42 PRESIDENTS ADDRESS—SECTION A. 
character. In 1858 this suggestion was worked out in all its 
details by Sir Wiliam Thomson, and Feddersen Schiller and 
others verified Sir William Thomson’s formule in cases where the 
oscillations were not unmanageably rapid. From the point of 
view which we have adopted, it does not require much imagination 
to enable us to see that since the charge of a condenser corresponds 
toadistribution of stress in the dielectric, that stress may be reversed 
if discharge is very rapid, at all events if we regard the stress as 
produced by a mechanism of any sort. This sort of guessing, 
however, clearly is out of place unless we are prepared to grant 
the mechanism something lke inertia. I do not know that we 
ought to go so far, and consequently the above considerations are 
only justified because they lead us to a known experimental result. 
As a matter of fact, Sir William Thomson’s calculation was exceed- 
ingly general, and only involved the principle of the conservation of 
energy and known experimental electrical relations. The result 
of the theory is simplicity itself. If the resistance in the discharge 
circuit is greater than a certain quantity depending on the 
capacity and selfinduction of the jar and circuit, the discharge 
will not be oscillatory at all, but the plates will fall back to zero- 
potential by jumps; in other words, the discharge will be 
intermittent. It is not dificult to see the reason of this by 
considering the induction tubes. When a discharge once begins, 
the resistance of the spark-gap enormously diminishes, the slope 
of electric and magnetic intensities in the neighbourhood gets very 
big, and the corresponding flow of energy “inuan is propor ronal to 
them enormously increases, consequently, that part of the field 
gets discharged ; and if the resistance in the other part of the 
circuit is sufficiently great, the discharge may have disappeared 
and the dielectric healed up before sufticient energy has again 
accumulated to break down the dielectric. If, on the other hand, 
the resistance is below a certain critical value, the discharge will 
4 x self-induction 
capacity, 
and when the circuit has exactly this resistance, the discharge will 
be dead beat, as it is ordinarly assumed to be in the elementary 
theory. We must recollect that as soon as the oscillations are 
set up, both the self-induction and resistance of the circuit are 
vastly different from what they are for steady currents. Lord 
R: ayleigh has shown that as the frequency increases the resistance. 
Ww ill, as a rule, increase, and the self-induction diminish. Returning 
to our concept of tubes of electric induction, we must imagine that 
at first the tubes coverge on the wire faster than they can be 
broken up, and consequently produce a state of induction which 
reverses the direction of the intensities ; and consequently of the 
current, when a short time afterwards the tubes do actually move 
in. This process may be repeated several times, or we may have. 
an oscillatory discharge of several oscillations before equilibrium 
be oscillatory. The critical resistance is 
