Its light is of a continuous nature,' although it may not prove, on further examination, to be genuinely stellar.

Curved furrows of light, such as it is agreed to call 'spiral,' have been traced in many planetary and annular nebulæ ; with still greater optical power than is now at the disposal of astronomers, they might possibly be brought to view in all. The kind of structure which they indicate seems indeed to characterise, in some degree, every form of cosmical agglomeration, and depends no doubt upon laws of their ordained development foreign to our terrestrial experience.

The tendency of winding nebulous bands to become knotted often proceeds so far that the knots all but completely absorb the bands. Multiple groups of nebulæ then appear ranged along curved lines, the intermediate faint luminosity becoming perceptible only with large telescopic apertures. Such is the curious double nebula in Perseus (M 76), noticed by the present Lord Rosse to constitute, with subordinate nodules and streamers, a system modelled on a 'reaping-hook' pattern.2 A gaseous spectrum is derived from it. Similar combinations are met with in the southern constellation Mensa (N. G. C., 2046) where five nebulæ are disposed along an oval line, and in a 'falcated' nebula with three knots situated in Cepheus (N. G. C., 7008).

All the diversities of double stars, it was pointed out by Sir John Herschel,3 have their counterparts in nebulæ ; besides which, the varieties of form and gradation of light in the latter afford room for combinations peculiar to this class of objects.' Its members are surprisingly numerous, one in sixteen of the 5079 nebulæ catalogued by Herschel in 1864 being in unmistakable connection with other adjacent objects. Double and triple nebulæ are usually spherical, centrally condensed, and with traces of mutual coherence. Their separation is thus still visibly incomplete; orbital revolutions can scarcely be assumed as probable; nor is there any sign of their being in progress.

1 Harvard Annals, vol. xiv. p. 288. The nebula is No. 242 of the Harvard Photometric Catalogue.

2 Trans. R. Dub. Soc. vol. ii. p. 21.

3 Outlines of Astronomy, p. 647.

A nebula in Ursa Major (N. G. C., 3690) divided by Swift in 1885,1 makes probably the closest pair known; and a curious reproduction, with greatly widened spatial intervals, of star systems like that of y Andromeda occurs in a triple nebula in Virgo consisting of a bright round nebula, attended, at a distance of 5', by an extremely faint one which is itself double 2 (N. G. C., 5813-14). Another compound object of a striking character was noticed by Sir John Herschel3 in Canes Venatici (N. G. C., 4631), where an enormously long ray of nebulosity has a round, dimly luminous companion, a tenthmagnitude star placed between serving perhaps as a centre of attraction for both.

5

A very close double nebula in Gemini (N. G. C., 2371-12) has also an intervening star symmetrically located in the line. joining their centres. Cirrus-like streaks of nebulosity partially encircle the two objects. Duplicity is, in other cases, still less clearly defined. Thus, a pair of nebulæ near y Leonis (N. G. C., 3226-27) are together enclosed in a faint luminous envelope, the effect recalling that of the celebrated 'Dumb-bell' nebula in Vulpecula (M. 27), which is only perceived to be essentially single when the 'neck' uniting two conspicuous hazy masses is brought into view with a powerful telescope. Sir John Herschel first observed the elliptical outline of the entire to be rounded out by faint luminosity, and thus saw it in its true aspect as a large, diversified oval disc, measuring about 5′ by 8'. It might indeed be called a magnified planetary nebula not devoid of annular inclinations. The possibility that, by the progress of the central contraction and marginal spreading indicated by its present hourglass shape, the chief part of its mass may, in the course of ages, become diffused into a ring, is strongly suggested by the analogy of the bright spots at either end of the minor axis of the ring-nebula in Lyra. A planetary in the southern hemisphere (N. G. C., 1365) appears, in fact, to

Sid. Mess. vol. iv. p. 39.

2 D'Arrest, Astr. Nach. No. 1369. 3 Phil. Trans. vol. cxxiii. p. 431. Lassell, Memoirs R. Astr. Soc. vol. xxiii. p. 62; Lord Rosse, Phil. Trans. vol. cxl. p. 512.

' D'Arrest, Abhandlungen, Leipzig, 1857, p. 325.

have already reached a more advanced stage on the same road, and several miniatures' of the Dumb-bell' are included among that class of objects. One especially in Cygnus (N. G. C., 6905), depicted by Vogel with the Vienna 27-inch, easily gives an impression of actual duplicity,' and showed at Parsonstown as a beautiful little spiral.' It has a central star, and four 'satellites.' Its spectrum, like that of the Dumb-bell nebula, is (so far as is yet known) absolutely monochromatic. The leading nebular ray at 5005 concentrates the whole of its light.

In a photograph of the Dumb-bell nebula, taken by Mr. Roberts with an exposure of three hours, October 3, 1888, the shaping-in towards the middle, so marked to the eye, is almost obliterated through the prolonged accumulation of chemical effects; but it intimates pretty clearly the approximate completeness of the oval bright border of the disc, as well as its superposition upon a fainter, more elliptical one, visible as a kind of effusion at the extremities of its longest diameter. Vogel's drawing too 2 suggests, though after a different fashion, the presence of two ellipses, one partially concealed behind the other; and there hence seems reason to think that this singular formation partakes, in more ways than one, of the compound character evident in many planetary nebulæ.

1 Potsdam Publicationen, No. 14, p. 36.

2 Ibid. p. 35.

THE elliptical and irregular classes of nebulæ are illustrated by such splendid examples that we have thought it well to devote a chapter to their separate consideration. One member especially of each towers above the rest, like Ajax among the Argive host, its rival alone excepted, and the two are so different that it is not easy to award the palm of superiority to either. Needless to say that we allude to the objects in Andromeda and Orion, the types respectively of the elliptical and irregular plans of nebular construction.

The former (M 31) is the only real nebula which can readily be detected with the unaided eye, and it is the only one, accordingly, which was discovered in pre-telescopic times. Al Sufi was familiar with the little cloud' near the most northern of the three stars in the girdle of Andromeda ; and its place was marked on a star-map brought from Holland to Paris by De Thou, and believed to date from the tenth century.2 Simon Marius, who was the first to turn a telescope upon it, December 15, 1612, called it 'stellam quandam admirandæ figuræ,' and compared its dull and pallid rays to those of a candle shining by night through a semi-transparent piece of horn. Yet this strange phenomenon was only rescued from neglect by Boulliaud, whose attention was directed to it by the passage of the comet of 1664 across that part of the sky. So surprising did the disregard of it by Hipparchus, Tycho, and Bayer then appear to him that he concluded it to vary in light, an hypothesis which, however, derives no support from recent observations.

With powerful light-concentration this most magnificent object' (to borrow Sir John Herschel's phrase) assumes vast proportions. They were extended by G. P. Bond, using the fifteen-inch refractor of Harvard College, to cover an area of 4° x 210, and he probably did not reach their absolute limits. Two adjacent nebulæ, one (M 32) descried by Le Gentil in 1749, the other (N. G. C., 205) by Caroline Herschel in 1783, undoubtedly fall within their compass.'1 The light of this nebula is of the most perfectly milky, absolutely irresolvable kind.' It does not collect into floccules,' and produces none of the scintillating effect giving to many gaseous nebulæ a delusive appearance of resolvability. From the circumference towards the centre, however, it gradually brightens, then abruptly condenses to a small nucleus of indistinct outline under high magnifying powers, and possibly (like the nuclei of many comets) granulated, but assuredly not stellar.

This progressive brightening inward shows, nevertheless, interruptions. On September 14, 1847, Bond discovered two long dark rifts running nearly parallel to one another, and to the axis of the nebula.3 Their detection was a consequence of the widened area of the luminosity perceived by him, the inner rift having been taken, until then, for its boundary in that direction. The outlines of Bond's drawing are given in the accompanying diagram by Mr. Wesley (fig. 40), in which the rifts are marked A and B. C represents Le Gentil's, D Miss Herschel's attendant nebula, E an exceptionally lucent region crowded (it has since been found) with hosts of minute stars.

These enigmatical appearances at last assumed an intelligible form in a photograph taken by Mr. Roberts, October 1, 1888.5 The view given by this magnificent picture of the Andromeda nebula as a symmetrical, though still inchoate structure, ploughed up by tremendous, yet not undisciplined