with a small telescope. But when the earth is on the opposite side
of the sun [Page 23] from Jupiter, at E', these eclipses at J' take
place sixteen and a half minutes too late. What is the reason? Is
the celestial chronometry getting deranged? No, indeed; these great
worlds swing never an inch out of place, nor a second out of time.
By going to the other side of the sun the earth is 184,000,000 miles
farther from Jupiter, and the light that brings the intelligence of
that eclipse consumes the extra time in going over the extra
distance. Divide one by the other and we get the velocity, 185,000
miles per second. That is probably correct to within a thousand
miles. Methods of measurement by the toothed wheel of Fizeau confirm
this result. Suppose the wheel, Fig. 5, to have one thousand teeth,
making five revolutions to the second. Five thousand flashes of
light each second will dart out. Let each flash travel nine miles to
a mirror and return. If it goes that distance in 1/10000 of a
second, or at the rate of 180,000 miles a second, the next tooth
will have arrived before the eye, and each returning ray be cut off.
Hasten the revolutions a little, and the next notch will then admit
the ray, on its return, that went out of each previous notch: the
eighteen miles having been traversed meanwhile. The method of
measuring by means of a revolving mirror, used by Faucault, is held
to be even more accurate.

[Illustration: Fig. 5.--Measuring the Velocity of Light.]

When we take instantaneous photographs by the exposure [Page 24] of
the sensitive plate 1/20000 part of a second, a stream of light nine
miles long dashes in upon the plate in that very brief period of
time.