ELECTRIC DISCHARGES IN GASES AT LOW PRESSURES. 
125 
will be seen that a potential difference of abont 365 volts is necessary to produce 
discharge when the spai-k length of 5 inicra is readied. 
Wliile the three types of curves whidi have been described all present different 
characteristics, it will be seen that all are confined to S}iark lengths above 5 micra, 
and to spark potentials greater than 350 volts. 
Earhart has shown that for spark leiigths below 5 inicra the spark potentials 
again decrease as the spark lengths are shortened, until finally the two electrodes 
come together and direct electrical contact is established. Tbroughont this lower 
range of spark lengths his results also show, for a series of pressures, that the spark 
potentials vary directly with the spark lengths. 
These experimental results of Earhart give an indication of the forms the curves 
in fig. 9 would have taken liad the experiments with the apparatus used in this 
investigation been extended to the shorter set of sjiark lengths. Had this been done. 
Fig. 10. 
it is highly probable that the curves in fig. 9, on reaching the 5-micra line, would 
have followed courses such as are indicated liy the dotted straight lines in the 
figure. 
On this view it is of interest to examine the character of the pressure-spark 
potential curves that can be drawn in fig. 8 for spark lengths shorter than 5 micra. 
VOL. coi.—A. 
