62 INDUCTION SHOCKS 
sparking, so that the sources of error for makes are not 
the same as for breaks. As a circuit is made the re- 
sistance falls from infinity to the resistance of the closed 
circuit itself. It is during the change from the first of 
these resistances to the second that the secondary cur- 
rent is induced. The more nearly instantaneous the 
change, the greater is the physiological intensity of the 
induced current. In hand-operated metal-contact keys 
there can be no assurance that the contact points will be 
pressed together with the same firmness twice in succes- 
sion, so that to secure uniformity of contact automatic 
keys are required for make shocks as well as for breaks. 
A further and more serious defect in metal-contact keys 
for make shocks is their liability to rebound slightly, or 
to slip sidewise, jthus giving not a single clean-cut make, 
but a succession of make, break, and make. So con- 
stantly has this defect shown itself in my experiments, 
even with carefully constructed automatic metal-contact 
keys, that I have found it necessary to use mercury con- 
tacts altogether in studying make shocks. 
The considerations stated above lead to the following 
conclusions: That hand-operated keys are not to be de- 
pended on for uniform makes and breaks ; that for break 
shocks platinum contacts are to be preferred to mercury 
because of their less volatilization, while for make 
shocks, on account of the rebound or side-slip of metal 
contacts, mercury affords the only trustworthy contact. 
