Last and most important of all are the spiral nebulæ. The finest example is in the constellation Canes Venatici, and its spiral configuration was first noted by Lord Rosse, an epoch-making discovery. The convolutions of its spiral are filled with numerous starlike condensations, themselves engulfed in nebulosity. Photography possesses a vast advantage over the eye in revealing the marvelous character of this object, an inconceivably vast celestial whirlpool. Naturally the central regions of the whorl would revolve most swiftly, but no comparison of drawings and photographs, separated by intervals of many years, has yet revealed even a trace of any such motion.
The number of large spiral nebulæ is not very great; the largest of all is the great nebula of Andromeda, whose length stretches over an arc of seven times the breadth of the moon, and its width about half as great. This nebula is a naked-eye object near Eta Andromedæ, and it is often mistaken for a comet. Optically it was always a puzzle, but photographs by Roberts of England first revealed the true spiral, with ringlike formations partially distinct, and knots of condensing nebulosity as of companion stars in the making. While its spectrum shows the nongaseous constitution of this nebula, no telescope has yet resolved it into component stars.
Systematic search for spiral nebulæ by Keeler, and later continued by Perrine, at the Lick Observatory, with the 36-inch Crossley reflector, disclosed the existence of vast numbers of these objects, in fact many hundreds of thousands by estimation; so that, next to the stars, the spiral nebulæ are by far the most abundant of all objects in the sky. They present every phase according to the angle of their plane with the line of sight, and the convolutions of the open ones are very perfectly marked. Many are filled with stars in all degrees of condensation, and the appearance is strongly as if stars are here caught in every step of the process of making.
The vast multitude of the spiral nebulæ indicates clearly their importance in the theory of the cosmogony, or science of the development of the material universe. Curtis of the Lick Observatory has lately extended the estimated number of these objects to 700,000. He has also photographed with the Crossley reflector many nebulæ with lanes or dark streaks crossing them longitudinally through or near the center. These remarkable streaks appear as if due to opaque matter between us and the luminous matter of the nebula beyond. Perhaps a dark ring of absorptive or occulting matter encircles the nebula in nearly the same plane with the luminous whorls. Duncan has employed the 60-inch Mount Wilson reflector in photographing bright nebulæ and star clusters in the very interesting regions of Sagittarius. One of these shows unmistakable dark rifts or lanes in all parts of the nebula, resembling the dark regions of the neighboring Milky Way.
Pease of Mount Wilson has recently employed the 60-inch and the 100-inch reflectors of the Mount Wilson Observatory to good advantage in photographing several hundred of the fainter nebulæ. Many of these are spirals, and others present very intricate and irregular forms. A search was made for additional spirals among the smaller nebulæ along the Galaxy, but without success. Several of the supposedly variable nebulæ are found to be unchanging. Many nights in each month when the moon is absent are devoted to a systematic survey of the smaller nebulæ and their spectra by photography. The visible spiral figure of all these objects is a double-branched curve, its two arms joining on the nucleus in opposing points, and coiling round in the same geometrical direction. The spiral nebulæ, as to their distribution, are remote from the Galaxy, and the north Galactic polar region contains a greater aggregation than the south. The distances of the spiral nebulæ are exceedingly great. They lie far beyond the planetary and irregular gaseous nebulæ, like that of Orion, which are closely related to the stars forming part of our own system. Possibly the spiral nebulæ are exterior or separate "island universes." If so, they must be inconceivably vast in size, and would develop, not into solar systems, but into stellar clusters. The enormous radial velocities of the spiral nebulæ, averaging 300 to 400 kilometers per second, or twenty-fold that of the stars, tend to sustain the view that they may be "island universes," each comparable in extent with the universe of stars to which our sun belongs.
Recent spectroscopic observations of the nebulæ applying the principle of Doppler have revealed high velocities of rotation. Slipher of the Lowell Observatory made the first discovery of this sort and Van Maanen of Mount Wilson has detected in the great Ursa Major spiral, No. 101 in Messier's catalogue,  a speed of rotation at five minutes of arc from the center that would correspond to a complete period in 85,000 years. As was to be expected, the nebula does not rotate as a rigid body, but the nearer the center the greater the angular velocity, and Van Maanen finds evidence of motion along the arms and away from the center.
These great velocities appear to belong to the spiral nebulæ as a class, and not to other nebulæ. Thirteen nebulæ investigated by Keeler are as a whole almost at rest relatively to our system, as are the large irregular objects in Orion, and the Trifid nebula. This would seem to indicate that the spiral nebulæ form systems outside our own and independent of it.
Quite different from the spirals in their distribution through space are the planetary nebulæ. The spirals follow the early general law of nebulæ arrangement, that is, they are concentrated toward the poles of the Galaxy; but the planetary nebulæ, on the other hand, are very few near the poles and show a marked frequency toward the Galactic plane. Campbell and Moore have found spectroscopic evidence of internal rotatory motion in a large proportion of the planetary nebulæ.
The distribution of the nebulæ throughout space, like that of the stars, is still under critical investigation, but the location of vast numbers of the more compact nebulæ on the celestial sphere is very extraordinary. The Milky Way appears to be the determining plane in both cases; the nearer we approach it the more numerous the stars become, whereas this is the general region of fewest nebulæ and they increase in number outward in both directions from the Galaxy, and toward both poles of the Galactic circle. Obviously this relation, or contra-relation of stars and nebulæ on such a vast scale is not accidental, and it also must be duly accounted for in the true theory of the cosmogony. The nebulæ which are found principally in and near the Milky Way are the large irregular nebulæ, and vast nebulous backgrounds, like those photographed by Barnard in Scorpio, Taurus and elsewhere, as well as the Keyhole, Omega, and Trifid nebulæ. Allied to these backgrounds are doubtless some of the dark Galactic spaces, radiating little or no intrinsic light, and absorbing the light of the fainter stars beyond them. A peculiar veiled or tinted appearance has been remarked in some cases visually, and examination of the photographs strongly confirms the existence of absorbing nebulosity.
The spiral nebulæ are so abundant, and so much attention is now being given to them, both by observers and mathematicians, that their precise relation to the stellar systems must soon be known; that is, whether they are comparatively small objects belonging to the stellar system, or independent systems on the borders of the stellar system, or as seems more likely, vast and exceedingly remote galaxies comparable with that of the Milky Way itself. Our knowledge of the motions of the spirals, both radial and angular, is increasing rapidly, and must soon permit accurate general conclusions to be drawn.