{eks-truh-guh-lak'-tik} Extragalactic systems are the enormous assemblages of stars found outside the Milky Way Galaxy. They may be normal galaxies, radio galaxies, or \Lquasar\ls, which are believed to be extremely active nuclei of very distant galaxies. Normal galaxies have diameters ranging from 2,000 to nearly 800,000 \Llight-year\ls, masses ranging from 1 million to 10 trillion solar masses, and luminosities ranging from 1 million to 100 billion suns. Our own Milky Way system is a very large galaxy with a visible diameter of about 100,000 light-years (see \TGalaxy, The\t). Quasars may be more than 100 times more luminous than the brightest known galaxies, and yet are much smaller. They range in distance from 700 million light-years to perhaps 15 billion light-years, if their \Lred shift\ls are proportional to their distances, as is the case for other galaxies. DISCOVERY OF OTHER GALAXIES Early telescopic observers noted in increasing numbers many faint patches of light that they called \Lnebula\le (clouds). These objects puzzled 17th-, 18th-, and 19th-century astronomers, and their true nature was a subject of great controversy until 1924. As early as 1755, Immanuel Kant speculated that the nebulae that were elliptical in form might be systems of stars analogous to the Milky Way. In 1781, the avid French comet hunter Charles \TMessier\t prepared a catalog of 103 nebulous objects but used the catalog in order that he might not mistake these objects for comets. During the same decade, using a reflecting telescope that he built himself, William Herschel added 2,000 nebulae to Messier's list and resolved some into stars. In the 19th century, spectroscopy revealed what Herschel had suspected; some nebulae were true gas clouds. Only in the 20th century were nebulae recognized as at least four different kinds of objects: distant star clusters; \Lplanetary nebula\le; luminous, irregular-shaped gas clouds; and galaxies. Observational Evidence In identifying and determining the distances of the nebulae, observations centered on those that were shaped like a spiral. The observational evidence, especially that obtained at the Mount Wilson Observatory with the 60-in (152-cm) reflector in 1908 and with the 100-in (254-cm) reflector in 1918, was confusing and discordant. A comparison of 60-in photographic plates, taken a number of years apart by Adriaan van MAANEN, strongly indicated that a few of the brighter spirals were rotating in the sense that their spiral arms were leading, or unwinding. These angular motions, if real, were so large that if the spirals had been at distances of 1 million light-years, the true motions would have been impossibly great, larger than the velocity of light. It seemed, therefore, that they were considerably closer. Moreover, the Mount Wilson astronomers were not even convinced that the stars they were photographing in the spiral arms were truly stars. They looked slightly fuzzy and were described as "nebulous stars." Another aspect of the puzzle was the appearance in 1885 of a 7th magnitude nova, a few seconds of arc from the center of the Andromeda nebula (M 31). If this were an ordinary nova, as its behavior and light curve strongly suggested, the distance of M 31 was calculated to be less than 30,000 light-years, within the boundaries of our own galaxy. In 1917, Harlow \TShapley\t of the Mount Wilson Observatory had determined the distances of 93 globular clusters and from these distances had derived the size of this system of globulars, which he equated to the size of the Milky Way, as 300,000 light-years. He also determined that the Sun's position was far from the galactic center. He believed especially the van Maanen evidence and thought that the spiral nebulae were relatively close, about 20,000 light-years away, and were part of the Milky Way system, which he envisioned as the entire universe. In 1899, however, James E. \LKeel\ler, director of the Lick Observatory, had shown, using a 36-in (91-cm) reflecting telescope, that a spiral nebula is the most common type of nebula in the universe and could be identified in great numbers, getting both smaller and fainter as one went down to the limits of the photographic plates. From this it could be argued that even if the nearer, larger spirals were only 20,000 light-years distant, then the fainter, smaller spirals would be up to 1,000 times farther away and hence far outside the Milky Way boundaries. Spectroscopic (radial velocity) evidence obtained by V. M. \TSlipher\t at the Lowell Observatory strongly indicated that the spirals rotated with their arms trailing, or winding tighter, contrary to van Maanen's evidence. A great debate between Shapley and H. D. \TCurtis\t of the Lick Observatory took place before the U.S. National Academy of Sciences on Apr. 26, 1920. Shapley argued that the Milky Way was 300,000 light-years in diameter, that the Sun was far from the center in the direction of a point in the constellation of Sagittarius on the galactic equator, and that the spiral nebulae were part of the Milky Way system. Curtis advanced arguments that the Milky Way was about 30,000 light-years in size with the Sun at, or close to, the center and that the spirals were island universes far away from our galaxy. Each protagonist was only partly right. What was needed was a reasonably reliable distance to one or more of the nearer galaxies. Role of the Cepheids Edwin P. Hubble of Mount Wilson Observatory filled the need for reliable distances by using \TCepheid\t variables--unstable, pulsating giant and supergiant stars whose light varies in a characteristic way by a factor of 2 in periods ranging from about a day up to about 50 days. If these objects could be found in a spiral nebula, their periods could be observed, and the luminosity of the Cepheid calculated from the \Tperiod-luminosity relation\t discovered by Henrietta \TLeavitt\t at the Harvard Observatory in 1912. When the luminosity was determined, the distance could be determined by comparing the luminosity, or absolute magnitude, with the apparent magnitude. Shapley had used short-period Cepheids, known as RR Lyrae stars, in which the light varies over periods from 1.5 hours to 1 day, to determine the distances to globular clusters. Hubble had to use the longest of the long-period Cepheids, because these stars are intrinsically about 1,000 times more luminous than the RR Lyraes and therefore could be photographed to distances 30 times greater. Hubble had first taken photographs with the 100-inch telescope under the best seeing conditions, centered on the spiral arms, rather than on the nucleus of a galaxy. Under these conditions, the so-called nebulous stars had images exactly similar to foreground stars, and it could safely be concluded that they were indeed individual stars. This conclusion was strengthened in 1922 when John C. Duncan discovered at Mount Wilson that three faint stars in M 33 were variable. Later, Hubble identified one of these three as a Cepheid variable and discovered many more such Cepheids in this and the other two galaxies. He showed that the fainter Cepheids in these galaxies had shorter periods of light variation as compared to the brighter Cepheids, in conformance with the relation between period and luminosity. The relation could therefore be applied to these objects to determine galactic distances. Hubble's pattern of research on galaxies involved four steps: discovery, measurement, identification, and calibration. After the discovery of a variable star in another possible galaxy, Hubble measured the apparent photographic magnitudes of the star on each of a series of dozens of plates. From these measurements the period of variation was deduced and the object was identified as a Cepheid variable. Calibration was achieved through the known relation between period and luminosity of Cepheids in the Magellanic Clouds. The zero point of this relationship is found from the luminosities (absolute magnitudes) of a few individual Cepheids that are members of relatively nearby open clusters, whose distances can be determined, accurately and independently, by other methods. In December 1924, Hubble presented a paper before the American Astronomical Society showing that the spirals M 31 and M 33 and an irregular galaxy NGC 6822 were at distances of about three-quarters of a million light-years. Although Hubble's distances were too small by a factor of 3, they proved the existence of extragalactic systems. For more distant galaxies, the Cepheids had faded from view, and Hubble had to use a mean magnitude of the brightest stars in a galaxy as a distance indicator. These supergiants are about 10 times more luminous than the brightest Cepheids. The mean absolute magnitudes of the brightest stars in a galaxy were determined from those nearer galaxies that contained observable Cepheids; Hubble found that this mean was constant from galaxy to galaxy. By this method of photometric continuation, Hubble extended his galactic distance scale out to the great Virgo cluster of galaxies at a distance of about 50 million light-years. This gave Hubble his third calibration: the absolute magnitude of the brightest galaxies in a cluster of galaxies. He then identified clusters of galaxies down to the plate limit of the 100-in (254-cm) reflector; they appeared as isolated groups of a few dozen wispy, sometimes oval-shaped images, which could be distinguished from stellar images. In about half a dozen years Hubble had extended his breakthrough into what he called "the realm of the nebulae" by a factor of nearly 1,000 to what then seemed to be the edge of the observable universe. MODERN MEASUREMENT OF EXTRAGALACTIC DISTANCES Distance determination is of utmost importance to extragalactic astronomy, and several reliable methods are now used (see \Tdistance, astronomical\t). Magnitudes and Sizes of Objects Today Cepheids can be photographed and identified to distances of about 23 million light-years. Not only Cepheids but also the magnitudes of globular clusters and novae and the sizes of ionized hydrogen (H II) regions in distant galaxies are used. The magnitudes of these very faint objects are determined by accurate photoelectric techniques rather than by photography. The greatest uncertainty, in Hubble's time and perhaps now, is in the distances to the nearest clusters of galaxies. Part of the problem lies in the different assumptions that can be made about the possible densities of intervening, unseen dust clouds. The Red Shift Effect A new method for the determination of galactic distances came in a most unexpected and exciting way, again associated with Hubble. Early radial velocity observations with small telescopes of a few galaxies by E. A. Fath at the Lick Observatory and especially by V. M. Slipher at the Lowell Observatory showed that some galaxies were receding from the Sun with enormous velocities up to 1,800 km/sec (1,120 mi/sec). The lines in the galaxian spectra, especially the strong twin absorption lines of ionized calcium, were all shifted to the red. These red shifts were thought to be proportional to the distance, but the evidence was not conclusive because the effect was small and the distances were too uncertain. By 1929, Hubble had enough well-determined distances to announce that the red shifts were closely proportional to the distances. This correlation had the most profound implications. It demonstrated that the universe was expanding, that it possibly originated in a \Tbig bang\t some billions of years ago, and that the distance of any isolated, extremely distant object could be determined from its red shift. This close correlation between the red shift, designated by the symbol z, and distance also lent confidence to the self-consistency of the Hubble distance scale. Whatever the basic reason might be for this close correlation with z, very poorly determined distances would give little or no correlation with the independently observed red shifts. Following Hubble's 1929 announcement of the red shift effect, a colleague, M. L. \THumason\t, began a large observational effort with the Mount Wilson 100-in reflector and by 1936 had observed a red shift of a galaxy in Ursa Major of 40,360 km/sec (25,000 mi/sec), corresponding to 13.5% of the velocity of light. All lines in the spectrum were shifted 13.5% of their normal wavelength to the red. In mathematical shorthand, z = 0.135, corresponding to the shift in wavelength divided by the normal (rest) wavelength. Humason's work extended Hubble's 1929 correlation by a factor of 37. In 1960, Rudolph Minkowski, using the 200-in Palomar reflector on his last night of observing before retirement, observed a z of 0.461 for a galaxy in a distant cluster. By the late 1980s a z of 3.8 had been determined for a very faint radio galaxy. This may be compared with the z's of greater than 4 that have been established for a number of \Lquasar\ls, making them the most distant objects known should their red shifts indeed be attributable to the Hubble concept and not to other, unknown mechanisms. Hubble's Law Hubble's law states that the radial velocity V of a cluster or an extragalactic object is equal to the product of the Hubble constant (a constant of proportionality) and the distance d of the object from the Earth, Hubble's original value of the constant was 500 km/sec (310 mi/sec) megaparsec, where 1 megaparsec is equal to 1 million parsecs, 3.26 million light-years, or 19 X 1,000,000,000,000,000,000 miles. The modern value of the Hubble constant is much smaller than 500 and still uncertain but almost certainly between 35 and 100. This uncertainty is mostly because of the question of the luminosity of the brightest galaxy in a cluster of galaxies. If z = 1.378 and the Hubble constant is taken to be 50, the corresponding distance is nearly 24 billion light-years. Strong reasons exist for believing this to be an overestimate, but we are looking backward in time to a galaxy that was formed in the early history of the expanding universe. The currently accepted age of the universe is between 10 and 20 billion years, zero time being that of the beginning of the enormous explosion postulated by the \Tbig bang theory\t. This theory, however, is also being subjected to proposed modifications (see \Tcosmology\t), and even basic challenges, and the age of the universe remains an open question. CHARACTERISTICS OF EXTRAGALACTIC SYSTEMS Whereas one star image on a photographic plate looks like any other star image, except for differences in brightness, galaxies have complex shapes and can be classified usefully into a logical sequence on this basis. Classification The three broad categories of galactic shapes are elliptical (E), spiral (S), and irregular (Irr). Elliptical galaxies are further classified as E0 through E7, where E0 is circular and E7 highly elliptical. The shape depends, however, on the galaxy's orientation to the observer. Spirals are subdivided into two types: normal (S) and barred (SB). An intermediate type between ellipticals and spirals is called SO. (Such SO galaxies are not yet well understood, but some show characteristics of both spirals and elliptical galaxies.) Two sequences follow: (1) Sa, Sb, and Sc and (2) SBa, SBb, and SBc. The subdivisions a, b, and c refer to the decreasing prominence of the nucleus and of the tightness of the spiral arms. In barrel galaxies, the spiral arms are wound. About one-quarter of all spirals are barred spirals, in which the spiral arms terminate at the ends of a central bar, which rotates as a unit. More than two-thirds of the brightest, most conspicuous galaxies are spirals, about 3% are irregulars, and the rest are ellipticals. In the universe as a whole, elliptical galaxies are far more numerous. Because of their low luminosity, dwarf elliptical galaxies are difficult to discover but are probably the most numerous type of galaxy in a given volume of space. Elliptical galaxies are both larger and more luminous and also smaller and less luminous than spirals. Clusters of Galaxies Galaxies come in pairs, triples, groups, and clusters. Clusters of galactic clusters also exist. The \TLocal Group of galaxies\t contains 22 known members and is spread out over a region about 3 million light-years in diameter. Among the group are 3 spirals, 13 ellipticals (including 9 dwarf ellipticals), and 6 irregular galaxies. The largest galaxy in the group is the Andromeda galaxy, an Sb about 30% larger than the Milky Way, which is either an Sb or SBb spiral. The Local Group is near the edge of a supercluster of galaxies, revolving about its center at about 400 km/sec (240 mi-sec). The supercluster, which has a diameter of about 130 million light-years, also contains the great Virgo cluster of about 2,500 galaxies, and evidence exists that the supercluster itself may in turn be part of a much larger supercluster of tens of thousands of galaxies called the "Great Attractor." George \TAbell\t cataloged 2,713 rich clusters of galaxies on the Palomar Sky Survey plates. These clusters group into "clumps" or second-order clusters with diameters typically 300 million light-years in diameter. Rich galaxian clusters are composed almost entirely of elliptical galaxies. The brightest galaxies in a cluster of galaxies are supergiant elliptical galaxies, perhaps ten times as luminous as the Andromeda galaxy; from the red shift evidence, they are nearly identical from cluster to cluster in both their size and luminosity (see \TComa cluster\t). Normal Galaxies A typical galaxy has a central nucleus of closely packed stars. The star density in the nucleus may exceed 1 million times that of the rest of the galaxy. It contains chiefly old, hot, low-mass stars known as population II stars. Outside of the nucleus, absent in elliptical galaxies but increasing in prominence in the spirals as they become more open, are the young, massive, and hotter stars of population I, usually found in connection with gas and dust. Population I stars are also found in the halo of a galaxy--the large, circular region surrounding--where the globular clusters are located. (See \Tpopulation, stellar\t.) Cosmological theories attempt to account for the observed structures of the galaxies. According to the standard big bang theory, matter came to predominate over radiation about one million years after the big bang and eventually fragmented into clouds of hot gas, the protogalaxies. The clouds contracted gravitationally and began to take a disk form if their spin was fast enough. Unincorporated gas formed a galactic halo of stars. Later generations of metal-enriched stars were formed in the disk as \Tinterstellar matter\t was recycled through the stars. The young blue supergiant stars in the disk are short-lived (see \Tsupernova\t) and are continuously being formed. The observation in 1986 of a galaxy apparently being formed about 12 billion light-years away suggests, however, that the time limit placed by the standard theory on galaxy formation may have to be extended. Standard theory also suggests that elliptical galaxies took their different form because they had little initial rotation. Stars were formed in an initial burst, and the shape of such galaxies changed very little over time. Some astronomers propose instead that elliptical galaxies may be the product of collisions between spiral galaxies, accounting for the greater average massiveness of ellipticals and their lack of interstellar matter, most of which was converted into stars through the collision process. A normal galaxy may shine for many billions of years until all its available gas and dust are contained in heavier elements, \Lwhite dwarf\ls, \Lneutron star\ls, and possibly \Lblack hole\ls. According to standard theory, no new material will be left that is capable of forming new stars, so the galaxies may eventually cool and fade away. Active Galaxies, Radio Galaxies, and Quasars Galaxies that exhibit abnormalities are generally classified as active galaxies. The abnormalities may range from extreme luminosity to variation in optical and radio brightness and colliding and exploding galaxies. Radio galaxies emit up to 10 million times more energy at radio wavelengths than normal galaxies. (See \Tradio astronomy\t.) \TSeyfert galaxies\t are galaxies that have extremely active, energetic nuclei, as do the so-called starburst galaxies, whose nuclei radiate strongly in the infrared and are thought to be sites of large-scale star formation. Quasars and \TBL Lacertae objects\t have very active nuclei and may be closely related to radio galaxies. Astronomers have suggested that all galaxies, from normal ones to quasars, may form an evolutionary sequence in which massive black holes, at the centers of the galaxies, play a critical role. Mass of Galaxies Gravitational forces between galaxies are enormous, much larger than the gravitational force that keeps the Earth in orbit. Solar gravitation, acting at a distance of 150 million km (93 million mi), causes the Earth to move at 30 km/sec (19 mi/sec) in its orbit. The mass of the local supercluster, acting at a distance 4 trillion times larger than the Earth-Sun distance, causes the Sun (and Earth) to move in space 13 times faster than the Earth moves around the Sun. Masses of galaxies can be determined from the velocities they induce on objects both outside and within themselves. The Sun moves in a nearly circular orbit about the galactic center 32,600 light-years distant with a velocity of 250 km/sec (155 mi/sec). A straightforward calculation indicates that our Galaxy contains 150 billion solar masses. When detailed corrections are made-- allowing for the nonspherical distribution of mass in the Galaxy--the galactic mass is estimated to be about 200 billion solar masses. Calculating the masses of other galaxies becomes more difficult because of their distance, but the effects of pairs of galaxies on one another can be used, as can the observation of bright objects within the nearer galaxies. The dispersion in galactic velocities within a cluster of galaxies can also be employed in estimating the total mass of the cluster. One major issue of modern cosmology is that of the total mass of the galaxies and other matter in the universe, because whether the universe will continue to expand or will collapse upon itself depends upon its average density. Current astronomical models call for a great deal more mass than has actually been observed. Some astronomers propose that this "missing mass" might be accounted for by the existence of many dim masses as yet unobserved. A few such dim galaxies have already been detected. John B. Irwin Bibliography: Hodge, P. W., Galaxies (1986); Kaufman, W. J., III, Universe (1985); Mardirossian, F., et al., eds., Clusters and Groups of Galaxies (1984); Miller, J. P., ed., Astrophysics of Active Galaxies and Quasi-Stellar Objects (1987); Prestwich, Andrea, "Starburst Galaxies," New Scientist, May 14, 1987; Vorontsov-Velyaminov, Boris, Extragalactic Astronomy, rev. ed. (1987); Wray, J. D., The Color Atlas of Galaxies (1988); Wright, Alan and Hilary, At the Edge of the Universe (1989).