It was in 1864, a year further made memorable astronomically by Sir William Huggins's discovery of the gaseous nature of some of the nebulæ, that the spectroscope was first applied to the study of a comet. The celestial visitor thus put to the question, a comet discovered by Tempel, was in nowise a distinguished object, appearing like a star of the second magnitude, or less, with a feeble though fairly long tail. When analyzed by Donati, it was found to yield a spectrum consisting of three bright bands, yellow, green, and blue, separated by dark spaces. This observation at once modified ideas as to cometary structure. Hitherto it had been supposed that comets shone by reflected light; but Donati's observation revealed beyond question that the light of the 1864 comet at all events was inherent, and that, so far as the observation went, the comet consisted of glowing gas.
Great Comet. Photographed May 5, 1901, with the 13-inch Astrographic Refractor of the Royal Observatory, Cape of Good Hope.
In 1868 Sir William Huggins carried the matter one step further by showing that the spectrum of Winnecke's comet of that year agreed with that of olefiant gas rendered luminous by electricity; and the presence of the hydrocarbon spectrum has since been detected in a large number of comets. The first really brilliant comet to be analyzed by the spectroscope was Coggia's (1874), and it presented not only the three bright bands that had been already seen, but the whole range of five bands characteristic of the hydrocarbon spectrum. In certain cases, however—notably, that of Holmes's comet of 1892 and that of the great southern comet of 1901 (Plate [XXV.])—the spectrum has not exhibited the usual bright band type, but has instead shown merely a continuous ribbon of colour. From these analyses certain facts emerge. First, that the gaseous surroundings of comets consist mainly of hydrogen and carbon, and that in all probability their luminosity is due, not to mere solar heat, but to the effect of some electric process acting upon them during their approach to the sun; and second, that, along with these indications of the presence of luminous hydrocarbon compounds, there is also evidence of the existence of solid particles, mainly in the nucleus, but also to some extent in the rest of the comet, which shine by reflected sunlight. It is further almost certain, from the observation by Elkin and Finlay of the beginning of the transit of Comet 1882 (iii.) across the sun's face, that this solid matter is not in any sense a solid mass. The comet referred to disappeared absolutely as soon as it began to pass the sun's edge. Had it been a solid mass or even a closely compacted collection of small bodies it would have appeared as a black spot upon the solar surface. The conclusion, then, is obvious that the solid matter must be very thinly and widely spread, while its individual particles may have any size from that of grains of sand up to that of the large meteoric bodies which sometimes reach our earth.
Thus the state of the case as regards the constitution of comets is, roughly speaking, this: They consist of a nucleus of solid matter, held together, but with a very slack bond, by the power of gravitation. From this nucleus, as the comet approaches perihelion, the electric action of the sun, working in a manner at present unknown, drives off volumes of luminous gas, which form the tail; and in some comets the waves of this vapour have been actually seen rising slowly in successive pulses from the nucleus, and then being driven backwards much as the smoke of a steamer is driven. It has been found also by investigation of Comet Wells 1882 and the Great Comet of 1882 that in some at least of these bodies sodium and iron are present.
The question next arises, What becomes of comets in the end? Kepler long ago asserted his belief that they perished, as silkworms perish by spinning their own thread, exhausting themselves by the very efforts of tail-production which render them sometimes so brilliant to observation; and this seems to be pretty much the case. Thus Halley's comet, which was once so brilliant and excited so much attention, was at its last visit a very inconspicuous object indeed. At its apparition in 1845-1846 Biela's comet was found to have split into two separate bodies, which were found at their return in 1852 to have parted company widely. Since that year it has never been observed again in the form of a comet, though, as we shall see, it has presented itself in a different guise. The same fate has overtaken the comets of De Vico (1844), and Brorsen (1846). The former should have returned in 1850, but failed to keep its appointment; and the latter, after having established a character for regularity by returning to observation on four occasions, failed to appear in 1890, and has never since been seen.
The mystery of such disappearances has been at least partially dispelled by the discovery, due to Schiaparelli and other workers in the same field, that various prominent meteor-showers travel in orbits precisely the same as those of certain comets. Thus the shower of meteors which takes place with greater or less brilliancy every year from a point in the constellation Perseus has been proved to follow the orbit of the bright comet of 1862; while the great periodic shower of the Leonids follows the track of the comet of 1866; the orbit of the star-shower of April 20—the Lyrids—corresponds with that of a comet seen in 1861; and the disappearance of Biela's comet appears to be accounted for by the other November shower whose radiant point is in the constellation Andromeda. In fact, the state of the matter is well summed up by Kirkwood's question: 'May not our periodic meteors be the débris of ancient but now disintegrated comets, whose matter has become distributed round their orbits?' The loosely compacted mass which forms the nucleus of the comet appears to gradually lose its cohesion under the force of solar tidal action, and its fragments come to revolve independently in their orbit, for a time in a loosely gathered swarm, and then gradually, as the laggards drop behind, in the form of a complete ring of meteoric bodies, which are distributed over the whole orbit. The Leonid shower is in the first condition, or, rather, was when it was last seen, for it seems to be now lost to us; the Perseid shower is in the second. The shower of the Andromedes has since confirmed its identity with the lost comet of Biela by displays in 1872, 1885, and 1892, at the seasons when that comet should have returned to the neighbourhood of the sun. It appears to be experiencing the usual fate of such showers, and becoming more widely distributed round its orbit, and the return in 1905 was very disappointing, the reason apparently being that the dense group in close attendance on the comet has suffered disturbance from Jupiter and Saturn, and now passes more than a million miles outside the earth's orbit.
In 1843 there appeared one of the most remarkable of recorded comets. It was not only of conspicuous brilliancy and size, though its tail at one stage reached the enormous length of 200,000,000 miles, but was remarkable for the extraordinarily close approach which it made to the sun. Its centre came as near to the sun as 78,000 miles, leaving no more than 32,000 miles between the surfaces of the two bodies; it must, therefore, have passed clear through the corona, and very probably through some of the prominences. Its enormous tail was whirled, or rather appeared to be whirled, right round the sun in a little over two hours, thus affording conclusive proof that the tail of a comet cannot possibly be an appendage, but must consist of perpetually renewed emanations from the nucleus. But in addition to these wonders, the comet of 1843 proved the precursor of a series of fine comets travelling in orbits which were practically identical. The great southern comet of 1880 proved, when its orbit had been computed, to follow a path almost exactly the same as that of its predecessor of thirty-seven years before. It seemed inconceivable that a body so remarkable as the 1843 comet should have a period of only thirty-seven years, and yet never previously have attracted attention. Before the question had been fairly discussed, it was accentuated by the discovery, in 1881, of a comet whose orbit was almost indistinguishable from that of the comet of 1807. But the 1807 comet was not due to return till A.D. 3346. Further, the comet of 1881 proved to have a period, not of seventy-four years, as would have been the case had it been a return of that of 1807, but of 2,429 years. The only possible conclusion was that here were two comets which were really fragments of one great comet which had suffered disruption, as Biela's comet visibly did, and that one fragment followed in the other's wake with an interval of seventy-four years.