s
seen before its report is heard, or by listening to the noise of a
flash of lightning. Sound takes five seconds to travel a mile--it has
about the same speed as a rifle bullet; but light is much quicker than
that.

The rude experiment suggested by Galileo was to send two men with
lanterns and screens to two distant watch-towers or neighbouring
mountain tops, and to arrange that each was to watch alternate displays
and obscurations of the light made by the other, and to imitate them as
promptly as possible. Either man, therefore, on obscuring or showing his
own light would see the distant glimmer do the same, and would be able
to judge if there was any appreciable interval between his own action
and the response of the distant light. The experiment was actually tried
by the Florentine Academicians,[22] with the result that, as practice
improved, the interval became shorter and shorter, so that there was no
reason to suppose that there was any real interval at all. Light, in
fact, seemed to travel instantaneously.

Well might they have arrived at this result. Even if they had made far
more perfect arrangements--for instance, by arranging a looking-glass at
one of the stations in which a distant observer might see the reflection
of his own lantern, and watch the obscurations and flashings made by
himself, without having to depend on the response of human
mechanism--even then no interval whatever could have been detected.

If, by some impossibly perfect optical arrangement, a lighthouse here
were made visible to us after reflection in a mirror erected at New
York, so that the light would have to travel across the Atlantic and
back before it could be seen, even then the appearance of the light on
removing a shutter, or the eclipse on interposing it, would seem to
happen quite instantaneously. There would certainly be an interval: the
interval would be the fiftieth part of a second (the time a stone takes
to drop 1/13th of an inch), but that is too short to be securely
detected without mechanism. With mechanism the thing might be managed,
for a series of shutters might be arranged like the teeth of a large
wheel; so that, when the wheel rotates, eclipses follow one another very
rapidly; if then an eye looked through the same opening as that by which
the light goes on its way to the distant mirror, a tooth might have
moved sufficiently to cover up this space by the time the light
returned; in which case the whole would ap