ywhere else in our system, we can feel but little
doubt that Uranus must rotate about an axis. The ordinary means of
demonstrating this rotation can be hardly available in a body whose
surface appears so small and so faint. The period of rotation is
accordingly unknown. The spectroscope tells us that a remarkable
atmosphere, containing apparently some gases foreign to our own, deeply
envelops Uranus.

There is, however, one feature about Uranus which presents many points
of interest to those astronomers who are possessed of telescopes of
unusual size and perfection. Uranus is accompanied by a system of
satellites, some of which are so faint as to require the closest
scrutiny for their detection. The discovery of these satellites was one
of the subsequent achievements of Herschel. It is, however, remarkable
that even his penetration and care did not preserve him from errors
with regard to these very delicate objects. Some of the points which he
thought to be satellites must, it would now seem, have been merely stars
enormously more distant, which happened to lie in the field of view. It
has been since ascertained that the known satellites of Uranus are four
in number, and their movements have been made the subject of prolonged
and interesting telescopic research. The four satellites bear the names
of Ariel, Umbriel, Titania, and Oberon. Arranged in order of their
distance from the central body, Ariel, the nearest, accomplishes its
journey in 2 days and 12 hours. Oberon, the most distant, completes its
journey in 13 days and 11 hours.

The law of Kepler declares that the path of a satellite around its
primary, no less than of the primary around the sun, must be an ellipse.
It leaves, however, boundless latitude in the actual eccentricity of the
curve. The ellipse may be nearly a circle, it may be absolutely a
circle, or it may be something quite different from a circle. The paths
pursued by the planets are, generally speaking, nearly circles; but we
meet with no exact circle among planetary orbits. So far as we at
present know, the closest approach made to a perfectly circular movement
is that by which the satellites of Uranus revolve around their primary.
We are not prepared to say that these paths are absolutely circular. All
that can be said is that our telescopes fail to show any measurable
departure therefrom. It is also to be noted as an interesting
circumstance that the orbits of the satellites of Uranus all lie i