EXCURSIONS OF THE CAPILLARY ELECTROMETER. 
91 
The decrease of internal electrical resistance as the meniscus approaches the tip of 
the capillary, tends to increase its velocity, while the increase of sensitiveness makes 
it move more slowly. But the latter has a much more powerful effect than the former, 
and may practically neutralize it, or even overpower it, without introducing a calibra¬ 
tion error great enough to make an appreciable difference between the scale readings 
and the true value of an excursion. 
The time-relations of the movement are conditioned partly by electrical resistance, 
but mainly by some other cause—probably mechanical friction. 
The Production and Analysis of Photographs of the Excursions of the Capillary 
Electrometer. 
The rapid movement of the sensitive plate required to bring out the details of the 
electric phenomena of muscle and nerve, necessitated an alteration in the form of 
the apparatus. The ingenious arrangement devised by the Rev. F. G. Smith for 
producing a rectilinear motion of uniform velocity was inapplicable, owing to lack 
of space in the dark room. I therefore made an apparatus, the details of which will 
be described at length elsewhere, in which the sensitive plate was caused to describe 
an arc of a circle. The dark-slide containing it was attached to a kind of balanced 
pendulum, which carried it, at a uniform velocity, past the slit on which the magnified 
image of the column of mercury was thrown. The requisite velocity was given to the 
pendulum by a weight which, as in Atwood’s machine, was caught by a stop just before 
the plate reached the slit. The return of the pendulum after the exposure was 
prevented by a catch, and the key producing the excursion was actuated at the right 
moment by an arrangement of electromagnets. Time was recorded upon the plates 
by a magnetic vibrator placed in front of the slit, and driven by a tuning-fork in the 
usual way. 
The exact instant of excitation was recorded by the signal-key before referred to, 
which was also placed in front of the slit. 
With this arrangement the normal curve is most easily expressed in polar coordi¬ 
nates. 
Time being recorded upon a circular arc, 
t becomes 6. 
Instead of the rectilinear asymptote, there is an asymptotic circle of radius = R. 
The expression for the radius vector is 
r = R ± y, 
the equation connecting y and 6 being 
y — ae~ cd . 
N 2 
