been projected from volcanoes on the earth, while the other assumes that they have come from volcanoes on the moon. Observation has shown that when Mount Etna is in full activity, the masses of stone thrown from its crater have a velocity of less than 1,600 feet per second, which is but one-112th part of the mean velocity with which shooting-stars are observed to move. The theory that falling-stars come from the moon was first propounded by Terzago, an Italian, in the seventeenth century. It appears, however, to have been not unknown in ancient times, since we learn that the Syrian astronomers were in the habit of looking for shooting-stars when the moon was full; while Greek astronomers considered the most favourable season to be at the time of lunar eclipse, that is, when the moon is full, but the sky dark. Bizarre as it may seem, this fanciful explanation has been thought worthy of strict mathematical examination by such astronomers as Laplace, Olbers, and Poisson. It appears, from their calculations, that the velocity with which stone-showers should be propelled from the moon in order to reach our earth with the velocities observed among shooting-stars may be considered to be utterly beyond the powers we could concede to lunar volcanoes, even if it were proved (which is far from being the case) that any active volcanoes now exist on the moon's surface.

The three theories just considered have been effectually overthrown by the simple observation of the height and velocities of shooting-stars. When we

add to this consideration the recurrence of star-showers, not in particular states of the earth's atmosphere, not connected in any way with the activity of terrestrial volcanoes, nor conceivably with the action of assumed lunar volcanoes, these theories appear yet more inadequate to explain observed phenomena. The phenomenon of radiant points, lastly, is so wholly inexplicable on any of these theories, that we may dismiss them finally, as utterly untenable.

We must, therefore, turn to the theory which had already been suggested by Greek philosophers-that shooting-stars and meteors are extraneous bodies dragged towards the earth by the force of her attractive influence. But modern scientific discoveries enable us to exhibit this theory in a more inviting form, and at the same time to offer analogues obviously tending to confirm the hypothesis. The discovery of a zone of planetoids, the inquiry into the nature of the zodiacal light, and the mathematical examination of the 'stability' of the Saturnian ring-system, have led astronomers to recognise the existence in the solar system of minute bodies travelling in zones or clusters around a central orb. There is, therefore, nothing unreasonable in the supposition that there are zones and clusters of such bodies travelling round the sun in orbits which intersect the earth's path. When in her course around the sun she encounters any of the bodies forming such zones and clusters, they are ignited by friction as they pass through the upper layers of the air, and become visible as shooting-stars or meteors according to their

dimensions; or they may even fall upon her surface as aërolites.

The recurrence of star-showers is a necessary consequence of the hypothesis we are considering. For, if we suppose the zones of meteors, or the orbits of meteor-clusters, to have a fixed position in the solar system, or to be subject to those slow, progressive or retrogressive shiftings with which the study of the solar system familiarises us, there will necessarily result a regular recurrence of showers either on fixed days, or on days uniformly shifting round among the This is precisely what is observed with the fifty-six recognised star-showers.

seasons.

The earth does not necessarily (or probably) pass centrally through a meteor-cluster every year, nor probably are the meteor-zones uniformly rich throughout. Thus we can readily understand periodic undulations in the intensity of star-showers, or even periodic intermittences.

The phenomenon of radiant points also is not merely reconcilable with, but obviously indicates the hypothesis we are considering. For during the brief interval occupied by the earth in passing through a well-marked zone or cluster, the bodies composing such zone or cluster may be considered to be moving (relatively to the moving earth) in parallel lines. Therefore, by a well-known law in perspective, their apparent paths, viewed from the earth, must have a vanishing point' on the celestial sphere, that is, a 'radiant point' among the fixed stars.

6

The remarkable velocity with which shooting-stars travel is satisfactorily accounted for by the modern theory. If we suppose zones and clusters of cosmical bodies (pocket-planets we may term them with Humboldt) to be travelling in different directions around the sun, it is clear that the members of those zones which travel in the same direction as the earth, will overtake, or be overtaken by her, with the difference of their respective velocities, while those which travel in the contrary direction will encounter the earth with the sum of their own and the earth's velocity. Now, just as in walking along a crowded road we meet many more people than we overtake, or are overtaken by; so, clearly, by far the larger number of observed shooting-stars must belong to the latter class named above, and therefore the average observed velocity will not fall very far short of the sum of the velocities of the earth and the shooting-star system.

Fairly considered, the modern theory may be looked upon as established: for, first, all other available hypotheses have been shown to be untenable; and, secondly, the most remarkable shooting-star phenomena are shown to be consistent with, or rather to point directly to, the modern hypothesis. It remains only that some minor peculiarities should be noticed.

It has been remarked that shooting-stars are much more commonly seen in the months from July to December, than in those from January to June. Remembering that this remark refers to observations

made in our northern hemisphere, it is easily reconciled with the modern theory, when we consider that the north pole is on the forward hemisphere of the earth (considered with reference to her orbital motion) during the first-named period, and on the rear (or sheltered) hemisphere during the second.

Again, it has been remarked that shooting-stars are seen more commonly in the hours after midnight, and that aërolites fall more commonly before noon. In other words, these extraneous bodies reach the earth (or her atmosphere) more frequently in the hours from midnight to noon than in those from noon to midnight. Humboldt suggests in explanation we know not what theory of variation in the ignition-powers of different hours. But it is clear that the true explanation is founded on the principle presented in the preceding paragraph, since the forward hemisphere contains places whose local time lies, roughly speaking, between midnight and noon, while places whose local hour lies between noon and midnight lie on the sheltered hemisphere.

If we remember that the earth is but a point in space, we may fairly conclude that the number of bodies composing meteor-zones is all but infinite. Large, therefore, as the numbers of these bodies which fall on the earth may be, there is no reason to suppose (perhaps if we knew the true functions of these bodies, we might say-there is no reason to fear) that the supply of meteors will ever be perceptibly diminished. Although the contrary opinion is often expressed, it is