214 J. A. POLLOCK AND A. B. B. RANCLAUD. 
addition, the temperature of the cool carbon when below 
the hot one is much lower than when above it; the value 
of [HA may, therefore, have to be higher for arcing to 
start in the former case than in the latter. If the change 
to the luminous discharge is considered to originate near 
the cathode surface, the observed differences in the arcing 
voltages, due to alteration in the relative positions of the 
hot and cool carbons, would be difficult to explain. 
7. Summary.—The flow of negative electricity from hot 
carbon, in a Circuit containing an air gap, up to three 
millimetres in length, between a hot and a cool carbon rod, 
has been investigated for temperatures of the hot rod from 
1100° ©. to 1800° O., and for various voltages up to the - 
point at which an are forms between the carbons, the 
experiments being made in air at natural pressure. 
If J is the current of negative ions, E the potential 
gradient at the anode surface, 4 the length of the last free 
run of the negative ions at the end of which they reach 
the anode, it is suggested that arcing commences when 
IE attains a value sufficient to raise a part of the 
anode surface to such a temperature that positive ions are 
somewhat freely emitted. 
