spark gap with this 
wire and with other sizes of copper wire are essentially similar. With 
a No. 18 wire of copper or, better, aluminum, the efficiencies are in- 
creased two or three times, while the concentrations are often doubled. 
In this figure the points of interest are the similarity of the curves for 
primary amperes and for tube amperes, a direct proportionality usu- 
ally existing. The voltage, on the other hand, as a rule requires a high 
value in order to start the spark gap and then falls to a minimum. In 
most cases, the efficiency of the process as expressed in g. per kilowatt- 
hour (mg. per watt-hour) runs closely parallel to the curve for the 
yield of ozone per coulomb as calculated from the hot-wire milliammeter 
in the tube cireuit. Usually, also, the concentration is increased at the 
expense of efficiency, although there are certain conditions where this 
generalization does not hold. In Fig. III the efficiency is constantly 
decreasing because in the short distance of the spark gap the tendency 
of the spark is to change over in characteristics toward a power arc, 
16 [ COULOMB 
9 
oo 
Q 
8 
% 8 3 
176 if WATTHOUR 
3 
GFILIANI LY NOWLYYNG 
Lip (ape) = Me | Liver 
Fig. IV. Enclosed air was subjected to discharge in the same tube provided with 
a No. 12 wire. The gross power input into the tube was 65 + 2 watts. The spark-gap 
was set at 2.0 mm. Note the reciprocal relation of the concentration and efficiency 
eurves, characteristic of nearly all chemical reactions in corona discharges. 
