The orbits of comets are, as we have seen, very elliptic. In some instances this ellipticity is so great as to take the bodies out into space to nearly six times the distance of Neptune from the sun. For a long time, indeed, it was considered that comets were of two kinds, namely, those which actually _belonged_ to the solar system, and those which were merely _visitors_ to it for the first and only time--rushing in from the depths of space, rapidly circuiting the sun, and finally dashing away into space again, never to return. On the contrary, nowadays, astronomers are generally inclined to regard comets as permanent members of the solar system. The difficulty, however, of deciding absolutely whether the orbits of comets are really always _closed_ curves, that is to say, curves which must sooner or later bring the bodies back again towards the sun, is, indeed, very great. Comets, in the first place, are always so diffuse, that it is impossible to determine their exact position, or, rather, the exact position of that important point within them, known as the centre of gravity. Secondly, that stretch of its orbit along which we can follow a comet, is such a very small portion of the whole path, that the slightest errors of observation which we make will result in considerably altering our estimate of the actual shape of the orbit. Comets have been described as so transparent that they can pass across the sky without dimming the lustre of the smallest stars, which the thinnest fog or mist would do. This is, indeed, true of every portion of a comet except the nucleus, which is, as its name implies, the densest part. And yet, in contrast to this ghostlike character, is the strange fact that when comets are of a certain brightness they may actually be seen in full daylight. As might be gathered from their extreme tenuity, comets are so exceedingly small in mass that they do not appear to exert any gravitational attraction upon the other bodies of our system. It is, indeed, a known fact that in the year 1886 a comet passed right amidst the satellites of Jupiter without disturbing them in the slightest degree. The attraction of the planet, on the other hand, so altered the comet's orbit, as to cause it to revolve around the sun in a period of seven years, instead of twenty-seven, as had previously been the case. Also, in 1779, the comet known as Lexell's passed quite close to Jupiter, and its orbit was so changed by that planet's attraction that it has never been seen since. The density of comets must, as a rule, be very much less than the one-thousandth part of that of the air at the surface of our globe; for, if the density of the comet were even so small as this, its mass would _not_ be inappreciable.