uired in selecting a consecutive series of drawings. The
only way is to make a considerable number of drawings of each stage, and
then to pick out a consecutive series. Now, whenever judgment has to be
used, there is room for error of judgment, and moreover, it is
impossible to put together the drawings so as to tell a consecutive
story, without being guided by some theory, such as I have already
sketched, as to the nature of the motion and the conditions that govern
it. You will therefore be good enough to remember that this chronicle of
the events of a tenth of a second is not a mechanical record but is
presented by a fallible human historian, whose account, like that of
any other contemporary observer, will be none the worse for independent
confirmation. That confirmation is fortunately obtainable. In an attempt
made eighteen years ago to photograph the splash of a drop of mercury, I
was unable to procure plates sufficiently sensitive to respond to the
very short exposures that were required, and consequently abandoned the
endeavour. But in recent years plates of exquisite sensitiveness have
been produced, and such photographs as those taken by Mr. Boys of a
flying rifle bullet have shown that difficulties on the score of
sensitiveness have been practically overcome. Within the last few weeks,
with the valuable assistance of my colleague at Devonport, Mr. R.S.
Cole, I have succeeded in obtaining photographs of various splashes.
Following Prof. Boys' suggestion, we employed Thomas's cyclist plates,
or occasionally the less sensitive "extra-rapid" plates of the same
makers, and as a developer, Eikonogen solution of triple strength, in
which the plates were kept for about 40 minutes, the development being
conducted in complete darkness.

A few preliminary trials with the self-induction spark produced at the
surface of mercury by the apparatus that you have seen at work, showed
that the illumination, though ample for direct vision, was not
sufficient for photography. When the current strength was increased, so
as to make the illumination bright enough for the camera, then the spark
became of too great duration, for it lasted for between 4 and 5
thousandths of a second, within which time there was very perceptible
motion of the drop and consequent blurring. It was therefore necessary
to modify the apparatus so as to employ a Leyden-jar spark whose
duration was probably less than 10-millionths of a second. A very
slight ch