the gravitation at that surface
which we see is about 2-1/2 times what it is on the surface of the
earth. In accordance, therefore, with the kinetic theory, we may expect
the planet to retain an extensive layer of gases around it; and this is
confirmed by the spectroscope, which gives evidence of the presence of a
dense atmosphere.

All things considered, it may be safely inferred that the interior of
Jupiter is very hot, and that what we call its surface is not the actual
body of the planet, but a voluminous layer of clouds and vapours driven
upwards from the heated mass underneath. The planet was indeed formerly
thought to be self-luminous; but this can hardly be the case, for those
portions of the surface which happen to lie at any moment in the
shadows cast by the satellites appear to be quite black. Again, when a
satellite passes into the great shadow cast by the planet it becomes
entirely invisible, which would not be the case did the planet emit any
perceptible light of its own.

In its revolutions around the sun, Jupiter is attended, so far as we
know, by seven[22] satellites. Four of these were among the first
celestial objects which Galileo discovered with his "optick tube," and
he named them the "Medicean Stars" in honour of his patron, Cosmo de
Medici. Being comparatively large bodies they might indeed just be seen
with the naked eye, were it not for the overpowering glare of the
planet.

It was only in quite recent times, namely, in 1892, that a fifth
satellite was added to the system of Jupiter. This body, discovered by
Professor E.E. Barnard, is very small. It circulates nearer to the
planet than the innermost of Galileo's moons; and, on account of the
glare, is a most difficult object to obtain a glimpse of, even in the
best of telescopes. In December 1904 and January 1905 respectively, two
more moons were added to the system, these being found by _photography_,
by the American astronomer, Professor C.D. Perrine. Both the bodies in
question revolve at a greater distance from the planet than the
outermost of the older known satellites.

Galileo's moons, though the largest bodies of Jupiter's satellite
system, are, as we have already pointed out, very small indeed when
compared with the planet itself. The diameters of two of them, Europa
and Io, are, however, about the same as that of our moon, while those of
the other two, Callisto and Ganymede, are more than half as large again.
The recently discove