Again, it seems that such falls occur periodically, or rather that at regular intervals great meteoric streams pour upon the sun's surface. For instance, the periodic increase and decrease in the number of sun-spots is accompanied (so far as we can judge by the observations made at Edinburgh and Greenwich) by an accession and diminution of the solar heat; and if the change is attributed to the passage of a meteoric stream athwart the sun, we should have to assign to such a stream a period of rather more than eleven years. This, from what we know about the association between meteors and comets, would correspond simply to the existence of a comet whose path intersects the sun's globe, and which is followed by a train of millions of large meteoric masses, many of which are consumed at each passage of the rich portion of the train athwart the globe of the sun. This comet must of necessity be inconspicuous, since it has hitherto escaped detection. In fact, its head and nucleus must long since have been entirely destroyed. Only the meteoric train, far more widely scattered, remains, simply because at each passage past the sun, though many are captured, far greater numbers get safely past. I am careful to remind the reader that though I have, for convenience, used the indicative mood in describing these matters, I am in reality presenting merely a theory. It may be that the solar spots and the accessions of heat are produced in some other way. But I must admit I find strong reasons for regarding as probable the general theory, that the alternations of solar activity (not the solar activity itself be it noted) are excited from without. And since we know, as a matter of fact, that meteors exist in enormous numbers within the solar system, and that they aggregate with rapidly increasing density in the sun's neighbourhood, we must believe that they fall upon the sun in enormous numbers. We also perceive that the supply cannot be uniform, but must vary greatly from time to time; while what we know about the periodicity of meteoric showers on our own earth suggests the belief, we may almost say the certainty, that there must be periodic downfalls of very heavy meteoric showers upon the sun's surface. We have, then, strong probability in favour of the belief that events may occur which, _if_ they occurred, might be expected, with a high degree of probability, to produce effects resembling those actually observed,--viz., the production of a heat more intense than usual, accompanied by signs of great disturbance like the sun-spots. It does, therefore, seem at least not improbable that these accessions of heat and these signs of great disturbance really are brought about in the way supposed.