THE MAKE AND BREAK OF THE PRIMARY CIRCUIT 6l 
made that this expression applies exactly only when the 
break is instantaneous, although it holds relatively so 
long as the time occupied by the break does not vary. 
Since this in turn depends on the duration of the spark, 
our present inquiry resolves itself, so far as break shocks 
are concerned, into a study of the conditions governing 
contact sparking. 
The duration of the spark at a broken primary con- 
tact depends in part upon the intensity of the primary 
current, in part upon the amount of volatilization occur- 
ring at the contact, and in part upon the speed with 
which the points are separated. This last factor ex- 
plains why keys operated by hand cannot be depended 
upon to give uniform results, and why some form of 
automatic key is required, since only thus can a uniform 
speed of separation be secured. Moreover, ordinary 
mercury keys cannot be depended on even when oper- 
ated automatically, because of the tendency of mercury 
when not absolutely clean to cling in drops and thus 
vary the speed with which the contact points actually 
separate. In practically all keys there is some volatil- 
ization; platinum contacts giving the least, ordinary 
mercury contacts the most. It is impracticable to use 
always primary currents of a single intensity; but, in 
primary currents not exceeding i ampere, the variation 
is too slight to be of practical importance. 
The making of a primary circuit is not attended with 
