1: %\documentstyle[draft,12pt,psfig]{nature}
2: %\documentstyle[psfig]{nature}
3: \documentstyle[psfig]{nature_local}
4:
5: \def\ltsima{$\; \buildrel < \over \sim \;$}
6: \def\simlt{\lower.5ex\hbox{\ltsima}}
7: \def\gtsima{$\; \buildrel > \over \sim \;$}
8: \def\simgt{\lower.5ex\hbox{\gtsima}}
9:
10: \title{Rapid Growth of Black Holes in Massive Star-Forming Galaxies}
11:
12: \author{D.~M.~Alexander\affiliation{Institute of Astronomy, Madingley Road, Cambridge CB3 0HA, UK},
13: I.~Smail\affiliation{Institute for Computational Cosmology, University of Durham, South Road, Durham DH1 3LE, UK},
14: F.~E.~Bauer $^{*}$,
15: S.~C.~Chapman\affiliation{California Institute of Technology, Pasadena, CA 91125, USA},
16: A.~W.~Blain $^{\ddag}$,
17: W.~N.~Brandt\affiliation{Department of Astronomy and Astrophysics, Pennsylvania State University, 525 Davey Laboratory, University Park, PA 16802, USA},
18: R.~J.~Ivison\affiliation{Astronomy Technology Centre, Royal Observatory, Blackford Hill, Edinburgh EH9 3HJ, UK}\affiliation{Institute for Astronomy, University of Edinburgh, Blackford Hill, Edinburgh EH9 3HJ, UK}
19: }
20:
21: \headertitle{}
22: \mainauthor{}
23:
24: \summary{\noindent The tight relationship between the masses of black
25: holes and galaxy spheroids in nearby galaxies$^{1}$ implies a causal
26: connection between the growth of these two components. Optically
27: luminous quasars host the most prodigious accreting black holes in
28: the Universe and can account for $\simgt30$~\% of the total
29: cosmological black-hole growth$^{2,3}$. As typical quasars are not,
30: however, undergoing intense star formation and already host massive
31: black holes ($>10^{8}$~$M_{\odot}$)$^{4,5}$, there must have been an
32: earlier pre-quasar phase when these black holes grew (mass range
33: $\approx$~$10^{6}$--$10^{8}$~$M_{\odot}$). The likely signature of
34: this earlier stage is simultaneous black-hole growth and star
35: formation in distant (i.e., $z>1$; $>8$~billion light years away)
36: luminous galaxies. Here we report ultra-deep X-ray observations of
37: distant star-forming galaxies that are bright at submillimetre
38: wavelengths. We find that the black holes in these galaxies are
39: growing almost continuously throughout periods of intense star
40: formation. This activity appears to be more tightly associated with
41: these galaxies than any other coeval galaxy populations. We show
42: that the black-hole growth from these galaxies is consistent with
43: that expected for the pre-quasar phase.} \dates{12 October 2004}{11
44: February 2005}
45:
46: \begin{document}
47: \maketitle
48:
49: The most intense sites of star formation at high redshift are
50: associated with submillimetre galaxies (SMGs)$^{6-9}$. SMGs are
51: amongst the most bolometrically luminous galaxies in the Universe
52: ($L_{\rm BOL}\approx10^{13}$~$L_{\odot}$) and the majority of the
53: population is coeval with the peak in quasar activity (i.e.,
54: $z\approx$~1.5--3.0)$^{7,9}$. The apparent association of SMGs with some
55: quasars and the similarity in the comoving space densities of SMGs and
56: optically luminous ($M_{\rm B}<-24$) quasars (when corrected for
57: probable source lifetimes) provides direct evidence for an
58: evolutionary connection between SMGs and quasars$^{5,10,11}$. However,
59: in contrast to quasars, the bolometric output of SMGs appears to be
60: dominated by powerful star-formation activity and any black-hole
61: accretion [i.e., Active Galactic Nuclei (AGN) activity] is
62: comparatively weak$^{12,13}$. Given the large available molecular gas
63: supply (typically $\approx3\times10^{10}$~$M_{\odot}$)$^{14}$, SMGs
64: can fuel this luminous star-formation activity for $\approx10^{8}$~yrs
65: (ref 14,15). Various pieces of complementary observational support
66: have also shown that SMGs are massive galaxies (typically
67: $\approx10^{11}$~$M_{\odot}$), suggesting that they will become
68: $\simgt$$M_{*}$ spheroid-dominated galaxies in the local
69: Universe$^{14,16,17}$.
70:
71: AGN activity has been identified in many SMGs$^{6,12}$. However, the
72: difficulty in obtaining high-quality optical spectra and reliable
73: source redshifts has hindered a complete census of AGN activity in
74: SMGs. We have initiated a project investigating the properties of AGNs
75: in SMGs using deep optical spectroscopic data (obtained with the 10-m
76: Keck telescope)$^{9}$ and ultra-deep X-ray observations (the 2 Ms
77: Chandra Deep Field-North; CDF-N)$^{18}$. As X-ray emission appears to
78: be an ubiquitous property of AGNs that is comparatively impervious to
79: obscuration, the combination of these deep datasets provides a
80: detailed census of AGN activity in SMGs. Here we focus on the role of
81: black-hole accretion in SMGs and the growth of black holes in massive
82: galaxies. The assumed cosmology is $H_0=65$~km~s$^{-1}$ Mpc$^{-1}$,
83: $\Omega_{\rm M}=1/3$, and $\Omega_{\Lambda}=2/3$.
84:
85: The parent SMG sample comprises 20 submillimetre detected (with the
86: SCUBA camera on the James Clerk Maxwell Telescope) sources in the
87: CDF-N region whose positions have been reliably located from their
88: radio emission and which could then be spectroscopically
89: targeted$^{9}$. These sources have overall properties consistent with
90: the general SMG population ($S_{\rm 850\mu m}\approx$~4--12~mJy and
91: $z=$~0.56--2.91, with the majority at $z>2$)$^{6-9}$. Seventeen
92: ($\approx$~85\%) of the 20 SMGs are detected at X-ray energies. A
93: detailed investigation of the nature of the X-ray emission has
94: revealed AGN activity in at least 15 ($\approx$~75\%) SMGs$^{19}$; the
95: properties of these sources are given in Table~1. The X-ray properties
96: of the other 5 ($\approx$~25\%) SMGs are consistent with those
97: expected from luminous star-formation activity, but we note that at
98: least one shows evidence for AGN activity at near-infrared
99: wavelengths$^{16}$.
100:
101: \begin{figure}[t]
102: \centerline{\psfig{file=alexfig1.ps,width=3.5in}}
103: \caption{Black-hole mass accretion rates. SMGs (filled circles), and a
104: comparison sample of $M_{\rm B}<-24$ quasars (dots) are shown$^{4}$.
105: The error bars represent the estimated uncertainty in the X-ray to
106: bolometric correction for AGN activity in SMGs; see Table~1. The
107: solid and dashed curves indicate the median and interquartile ranges
108: for the comparison quasar sample. The dotted line indicates the
109: approximate Eddington-limited mass accretion rate for an
110: $\approx10^{8}$~$M_{\odot}$ black hole ($M_{\rm BH (Edd)}$). The
111: SMGs have mass accretion rates approximately an order of magnitude
112: lower than those of coeval quasars$^{4}$, suggesting that their
113: black holes are smaller (typically $\approx
114: 10^{7}$--$10^{8}$~$M_{\odot}$); see text for additional black-hole
115: mass constraints.}
116: \end{figure}
117:
118: Two different observing modes were used in constructing the SMG
119: sample: 14 of the radio sources were specifically targeted with SCUBA
120: observations and 6 of the radio sources were detected in blank-field
121: SCUBA maps. Although targeting known radio sources with SCUBA
122: observations could potentially bias our AGN fraction owing to
123: contributions to the radio emission from AGN activity, a statistical
124: analysis does not reveal a strong AGN bias in our sample (P=1.0,
125: two-sided Fisher's exact test); see Table~1 for the observing modes of
126: the X-ray classified AGNs. However, due to the large positional
127: uncertainties of SCUBA sources ($\approx$~6--7 arcseconds), redshifts
128: could not be obtained for the $\approx$~35--50\% of the
129: radio-undetected SMG population, some of which might host AGN
130: activity$^{6,9}$. Making the conservative assumption that none of the
131: radio-undetected SMGs hosts AGN activity, our X-ray data alone suggest
132: an AGN fraction in the SMG population of $>38^{+12}_{-10}$\%. We note
133: that a similar fraction of comparably luminous galaxies in the local
134: Universe host AGN activity$^{22}$, although their space density is
135: approximately 3 orders of magnitude smaller than SMGs.
136:
137: The large AGN fraction in the SMG population indicates that their
138: black holes are growing almost continuously throughout the intense
139: star-formation phase of these galaxies [i.e., assuming that all SMGs
140: are growing their black holes, the AGN duty cycle is
141: $>38^{+12}_{-10}$\%]. This suggests that the black holes and galaxy
142: spheroids are growing concordantly in SMGs$^{1}$. Such a close association
143: between AGN and star-formation activity is not seen in the coeval
144: field galaxy population or other coeval star-forming galaxy
145: populations ($\approx$~3--15\% AGN fraction)$^{23,24}$. The almost
146: continuous black-hole growth in SMGs suggests that there is an
147: abundance of available fuel, hinting that the accretion may be
148: occurring efficiently (i.e., at or close to the Eddington limit), as
149: predicted by theoretical studies of the growth of massive black
150: holes$^{25-27}$. The majority ($\approx$~85\%) of the AGNs are
151: obscured (see Table~1), as also predicted for black holes undergoing
152: efficient growth$^{25}$.
153:
154: The unobscured X-ray luminosities of the AGNs ($L_{\rm
155: X}\approx10^{43}$--$10^{44}$~erg~s$^{-1}$) suggest that the mass
156: accretion rates are modest ($\simlt$~1~$M_{\odot}$~yr$^{-1}$); see
157: Table~1 and Figure~1. By comparison to coeval quasars, the mass
158: accretion rates in SMGs are approximately an order of magnitude lower,
159: also suggesting modest black-hole masses (i.e.,
160: $\approx10^{7}$--$10^{8}$~$M_{\odot}$; see Table 1 and Figure 1).
161: Although these black-hole mass estimates are uncertain, the narrowness
162: of the broad emission lines (BELs; typical full-width half maximum
163: velocities of $\approx$~2500~km~s$^{-1}$)$^{16}$, when they are
164: visible, also suggest black-hole masses of $<10^{8}$~$M_{\odot}$ (ref
165: 4). As well-studied AGNs with narrow BELs are found to be accreting at
166: (or even beyond) the Eddington limit$^{4}$, this provides further
167: support for the idea that the mass accretion rate in SMGs is
168: approximately Eddington limited. Under the assumption of
169: Eddington-limited accretion, the black holes in the SMGs will grow by
170: up to an order of magnitude over their $\approx10^{8}$~yr
171: star-formation lifetime. These overall properties are consistent with
172: those expected for massive galaxies undergoing rapid black-hole
173: growth.
174:
175: The total black-hole growth from SMGs can be calculated by integrating
176: the average accretion density over the SMG redshift range; see
177: Figure~2. Over the redshift interval $z=$~1.8--3.0 (corresponding to
178: 80\% of the SMGs studied here), the black-hole density produced by
179: SMGs is $\approx6^{+11}_{-4}\times10^{3}$~$M_{\odot}$~Mpc$^{-3}$. To
180: put this quantity into context we need to compare it to the black-hole
181: growth from coeval quasar activity. Considering only quasars in the
182: luminosity range of $M_{\rm B}=$~--24 to --27 (the majority of which
183: reside in $\simgt$$M_{*}$ spheroid-dominated galaxies$^{28}$ and have
184: the properties consistent with being the progeny of SMGs$^{5}$), the
185: total SMG black-hole density is $\approx13^{+27}_{-9}$\% of the
186: black-hole growth from quasar activity$^{2}$. These constraints should
187: be considered lower limits since further pre-quasar growth at
188: $z=$~1.8--3.0 may be produced by radio-undetected SMGs and galaxies
189: with fainter submillimetre fluxes, if they host AGN activity. However, these
190: results are consistent with, for example, SMGs undergoing an intense
191: black-hole growth phase, where the black hole is grown from
192: $\approx10^{7}$ to $\approx10^{8}$~$M_{\odot}$, prior to a high
193: accretion-rate quasar phase, where the black hole is grown from
194: $\approx10^{8}$ to $\approx8\times10^{8}$ $M_{\odot}$. In this
195: scenario, although SMGs do not produce a large fraction of the
196: cosmological black-hole density, they are responsible for the crucial
197: pre-quasar phase where the black holes in massive galaxies are rapidly
198: grown. Indeed, the total black-hole growth from SMGs could only exceed
199: that from quasar activity if the quasar lifetime is insufficient to
200: double the mass of the black hole (i.e., $<3\times10^{7}$~yrs assuming
201: Eddington-limited accretion). Overall, this picture is in good
202: agreement with direct theoretical predictions of the black-hole growth
203: of SMGs and quasars$^{26,27}$.
204:
205: What was the catalyst for the rapid black hole and stellar growth seen
206: in SMGs? Rest-frame ultra-violet images taken by the {\it Hubble Space
207: Telescope} have shown that a considerably larger fraction of SMGs
208: are in major mergers (i.e., the merging of two galaxies of comparable
209: masses) than has been found in the coeval galaxy population$^{17,29}$.
210: Hydrodynamical simulations have shown that major mergers can
211: efficiently transport material toward the central regions of galaxies,
212: providing an effective mechanism to trigger nuclear star formation and
213: fuel the black hole$^{30}$. The result of these major mergers is
214: thought to be massive spheroid-dominated galaxies. Ultra-deep X-ray
215: observations of SMGs undergoing major mergers have shown that AGN
216: activity can be ongoing in both galactic components$^{12}$. Presumably
217: in these major-merger events the black holes will eventually coalesce,
218: further increasing the mass of the black hole in the resultant galaxy.
219:
220: \begin{figure}[t]
221: \centerline{\psfig{file=alexfig2.ps,width=3.5in}}
222: \caption{Cosmological black-hole accretion density. SMGs (filled
223: circle), and $M_{\rm B}=$~--24 to --27 quasars (solid line)$^{2}$ are
224: shown. Error bars in the $x$-axis direction indicate the redshift
225: range and error bars in the $y$-axis direction represent $1\sigma$
226: Poisson counting uncertainties. The shaded region indicates the
227: uncertainty in the mass accretion rate conversion for the SMGs; see
228: Table~1. The mass accretion density from SMGs is
229: $\approx13^{+27}_{-9}$\% of that produced by coeval $M_{\rm B}=$~--24
230: to --27 quasar activity over the redshift interval $z=$~1.8--3.0.}
231: \end{figure}
232:
233: %
234: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
235: % TABLE
236: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
237: %
238: \begin{table*}[!t]
239: \begin{center}
240: \normalsize
241: \caption{The Properties of the X-ray Classified AGNs}
242: \begin{tabular}{ccccccccc}
243: \hline
244: \\
245: & $S_{\rm 850\mu m}$ & &log$(L_{\rm X})$ & log($L_{\rm BOL})$ & ${dM}\over{dt}$ & $M_{\rm BH (Edd)}$ & X-ray & Observation\\
246: SMG name & (mJy) & $z$ &(erg~s$^{-1}$)&($L_{\odot}$) & ($M_{\odot}$~yr$^{-1}$) & ($10^7$~$M_{\odot}$) & Obscuration? & Mode? \\
247: \\
248: \hline
249: \\
250: 123549.4 +621536 & 8.3$\pm$2.5 & 2.20 & 44.0 & 12.9 & $0.28^{+0.57}_{-0.19}$ & $1.3^{+2.6}_{-0.9}$ & Y & T \\
251: 123555.1 +620901 & 5.4$\pm$1.9 & 1.88 & 44.4 & 13.2 & $0.80^{+1.60}_{-0.53}$ & $3.6^{+7.3}_{-2.4}$ & Y & T \\
252: 123606.7 +621550 & 4.4$\pm$1.4 & 2.42 & 43.8 & 12.6 & $0.21^{+0.42}_{-0.14}$ & $1.0^{+1.9}_{-0.6}$ & N & M \\
253: 123606.8 +621021 & 11.6$\pm$3.5 &2.51 & 43.7 & 13.1 & $0.15^{+0.31}_{-0.10}$ & $0.7^{+1.4}_{-0.5}$ & Y & T \\
254: 123616.1 +621513 & 5.8$\pm$1.1 & 2.58 & 43.7 & 13.0 & $0.16^{+0.33}_{-0.11}$ & $0.7^{+1.5}_{-0.5}$ & Y & T \\
255: 123622.6 +621629 & 7.7$\pm$1.3 & 2.47 & 44.0 & 13.1 & $0.35^{+0.70}_{-0.23}$ & $1.6^{+3.2}_{-1.1}$ & Y & T \\
256: 123629.1 +621045 & 5.0$\pm$1.3 & 1.01 & 43.2 & 12.1 & $0.05^{+0.09}_{-0.03}$ & $0.2^{+0.4}_{-0.1}$ & Y & T \\
257: 123632.6 +620800 & 5.5$\pm$1.3 & 1.99 & 43.9 & 12.9 & $0.24^{+0.47}_{-0.16}$ & $1.1^{+2.2}_{-0.7}$ & Y & M \\
258: 123635.5 +621424 & 5.5$\pm$1.4 & 2.01 & 44.0 & 12.9 & $0.30^{+0.60}_{-0.20}$ & $1.4^{+2.7}_{-0.9}$ & Y & T \\
259: 123636.7 +621156 & 7.0$\pm$2.1 & 0.56 & 42.7 & 11.1 & $0.02^{+0.03}_{-0.01}$ & $0.1^{+0.1}_{-0.1}$ & N & M \\
260: 123707.2 +621408 & 4.7$\pm$1.5 & 2.48 & 43.8 & 12.9 & $0.18^{+0.36}_{-0.12}$ & $0.8^{+1.6}_{-0.5}$ & Y & T \\
261: 123711.9 +621325 & 4.2$\pm$1.4 & 1.99 & 43.6 & 12.7 & $0.13^{+0.25}_{-0.08}$ & $0.6^{+1.2}_{-0.4}$ & Y & T \\
262: 123712.0 +621212 & 8.0$\pm$1.8 & 2.91 & 43.4 & 12.7 & $0.09^{+0.17}_{-0.06}$ & $0.4^{+0.8}_{-0.3}$ & Y & M \\
263: 123716.0 +620323 & 5.3$\pm$1.7 & 2.04 & 44.3 & 13.0 & $0.67^{+1.33}_{-0.44}$ & $3.0^{+6.1}_{-2.0}$ & Y & T \\
264: 123721.8 +621035 & 12.0$\pm$3.9 &0.98 & 43.7 & 11.7 & $0.16^{+0.31}_{-0.10}$ & $0.7^{+1.4}_{-0.5}$ & Y & M \\
265: \\
266: \hline
267: \end{tabular}
268: \end{center}
269: \vskip 2pt
270: \footnotesize
271: {\sc Notes. ---}
272: SMG name, submillimetre flux density ($S_{\rm 850\mu m}$), redshift (z), and
273: total bolometric luminosity ($L_{\rm BOL}$) are taken from ref 9. Unobscured
274: rest-frame 0.5--8.0 keV luminosity ($L_{\rm X}$), and the presence of
275: X-ray obscuration ($N_{\rm H}>10^{22}$~cm$^{-2}$) are taken from ref
276: 19. The observation mode refers to whether the source was specifically
277: targeted with a SCUBA observation (``T'') or whether the source was
278: detected in a blank-field SCUBA map (``M''): 10 of the AGNs were
279: identified in targeted SCUBA observations and 5 of the AGNs were
280: identified in blank-field SCUBA maps. The mass accretion rate (dM/dt)
281: is estimated by converting the X-ray luminosity to the AGN bolometric
282: luminosity, under the assumption that the X-ray luminosity accounts
283: for $6^{+12}_{-4}$\% of the bolometric luminosity of the AGN (i.e.,
284: the range in conversion factors found in ref 20). Although our assumed
285: bolometric conversion factors were originally derived from unobscured
286: AGNs, while the majority of the AGNs in our sample are obscured, the
287: important issue is the relationship between the intrinsic X-ray
288: emission and the ultraviolet-optical accretion disk emission, which is
289: unlikely to be obscuration dependent. We also note that our bolometric
290: conversion factors agree with those estimated for obscured AGNs with
291: similar luminosities (and therefore similar mass accretion rates) to
292: the sources studied here$^{3,21}$. When converting from the bolometric
293: luminosity to the accreted mass we assumed the canonical efficiency of
294: 10\% ($\epsilon=0.1$). The black-hole mass ($M_{\rm BH (Edd)}$) is
295: calculated from the mass accretion rate under the assumption that the
296: accretion is Eddington limited; see text for justification. The
297: black-hole masses will be higher if the accretion is sub Eddington;
298: see text for additional black-hole mass constraints.
299: \\
300: \end{table*}
301: \normalsize
302:
303: %\bibliographystyle{nature}
304: %\bibliography{J1808burst}
305: \begin{thebibliography}{10}
306:
307: \bibitem[<1>]{}
308: Tremaine, S., et al. The Slope of the Black Hole Mass versus Velocity
309: Dispersion Correlation, {\it Astrophys. J.}, {\bf 574}, 740--753 (2002).
310:
311: \bibitem[<2>]{}
312: Yu, Q. \& Tremaine, S. Observational constraints on the growth of
313: massive black holes, {\it Mon. Not. R. Astron. Soc.}, {\bf 335},
314: 965--976 (2002).
315:
316: \bibitem[<3>]{}
317: Barger, A.J., Cowie, L.L., Mushotzky, R.F., Yang, Y., Wang, W.-H.,
318: Steffen, A.T., \& Capak, P. The Cosmic Evolution of Hard X-ray Selected
319: Active Galactic Nuclei, {\it Astron. J.}, {\bf 129}, 578--609 (2005).
320:
321: \bibitem[<4>]{}
322: McLure, R.J. \& Dunlop, J.S. The cosmological evolution of quasar black
323: hole masses, {\it Mon. Not. R. Astron. Soc.}, {\bf 352}, 1390--1404 (2004).
324:
325: \bibitem[<5>]{}
326: Page, M.J., Stevens, J.A., Ivison, R.J., \& Carrera, F.J. The
327: Evolutionary Sequence of Active Galactic Nuclei and Galaxy Formation
328: Revealed, {\it Astrophys. J.}, {\bf 611}, L85--L88 (2004).
329:
330: \bibitem[<6>]{}
331: Smail, I., Ivison, R.J., Blain, A.W., \& Kneib, J.-P. The nature of
332: faint submillimetre-selected galaxies, {\it
333: Mon. Not. R. Astron. Soc.}, {\bf 331}, 495--520 (2002).
334:
335: \bibitem[<7>]{}
336: Chapman, S.C., Blain, A.W., Ivison, R.J., \& Smail, I.R. A median
337: redshift of 2.4 for galaxies bright at submillimetre wavelengths, {\it
338: Nature}, {\bf 422}, 695--698 (2003).
339:
340: \bibitem[<8>]{}
341: Hughes, D.H., et al. High-redshift star formation in the Hubble Deep
342: Field revealed by a submillimetre-wavelength survey, {\it Nature},
343: {\bf 394}, 241--247 (1998).
344:
345: \bibitem[<9>]{}
346: Chapman, S.C., Blain, A.W., Smail, I.R., \& Ivison, R.J. A redshift
347: survey of the submillimeter galaxy population, {\it Astrophys. J.},
348: {\bf 622}, 2--26 (2005).
349:
350: \bibitem[<10>]{}
351: Stevens, J.A., Page, M.J., Ivison, R.J., Smail, I., \& Carrera, F.J. A
352: Filamentary Structure of Massive Star-forming Galaxies Associated with
353: an X-Ray-absorbed QSO at $z=1.8$, {\it Astrophys. J.}, {\bf 604},
354: L17--L20 (2004).
355:
356: \bibitem[<11>]{}
357: Croom, S.M., et al. The 2dF QSO Redshift Survey -- XII. The
358: spectroscopic catalogue and luminosity function, {\it
359: Mon. Not. R. Astron. Soc.}, {\bf 349}, 1397--1418 (2004).
360:
361: \bibitem[<12>]{}
362: Alexander, D.M., et al. The Chandra Deep Field North
363: Survey. XIV. X-Ray-Detected Obscured AGNs and Starburst Galaxies in
364: the Bright Submillimeter Source Population, {\it Astron. J.}, {\bf
365: 125}, 383--397 (2003).
366:
367: \bibitem[<13>]{}
368: Ivison, R.J., et al. Spitzer observations of MAMBO galaxies: weeding
369: out active nuclei in starbursting proto-ellipticals, {\it
370: Astrophys. J. Suppl.}, {\bf 154}, 124--129 (2004).
371:
372: \bibitem[<14>]{}
373: Greve, T.R., et al. An interferometric CO survey of luminous submm
374: galaxies, {\it Mon. Not. R. Astron. Soc.}, in press; astroph/0503055
375: (2005).
376:
377: \bibitem[<15>]{}
378: Tecza, M., et al. SPIFFI Observations of the Starburst SMM
379: J14011+0252: Already Old, Fat, and Rich by $z=2.565$, {\it
380: Astrophys. J.}, {\bf 605}, L109--L112 (2004).
381:
382: \bibitem[<16>]{}
383: Swinbank, A.M., et al. The rest-frame optical spectra of SCUBA
384: galaxies, {\it Astrophys. J.}, {\bf 617}, 64--80 (2004).
385:
386: \bibitem[<17>]{}
387: Smail, I., Chapman, S.C., Blain, A.W., \& Ivison, R.J. The Rest-Frame
388: Optical Properties of SCUBA Galaxies, {\it Astrophys. J.}, {\bf 616}, 71-85
389: (2004).
390:
391: \bibitem[<18>]{}
392: Alexander, D.M., et al. The Chandra Deep Field North Survey. XIII. 2
393: Ms Point-Source Catalogs, {\it Astron. J.}, {\bf 126}, 539--574
394: (2003).
395:
396: \bibitem[<19>]{}
397: Alexander, D.M., et al. The X-ray Spectral Properties of SCUBA galaxies, {\it
398: Astrophys. J.}, submitted.
399:
400: \bibitem[<20>]{}
401: Elvis, M., et al. Atlas of quasar energy distributions, {\it
402: Astrophys. J. Suppl.}, {\bf 95}, 1--68 (1994).
403:
404: \bibitem[<21>]{}
405: Marconi, A., et al. Local supermassive black holes, relics of active
406: galactic nuclei and the X-ray background, {\it
407: Mon. Not. R. Astron. Soc.}, {\bf 351}, 169--185 (2004).
408:
409: \bibitem[<22>]{}
410: Veilleux, S., Kim, D.-C., \& Sanders, D.B. Optical Spectroscopy of the
411: IRAS 1 JY Sample of Ultraluminous Infrared Galaxies, {\it
412: Astrophys. J.}, {\bf 522}, 113--138 (1999).
413:
414: \bibitem[<23>]{}
415: Steidel, C.C., et al. A Survey of Star-forming Galaxies in the
416: $1.4<z<2.5$ Redshift Desert: Overview, {\it Astrophys. J.}, {\bf 604},
417: 534--550 (2004).
418:
419: \bibitem[<24>]{}
420: Fadda, D., et al., The AGN contribution to mid-infrared surveys. X-ray
421: counterparts of the mid-IR sources in the Lockman Hole and HDF-N, {\it
422: Astron. Astrophys.}, {\bf 383}, 838--853 (2002).
423:
424: \bibitem[<25>]{}
425: Fabian, A.C. The obscured growth of massive black holes, {\it
426: Mon. Not. R. Astron. Soc.}, {\bf 308}, L39--L43 (1999).
427:
428: \bibitem[<26>]{}
429: Archibald, E.N., et al. Coupled spheroid and black hole formation, and
430: the multifrequency detectability of active galactic nuclei and
431: submillimetre sources, {\it Mon. Not. R. Astron. Soc.}, {\bf 336},
432: 353--362 (2002).
433:
434: \bibitem[<27>]{}
435: Granato, G.L., De Zotti, G., Silva, L., Bressan, A., \& Danese, L. A
436: Physical Model for the Coevolution of QSOs and Their Spheroidal Hosts,
437: {\it Astrophys. J.}, {\bf 600}, 580--594 (2004).
438:
439: \bibitem[<28>]{}
440: Dunlop, J.S., McLure, R.J., Kukula, M.J., Baum, S.A., O' Dea, C.P., \&
441: Hughes, D.H. Quasars, their host galaxies and their central black
442: holes, {\it Mon. Not. R. Astron. Soc.}, {\bf 340}, 1095--1135 (2003).
443:
444: \bibitem[<29>]{}
445: Conselice, C.J., Chapman, S.C., \& Windhorst, R.A. Evidence for a Major
446: Merger Origin of High-Redshift Submillimeter Galaxies, {\it
447: Astrophys. J.}, {\bf 596}, L5--L8 (2003).
448:
449: \bibitem[<30>]{}
450: Springel, V., Di Matteo, T., \& Hernquist, L. Black holes in galaxy
451: mergers: the formation of red elliptical galaxies, {\it
452: Astrophys. J.}, {\bf 620}, L79--L82 (2005).
453:
454: \end{thebibliography}
455:
456:
457: \smallskip
458: \noindent {\small {\bf Acknowledgements.} We are grateful to Ross
459: McLure, Matt Page, Francesco Shankar, and Qingjuan Yu for providing
460: important data and useful scientific insight. We thank the Royal
461: Society (DMA; IS), PPARC (FEB), and NASA (SCC; WNB) for support. Data
462: presented herein were obtained using the W.M. Keck Observatory, which
463: is operated as a scientific partnership among Caltech, the University
464: of California and NASA.}
465:
466: \medskip
467: \noindent {\small {\bf Competing interests statement.} The authors
468: declare that they have no competing financial interests.}
469:
470: \medskip
471: \noindent {\small {\bf Correspondence} and requests for materials should be
472: addressed to DMA. (e-mail: dma@ast.cam.ac.uk).}
473:
474: \end{document}
475: