would produce; and as fast as the cloud behind the head came, through the comet's motion, under full solar action, it would be destroyed. So the tail would always be behind the head.

It appeared to Professor Tyndall that the curvature of a comet's tail, or the existence of more tails than one, as in Donati's Comet (figs. 6 and 7), was not inconsistent with this interpretation. For he noticed that, according to the gas dealt with, the luminous cloud would take a longer or shorter time in becoming visible. And he suggested that when the cloud formed slowly, the tail would be curved, the part near the head being behind the position which the head had recently passed through, while the part near the end of the tail would be behind the regions through which the comet had passed much earlier. Such luminous trai s as were formed more quickly would account, he considered, for the straighter tails. He overlooked, I think, the circumstance that the shape of the luminous cloud-trail would not in reality depend at all upon the length of time which the cloud might take in becoming visible. Light would pass with the same velocity through the different kinds of tenuous gas, and whether the cloud became visible at once along the space thus passed through, or did not become visible for several seconds, or minutes, or even hours, it would become visible at the farther end of its course only just so long after it had become visible at the nearer end, as light had taken in traversing the length of cloud so formed. This interval of time would be the same for the quickly-appearing as for the slowly-appearing luminous cloud, and there would, therefore, be no difference between their forms. It would be necessary to account in this way for the curvature of the larger tail in the figure, as compared with the straightness of the smaller tails, that the curved tail should have been more slowly extended from the head; whereas the theory gives the same rate of extension for both, namely, the rate at which light travels.

We seem almost forced, by the phenomena of such a

comet as Donati's, to adopt the theory of the actual repulsion of matter from the head of the comet into the tails-matter repelled most swiftly forming the straighter tails, while

matter repelled more slowly, and seemingly in greater abundance, forms the great curved tail.

We shall proceed to consider further on the evidence which seems to show that, strange though this theory

FIG. 6.-Donati's Comet, September 24, 1858

of material repulsion may be, it is in point of fact the only admissible theory. If this shall be established, we shall have to admit the existence of a repulsive force, whose

action on the grosser material of planetary bodies is insensible.

Before we proceed to consider the theory by which, so far as can be judged at present, the phenomena of comets'

FIG. 7.--Donati's Comet, September 26, 1858.

tails can be explained, it may be well that we should consider the evidence derived from other comets than those hitherto considered.

In the first place we would direct special attention to the comet of 1811. In this comet, as may be seen from its picture in fig. 8, the various parts of the comet and its tail could be distinguished by the naked eye. There was the condensed part, called the nucleus, which in this case was apparently globular in form; the nebulous envelope which surrounds the nucleus, the so-called coma; the bright sideparts of the tail where it seems to be swept away from the coma, leaving a comparatively dark region behind the head; and the tail, widening and growing fainter with distance from the head. No one, we think, who considers this picture will for a moment imagine that the comet is a mere lens, and its tail merely the track of light condensed by this lens along the region behind the head. Here, again, the hollow structure of the tail seems indicated by the bright tracks on either side, though, as we shall endeavour to show later, the exceedingly well-defined nature of the dark track behind the nucleus in many comets seems to force upon us a different interpretation of this singular and characteristic feature,

In some respects the comet of 1811 tells us more of cometic possibilities, so to speak, than any other comet that has ever yet been observed. Discovered on March 26, 1811, this comet remained visible for a longer time than any yet seen, viz. for 16 months, 22 days. It had a tail 120 millions of miles in length, and 15 millions of miles in diameter at the widest part. The diameter of the nucleus was about 127,000 miles; that of the envelope round the head about 643,000 miles. But what was so remarkable about this comet was, that it obtained this remarkable development without approaching the sun, as other comets have done. The usual rule with comets is that the nearer they approach to the sun, the more their head and tails are developed. But the least distance of the comet of 1811 from the sun was little less than 100 millions of miles. Again, although

it had so remarkable an appearance, as seen from the earth, the distance of that comet from us was at no time less than 10 millions of miles. Its true magnitude, therefore, as Professor Kirkwood well remarks, 'has probably not been surpassed by that of any other comet which has yet been observed.' If its path had carried it nearer to the

sun, its appearance would probably have been terrible in the extreme. If we consider the enormous volume occupied by this comet and its tail, its

Million cubic miles of head,

Ten billion leagues of tail,

we shall see that the phenomena we have to interpret ought not to escape us because of minuteness of scale.