were taken by stimulating A and B
alternately, on mounting in the cell, which was filled with water. Hot
water was then substituted, and the cell was allowed to cool down to
its original temperature. The six following pairs of responses were then
taken. That this beneficial effect of annealing was not due to any
accidental circumstance will be seen from the fact that _both_ wires
have their sensitiveness equally enhanced.

(4) In addition to this mode of annealing, both wires may be
short-circuited and vibrated for a time. Lastly (5) slight stretching
_in situ_ will also sometimes be found beneficial. For this purpose I
have a screw arrangement.

By one or all of these methods, with a little practice, it is always
possible to bring the wires to a normal condition. The responses
subsequently obtained become extraordinarily consistent. There is
therefore no reason why perfect results should not be arrived at.

[Illustration: FIG. 59.--UNIFORM RESPONSES IN TIN]

#Effect of single stimulus.#--The accompanying figure (fig. 59) gives a
series, each of which is the response curve for a single stimulus of
uniform intensity, the amplitude of vibration being kept constant. The
perfect regularity of responses will be noticed in this figure. The wire
after a long period of rest may be in an abnormal condition, but after a
short period of stimulation the responses become extremely regular, as
may be noticed in this figure. Tin is, usually speaking, almost
indefatigable, and I have often obtained several hundreds of successive
responses showing practically no fatigue. In the figure it will be
noticed that the rising portion of the curve is somewhat steep, and the
recovery convex to the abscissa, the fall being relatively rapid in its
first, and less rapid in its later, parts. As the electric variation is
the concomitant effect of molecular disturbance--a temporary upset of
the molecular equilibrium--on the cessation of the external stimulus,
the excitatory state, and its expression in electric variation,
disappear with the return of the molecules to their condition of
equilibrium. This process is seen clearly in the curve of recovery.

Different metals exhibit different periods of recovery, and this again
is modified by any influence which affects the molecular condition.

That the excitatory state persists for a time even on the cessation of
stimulus can be independently shown by keeping the galvanometer circuit
open during th