In a hasty consideration of the subject, it might be thought that as the comet was dashing along with enormous velocity the tail was merely streaming out behind, just as the shower of sparks from a rocket are strewn along the path which it follows. This would be an entirely erroneous analogy; the comet is moving not through an atmosphere, but through open space, where there is no medium sufficient to sweep the tail into the line of motion. Another very remarkable feature is the gradual growth of the tail as the comet approaches the sun. While the body is still at a great distance it has usually no perceptible tail, but as it draws in the tail gradually develops, and in some cases reaches stupendous dimensions. It is not to be supposed that this increase is a mere optical consequence of the diminution of distance. It can be shown that the growth of the tail takes place much more rapidly than it would be possible to explain in this way. We are thus led to connect the formation of the tail with the approach to the sun, and we are accordingly in the presence of an enigma without any analogy among the other bodies of our system. That the comet as a whole is attracted by the sun there can be no doubt whatever. The fact that the comet moves in an ellipse or in a parabola proves that the two bodies act and react on each other in obedience to the law of universal gravitation. But while this is true of the comet as a whole, it is no less certain that the tail of the comet is _repelled_ by the sun. It is impossible to speak with certainty as to how this comes about, but the facts of the case seem to point to an explanation of the following kind. We have seen that the spectroscope has proved with certainty the presence of hydrocarbon and other gases in comets. But we are not to conclude from this that comets are merely masses of gas moving through space. Though the total quantity of matter in a comet, as we have seen, is exceedingly small, it is quite possible that the comet may consist of a number of widely scattered particles of appreciable density; indeed, we shall see in the next chapter, when describing the remarkable relationship between comets and meteors, that we have reason to believe this to be the case. We may therefore look on a comet as a swarm of tiny solid particles, each surrounded by gas.