916 J. A. POLLOCK, E. M. WELLISCH AND A. B. B. RANCLAUD. 
remaking the connections, is astonishingly well defined, 
and could in our observations be determined to ‘002 second. 
If the interval between the break and the make of the 
circuit exceeds what may be called the critical time for 
the given circumstances, after reclosing the circuit a small 
non-luminous current passes between the carbons; the 
heating effects associated with this current are not sufficient 
to maintain the electrodes at their high temperatures, and 
the current soon dies away as the temperatures diminish. 
From considerations advanced in a previous paper’ with 
reference to the establishment of the cathode fall of 
potential which is such a characteristic feature of the 
developed arc, one is led to think that, on reclosing the 
circuit, this smaller current always precedes the larger 
one of the fully formed discharge; the problem of critical 
relighting is then essentially that of the change from a 
non-luminous to a luminous current under the circumstances 
of the experiments. 
In the relighting of the arc, both carbons are at a high 
temperature, and the conditions are complicated by the 
presence, at the moment of reclosing the circuit, of ions at 
the anode surface as well as near that of the cathode. 
Simpler conditions are associated with the change of 
current régime when only the negative carbon is incand- 
escent; this case, involving, previous to the formation of 
the arc, the flow of negative electricity from a hot toa 
cool carbon, has been investigated by two of us,’ and an 
explanation reached which seems to account for the 
phenomena observed. 
In the present experiments it will be seen that the flow 
of negative electricity at the moment of reclosing the 
circuit is not always from a hot to a cooler carbon; the 
+ Pollock and Ranclaud, this Journal, p. 201. 
2 Pollock and Ranclaud, loc. ctt. 
