228 J. A. POLLOCK, E. M. WELLISCH AND A. B. B. RANCLAUD. 
tions, a fact which seems to support the view taken in this 
and the previous paper as to the mode of growth of the arc 
from the non-luminous discharge. Figure 12 isan example 
of many of the photographs, showing that, in the case of 
the reversal of the connections, the new crater commences 
on cool rather than hot carbon. 
7. Summary.—In connection with the relighting of the 
carbon arc, without movement of the electrodes, when the 
circuit is opened and reclosed, the relation between the 
potential difference, established between the carbons at 
the moment of the remaking of the connections, and the 
maximum time of interruption of the circuit, within which 
the arc will reform, has been investigated for cases differ- 
ing as to the relative positions of the carbons before the 
opening of the circuit, and as to the direction of the 
potential difference after the reclose of the connections. 
The problem is that of the change from a non-luminous to 
a luminous discharge in air at normal pressure, of which an 
explanation, in the case where negative electricity flows 
from a hot to a cool carbon, has been given in a previous 
paper. In the present experiments both carbons are at a 
high temperature and the conditions of the change 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. In some of the cases examined, the 
relations show a critical characteristic, but sufficient data 
are not available to enable an explanation of this result to 
be given. Photographs have been taken showing stages in 
the formation of the arc; it is seen that the glow grows from 
the electrode which is positive on the remake of the circuit, 
a fact which seems to support the view taken in this and 
the previous paper as to the mode of development of the 
arc from the non-luminous discharge. 
We are indebted to Mr. H. L. Watkins, B.a., and Mr. 
L. A. Cotton, B.a., B.sc.,for help in connection with the 
earlier part of the investigation. 
