is rotation of these appendages
of Saturn. If the ring were composed of a vast number of small bodies,
then the third law of Kepler will enable us to calculate the time which
these tiny satellites would require to travel completely round the
planet. It appears that any satellite situated at the outer edge of the
ring would require as long a period as 13 hrs. 46 min., those about the
middle would not need more than 10 hrs. 28 min., while those at the
inner edge of the ring would accomplish their rotation in 7 hrs. 28 min.
Even our mightiest telescopes, erected in the purest skies and employed
by the most skilful astronomers, refuse to display this extremely
delicate phenomenon. It would, indeed, have been a repetition on a grand
scale of the curious behaviour of the inner satellite of Mars, which
revolves round its primary in a shorter time than the planet itself
takes to turn round on its own axis.

[Illustration: Fig. 66.--Prof. Keeler's Method of Measuring the
Rotation of Saturn's Ring.]

But what the telescope could not show, the spectroscope has lately
demonstrated in a most effective and interesting manner. We have
explained in the chapter on the sun how the motion of a source of light
along the line of vision, towards or away from the observer, produces a
slight shift in the position of the lines of the spectrum. By the
measurement of the displacement of the lines the direction and amount of
the motion of the source of light may be determined. We illustrated the
method by showing how it had actually been used to measure the speed of
rotation of the solar surface. In 1895 Professor Keeler,[26] Director of
the Allegheny Observatory, succeeded in measuring the rotation of
Saturn's ring in this manner. He placed the slit of his spectroscope
across the ball, in the direction of the major axis of the elliptic
figure which the effect of perspective gives the ring as shown by the
parallel lines in Fig. 66 stretching from E to W. His photographic
plate should then show three spectra close together, that of the ball of
Saturn in the middle, separated by dark intervals from the narrower
spectra above and below it of the two handles (or ansae, as they are
generally called) of the ring. In Fig. 67 we have represented the
behaviour of any one line of the spectrum under various suppositions as
to rotation or non-rotation of Saturn and the ring. At the top (1) we
see how each line would look if there was no rotatory motion; the