from the Sun's surface even in vacuo to the observed limits of the corona. But as such matter, when first ejected from the Sun, must necessarily be in the state of vapour, and as it has to pass through the Sun's vaporous atmosphere (which close by the photosphere must be of no inconsiderable density), there must be enormous retardative effect, and consequently a much greater initial velocity would be required.

We have, however, in the jet prominences indisputable evidence of velocities of ejection as fully beyond our powers of conception as the enormous initial velocities just referred to; and I think these prominences supply strong reason for believing that movements of far greater velocity must take place within the solar orb. The jet prominences quite commonly reach to a height exceeding 50,000 or 60,000 miles, and some of them have been 80,000 or 90,000 miles in height. Now to propel a projectile in vacuo to a height of 80,000 miles, an initial velocity of about 140 miles per second would be required. But the erupted matter is not a solid projectile flung through empty space, but glowing hydrogen (and other vaporous elements) flung through a hydrogen atmosphere. Now a cannon ball, though propelled with a velocity of less than a third of a second per mile, is yet so retarded by atmospheric resistance, that its range is a mere fraction of that calculated on the supposition that there is no resistance. It will be understood how enormously the initial velocity of these hydrogen projectiles (so to describe them) must exceed that inferred from the hypothesis of an

ordinary projectile travelling in vacuo. So that it cannot but be regarded as probable (though it may be perfectly inexplicable) that the glowing hydrogen of the jet prominences is ejected with a velocity many times exceeding that which would be necessary to carry matter altogether away from the Sun, if he were not surrounded by an atmosphere.* Now since other matter is ejected along with the hydrogen, and some of this other matter is certainly denser than hydrogen, under equal circumstances of pressure and temperature, we cannot but believe that such matter undergoes far less retardation than the hydrogen in passing through the solar atmosphere. How much less we cannot say; but it seems no rash assumption to suppose that the metallic vapours (à fortiori if they condensed into metallic globules) would retain a velocity sufficient to carry them even beyond the observed limits of the

corona.

But if any evidence were really required to prove that the power of the solar eruptions is sufficient for our theory, we need only consider the circumstances of the eruption witnessed by Professor Young on September 7, 1871 (see pp. 318-322). For it will be noticed that he saw matter carried from a height of less than 100,000 miles to a height of more than 200,000 miles

✦ We could scarcely expect to detect spectroscopic evidence of these enormous initial velocities-first, because it is probable that the real jets have a relatively minute cross-section; and, secondly, because by communicating motion (gradually diminishing with distance) to the surrounding photospheric and chromatospheric matter, the outflowing hydrogen tends to mask (by widening the lines) the spectroscopic evidence of its own motion.

in ten minutes. Now I have gone through the requisite calculation for determining in what interval of time this distance would be traversed by a projectile in vacuo, having 200,000 miles as the upper limit of its flight; and I find that no less than twenty-five minutes fifty-six seconds would be required. Now on the supposition that there was no retardation, we should have to suppose that when the hydrogen wisps became invisible (doubtless through cooling), they still continued to move upwards, and that they reached a height of about 350,000 miles before their upward motion ceased. For in this case the observed distance would be traversed in the observed time. Even this altogether incredible supposition forces us to believe in a velocity of ejection of about 255 miles per second. But if we assume (what is altogether more likely) that the hydrogen ceased to move upwards after reaching a height of about 200,000 miles, we must infer that it was projected with a much greater velocity than 255 miles per second. For this velocity would only carry it across the observed range in the observed time when not subject to any retardation within that range. But it was subjected in reality to such an enormous retardation that, instead of passing to a height of 350,000 miles, it was brought to rest at a height of little more than 200,000 miles. Therefore when at a height of 100,000 miles (or at the beginning of its ob

*

The whole reasoning is very much strengthened if we suppose that the real seat of eruption lay some distance from the Sun's limb,the motion of the hydrogen was so seen as to be foreshortened.

-or that

served course) it must have been travelling much faster than a projectile, in vacuo, making for a height of 350,000 miles. So that even if projected to that height through a vacuum the uprushing hydrogen would have required a much greater initial velocity than 255 miles per second. But so far from the retardation being nil in the lower half of the hydrogen's course, it must have enormously exceeded the retardation in the upper half (since the atmospheric pressure increases not only rapidly, but with constantly increasing rapidity, as the Sun's surface is approached). Thus it is not merely probable, but certain, that in the case of this outburst, the ejectional velocity exceeded enormously-exceeded many times -the utterly inconceivable velocity of 255 miles per second.

Furthermore, since the velocity due to a projectile in vacuo, flung from the Sun's surface to a height of 200,000 miles amounts to but 212 miles per second, it is not only probable but certain, that in this instance the velocity thus estimated falls enormously short of the real velocity of ejection. Hence, in all jet prominences the like must be inferred. It may be concluded that the matter of the smallest jet prominence actually recognisable as such must have been ejected with a velocity sufficient, if there were no atmosphere, to carry the matter for ever away from the Sun.

Now when we remember what the spectroscope has taught us respecting the constitution of the corona, we cannot but see that the inference deducible from the

EE

preceding reasoning-the inference, namely, that erupted metallic vapours pass away to a far greater distance than the erupted hydrogen-is very much strengthened. It is noteworthy also that the polariscopic evidence has been regarded by Professor Stokes (the highest English authority on the subject) as seeming to imply the condensation of metallic vapour moving on radial lines from the Sun, into liquid or solid particles of exceeding minuteness.*

And here we are brought to consider several matters which bring our present subject into somewhat unexpected relationship with other departments of cosmical research. Space will only permit me to touch very briefly upon them.

If matter is thus propelled from the Sun so constantly as to account for the permanence of the corona, it can scarcely be doubted that, exceptionally perhaps, or perhaps not uncommonly, matter is projected as far as the Earth's orbit, or even beyond. One would even expect that no inconsiderable proportion of the Sun's substance might be ejected with such velocity as to pass for ever away from his dominion.

Now in the case first mentioned, it would be reasonable to anticipate that, as a direct result, the number of aerolitic masses falling on our earth by day would somewhat exceed the number falling at night. For if

He conceives that their dimensions fall short of the wave-lengths of light; and he suggests magnesium as the element principally in question. It does not appear, however, that he lays stress (so far as optical considerations are concerned) on the nature of the elementary constitution of these metallic clouds.