0705.1537/ms.tex
1: % cup2egui.tex (LaTeX 2e version)
2: % v1.01 --- released 9th July 1997
3: % v1.0  --- released 9th May 1997
4: %           based on cupguide.tex v1.2 (for LaTeX2.09, 27.4.95)
5: 
6: \NeedsTeXFormat{LaTeX2e}[1996/06/01]
7: 
8: \documentclass[cup6a]{cupbook}
9: \usepackage[dvips]{graphicx}
10: 
11: %\title[ \LaTeXe\ Style Guide for Authors]
12: %      {CUP Standard Designs}
13: %\author{Cambridge \TeX-to-type}
14: %\date{9th July 1997}
15: 
16: \begin{document}
17: 
18: \pagenumbering{roman}
19: \maketitle
20: \tableofcontents
21: \cleardoublepage
22: \pagenumbering{arabic}
23: 
24: \chapter{Supermassive Black Holes}
25: 
26: Supermassive black holes have generally been recognized as the most destructive 
27: force in nature. But in recent years, they have undergone a dramatic shift in 
28: paradigm. These objects may have been critical to the formation of structure in 
29: the early universe, spawning bursts of star formation and nucleating proto-galactic 
30: condensations. Possibly half of all the radiation produced after the Big Bang
31: may be attributed to them, whose number is now known to exceed 300 million. 
32: The most accessible among them is situated at the Center of Our Galaxy. In
33: the following pages, we will examine the evidence that has brought us to this 
34: point, and we will understand why many expect to actually image the event horizon
35: of the Galaxy's central black hole within this decade.
36: 
37: The supermassive black hole story begins in 1963, at the Mount Palomar observatory,
38: where Schmidt (1963) was pondering over the nature of a star-like object with truly 
39: anomalous characteristics.  Meanwhile, at the university of Texas, Kerr (1963) was 
40: making a breakthrough discovery of a solution to Einstein's field equations of
41: general relativity. Kerr's work would eventually produce a description of space 
42: and time surrounding a spinning black hole, now thought to power the compact 
43: condensations of matter responsible for producing the mystery on Schmidt's desk 
44: that year.
45: 
46: The development and use of radio telescopes in the 1940s had led to the gradual
47: realization that several regions of the sky are very bright emitters of 
48: centimeter-wavelength radiation. In the early 1960's, the British astronomer 
49: Cyril Hazard's idea of using lunar occultation to determine with which, if any, 
50: of the known visible astronomical objects the emitter of centimeter-wavelength 
51: radiation was associated, lead to the successful identification of 3C~273 as a 
52: star-like object in Virgo. Its redshifted lines, however, indicated that this was 
53: not a star at all, but rather an object lying at great cosmological distances.  
54: A recent image of this historic source was made with the {\sl Chandra} X-ray 
55: telescope, and is shown in figure~1.
56: 
57: \begin{figure}[h]
58: \centerline{\includegraphics[width=4in]{f1bw.eps}}
59: \vskip0.15in
60: \caption{The quasar 3C~273 (the bright object in the upper-left-hand corner) 
61: was one of the first objects to be recognized as a ``quasi-stellar-radio-source'' 
62: (quasar), due to its incredible optical and radio brightness. Insightful analysis
63: led to the realization that 3C~273 is actually an incredibly powerful, distant 
64: object.  This {\sl Chandra} image has a size $\approx 22\times 22$ square
65: arcseconds, which at the distance to 3C~273, corresponds to about $2,000\times 2,000$
66: square light-years.  (Photograph courtesy of H. L. Marshall et al., NASA, and MIT)}
67: \end{figure}
68: 
69: 3C~273's total optical output varies significantly in only 10 months or so, implying 
70: that its size cannot exceed a few light-years---basically the distance between the 
71: Sun and its nearest stellar neighbor. So it was clear right from the beginning that 
72: the quasar phenomenon must be associated with highly compact objects. Even more
73: impressively, their X-ray output has now been seen to vary in a matter of only hours,
74: corresponding to a source size smaller than Neptune's orbit.  Each quasar typically 
75: releases far more energy than an entire galaxy, yet the central engine that drives 
76: this powerful activity occupies a region smaller than our solar system.
77: 
78: The idea that such small volumes could be producing the power of a hundred billion
79: Suns led to their early identification as radiative manifestations of supermassive
80: black holes (see, e.g., Salpeter 1964, Zel'dovich and Novikov 1967, and
81: Lynden-Bell 1969).  But are they ``naked''---deep, dark pits of matter floating 
82: aimlessly across the primeval cosmic soup---or are they attached to more 
83: recognizable structures in the early universe?
84: 
85: In recent years, the task of source identification has been made easier using the Hubble
86: Space Telescope (see figure~2).  The most widely accepted view now is that quasars 
87: are found in galaxies with active, supermassive black holes at their centers. 
88: Because of their enormous distance from Earth, the ``host"  galaxies appear very 
89: small and faint, and are very hard to see against the much brighter quasar light 
90: at their center.  
91: 
92: \vskip 0.1in
93: \begin{figure}[h]
94: \centerline{\includegraphics[width=4in]{f2bw.eps}}
95: \vskip -0.0in
96: \caption{This Hubble Space Telescope (HST) image reveals the faint host galaxy within
97: which dwells the bright quasar known as QSO 1229+204. The quasar is seen to lie in the 
98: core of an ordinary-looking galaxy with two spiral arms of stars connected by a bar-like
99: feature.  (Photograph courtesy of John Hutchings, Dominion Astrophysical Observatory, 
100: and NASA)}
101: \end{figure}
102: 
103: Quasars actually reside in the nuclei of many different types of galaxy, from 
104: the normal to those highly disturbed by collisions or mergers.  A supermassive 
105: black hole at the nucleus of one of these distant galaxies ``turns on" when it 
106: begins to accrete stars and gas from its nearby environment; the rate at which 
107: matter is converted into energy can be as high as 10 solar masses per year.  So 
108: the character and power of a quasar depend in part on how much matter is available
109: for consumption.  Disturbances induced by gravitational interactions with
110: neighboring galaxies can trigger the infall of material toward the center of the
111: quasar host galaxy (see figure~3).  However, many quasars reside in apparently 
112: undisturbed galaxies, and this may be an indication that mechanisms other than 
113: a disruptive collision may also be able to effectively fuel the supermassive 
114: black hole residing at the core.
115: 
116: \begin{figure}[h]
117: \centerline{\includegraphics[width=4in]{f3bw.eps}}
118: \vskip 0.15in
119: \caption{The collision between two galaxies begins with the unraveling of
120: the spiral disks.  This HST image shows the interacting pair of galaxies NGC~2207 (the
121: larger, more massive object on the left) and IC~2163 (the smaller one on the right),
122: located some 114 million light-years from Earth.  By this time, IC~2163's stars have
123: begun to surf outward to the right on a tidal tail created by NGC~2207's strong
124: gravity. (Image courtesy of Debra M. Elmegreen at Vassar College, Bruce G. Elmegreen
125: at the IBM Research Division, NASA, M. Kaufman at Ohio State U., E. Brinks at U. de
126: Guanajuato, C. Struck at Iowa State U., M.  Klaric at Bell South, M. Thomasson at
127: Onsala Space Observatory, and the Hubble Heritage Team based at the Space Telescope
128: Space Institute and AURA)}
129: \end{figure}
130: 
131: Some supermassive black holes may not be visible as quasars at all, but rather just
132: sputter enough to become the fainter galactic nuclei in our galactic neighborhood.
133: Our Milky Way galaxy and our neighbor, the Andromeda galaxy, harbor supermassive 
134: black holes with very little nearby plasma to absorb.  The question concerning how 
135: the undisturbed galaxies spawn a quasar is still not fully answered.  Perhaps the 
136: Next Generation Space Telescope, now under development and expected to fly soon
137: after 2010, will be able to probe even deeper than the Hubble Space Telescope has 
138: done, and expose the additional clues we need to resolve this puzzle.
139: 
140: \section{The Quasar/Supermassive Black Hole Census}
141: By now, some 15,000 distant quasars have been found, though the actual number of 
142: supermassive black holes discovered thus far is much greater.  Because of their 
143: intrinsic brightness, the most distant quasars are seen at a time when the universe 
144: was a mere fraction of its present age, roughly one billion years after the Big Bang.  
145: The current distance record is held by an object found with the Sloan Digital Sky 
146: Survey (SDSS), with a redshift of $\sim 6.3$, corresponding to a time roughly 700 
147: million years after the Big Bang.
148: 
149: The SDSS has shown that the number of quasars rose dramatically from a billion years 
150: after the Big Bang to a peak around 2.5 billion years later, falling off sharply at 
151: later times toward the present.  Quasars turn on when fresh matter is brought into 
152: their vicinity, and then fade into a barely perceptible glimmer not long thereafter. 
153: 
154: However, not all the supermassive black holes in our midst have necessarily
155: grown through the quasar phase. Quasars typically have masses $\sim 10^9\;
156: M_\odot$. Yet the black hole at the center of our galaxy is barely $3.4\times
157: 10^6\;M_\odot$. In other words, not all the supermassive black holes in our 
158: vicinity are dormant quasars. 
159: 
160: A recent discovery suggests how some of these ``smaller'' black holes may
161: have gotten their start. {\sl Chandra} has identified what appears to be a 
162: mid-sized black hole 600 light-years from the center of M82 (Matsumoto et al. 
163: 2001).  With a mass of $\sim 500\;M_\odot$, this object could conceivably 
164: sink to the center of M82, and then grow to become a supermassive black hole 
165: in its own right, without having passed through the rapid accretion phase of
166: a quasar.
167: 
168: So the class of known quasars may be a good tracer of supermassive black holes,
169: but it clearly does not encompass all of them. Taking advantage of two patches
170: of sky relatively devoid of nearby objects, {\sl Chandra} produced two of the 
171: deepest images ever made of the distant cosmos at X-ray energies, one in the 
172: southern hemisphere and the other in the north---the latter, called the 
173: {\sl Chandra} Deep Field North, is shown in figure~4. Based on the number of
174: suspected supermassive black holes in these images, one infers an overall
175: population of $\sim 300$ million throughout the cosmos.
176: 
177: \begin{figure}[h]
178: \centerline{\includegraphics[width=4in]{f4bw.eps}}
179: \vskip -0.0in
180: \caption{The {\sl Chandra} Deep Field North X-ray image. The vast majority of 
181: the 500 or so sources in this view (spanning a region approximately 28 
182: arcminutes wide) are supermassive black holes.  Extrapolating this number to 
183: the whole sky, one infers $\sim 300$ million such objects spread across the 
184: universe. (Image courtesy of D. M. Alexander, F. E. Bauer, W. N. Brandt, 
185: et al., NASA and PSU)}
186: \vskip 0.0in
187: \end{figure}
188: 
189: And yet, these X-ray detections speak only of those particular supermassive 
190: black holes whose orientation facilitates the transmission of their high-energy 
191: radiation toward Earth. Their actual number must be higher than this;  indeed, 
192: there is now growing evidence that many---perhaps the majority---of the 
193: supermassive black holes in the universe are obscured from view.  The faint 
194: X-ray background pervading the intergalactic medium has been a puzzle for many years.
195: Unlike the cosmic microwave background radiation left over from the Big Bang, the 
196: photons in the X-ray haze are too energetic to have been produced at early times.  
197: Instead, this radiation field suggests a more recent provenance associated with 
198: a population of sources whose overall radiative output may actually dominate over 
199: everything else in the cosmos.  Stars and ordinary galaxies simply do not radiate 
200: profusely at such high energy, and therefore cannot fit the suggested profile.
201: 
202: A simple census shows that in order to produce such an X-ray glow with quasars alone, 
203: for every known source there ought to be ten more obscured ones.  This would also mean 
204: that the growth of most supermassive black holes by accretion is hidden from the view 
205: of optical, UV, and near infrared telescopes.  Fabian et al. (2000) have reported the 
206: discovery of an object they call a Type-2 quasar.  Invisible to optical light telescopes, 
207: the nucleus of this otherwise normal looking galaxy betrayed its supermassive guest with 
208: a glimmer of X-rays.  The implication is that many more quasars, and their supermassive 
209: black-hole power sources, may be hidden in otherwise innocuous-looking galaxies.
210: 
211: And so, the all-pervasive X-ray haze, in combination with the discovery of gas-obscured 
212: quasars, now point to supermassive black holes as the agents behind perhaps {\it half} 
213: of all the universe's radiation produced after the Big Bang. Ordinary stars no longer 
214: monopolize the power as they had for decades prior to the advent of space-based astronomy.
215: 
216: \section{Black Holes in the Nuclei of Normal Galaxies}
217: Much closer to Earth, within hundreds of thousands of light-years as opposed to the 
218: 11 billion-light-year distance to the farthest quasars, supermassive black holes 
219: accrete at a lower rate than their quasar brethren and are therefore much fainter. 
220: An archetype of this group, Centaurus A, graces the southern constellation of Centaurus
221: at a distance of 11 million light-years (see figure~5). At the center of the dark bands of dust,
222: HST recently uncovered a disk of glowing, high-speed gas, swirling about a concentration of matter
223: with the mass of $\sim 2\times 10^6\;M_\odot$.  This enormous mass within the central cavity 
224: cannot be due to normal stars, since these objects would shine brightly, producing an intense 
225: optical spike toward the middle, unlike the rather tempered look of the infrared image shown here.
226: 
227: \begin{figure}[h]
228: \centerline{\includegraphics[width=4in]{f5bw.eps}}
229: \vskip -0.0in
230: \caption{This image of Centaurus A is a composite of three photographs taken by the 
231: European Southern Observatory. The dramatic dark band is thought to be the remnant 
232: of a smaller spiral galaxy that collided, and ultimately merged, with a large 
233: elliptical galaxy.  Imaging this galaxy at radio wavelengths, we would see two 
234: jets of plasma spewing forth from the central region in a direction perpendicular 
235: to the dark dust lanes (see figure~6 for the corresponding configuration in
236: Cygnus A).  These relativistic expulsions of plasma share much in common with the
237: X-ray glowing stream shown in figure~1. (Photograph courtesy of Richard M. West and the
238: European Southern Observatory)}
239: \end{figure}
240: 
241: Centaurus A is apparently funneling highly energetic particles into beams perpendicular
242: to the dark strands of dust. It may therefore have much in common with the X-ray
243: jet-producing black hole in 3C 273 (see figure~1), and another well-known active
244: galactic nucleus, Gygnus A, shown in figure~6.  The luminous extensions in this figure
245: project out from the nucleus of Cygnus A, an incredible distance three times the size of
246: the Milky Way.  Yet located 600 million light-years from Earth, they cast an aspect
247: only one-tenth the diameter of the full moon. Radio and X-ray observations show that 
248: objects such as this accelerate plasma to relativistic speeds on scales of 10 to 100 
249: Schwarzschild radii (see figure~7), and that these jets are not rare.
250: 
251: \begin{figure}[h]
252: \centerline{\includegraphics[width=\textwidth]{f6bw.eps}}
253: \vskip -0.0in
254: \caption{A VLA image of the powerful central engine and its relativistic ejection of plasma 
255: in the nucleus of Cygnus A.  Taken at $6$ cm, this view reveals the highly ordered structure 
256: spanning over 500,000 light-years, fed by ultra-thin jets of energetic particles beamed from 
257: the compact radio core between them. The giant lobes are themselves formed when these jets 
258: plow into the tenuous gas that exists between galaxies.  Despite its great distance from us 
259: (over 600 million light-years), it is still by far the closest powerful radio galaxy and one 
260: of the brightest radio sources in the sky. The fact that the jets must have been sustained 
261: in their tight configuration for over half a million (possibly as long as ten million) 
262: years means that a highly stable central object---probably a rapidly spinning supermassive 
263: black hole acting like an immovable gyroscope---must be the cause of all this activity.  
264: (Photograph courtesy of Chris Carilli and Rick Perley,
265: NRAO, and AUI)}
266: \end{figure}
267: 
268: An important conclusion to draw from the morphology of jets like those in Cygnus A is that
269: the process responsible for their formation must be stable for at least as long as it takes 
270: the streaming particles to journey from the center of the galaxy to the extremities of the 
271: giant radio lobes. Evidently, these pencil-thin jets of relativistic plasma have retained 
272: their current configuration for over one million years.  The most conservative view regarding
273: the nature of these objects is that a spinning black hole is ultimately responsible for this
274: activity. The axis of its spin functions as a gyroscope, whose direction determines the 
275: orientation of the jets.  Although the definitive mechanism for how the ejection takes place 
276: is yet to be determined, almost certainly the twisting motion of magnetized plasma near the 
277: black hole's event horizon is causing the expulsion.  The Kerr spacetime, which describes the 
278: dragging of inertial frames about the black hole's spin axis, provides a natural setting for 
279: establishing the preferred direction for this ejective process.
280: 
281: \begin{figure}[h]
282: \centerline{\includegraphics[width=\textwidth]{f7bw.eps}}
283: \vskip -0.05in
284: \caption{Streaming out from the center of M87 is a black-hole-powered jet of 
285: plasma traveling at near lightspeed.  Its source is a powerful central object 
286: with a mass of $\sim 3\times 10^9\;M_\odot$.  This sequence of photographs shows 
287: progressively magnified views: {\sl top left:} a VLA image showing the full extent 
288: of the jets and the blobs at the termination points; {\sl top right:} a visible 
289: light image of the giant elliptical galaxy M87, taken with NASA's Hubble Space 
290: Telescope;  {\sl bottom:} a Very Long Baseline Array (VLBA) radio image of the 
291: region surrounding the black hole.  This view shows how the extragalactic jet 
292: is formed into a narrow beam within a few tenths of a light-year of the nucleus 
293: (the red region is only a tenth of a light-year across), corresponding to only
294: 100 Schwarzschild radii for a black hole of this mass. (Photographs courtesy of
295: the National Radio Astronomy Observatory and the National Science Foundation
296: [top left and bottom], and John Biretta at the Space Telescope Science Institute,
297: and NASA [top right])}
298: \end{figure}
299: 
300: \section{Weighing Supermassive Black Holes}
301: Black hole masses are measured with a variety of techniques, though all have to
302: do with the dynamics of matter within their gravitational influence. Knowledge
303: of the radiating plasma's distance from the central object, and the force required 
304: to sustain its motion at that distance, is sufficient for us to extract the 
305: central mass.
306: 
307: One of the more compelling applications of this technique has been made to the 
308: spiral galaxy NGC~4258. Using global radio interferometry, Miyoshi et al. (1995) 
309: observed a disk of dense molecular material orbiting within the galaxy's nucleus 
310: at speeds of up to 650 miles per second. This disk produces sufficient radiation 
311: to excite condensations of water molecules, leading to strong maser emission at 
312: radio wavelengths. The disk within which these water molecules are trapped is
313: small compared to the galaxy itself, but it happens to be oriented fortuitously 
314: so that beams of microwaves are directed along our line-of-sight.
315: 
316: The maser clouds appear to trace a very thin disk, with a motion that follows 
317: Kepler's laws to within one part in 100, reaching a velocity (inferred from
318: the Doppler shift of the lines) of about 650 miles per second at a distance of 0.5 
319: light-years from the center.  The implied central mass is $\sim (35-40)\times
320: 10^6\;M_\odot$, concentrated within 0.5 light years of the center in NGC~4258.
321: This points to a matter density of at least $10^8\;M_\odot$ per cubic light-year.  
322: If this mass were simply a highly concentrated star cluster, the stars would be 
323: separated by an average distance only somewhat greater than the diameter of the 
324: solar system, and with such proximity, they would not be able to survive the 
325: inevitable catastrophic collisions and mergers with each other. Because of the 
326: precision with which we can measure this concentration of dark mass, we regard 
327: the object in the nucleus of NGC~4258 as one of the two most compelling 
328: supermassive black holes now known, the other being the object sitting at the 
329: center of our own galaxy, about which we will have more to say shortly.
330: 
331: The fortuitous arrangement of factors that permits the use of this particular 
332: technique does not occur often, however, so other methods must be used. 
333: Often, clouds of gas orbiting the nucleus are irradiated by the central engine, 
334: and they in turn produce a spectrum with emission lines indicative of the plasma's 
335: ionized state.  The method used to determine the distance of these ionized clouds 
336: from the black hole is known as reverberation.  By monitoring the light emitted by 
337: the supermassive black hole and, independently, the radiation from its halo of 
338: irradiated clouds of gas, one can determine when a variation in the radiative 
339: output has occurred.  When the quasar varies its brightness, so does the 
340: surrounding matter---but only after a certain time delay.  The lag is clearly 
341: due to the time it took the irradiating light from the center to reach the clouds, 
342: and using the speed of light, this delay provides a measure of the distance between 
343: the nucleus and the orbiting plasma.  Again, this procedure provides the speed of
344: matter and its distance from the center, which together yield a determination of
345: the gravitating mass.
346: 
347: Having said this, the best mass determination one can make is still based on kinematic
348: studies of stars orbiting the central object. The center of our Galaxy is close enough
349: for this method to work spectacularly.  Known as Sagittarius A*, the black hole
350: at the center of the Milky Way may not be the most massive, nor the most energetic, 
351: but it is by far the closest, only 28,000 light-years away. Figure~10 shows an infrared 
352: image of Galactic center produced recently with the 8.2-meter VLT YEPUN telescope at the
353: European Southern Observatory in Paranal, Chile.  The image we see here is sharp because 
354: of the use of adaptive optics, in which a mirror in the telescope moves constantly to 
355: correct for the effects of turbulence in the Earth's atmosphere.  Sagittarius A* is so 
356: close to us compared to other supermassive black holes, that on an image such as this,
357: we can identify individual stars orbiting a mere seven to 10 light-days from the source 
358: of gravity. In the nucleus of Andromeda, the nearest major galaxy to the Milky Way, the 
359: best we could do right now is about two light-years.
360: 
361: \begin{figure}[h]
362: \centerline{\includegraphics[width=4in]{f8bw.eps}}
363: \vskip 0.2in
364: \caption{This ($1.6-3.5$ micron) image, taken by the European Southern Observatory's 
365: 8.2-meter telescope atop Paranal, Chile, provides one of the sharpest views of the 
366: stars surrounding the supermassive black hole at the Galactic center.  The location 
367: of the black hole itself is indicated by the two central arrows.  This view represents 
368: a scale of approximately $2\times 2$ square light-years.  (Photograph courtesy of R. 
369: Genzel et al. at the Max-Planck-Institut f\"ur Extraterrestrische Physik, and the 
370: European Southern Observatory)}
371: \end{figure}
372: 
373: In the Galactic center, stars orbit Sagittarius A* at speeds of up to five million 
374: kilometers per hour. This motion is rapid, in fact, that we can easily detect their
375: proper motion on photographic plates taken only several years apart. Some of them
376: complete an orbit about the center in only 15 years (Sch\"odel, Ott, Genzel, et al. 
377: 2002).  In the middle of the photograph in figure~8, it appears that one of the 
378: fainter stars---designated as S2---lies right on top of the position where the 
379: black hole is inferred to be.  S2 is an otherwise ``normal" star, though some 
380: 15 times more massive and seven times larger than the Sun. This star, S2, has
381: now been tracked for over ten years and the loci defining its path trace a perfect 
382: ellipse with one focus at the position of the supermassive black hole.  This 
383: photograph, taken near the middle of 2002, just happens to have caught S2 at the 
384: point of closest approach (known as the perenigricon), making it look like it was
385: sitting right on top of the nucleus.
386: 
387: At perenigricon, S2 was a mere 17 light-hours away from the black hole---roughly 
388: three times the distance between the Sun and Pluto, while traveling with a speed 
389: in excess of 5,000 kilometers per second, by far the most extreme measurements
390: ever made for such an orbit and velocity. We infer from these data that the mass
391: of Sagittarius A* is $\sim 3.4\times 10^6\;M_\odot$, compressed into a region no 
392: bigger than $\sim 17$ light-hours.  For this reason, Sagittarius A*, and the
393: central object in NGC~4258, are considered to be the most precisely ``weighed'' 
394: supermassive black holes yet discovered.
395: 
396: \section{The Formation of Supermassive Black Holes}
397: An increasingly important question being asked in the context of supermassive black holes
398: is which came first, the central black hole, or the surrounding galaxy?  Quasars seem
399: to have peaked 10 billion years ago, early in the universe's existence.  The light from 
400: galaxies, on the other hand, originated much later---after the cosmos had aged another 
401: 2 to 4 billion years. Unfortunately, both measurements are subject to uncertainty, and 
402: no one can be sure we are measuring {\it all} of the light from quasars and galaxies, so 
403: this argument is not quite compelling.  But we do see quasars as far out as we can look,
404: and the most distant among them tend to be the most energetic objects in the
405: universe, so at least {\it some} supermassive black holes must have existed near
406: the very beginning.  At the same time, images such as figure~3 provide evidence of mergers 
407: of smaller structures into bigger aggregates, but without a quasar.  Perhaps not every 
408: collision feeds a black hole or, what is more likely, at least some galaxies must have 
409: formed first. Several scenarios for the formation of supermassive black holes are currently 
410: being examined.  In every case, growth occurs when matter condenses following either the 
411: collapse of massive gas clouds, or the catabolism of smaller black holes in collisions and mergers.
412: 
413: All of the structure in the universe traces its beginnings to a brief era
414: shortly after the Big Bang. Very few ``fossils'' remain from this period; one
415: of the most important is the cosmic microwave background radiation.
416: The rapid expansion that ensued lowered the matter density and temperature,
417: and about one month after the Big Bang, the rate at which photons were created
418: and annihilated could no longer keep up with the thinning plasma. The radiation
419: and matter began to fall out of equilibrium with each other, forever imprinting
420: the conditions of that era onto the radiation that reaches us to this day from all
421: directions in space.
422: 
423: We now know that the temperature anisotropies are smaller than one part in
424: a thousand, a limit below which density perturbations associated with
425: ordinary matter would not have had sufficient time to evolve freely
426: into the nonlinear structures we see today. Only a gravitationally
427: dominant dark-matter component could then account for the strong
428: condensation of mass into galaxies and supermassive black holes.
429: The thinking behind this is that whereas the cosmic microwave
430: background radiation interacted with ordinary matter, it would
431: retain no imprint at all of the dark matter constituents in the
432: universe.  The nonluminous material could therefore be condensed
433: unevenly (sometimes said to be `clumped') all the way back to the
434: Big Bang and we simply wouldn't know it.
435: 
436: The first billion years of evolution following the Big Bang must have been
437: quite dramatic in terms of which constituents in the universe would eventually
438: gain primacy and lasting influence on the structure we see today.  The issue of how
439: the fluctuations in density, mirrored by the uneven cosmic microwave background radiation,
440: eventually condensed into supermassive black holes and galaxies is currently
441: a topic of ongoing work.  This question deals with the fundamental contents of 
442: the universe, and possibly what produced the Big Bang and what came before it. 
443: The evidence now seems to be pointing to a coeval history for these two dominant 
444: classes of objects---supermassive black holes and galaxies---though as we have 
445: already noted, at least some of the former must have existed quite early.
446: 
447: One possibility proposed by Balberg and Shapiro (2002) is that the first 
448: supermassive objects formed from the condensation of dark matter alone; only
449: later would these seed black holes have imposed their influence on the latter.
450: But this dark matter has to be somewhat peculiar, in the sense that its
451: constituents must be able to exchange heat with each other.  As long as this
452: happens, a fraction of its elements evaporate away from the condensation,
453: carrying with them the bulk of the energy, and the rest collapse and create
454: an event horizon.  The net result is that the inner core of such a clump forms
455: a black hole, leaving the outer region and the extended halo in equilibrium
456: about the central object.  Over time, ordinary matter gathers around it,
457: eventually forming stars, and planets.
458: 
459: Ordinary matter could not have achieved this early condensation because
460: it simply wasn't sufficiently clumped initially.  Perhaps this material 
461: formed the first stars, followed by more stars, eventually assembling a 
462: cluster of colliding objects.  Over time, the inner core of such an assembly 
463: would have collapsed due to the evaporation of some of its members and the 
464: ensuing loss of energy into the extended halo, just as the dark matter did 
465: (see, e.g., Haehnelt and Rees 1993).  A seed black hole might have formed
466: in the cluster's core. Estimates show that once formed, such an object 
467: could have doubled its mass every 40 million years, so over the age of the 
468: universe, even a modestly appointed black hole could have grown into a 
469: billion-solar-mass object. The problem is that this could not have happened
470: in only 700 million years, when the first supermassive black holes appeared.
471: 
472: Yet another method leading to black hole growth results from ongoing
473: collisions between galaxies, which eventually lead to the merger of the
474: black holes themselves. An example of such a process occurring right now 
475: is shown in figure~9. 
476: 
477: \begin{figure}[h]
478: \centerline{\includegraphics[width=4in]{f9bw.eps}}
479: \vskip -0.0in
480: \caption{NGC~6240 is a butterfly-shaped galaxy, believed to be the
481: product of a collision between two smaller galaxies some 30 million
482: years ago.  This {\sl Chandra} X-ray image (34,000 light-years across)
483: shows the heat generated by the merger activity, which created the extensive
484: multimillion degree Celsius gas. We see here for the first time direct
485: evidence that the nucleus of such a structure contains not one, but
486: two active supermassive black holes, drifting toward each other across 
487: their 3,000-light-year separation; they're expected to merge into a bigger
488: object several hundred million years hence. (Image courtesy of Susan Komossa, 
489: Gunter Hasinger, and Joan Centrella, and the Max-Planck-Institut f\"ur 
490: Extraterrestrische Physik and NASA)}
491: \end{figure}
492: 
493: Almost every large, normal galaxy harbors a supermassive black hole at its 
494: center. Some evidence for this has been provided by a recently completed survey
495: of 100 nearby galaxies using the VLA, followed by closer scrutiny
496: with the Very Long Baseline Interferometry array. At least 30 percent of
497: this sample showed tiny, compact central radio sources bearing the
498: unique signature of the quasar phenomenon. Of the hundreds of millions
499: of supermassive black holes seen to pervade the cosmos, none of them
500: appear to be isolated. And even more compelling is the work of Kormendy 
501: and Richstone (1995), who set about the task of systematically measuring as 
502: many black hole masses as is currently feasible.  Direct measurements of 
503: supermassive black holes have been made in over 38 galaxies, based on the 
504: large rotation and random velocities of stars and gas near their centers.  
505: These objects are all relatively nearby because these direct methods don't 
506: work unless we can see the individual stars in motion about the central 
507: source of gravity. Curiously, none of the supermassive black holes have 
508: been found in galaxies that lack a central bulge (figure~10).  Galaxies 
509: with a central bulge may have undergone one or more mergers in their past. 
510: Thus, a collision like that seen in figures 3 and 11 may have been
511: required to create a central supermassive object. 
512: 
513: \begin{figure}[h]
514: \centerline{\includegraphics[width=4in]{f10bw.eps}}
515: \vskip -0.0in
516: \caption{This image of the spiral galaxy NGC~2613 was captured on February 26, 2002, by the Very
517: Large Telescope in Paranal, Chile.  A search for a supermassive black hole in its core produced
518: a null result (Bower et al. 1993).  By now the preponderance of evidence suggests that highly
519: flattened disk galaxies lacking a significant central hub, or spheroidal component, also lack
520: a supermassive object in the nucleus.  On the other hand, every galaxy that does possess
521: a central bulge also harbors a supermassive black hole. (Image courtesy of S. D'Odorico et al.,
522: who obtained it during the test phase of the VIMOS instrument at the European Southern Observatory's
523: 8-meter MELIPAL telescope, and courtesy also of the European Southern
524: Observatory)}
525: \end{figure}
526: 
527: Another recently inferred correspondence between supermassive black holes and their 
528: host galaxies seems to have clinched the case for a coeval growth (Ferrarese and Merritt 
529: 2000; Gebhardt et al. 2000).  The data displayed in figure~11 show the relationship 
530: between the mass of the black hole and the velocity dispersion of stars within the 
531: spheroidal (or bulge) component of stars in the host galaxy. Based on the very tight 
532: correlation evident in this graph, it appears that the mass of the central black hole
533: can be predicted with remarkable accuracy simply by knowing the velocity of stars 
534: orbiting so far from it that its gravitational influence could not possibly be affecting 
535: their motion at the present time. Pinpointing the trajectory of a single object in the 
536: nuclei of the host galaxies is difficult and impracticable.  Instead, these measurements 
537: are based on the accumulated light from a limited region of the galaxy's central hub,
538: from which one may extract the collective Doppler shift, and thereby an average speed 
539: of the stars as a group. What these studies reveal is that the ratio of the black hole's 
540: mass to this average speed is constant across the whole sample of galaxies surveyed.
541: 
542: \begin{figure}[h]
543: \centerline{\includegraphics[width=\textwidth]{f11bw.eps}}
544: \vskip -0.1in
545: \caption{Plotted against the velocity dispersion of stars orbiting within the
546: spheroidal component of the host galaxy, the mass of a supermassive black hole
547: is tightly correlated with the former, hinting at a coeval history.
548: The stars are simply too far away from the nucleus to be directly affected by
549: the black hole's gravity at the present time (Ferrarese and Merritt 2000).}
550: \vskip 0.0in
551: \end{figure}
552: 
553: This result is one of the most surprising correlations yet discovered in the 
554: study of how the universe acquired its structure.  Supermassive black holes, 
555: it seems, ``know" about the motion of stars that are too distant to directly feel
556: their gravity.  This tight connection suggests an entangled history between a 
557: central black hole and the stellar activity in its halo. Although they may not 
558: be causally bonded today, they must have had an overlapping genesis in the past.
559: 
560: Additional information on the possible coeval history of supermassive black holes 
561: and their host galaxies is provided by images such as figure~4.  If we adopt a 
562: canonical efficiency of $10\;\%$ for the conversion of rest mass energy into 
563: radiation, we obtain the mass inflow rate versus redshift relation shown in 
564: figure~12.  This rate was as high as $10$ solar masses per year at a redshift 
565: of $3$ or so, and has declined monotonically to its much lower value of around 
566: $0.01$ solar masses per year in the current epoch.  An even more interesting 
567: evolutionary trait is that obtained by integrating the data in figure~12 over 
568: the black hole spatial distribution, which produces the redshift dependence 
569: of the so-called accretion rate density (figure~13).  This quantity effectively 
570: gives us the rate at which black hole mass is increasing with time.  The apparent 
571: $(1+z)^3$ dependence of this quantity (exhibited in figure~13) is identical to the
572: universe's star formation history, as deduced from the comoving UV intensity as 
573: a function of redshift.
574: 
575: \begin{figure}[h]
576: \centerline{\rotatebox{90}{\includegraphics[width=3.5in]{f12bw.eps}}}
577: \vskip -0.0in
578: \caption{The black hole accretion rate, in units of solar masses per year, versus
579: redshift, assuming a canonical efficiency rate of $10\,\%$ conversion of rest mass
580: energy into radiation. To generate a $10^9\;M_\odot$ black hole over an accretion
581: period of order $0.5$ Gyr, a rate of order $2\;M_\odot$ yr$^{-1}$ is required.
582: (From Barger et al. 2001)}
583: \vskip 0.0in
584: \end{figure}
585: 
586: \begin{figure}[h]
587: \centerline{\rotatebox{90}{\includegraphics[width=3.5in]{f13bw.eps}}}
588: \vskip -0.0in
589: \caption{The time history of the accretion rate density (effectively, the
590: rate at which black hole matter is increasing, in units of solar masses per year per megaparsec
591: cubed), integrated from the data exhibited in figure~12.  The upper (solid) and lower (open)
592: bounds in this figure correspond to the accretion rate density calculated from the bolometric
593: and X-ray luminosities, respectively.  The curves illustrate a $(1+z)^3$ dependence, normalized
594: to the $1<z<1.5$ redshift bin.  (From Barger et al. 2001)}
595: \vskip 0.0in
596: \end{figure}
597: 
598: Integrating this curve over time, we infer that about $0.006$ of the all
599: the mass contained within the bulge of any given galaxy should be in the
600: form of a central supermassive black hole.  This is close to the value of 
601: $\sim  0.002$ found in the local neighborhood of our Galaxy.  With these 
602: facts in hand, and satisfied that most supermassive black holes and their 
603: host galaxies grew interdependently, one may hypothesize that once created,
604: a primordial condensation of matter continues to grow with a direct
605: feedback on its surroundings.  This may happen either because the
606: quasar heats up its environment and controls the rate at which
607: additional matter can fall in from its cosmic neighborhood,
608: or because mergers between galaxies affect the growth of colliding
609: black holes in the same way that they determine the energy (and
610: therefore the average speed) of the surrounding stellar distribution.
611: In support of the idea that massive black holes may have formed
612: prior to the epoch of galaxy definition, Silk and Rees (1998)
613: suggest further that protogalactic star formation would have been
614: influenced significantly by the quasar's extensive energy outflows.
615: The ensuing feedback on the galaxy's spheroidal component could
616: be the reason we now see such a tight correlation between the
617: mass of the central object and the stellar velocities much farther
618: out.
619: 
620: Another possible reason why black hole masses are correlated with
621: the velocity dispersion of the host galaxy bulges may be the importance
622: of galaxy mergers to the hierarchical construction of elaborate elliptical,
623: or disk-plus-bulge profiles.  Most large galaxies have experienced at 
624: least one major merger during their lifetime. Larger galaxies have
625: grown with a succession of collisions and mergers, a process contributing
626: significantly to the variety of shapes encompassed by the Hubble sequence. 
627: Meanwhile, the turbulence generated in the core of the collision drives 
628: most of the gas into the middle, where it forms new stars and feeds
629: a central black hole, or a pair of black holes. The tight correlation 
630: between the black-hole mass and the halo velocity dispersion may be 
631: a direct consequence of this cosmic cascade.
632: 
633: \section{Imaging a Supermassive Black Hole}
634: Of course, the most exciting development in black hole research would be
635: the actual imaging of its event horizon (Melia 2003b).  This may be feasible
636: in the next several years, with the development of a global mm-VLBA. The spectrum 
637: of Sagittarius A* at the galactic center displays the well-known profile of a 
638: self-absorbed emitter longward of $3-7$ mm, and a transparent medium at higher 
639: frequencies (see figure~14). The fact that Sagittarius A* becomes transparent 
640: across the mm/sub-mm bump in its spectrum means that a depression in intensity
641: (literally, a `black hole') should arise shortward of $\sim 1$ mm, due to the
642: effects of strong light bending near the event horizon. The `shadow' cast by the 
643: black hole in this fashion should have an apparent diameter of about 5 Schwarzschild 
644: radii (Falcke, Melia, and Agol 2000), nearly independent of the black hole
645: spin or its orientation. The distance to Sagittarius A* being roughly $8$ kpc, this diameter
646: corresponds to an angular size of $\sim 30\,\mu$as, which approaches the resolution of
647: current radio interferometers.
648: 
649: \begin{figure}[h]
650: \centerline{\includegraphics[width=4in]{f14bw.eps}}
651: \vskip -0.1in
652: \caption{Spectra showing total flux density and polarization from Sagittarius A*.
653: The high-frequency data points are from Aitken et al. (2000) and the curves
654: are from the magnetohydrodynamic model discussed in Bromley, Melia, and Liu  (2001).  The
655: limit at 84 GHz is from Bower et al. (1999).  At even lower frequencies, the best
656: fit polarization is consistent with $0$.  In this figure, `negative' polarization
657: corresponds to the polarization vector being aligned with the spin axis of the black
658: hole, whereas positive is for a perpendicular configuration.  The polarization
659: crosses $0$ when this vector flips by $90^\circ$ (Bromley, Melia, and Liu 2001).}
660: \end{figure}
661: 
662: The fortunate aspect of this transition from optically-thick to optically-thin
663: emission across the mm/sub-mm bump is that it happens to occur at roughly the 
664: same frequency where scatter broadening in the interstellar medium decreases 
665: below the intrinsic source size.  Taking this effect into account, and the finite 
666: telescope resolution, one infers (Falcke, Melia, and Agol 2000) that the shadow of 
667: Sagittarius A* should be observable below $\sim 1$ mm. 
668: 
669: Most recently, ray-tracing simulations have taken into account the magnetized 
670: Keplerian flow and the effects of light-bending and area amplification 
671: to produce the most detailed theoretical predictions of the images that mm-astronomy 
672: will produce in the near future.  The polarimetric images shown in figure~15 
673: demonstrate that future mm/sub-mm interferometry can directly reveal material 
674: flowing near the horizon in Sagittarius A*.  
675: 
676: \begin{figure}[h]
677: \centerline{\includegraphics[width=\textwidth]{f15bw.eps}}
678: \vskip -0.0in
679: \caption{Polarization maps at three wavelengths near the
680: peak of the mm to sub-mm emission from Sagittarius A*. The top row shows
681: emission at 1.5~mm, the middle row is at 1~mm, and the
682: bottom row corresponds to 0.67~mm. The images in each row
683: show the raw ray-tracing output (first column on the left), and
684: an image blurred to account for finite VLBI resolution and
685: interstellar scattering (second column). The two rightmost
686: columns give the vertical and horizontal components of the polarized emission.
687: The pixel brightness in all images scales linearly with flux (Bromley, Melia, and Liu 2001).}
688: \vskip 0.0in
689: \end{figure}
690: 
691: The blurring in these images is modeled by convolution with two Gaussians: the first
692: is an approximate model of the scattering effects with an ellipsoidal filter
693: whose major- and minor-axis FWHM values are 24.2~$\mu$as~$\times$ $(\lambda/1.3\;\hbox{\rm
694: mm})^2$ and 12.8~$\mu$as~$\times (\lambda/1.3\;\hbox{\rm mm})^2$, respectively,
695: for emission at wavelength $\lambda$. In the simulations
696: shown in Figure~15, the spin axis of the disk was chosen to lie along
697: the minor axis of the scattering ellipsoid, and a global interferometer
698: was assumed with a $8,000$ km baseline.  The second is a spherically symmetric 
699: filter with a FWHM of 33.5~$\mu$as~$\times (\lambda/1.3\;\hbox{\rm mm})$ to 
700: account for the resolution effects of an ideal interferometer.  These images 
701: demonstrate clearly the viability of conducting polarimetric imaging of the
702: black hole at the Galactic center with upcoming VLBI techniques.
703: 
704: \section{Conclusion}
705: \noindent
706: We have come a long way since 1963, when Roy Kerr discovered the spacetime
707: that is now synonymous with black hole research. The past decade, in particular,
708: has seen a dramatic shift in our view of the role played by supermassive black
709: holes in the formation of structure in the universe. We now appear to be on the 
710: verge of finally ``seeing" some of these objects, providing more direct 
711: evidence for their existence, and a more satisfying confirmation of their 
712: properties predicted by general relativity.  Within the framework of the Kerr
713: metric, Sagittarius A* at the Galactic center, and possibly even the central
714: black hole in the relatively nearby M81 galaxy, are expected to produce distinct
715: shadows with a precise shape and size of the region where light bending
716: and capture are important.  This coming decade may finally give us a view into
717: one of the most important and intriguing predictions of general relativity.
718: 
719: \begin{thereferences}
720: 
721: \bibitem{Aitken}
722: 	Aitken, D. K., et al. (2000). Detection of Polarized Millimeter and Submillimeter Emission 
723: 	from Sagittarius A*, \textit{The Astrophysical Journal Letters} \textbf{534}, L173--L176.
724: 
725: \bibitem{Baganoff}
726: 	Baganoff, F. et al. (2001). Rapid X-ray flaring from the direction of the supermassive 
727: 	black hole at the Galactic Centre, \textit{Nature} \textbf{413}, 45--48.
728: 
729: \bibitem{Balberg}
730: 	Balberg, S., and S. L. Shapiro (2002). Gravothermal Collapse of Self-Interacting 
731: 	Dark Matter Halos and the Origin of Massive Black Holes, \textit{Physical Review Letters} 
732: 	\textbf{88}, 101301.1--101301.4.
733: 
734: \bibitem{barger}
735: 	Barger, A. J. et al. (2001). Supermassive Black Hole Accretion History Inferred from a 
736: 	Large Sample of Chandra Hard X-ray Sources, \textit{Astronomical Journal} \textbf{122}, 2177--2194.
737: 
738: \bibitem{bower}
739: 	Bower, G. A., D. O. Richstone, G. D. Bothun, and T. M. Heckman (1993).
740: 	A Search for Dead Quasars Among Nearby Luminous Galaxies. I.  The Stellar Kinematics in the
741: 	Nuclei NGC~2613, NGC~4699, NGC~5746, and NGC~7331, \textit{Astrophysical Journal} \textbf{402}, 76--94.
742: 
743: \bibitem{bower99}
744: 	Bower, G. C., et al. (1999). The Linear Polarization of Sagittarius A*. I. VLA 
745: 	Spectropolarimetry at 4.8 and 8.4 GHZ, \textit{The Astrophysical Journal} \textbf{521}, 582--586.
746: 
747: \bibitem{bromley}
748: 	Bromley, B., F. Melia, and S. Liu (2001). Polarimetric Imaging of the Massive Black Hole at the 
749: 	Galactic Center, \textit{Astrophysical Journal Letters} \textbf{555}, L83--L86.
750: 
751: \bibitem{coker}
752: 	Coker, R., and F. Melia (1997). Hydrodynamical Accretion onto Sagittarius A* from Distributed 
753: 	Point Sources, \textit{The Astrophysical Journal Letters} \textbf{488}, L149--L152.
754: 
755: \bibitem{connors}
756: 	Connors, P. A., and R. F. Stark (1977). Observable gravitational effects on polarised radiation 
757: 	coming from near a black hole, \textit{Nature} \textbf{269}, 128--129.
758: 
759: \bibitem{fabian}
760: 	Fabian, A. C., et al. (2000). Testing the Connection Between the
761: 	X-ray and Submillimeter Source Populations Using Chandra,
762: 	\textit{Monthly Notices of the Royal Astronomical Society} \textbf{315}, L8--L12.
763: 
764: \bibitem{ferrarese}
765: 	Ferrarese, L., and D. Merritt (2000). A Fundamental Relation Between
766: 	Supermassive Black Holes and Their Host Galaxies, \textit{The Astrophysical
767: 	Journal Letters} \textbf{539}, L9--L12.
768: 
769: \bibitem{gebhardt}
770: 	Gebhardt, K., J. Kormendy, L. C. Ho, R. Bender, G. Bower, A. Dressler, et al. (2000).
771: 	A Relationship Between Nuclear Black Hole Mass and Galaxy Velocity Dispersion,
772: 	\textit{The Astrophysical Journal Letters} \textbf{539}, L13--L16.
773: 
774: \bibitem{gebhardt1}
775: 	Gebhardt, K., R. M. Rich, and L. C. Ho (2002). A 20,000 Solar Mass Black Hole in the 
776: 	Stellar Cluster G1, \textit{The Astrophysical Journal Letters} \textbf{578}, L41--L45.
777: 
778: \bibitem{falcke}
779: 	Falcke, H., F. Melia, and E. Agol (2000). Viewing the Shadow of the Black Hole at the 
780: 	Galactic Center, \textit{The Astrophysical Journal Letters} \textbf{528}, L13--L16.
781: 
782: \bibitem{green}
783: 	Green, P. J., T. L. Aldcroft, S. Mathur, B. J. Wilkes, and M. Elvis (2001)
784: 	A Chandra Survey of Broad Absorption Line Quasars, \textit{The Astrophysical Journal} 
785: 	\textbf{558}, 109--118.
786: 
787: \bibitem{haehnelt}
788: 	Haehnelt, M. G., and M. J. Rees (1993). The Formation of Nuclei in
789: 	Newly Formed Galaxies and the Evolution of the Quasar Population, \textit{Monthly Notices 
790: 	of the Royal Astronomical Society} \textbf{263}, 168--178.
791: 
792: \bibitem{kerr}
793: 	Kerr, R. P. (1963). Gravitational Field of a Spinning Mass as an Example of 
794: 	Algebraically Special Metrics, \textit{Physical Review Letters} \textbf{11}, 237--238.
795: 
796: \bibitem{kormendy}
797: 	Kormendy, J., and D. Richstone (1995). Inward Bound---The Search for
798: 	Supermassive Black Holes in Galactic Nuclei, \textit{Annual Reviews of 
799: 	Astronomy and Astrophysics} \textbf{33}, 581--624.
800: 
801: \bibitem{laor}
802: 	Laor, A., and H. Netzer (1990). Massive thin accretion discs. II - Polarization,
803: 	\textit{Monthly Notices of the Royal Astronomical Society} \textbf{242}, 560--569.
804: 
805: \bibitem{liu}
806: 	Liu, S., V. Petrosian, and F. Melia (2004). Electron Acceleration around the 
807: 	Supermassive Black Hole at the Galactic Center, \textit{The Astrophysical Journal 
808: 	Letters} \textbf{611} L101--L104. 
809: 
810: \bibitem{lynden}
811: 	Lynden-Bell, D. (1969). Galactic Nuclei as Collapsed Old Quasars, 
812: 	\textit{Nature} \textbf{223}, 690.
813: 
814: \bibitem{malizia}
815: 	Malizia, A., L. Bassani, et al. (2001). Hard X-Ray Detection of the High-Redshift 
816: 	Quasar 4C 71.07, \textit{The Astrophysical Journal} \textbf{531}, 642--646.
817: 
818: \bibitem{martini}
819: 	Martini, P., and R. W. Pogge (1999). Hubble Space Telescope Observations of the 
820: 	CFA Seyfert 2 Galaxies: The Fueling of Active Galactic Nuclei, \textit{Astronomical 
821: 	Journal} textbf{118}, 2646--2657.
822: 
823: \bibitem{matsumoto}
824: 	Matsumoto, H., et al. (2001). Discovery of a Luminous, Variable, Off-Center Source in
825: 	the Nucleus of M82 with the Chandra High-Resolution Camera, \textit{Astrophysical Journal 
826: 	Letters} \textbf{547}, L25--L28.
827: 
828: \bibitem{meliaa}
829: 	Melia, F. (2003a). \textit{The Edge of Infinity. Supermassive Black Holes in the 
830: 	Universe}. Cambridge: Cambridge University Press.
831: 
832: \bibitem{meliab}
833: 	Melia, F. (2003b). \textit{The Black Hole at the Center of Our Galaxy}. Princeton: 
834: 	Princeton University Press.
835: 
836: \bibitem{melia1}
837: 	Melia, F., S. Liu, and R. Coker (2000). Polarized Millimeter and Submillimeter 
838: 	Emission from Sagittarius A* at the Galactic Center, \textit{The Astrophysical 
839: 	Journal Letters} \textbf{545}, L117--L120.
840: 
841: \bibitem{melia2}
842: 	Melia, F., S. Liu, and R. Coker (2001). A Magnetic Dynamo Origin for the 
843: 	Submillimeter Excess in Sagittarius A*, \textit{The Astrophysical Journal} \textbf{553}, 
844: 	146--157.
845: 
846: \bibitem{merrifield}
847: 	Merrifield, M. R., D. A. Forbes, and A. I. Terlevich (2000). The Black Hole
848: 	Mass-galaxy Age Relation, \textit{Monthly Notices of the Royal Astronomical Society}
849: 	\textbf{313}, L29--L32.
850: 
851: \bibitem{miyoshi}
852: 	Miyoshi, M., et al. (1995). Evidence for a Black Hole from High Rotation
853: 	Velocities in a Sub-Parsec Region of NGC~4258, \textit{Nature} \textbf{373}, p. 127.
854: 
855: \bibitem{salpeter}
856: 	Salpeter, E. E. (1964). Accretion of Interstellar Matter by Massive Objects,
857: 	\textit{Astrophysical Journal} \textbf{140}, 796--800.
858: 
859: \bibitem{schmidt}
860: 	Schmidt, M. (1963). 3C~273: A Star-like Object with Large Red-shift,
861: 	\textit{Nature} \textbf{197}, 1040.
862: 
863: \bibitem{schodel}
864: 	Sch\"odel, R., T. Ott, R. Genzel, et al. (2002). A Star in a 15.2-year
865: 	Orbit Around the Supermassive Black Hole at the Centre of the Milky Way,
866: 	\textit{Nature} \textbf{419}, 694--696.
867: 
868: \bibitem{silk}
869: 	Silk, J., and M. J. Rees (1998). Quasars and Galaxy Formation,
870: 	\textit{Astronomy and Astrophysics} \textbf{331}, L1--L4.
871: 
872: \bibitem{marel}
873: 	van der Marel, R. P. (1999). Structure of the Globular Cluster M15 and
874: 	Constraints on a Massive Central Black Hole, in \textit{Black Holes
875: 	in Binaries and Galactic Nuclei,} L. Kaper, E.P.J. van den Heuvel,
876: 	and P. A. Woudt (eds.), 246.
877: 
878: \bibitem{yuan}
879: 	Yuan, F., E. Quataert, and R. Narayan (2003). Nonthermal Electrons in 
880: 	Radiatively Inefficient Accretion Flow Models of Sagittarius A*,
881: 	\textit{The Astrophysical Journal} \textbf{598}, 301--312.
882: 
883: \bibitem{zeldovich}
884: 	Zel'dovich, Ya. B., and I. D. Novikov (1967). The Hypothesis
885: 	of Cores Retarded During Expansion and the Hot Cosmological Model, 
886: 	\textit{Soviet Astronomy} \textbf{10}, 602.
887: 
888: \end{thereferences}
889: 
890: \end{document}
891: