1: \documentclass{mn2e}
2: \usepackage{epsfig}
3:
4: \usepackage{graphicx}
5: %\usepackage{epstopdf}
6: %\DeclareGraphicsRule{.tif}{png}{.png}{`convert #1 `basename #1 .tif`.png}
7:
8: \newif\ifAMStwofonts
9: \AMStwofontstrue
10:
11:
12: \title[The K-band Hubble diagram of sub-mm galaxies and
13: hyperluminous galaxies]
14: {The K-band Hubble diagram of sub-mm galaxies and
15: hyperluminous galaxies}
16: \author[Serjeant, et al.]
17: {Stephen Serjeant$^1$, Duncan Farrah$^{2}$,
18: James Geach$^{3}$, Toshinobu Takagi$^1$,
19: \vspace*{0.3cm}\\
20: {\LARGE\rm
21: Aprajita Verma$^4$, Ali Kaviani$^{3}$,
22: Matt Fox$^{3}$}\\
23: \vspace*{0.1cm}\\
24: $^1$Centre for Astrophysics \& Planetary Science, School of Physical
25: Sciences, University of Kent, Canterbury, Kent CT2 7NR, UK\\
26: $^2$SIRTF Science Center, Jet Propulsion Laboratory, California
27: Institute of Technology, Pasadena, 91125, USA\\
28: $^3$Astrophysics Group, Imperial College London, Blackett Laboratory,
29: Prince Consort Road, London SW7 2BW\\
30: $^4$MPI fuer extraterrestrische Physik, Giessenbachstr. 1, D-85748,
31: Garching, Germany
32: }
33: \date{Received 2003}
34: \pubyear{2003}
35:
36:
37: \begin{document}
38:
39: %\input /Users/cpp/chris/IC-ASTRO-F/MP/IRCsurvey/IRCpaper/paper/july30paper/BoxedEPS.tex
40: %\SetEPSFDirectory{/Users/cpp/chris/IC-ASTRO-F/MP/IRCsurvey/IRCpaper/paper/july30paper/}
41:
42: \input BoxedEPS.tex
43: \SetEPSFDirectory{./}
44:
45: \SetRokickiEPSFSpecial
46: \HideDisplacementBoxes
47:
48: \label{firstpage}
49:
50: \maketitle
51:
52:
53: \begin{abstract}
54: We present the $K$-band Hubble diagrams ($K-z$ relations) of
55: sub-mm-selected galaxies and hyperluminous galaxies (HLIRGs).
56: We report the discovery of a remarkably tight $K-z$ relation
57: of HLIRGs,
58: indistinguishable from that of the most luminous radiogalaxies.
59: Like radiogalaxies, the HLIRG K-z relation at $z\stackrel{<}{_\sim}3$
60: is consistent with a passively evolving
61: $\sim3L_*$ instantaneous starburst starting from a redshift of $z\sim10$.
62: In contrast, many sub-mm selected galaxies are $\stackrel{>}{_\sim}2$
63: magnitudes fainter, and the population has a much larger
64: dispersion. We argue that dust obscuration and/or a larger mass range
65: may be responsible for this scatter.
66: The
67: galaxies so far proved to be hyperluminous may have been biased
68: towards higher AGN bolometric contributions than sub-mm-selected galaxies
69: due to the $60\mu$m
70: selection of some, so the location on the $K-z$ relation
71: may be related to the presence of the most massive AGN. Alternatively,
72: a particular host galaxy
73: mass range may be responsible for both extreme star formation and the
74: most massive active nuclei.
75: %The hyperluminous $K-z$ relation also makes it easier to estimate
76: %photometric
77: %redshifts for hyperluminous candidates.
78: %Our data is consistent with an evolutionary connection
79: %between the most extreme star formation and AGN activity at the sites
80: %of the largest supermassive black holes.
81: \end{abstract}
82:
83: \begin{keywords}
84: cosmology: observations -
85: galaxies: evolution -
86: galaxies:$\>$formation -
87: galaxies: star-burst -
88: infrared: galaxies -
89: submillimetre
90: \end{keywords}
91:
92: \section{Introduction}\label{sec:introduction}
93: Hyperluminous galaxies (HLIRGs, $L>10^{13}L_\odot$, as distinct from
94: the less-luminous ultraluminous population with $L=10^{12-13}L_\odot$),
95: were first identified from follow-ups of the IRAS mission
96: (e.g. Kleinmann et al. 1988, Rowan-Robinson et al. 1991).
97: Gravitational lensing was found to be responsible for some of
98: the extreme
99: luminosity of at least one HLIRG, IRAS F10214+4724
100: (Graham \& Liu 1995, Serjeant et
101: al. 1995, Broadhurst \& Lehar 1995, Eisenhardt et al. 1996), but
102: subsequent HST imaging of more HLIRGs showed no further lens
103: candidates (Farrah et al. 2002a). The morphologies were
104: found to be diverse, from interacting to quiescent.
105: Although active nuclei have been found in all HLIRGs to date, the
106: enormous gas and dust masses (e.g. Downes et al. 1993, Clements et
107: al. 1992, Farrah et al. 2002b) are indicative of violent, possibly
108: bolometrically-dominant, star formation.
109: By fitting multi-wavelength photometry of HLIRGs, several authors
110: have found comparable bolometric contributions from star formation and
111: active nuclei in many HLIRGs
112: Hyperluminous galaxies appear to be a population of
113: galaxies undergoing their major star formation episode (Rowan-Robinson
114: 2000), but at an epoch in which AGN activity is also present
115: (e.g. Rowan-Robinson 2000, Farrah et al. 2002b, Verma et
116: al. 2002).
117: The sub-mm detections of radiogalaxies (Archibald et al. 2001)
118: and quasars (e.g. Priddey et al. 2003) further
119: supported a link between violent star formation and AGN activity,
120: though quasar-heated dust has also been raised as a possibility
121: (Willott et al. 2002). In this paper we will present further evidence
122: for a link between AGN activity and extreme star formation, using the
123: $K$-band Hubble diagram.
124:
125: The tight dispersion in the
126: $K$-band Hubble diagram ($K-z$ relation) of radiogalaxies
127: has long been held to suggest a high
128: formation epoch for radiogalaxy hosts (Lilly \& Longair
129: 1984). Redshifted emission line contributions (Eales \& Rawlings 1993)
130: complicate the interpretation at redshifts $z>2$, but largely
131: only for the most
132: luminous radiogalaxies (e.g. Jarvis et al. 2001). The current
133: consensus is that the tight $\pm0.5$ magnitude
134: dispersion in the radiogalaxy $K-z$
135: relation persists at $z>2$, and is still consistent with a passively
136: evolving stellar population with a formation
137: epochs at $z>2.5$. There is also a weak correlation of $K$-band
138: luminosity with radio luminosity at any epoch (e.g. Willott et
139: al. 2003) which has been attributed to mutual correlations with
140: central nuclear black hole masses.
141: Furthermore, the host galaxies of radio-loud AGN tend to be restricted
142: to a more luminous population than their radio-quiet counterparts
143: (Dunlop et al. 2003a), suggesting that it is only the most massive
144: ($>10^9M_\odot$) nuclear black holes which give rise to radio-loud
145: AGN. Finally, the similarity of the $K$-band morphologies of
146: sub-mm-selected galaxies to those of high-$z$ radiogalaxies, the high
147: star formation rates in sub-mm galaxies (sufficient to assemble a
148: giant elliptical in $\sim10^8$ years), and the presence of radio-loud
149: AGN in local ellipticals has suggested to some authors that both
150: high-$z$ radiogalaxies and sub-mm selected galaxies are the
151: progenitors of the most massive spheroids (e.g. Dunlop 2002, Scott et
152: al. 2002).
153:
154: %K-z relation of radiogalaxies: original discovery; use as
155: %redshift estimator; dispersion at high redshift; latest news.
156: %Interpretation as high-z Magorrian.
157:
158: %Discovery of HLIRGs: IRAS F15307, IRAS F10214 and the subsequent
159: %discovery of lensing; review of what we've done ourselves; starburst
160: %dominance or AGN dominance.
161: %Suggestions of a link between hyperluminous activity and AGN by
162: %Michael, the Oxford group and the Edinburgh group.
163:
164:
165: In this paper we report the discovery of a remarkably tight $K-z$
166: relation of HLIRGs, and the surprising lack of a tight $K-z$ relation
167: for coeval sub-mm-selected galaxies.
168: Section \ref{sec:method} describes the compilation
169: of $K$-band magnitudes, and the $K-z$ relations are presented in
170: section \ref{sec:results}. Section \ref{sec:discussion} places the
171: results in the context of other high-$z$ populations, and discusses
172: the physical implications and possible applications of this relation.
173: Throughout this paper, ``quasars'' are taken to mean objects with
174: broad ($\stackrel{>}{_\sim}2000$ km s$^{-1}$) unpolarised emission
175: lines, regardless of the presence or absence of a host galaxy in
176: imaging data, and we assume $\Omega_{\rm M}=0.3$,
177: $\Omega_\Lambda=0.7$ and
178: $H_0=70$ km s$^{-1}$ Mpc$^{-1} = 100h$ km s$^{-1}$ Mpc$^{-1}$.
179: In this cosmology, a minority of the
180: hyperluminous galaxy compilation of
181: Rowan-Robinson (2000) slip just below the hyperluminous threshold, and
182: others attain hyperluminous status,
183: but for consistency with previous works we restrict ourselves to this
184: compilation. This choice does not affect the results in this paper.
185:
186: \section{Method}\label{sec:method}
187:
188: \begin{table}
189: \begin{tabular}{llllll}
190: Name & $z$ & $\log_{10}L_{\rm bol}$ & $K$ & Aper & Notes\\
191: IRAS F00235+1024 & 0.58 & 13.15 & 17.19 & $3''$ & (4)\\ %SB=AGN
192: SMM J02399-0136 & 2.803 & 13.08 & 18.79 & $3''$ & (1,7)\\ %
193: 4C41.17 & 3.8 & 13.12 & 19.6 & $4''$ & (11)\\ %
194: IRAS 09104+4109 & 0.44 & 13.24 & 15.41 & $3''$ & (4)\\ %
195: %IRASF10026+4949 & 1.12 16.85 Q Farrah
196: %IRASF10119+1429 & 1.55 16.37 Q Farrah
197: %LBQS1220+0939 & 0.68 17.54 Q Farrah
198: %IRASF14218+3854 & 1.21 17.21 Q Farrah
199: %B2 0902+343 & 3.391 & 12.89 & 19.9 & $4''$ & (6)\\ %
200: IRAS F10214+4724 & 2.286 & 13.54 & 20.10 & $\sim3''$ & (2,9)\\ %SB=AGN
201: IRAS F12514+1027 & 0.30 & 13.00 & 13.48 & $10''$ & (5)\\ %
202: SMM J14011+0252 & 2.55 & 13.18 & 18.71 & $3''$ & (3,7)\\
203: IRAS F15307+3252 & 0.93 & 13.50 & 16.59 & $2''$ & (10)\\
204: FFJ1614+3234 & 0.710 & 13.07 & 16.6 & $3''$ & (8)\\ %SB dominated
205: \end{tabular}
206: \caption{\label{tab:kmags}
207: K-band magnitudes for non-quasar hyperluminous galaxies of
208: Rowan-Robinson (2000). Following Dunlop et al. (2003) we neglect the
209: corrections between $K'$, $K$ and $K_{\rm s}$ as these are smaller
210: than the typical photometric errors ($\sim0.1-0.2$ magnitudes).
211: Bolometric luminosities are taken from Farrah et al. (2003), Farrah et
212: al. (2002b) and Rowan-Robinson (2000).
213: Notes: (1) corrected for $\times1.3$ lens
214: magnification factor; (2) corrected for emission line contribution and
215: $\times10$ magnification factor; (3) corrected for $\times2.8$ lens
216: magnification; (4) photometry from the data of
217: Farrah et al. 2003 (note that those authors used slightly different
218: photometric apertures); (5) photometry
219: from 2MASS (Cutri et al. 2003); (6) photometry from Eisenhardt \&
220: Dickinson 1992;
221: (7) photometry from
222: Smail, Ivison, Blain \& Kneib 2002b;
223: (8) photometry from Stanford et al. 2000;
224: (9) photometry
225: from Graham \& Liu 1995;
226: (10) photometry from Liu et al. 1996;
227: (11) photometry from Graham et al. 1994
228: }
229: \end{table}
230:
231: % 00235 R<22 (not on APM)
232: % 02399 R=21.2/mag = 21.48
233: % 4C41.17 R=21.7
234: % 09104 R=17.6
235: % 0902 V=22.5
236: % 10214 B=21.5/mag = 26.5
237: % 12514 R=17.34
238: % 14011 R=21.3/mag = 22.4
239: % 15307 R=19.03
240: % 1614 R=18.89
241: %
242: %
243:
244:
245:
246: \begin{table}
247: \begin{tabular}{lllll}
248: Name & $z$ & $K$ & Aper & Notes\\
249: SMMJ02399-0134 & 1.06 & 17.29 & $3''$ & (7)\\
250: SMMJ02399-0136 & 2.803 & 18.79 & $3''$ & (7)\\
251: CUDSS10A & 0.550 & 17.11 & $3''$ & (16)\\
252: LH850.11 & 2.610 & $>$20.6 & $4''$ & (13,21)\\
253: LH850.12 & 2.698 & $>$20.6 & $4''$ & (13,21)\\
254: LH850.15 & 2.429 & $>$20.4 & $4''$ & (13,21)\\
255: LH850.18 & 3.699 & $>$20.4 & $4''$ & (13,20)\\
256: HDF850.4 & 0.475 & 18.14 & $6''$ & (14,20)\\
257: HDF850.6 & 0.884 & 19.85 & $6''$ & (14,20)\\
258: HDF850.8 & 1.355 & 19.69 & total & (15,20)\\
259: SMMJ14011+0252 & 2.55 & 18.71 & $3''$ & (7)\\
260: CUDSS14F & 0.660 & 16.81 & $3''$ & (16)\\
261: N2850.1 & 0.840 & 19.48 & $4''$ & (13,20)\\
262: N2850.4 & 2.376 & 18.43 & $4''$ & (13,20)\\
263: N2850.8 & 1.189 & 18.15 & $4''$ & (13,20)\\
264: SMMJ17142+5016 & 2.39 & 21.80 & total & (12)\\
265: \hline
266: SMMJ00266+1708 & 2.0-5.0 & 23.45 & $3''$ & (7,19)\\
267: SMMJ09429+4658 & 2.1-4.5 & 20.15 & $3''$ & (7,19)\\
268: LH850.1 & 2.0-3.0 & 19.8 & $4''$ & (13,19)\\
269: LH850.3 & 1.2-3.5 & 18.86 & $4''$ & (13,19)\\
270: LH850.8 & 2.4-5.2 & 18.82 & $4''$ & (13,19)\\
271: HDF850.1 & 3.5-4.7 & 23.5 & $1''$ & (17,19)\\
272: SMMJ14099+0252 & 2.6-5.1 & 21.44 & $3''$ & (7,19)\\
273: CUDSS14.1 & 2.0-4.6 & 19.55 & $4''$ & (18,19)\\
274: \end{tabular}
275: \caption{\label{tab:kmags_scuba}
276: K-band magnitudes for sub-mm selected galaxies with spectroscopic
277: redshifts or $\geq3$ band photometric redshifts. Notes as table
278: \ref{tab:kmags}, and in addition: (12) from Smail et al. 2003 and
279: Keel et al. 2002, and corrected for emission line contributions; (13)
280: from Ivison et al. 2002; (14) Barger et
281: al. 2000; (15) Fernandez-Soto et al. 1999; (16) Lilly et al. 1999;
282: (17) from Dunlop et al. 2003, photometric aperture is small but source
283: is not significantly more extended, and no account taken of lensing;
284: (18) Gear et al. 2000; (19) Aretxaga et al. 2003; (20) Serjeant et
285: al. 2003; (21) Chapman et al. 2003. Non-sub-mm selected
286: galaxies (e.g. HR10) that are nevertheless detected in sub-mm
287: photometry are excluded from this table.
288: }
289: \end{table}
290:
291:
292: Table \ref{tab:kmags} lists the $K$ magnitudes for all non-quasar
293: hyperluminous galaxies in the compilation of Rowan-Robinson (2000),
294: except IRAS F14481+4454 and IRAS F14537+1950 for which no data is
295: available. Several quasar HLIRGs have optical host galaxy
296: measurements (e.g. Farrah et al. 2002a), but we exclude these on the
297: grounds that the $K$-band nuclear contribution will
298: differ.
299: %Nevertheless, the hyperluminous quasars IRAS F10026+4949 and
300: %IRAS F14218+3845 have nuclear components enhancing the $I$-band fluxes
301: %by $2.1$ and $1.5$ magnitudes respectively, which if applied to the
302: %$K$-band
303: %fluxes reported by Farrah et al. (2003) would place them well within
304: %the locus of the narrow-line HLIRGs discussed below.
305:
306: There are great difficulties with associating $14''$ resolution
307: sub-mm blank-field survey galaxies with
308: sources at other wavebands (e.g. Serjeant et al. 2003), especially
309: given the non-Poissonian distribution of these and other populations
310: (e.g. Almaini et al. 2003).
311: Most claimed spectroscopic redshifts
312: report spectroscopic evidence of star formation (e.g. Chapman et
313: al. 2003), but one should note that
314: there are precedents for
315: identifications of sub-mm sources changing with the advent of new data
316: (e.g. Dunlop et al. 2003b, Serjeant et al. 2003, Smail et al. 1999).
317: %The $16$ sources with published $K$
318: %magnitudes or limits are from the SCUBA 8mJy survey
319: %(Scott et al. 2002, Fox et al. 2002, Ivison et al. 2002), the
320: %cluster lens survey (Smail et al. 2002b), the signposts survey
321: %(Smail et al. 2003) and the SCUBA-HDF survey (Serjeant et al. 2003,
322: %Hughes et al. 1998).
323: %We also use the $\geq3$ band photometric redshifts of sub-mm selected
324: %galaxies of Aretxaga et
325: %al. (2003).
326: The $K$ magnitudes of all published sub-mm galaxies with spectroscopic
327: redshifts are listed in table \ref{tab:kmags_scuba}. We also include
328: galaxies with $\geq3$ band photometric redshifts, though the accuracy
329: of these redshifts will depend on how closely the high-$z$ population
330: resembles local templates (yet to be determined).
331:
332: In addition to HLIRGs and sub-mm galaxies,
333: we have compiled comparative data from the
334: literature on other, ostensibly related populations.
335: %We opt not
336: %to use photometric redshift estimates from the radio:sub-mm spectral
337: %index (Yun \& Carilli 2002, Carilli \& Yun 2000a,b, Dunne et al. 2000)
338: %because of the problems with the cross-identifications, and the
339: %additional uncertainty that the far-infrared:radio relation holds at
340: %all redshifts.
341: We use the photometry and redshifts of non-quasar Chandra sources
342: from the Hubble Deep Field North Chandra $1$ megasecond catalogue
343: of Barger et al. (2002), which is an improvement on the inhomogeneous
344: compilations available to Willott et al. (2001). We also use the
345: $K$-band photometry and spectral classifications
346: of ultraluminous infrared galaxies of Kim et al. (2002) and Veilleux
347: et al. (2002).
348: %$\gamma$-ray burst host galaxies from Le Floch
349: %et al. (2003).
350: Finally, we use the radiogalaxy photometry compilation in Willott et
351: al. (2003).
352: Because the 3CRR flux limit remains
353: close to the radio $L_*$ ($L_*$(radio)), 3CRR radio galaxies lie at
354: $\sim3-4L_*$(radio)
355: (Willott et al. 2003) at all redshifts. Note also that 3CRR sources
356: are also the highest luminosity
357: radio sources in the Hubble volume at $z<2$.
358:
359: \section{Results}\label{sec:results}
360:
361: \begin{figure}
362: \centering
363: \ForceWidth{5in}
364: \hSlide{-1.8cm}
365: \vspace*{-1cm}
366: \BoxedEPSF{kz1.ps}
367: \vspace*{-1cm}
368: \caption{\label{fig:kz1}$K$-band Hubble diagram of non-quasar
369: hyperluminous galaxies (filled circles), 3CRR radiogalaxies (open
370: triangles), 6CE radiogalaxies (open squares), and sub-mm selected
371: galaxies with spectroscopic redshifts (large open
372: circles). The
373: hyperluminous radiogalaxy 4C41.17 (also the highest redshift HLIRG) is
374: plotted with a
375: slightly smaller symbol, to distinguish it from the other HLIRGs.
376: %Radiogalaxy photometry is from the compilation in
377: %Willott et al. (2003), HLIRG
378: %photometry is given in table \ref{tab:kmags}, and sub-mm galaxy
379: %data is from
380: %Ivison et al. (2002) and Chapman et al. (2003).
381: B2 0902+343 is excluded because it
382: is not clear whether the sub-mm flux is thermal, Downes et al. 1996.
383: Note that two sub-mm selected galaxies are also hyperluminous.
384: Also plotted are: a no evolution curve (short dash), instantaneous
385: starburst at $z=10$ followed by passive stellar evolution (full line),
386: $z=5$ starburst (long dash) and $z=3$ starburst, from Willott et
387: al. (2003). Also plotted are obscured starburst models of Takagi et
388: al. (2003): age $t/t_0=2$, compactness $\Theta=0.4$ (dash-dot) and
389: $\Theta=0.5$ (dash-dot-dot-dot). All models are for a
390: $3L_*(K)$ galaxy.}
391: \end{figure}
392:
393: Figure \ref{fig:kz1} shows the $K-z$ relation of HLIRGs, compared to
394: that of bright radio sources from the 3CRR survey (Laing et al. 1983)
395: and 6CE survey
396: (Eales et al. 1997). Willott et al. (2003) quotes a fit to the
397: radiogalaxy $K-z$
398: relation of
399: \begin{equation}
400: K_{pred}=17.37 + 4.53\log_{10}(z) - 0.31(\log_{10}(z))^2 .
401: \end{equation}
402: 3CRR has dispersion around this relation of only $\pm0.5$
403: magnitudes. The HLIRGs are statistically indistinguishable from the
404: 3CRR radiogalaxies: a Kolmogorov-Smirnov test of the distributions of
405: $K-K_{pred}$ of HLIRGs and 3CRR radiogalaxies rejects the null
406: hypothesis of same distributions at only $23\%$
407: confidence.
408: \footnote{Although this test is only asymptotically
409: distribution-free, we
410: verified the test in this case with a bootstrap analysis. We randomly
411: selected $10$ 3CRR radiogalaxies, and compared $K-K_{pred}$ for this
412: subset against the remainder. By repeating this, we found that the
413: confidence levels returned
414: are well-represented by a uniform distribution on the interval $[0,1]$
415: as required.}
416: %Excluding HLIRGs which are radiogalaxies does not affect
417: %this conclusion ($23\%$ confidence of rejecting the null hypothesis).
418: (We have excluded the hyperluminous radiogalaxy 4C41.17 from this test,
419: though its magnitude is consistent with the $K-z$ relation of other
420: HLIRGs table \ref{tab:kmags}).
421:
422: Curiously
423: however, the same cannot be said of sub-mm selected galaxies, also
424: plotted in
425: figure \ref{fig:kz1}. Even assuming the $K$-band limits are obtained
426: for these galaxies, the distributions are still dissimilar at $99.98\%$
427: confidence. Figure \ref{fig:kz2} shows the HLIRG $K-z$ relation in the
428: context of other potentially
429: related populations.
430: %sub-mm selected galaxies with $\geq3$-band photometric redshifts
431: %(Aretxaga et al. 2003),
432: %%$\gamma$-ray burst host galaxies
433: %%(Le Floch et al. 2003),
434: %Chandra HDF-N $1$Ms sources (Barger et al. 2002), and
435: %ultraluminous infrared galaxies (Kim et al. 2002, Veilleux et
436: %al. 2002).
437: %%from 6C* (Jarvis et al. 2001a,b), 7CI+II (Willott et
438: %%al. 2003) and 7CIII (Lacy et al. 2000), though these lie in almost the
439: %%same region as 6CE (see Willott et al. 2003).
440: Note that the
441: photometric redshifts of three sub-mm selected galaxies and at least
442: $5$ with spectroscopic redshifts place them
443: securely away from the radiogalaxy locus.
444:
445: Neither the sub-mm galaxies nor the HLIRG samples from Rowan-Robinson
446: 2000 are spectroscopically
447: complete. Could this selection bias cause our HLIRG $K-z$ relation? In
448: a sample with a flux limit at two wavelengths, correlating the
449: luminosity at one wavelength with that at
450: the other is essentially a distance vs. distance correlation
451: (e.g. Serjeant et al. 1998). However
452: this cannot be the cause of our HLIRG $K-z$ relation because
453: (a) the apparent $K$ magnitudes
454: are not well-represented by a $K$ flux limit: the magnitude histogram of
455: table \ref{tab:kmags} is more or less uniform from $K\sim13$ to
456: $K>20$; (b) more importantly, the optical selection ($R$ or $B$)
457: shows no greater evidence for clustering around a particular apparent
458: magnitude. If Malmquist bias (distance vs. distance correlations)
459: were responsible, we would expect {\it more} clustering
460: around a particular apparent magnitude in the optical, compared to
461: the $K$-band. As discussed in Rowan-Robinson 2000,
462: there are probably
463: optically-fainter HLIRGs still to be found in the IRAS database, and
464: our results make very specific predictions for their K magnitudes.
465: However, if an as-yet-undiscovered selection effect results in the
466: undiscovered HLIRGs having fainter $K$ magnitudes, this would change
467: the results of this study.
468:
469: In both figure \ref{fig:kz1} and \ref{fig:kz2} we overplot passive
470: stellar evolution tracks of a $3L_*(K)$ galaxy
471: (where $L_*(K)$ is the K-band $L_*$) for an instantaneous starburst at
472: $z=10$, $z=5$, and $z=3$, as well as a no-evolution curve, as derived
473: by Willott et al. (2003), and two obscured starburst models of Takagi
474: et al. (2003). We adopt $M_*(K)=-24.5$ in accordance with
475: the recent determination by Huang et al. (2003). We refer to the
476: passively evolving curve starting from a $z=10$ instantaneous
477: starbust, normalised to $M(K)=-24.5$ at zero redshift, as $K_*(z)$.
478:
479: \section{Discussion}\label{sec:discussion}
480:
481: \begin{figure}
482: \centering
483: \ForceWidth{5in}
484: \hSlide{-1.8cm}
485: \vspace*{-1cm}
486: \BoxedEPSF{kz2.ps}
487: \vspace*{-1cm}
488: \caption{\label{fig:kz2}
489: $K$-band Hubble diagram for HLIRGs (filled circles; 4C41.17 has
490: smaller symbol), sub-mm galaxies
491: with spectroscopic redshifts (large open circles)
492: %the populations and models shown in figure
493: %\ref{fig:kz1}, and in addition
494: %$\gamma$-ray burst host galaxies ($+$)
495: %from Le Floch et al. (2003),
496: %sub-mm selected sources with
497: or photometric redshifts from detections in
498: $\geq3$ bands (large diamonds),
499: %from Aretxaga et al. 2003,
500: non-quasar ultraluminous infrared galaxies ($\times$),
501: %from Kim et al. 2003 and
502: %Veilleux et al. 2003,
503: and Chandra HDF-N $1$Ms non-quasar
504: sources ($+$) from Barger et al. (2002).
505: %radiogalaxies from 6C* (open downward
506: %triangles),
507: %7CI+II (open diamonds) and 7CIII (open upward triangles).
508: %Radiogalaxy photometry is from the compilation in Willott et al. (2003).
509: Chandra source photometry assumes $HK'-K=0.13+0.05(I-K)$ where
510: $I$-band photometry is available, and $HK'-K=0.3$ otherwise (Barger et
511: al. 2002).
512: For clarity we omit the
513: radiogalaxies. The same models are plotted as figure \ref{fig:kz1}.
514: }
515: \end{figure}
516:
517: The difference between the HLIRG and sub-mm galaxy $K-z$ relations is
518: all the more puzzling given our discovery that some HLIRGs would have
519: comparable sub-mm fluxes to sub-mm-selected galaxies, if redshifted to
520: $z\sim3$ (Farrah et al. 2002b), including two of the HLIRGs in the
521: present paper. Given a template spectral energy
522: distribution with a sufficiently warm colour temperature, many of the
523: sub-mm-selected galaxies could have bolometric luminosities
524: approaching $10^{13}L_\odot$.
525: Five of the ten HLIRGs in this paper were selected at
526: $60\mu$m, and two of the remainder were discovered by
527: follow-ups of radiogalaxies, either of which may have introduced a
528: bias towards high AGN bolometric
529: contributions.
530: If so, this would suggest that the position on the
531: $K-z$ relation may be related to the presence of the most massive
532: AGN. A prediction of this interpretation is that the multi-wavelength
533: data on sub-mm galaxies from SIRTF (e.g. Lonsdale et al. 2003) should
534: find an anticorrelation between $K-K_*$ magnitudes and the bolometric
535: contributions in the mid-infrared.
536:
537:
538: Alternatively, the HLIRG $K-z$ relation may be intrinsic, rather than
539: the product of a subtle AGN bias.
540: All of the galaxies currently {\it proved} to be hyperluminous
541: currently lie on a tight $K-z$ relation, including those selected in
542: the sub-mm.
543: Also, radiative transfer models of the Rowan-Robinson (2000) HLIRGs do
544: not find the AGN to be bolometrically dominant in all cases
545: (e.g. Verma et al. 2002, Farrah et al. 2002b).
546: There is
547: no prima facie reason to suppose that HLIRGs should necessarily follow
548: the radiogalaxy $K-z$ relation.
549: For example, the
550: obscured high-redshift AGN detected by Chandra are not found with similar
551: $K$ magnitudes to radiogalaxies (or HLIRGs, figure \ref{fig:kz2}),
552: which Willott et al. (2001)
553: argued was due to the Chandra sources hosting smaller mass nuclear black
554: holes.
555: However, the relative number densities
556: of radiogalaxies and HLIRGs lend plausibility to a physical link
557: between the populations, or at least a common progenitor.
558: Rowan-Robinson (2000) lists $16$ HLIRGs with $z\leq1$, implying a
559: lower limit to the $z\leq1$ HLIRG number density of
560: $\geq3\times10^{-10}h^3$Mpc$^{-3}$, and also
561: estimates
562: that only $\sim10-20\%$ of HLIRGs have been identified to date.
563: These number densities are comparable to the
564: $z\leq1$ number
565: density of 3CRR radio galaxies ($1.1\times10^{-9}h^3$Mpc$^{-3}$) and
566: significant compared to $z\leq1$ 3CRR active galaxies as a whole
567: ($2.5\times10^{-9}h^3$Mpc$^{-3}$). Notably, both HLIRGs and 3CRR are
568: the most luminous in the Hubble volume of their class.
569:
570: This does not necessarily imply that radiogalaxies and HLIRGs must be
571: the most massive galactic systems in the Hubble volume. Huang et
572: al. (2003) measured the space density of $10L_*(K)$ galaxies
573: to be
574: $2.1\times10^{-8}h^3$Mpc$^{-3}$ at $z<0.4$.
575: These galaxies are $2.5$ magnitudes brighter than the HLIRG and
576: radiogalaxy host galaxies, and $\sim10\times$ more numerous than
577: radio-loud AGN. However, the space density of $>10L_*(K)$ galaxies has
578: not been determined at higher redshifts.
579:
580: There is
581: evidence in figure \ref{fig:kz3} that infrared luminosity scales with
582: host luminosity, in a manner reminiscent of the trend of host
583: luminosity with radio lobe luminosity in radiogalaxies (Willott et
584: al. 2003). The difference in the mean $K-K_*$ for the
585: $1-3\times10^{12}L_\odot$ and $>10^{13}L_\odot$ bins is significant
586: at $99.6\%$ confidence using Student's T statistic, though the objects
587: in these bins span
588: different redshifts, so differential evolution may also be a factor
589: (e.g. Serjeant et al. 1998). Samples of infrared-luminous galaxies
590: spanning a narrow range in redshift, but a wide range in luminosity,
591: are needed to test this relationship definitively. Such samples may
592: become available with the advent of SIRTF.
593:
594: If radiogalaxy and HLIRG activity is both short lived and rare (in the
595: sense that a $5L_*(K)$ galaxy has a low probability of ever hosting a
596: radiogalaxy or HLIRG), then the lack of $>10L_*(K)$ HLIRGs might be
597: explained by the finite size of the volume surveyed so far for
598: HLIRGs.
599: If we assume that $K$-band galaxies in the range $0.5-5L_*(K)$ are the
600: hosts of short-lived HLIRG activity with $10^{13-14}L_\odot$, and we
601: interpolate from the
602: $K$-band luminosity function (Huang et al. 2003)
603: keeping the HLIRG duty cycle constant,
604: then the space density of $10^{14-15}L_\odot$ galaxies should be
605: around a factor of $\sim1000$ lower.
606: %``\"{U}ber-luminous''
607: %galaxies with luminosities $>10^{15}L_\odot$ in
608: %the Hubble volume.
609: This is, of course, only a toy model: the $K$-band
610: luminosity function is unlikely to
611: keep the same shape at all redshifts, the luminosities may not scale
612: with the host galaxy masses, and the duty cycle assumptions
613: may be ill-founded, but this does raise the interesting question of
614: the existence of still more extreme
615: %%``\"{U}ber-luminous''
616: populations of infrared galaxies.
617: Source count models differ widely in their predictions at these
618: luminosities (Pearson 2001, Rowan-Robinson 2001).
619: Whether such systems do in fact exist may be
620: testable with the next generation of sub-mm/mm-wave survey
621: facilities, such as SCUBA-2 (Holland et al. 2003) or the LMT (Baars et
622: al. 1998).
623:
624: %The only population of objects currently known to be more luminous
625: %than HLIRGs is $\gamma$-ray bursts.
626: %Figure \ref{fig:kz2} shows that
627: %the host galaxies of $\gamma$-ray
628: %bursts are under-luminous, as reported by Le Floch et al. (2003),
629: %rather than being found in the most massive
630: %(and potentially the most violently star forming) galaxies, or even in
631: %$\sim L_*(K)$ systems which dominate the comoving stellar mass.
632:
633: In short, there are plausible precedents for
634: abundant populations of galaxies with evidence of dust-enshrouded AGN,
635: extreme luminosities and/or large stellar masses, which are
636: many times more luminous in $K$, or
637: many times less luminous, than HLIRGs.
638: The fact that these populations
639: do not host HLIRG activity suggests
640: that the similarity of the HLIRG and radiogalaxy
641: $K-z$ relations is due to a direct physical link between the two
642: phenomena, such as an evolutionary connection, or a common progenitor
643: population.
644: Alternatively, the position of HLIRGs on the $K-z$ relation may be
645: solely related to the presence of the most massive AGN (see above),
646: for which a key test is whether the SIRTF detects hyperluminous
647: AGN activity in sub-mm selected galaxies.
648:
649:
650: \begin{figure}
651: \centering
652: \ForceWidth{5in}
653: \hSlide{-1.8cm}
654: \vspace*{-2cm}
655: \BoxedEPSF{kz3.ps}
656: \vspace*{-1cm}
657: \caption{\label{fig:kz3}
658: $K$-band luminosity relative to $L*$ ($z=10$ starburst model), plotted
659: against bolometric
660: luminosity. HLIRGs: filled circles; ultraluminous galaxies: crosses.
661: %Hyperluminous galaxies from Rowan-Robinson (2000) are
662: %plotted as filled circles, and ultraluminous galaxies from Kim et
663: %al. (2002) and Veilleux et al. (2002) are plotted as crosses.
664: Also plotted are the mean $K-K_*$ values in binned luminosity ranges
665: (horizontal lines).
666: The standard deviation is $\sim\pm0.6$ magnitudes in each bin, and the
667: errors on the means are plotted as vertical lines.
668: Note that for our cosmology, some of the Rowan-Robinson (2000)
669: compilation slip just below the HLIRG threshold, and one of
670: the ultraluminous galaxies becomes HLIRGs. For consistency with
671: previous work, we restrict our HLIRGs to the Rowan-Robinson (2000)
672: compilation, but as can be seen from this figure this choice does not
673: affect our conclusions.
674: }
675: \end{figure}
676:
677: Surprisingly, sub-mm galaxies have a very different distribution in
678: the $K-z$ plane.
679: %The evidence for their divergence from the radiogalaxy $K-z$
680: %relation is currently based on only two outlier galaxies (or more if
681: %the $K$-band limits are not obtained).
682: This is contrary to the result of Dunlop (2002)
683: due to the subsequent increase in (ostensibly) reliable spectroscopic
684: redshifts.
685: %used radio:sub-mm photometric redshifts to
686: %place sub-mm sources on the $K-z$ relation, and found a good agreement
687: %with radiogalaxies. His sample is similar to ours, except that we
688: %have excluded LH850.1, CUDSS14A and ELAIS N2850.2 as these lack
689: %spectroscopic redshifts, and we have the benefit of new spectroscopic
690: %redshifts since the Dunlop (2002) study (Chapman et al. 2003).
691: The spectroscopic redshift of ELAIS N2850.1
692: is anomalous compared to its radio:sub-mm ratio, which led Chapman et
693: al. (2002) to suggest that the system may be lensed. Placing the
694: system at higher redshift may well restore ELAIS N2850.1 to closer to
695: the locus of the radiogalaxy $K-z$ relation.
696: Nevertheless, several photometric or spectroscopic
697: redshifts place sub-mm
698: galaxies away from
699: the radiogalaxy $K-z$ relation (figure \ref{fig:kz2}).
700: %Aretxaga et al. (2003)
701: %lists photometric redshifts of sub-mm selected galaxies based on
702: %detections in at least $3$ passbands. Eight of these still lack a
703: %spectroscopic redshift, but of these three are clear outliers
704: %on the $K-z$ plane on the basis of their photometric redshifts
705: %(HDF850.1, SMMJ00266+1708, SMMJ14099+0252).
706: Plausibly, these may represent separate populations; there is no
707: reason to suppose that sub-mm selected galaxies represent a single
708: homogeneous population of objects (e.g. Dannerbauer et al. 2002), as
709: with Extremely Red Objects (e.g. Smail et al. 2002a).
710: %Our
711: %data in figure \ref{fig:kz2} is consistent with a heavily obscured
712: %sub-population.
713:
714: One possibility is that some sub-mm galaxies are less massive systems;
715: another is that
716: not all of them
717: are the progenitors of the most massive ellipticals
718: (Efstathiou \& Rowan-Robinson 2003,
719: Kaviani et al. 2003) but rather are cool cirrus-dominated objects.
720: Alternatively, some sub-mm galaxies may be heavily extinguished even
721: in the observed-frame $K$-band.
722: Such an interpretation is physically plausible: figures \ref{fig:kz1}
723: and \ref{fig:kz2}
724: show the predictions of such a model from
725: Takagi et al. 2003.
726:
727: %The similarity of the $K$-band morphologies of at least
728: %some sub-mm selected galaxies and radiogalaxies has been argued to
729: %suggest a common
730: %connection with the assembly of the most massive ellipticals (Lutz et
731: %al. 2001), as has the similarity of the comoving number densities of
732: %sub-mm selected galaxies, extremely red objects at $z\sim1.5$, and
733: %present-day $2-3L_*$ ellipticals (Scott et al. 2002). The clustering
734: %of sub-mm galaxies will be a powerful test of this interpretation (van
735: %Kampen et al. in prep.) and others
736: %(Efstathiou \& Rowan-Robinson 2003,
737: %Kaviani et al. 2003).
738:
739: %Our HLIRG
740: %data is also consistent with models in which the dust responsible
741: %for the far-infrared emission in HLIRGs is almost entirely
742: %heated by the active
743: %nucleus (Willott et al. 2002). While radiative transfer models
744: %of our HLIRGs typically suggest comparable contributions from
745: %starburst and active nucleus (e.g. Rowan-Robinson 2000), at
746: %least half the sample is starburst-dominated and only two have AGN
747: %contributions which models imply exceed the starburst (IRAS 09104+4109
748: %and IRAS F10214+4724, though differential lensing may play a role in the
749: %latter, Lacy et al. 1998). Another problem of quasar-heated dust
750: %models in HLIRGs is
751: %that every HLIRG so far
752: %has shown some evidence of an active nucleus, but
753: %not every radio-loud AGN
754: %has hyperluminous far-infrared emission. This might be explained by
755: %variations in the dust content and distribution in active nuclei, and
756: %in the central black hole masses and Eddington ratios, but
757: %at the cost of a natural explanation for the similarity of HLIRG and
758: %radiogalaxy $K-z$ relations. Notably, at least $50\%$ of
759: %normal galaxies also show evidence of active nuclei (Ho et al. 1997),
760: %without requiring these nuclei to be bolometrically dominant.
761:
762: The HLIRG $K-z$ relation could be used to estimate redshifts of
763: HLIRG candidates, as was the early
764: practice for radiogalaxies (e.g. Dunlop \&
765: Peacock 1990). Based on the 3CRR radiogalaxy $K-z$ relation, the $K$
766: magnitude is sufficient to determine the redshift of HLIRGs
767: to better than $\pm10\%$ in $(1+z)$ at all redshifts, provided the
768: systems are indeed hyperluminous and are not quasars. Regarding
769: the hyperluminous quasars in Farrah et al.
770: (2003), we can predict that the
771: hyperluminous quasar LBQS 1220+0939 should be
772: dominated by the host galaxy
773: flux in $K$, while the hyperluminous quasar
774: IRAS F10119+1429 should be
775: dominated by the nuclear component in $K$. The remaining cases in
776: Farrah et al. (2003) should be intermediate between these two cases.
777:
778: Only two of the known non-quasar HLIRGs lack $K$-band photometry, so
779: further tests of the HLIRG $K-z$ relation using HLIRGs discovered to
780: date must rely on
781: sub-arcsecond near-infrared imaging of hyperluminous quasar hosts.
782: Alternatively, both SIRTF and ASTRO-F will have sensitive $L$ and $M$-band
783: cameras, which can further reduce the contribution from the active
784: nucleus by sampling closer to the rest-frame $K$-band.
785:
786: \section{Acknowledgements}
787: We would like to thank Chris Willott for kindly supplying the
788: stellar evolution curves, and the anonymous referee for very
789: helpful
790: suggestions, including the possibility that an AGN bias in the
791: Rowan-Robinson (2000) sample may be related to their position on the
792: $K-z$ relation.
793: This research was supported by PPARC grant PPA/G/0/2001/00116, and
794: Nuffield Foundation grant NAL/00529/G.
795: This publication makes use of data products from the Two Micron All
796: Sky Survey, which is a joint project of the University of
797: Massachusetts and the Infrared Processing and Analysis
798: Center/California Institute of Technology, funded by the National
799: Aeronautics and Space Administration and the National Science
800: Foundation. This research made use of the NASA/IPAC Extragalactic
801: Database (NED) which is operated by the Jet Propulsion Laboratory,
802: California Institute of Technology, under contract with the National
803: Aeronautics and Space Administration.
804:
805: \begin{thebibliography}{}
806:
807: \bibitem{} Almaini, A., et al., 2003, MNRAS 338, 303
808: \bibitem{} Aretxaga, I., Hughes, D.H., Chapin, EL., Gazta\~{n}aga, E.,
809: Dunlop, J.S., Ivison, R.J., 2003, MNRAS 342, 759
810: \bibitem{} Archibald, E.N., Dunlop, J.S., Hughes, D.H., Rawlings, S.,
811: Eales, S.A., Ivison, R.J., 2001, MNRAS 323, 417
812: \bibitem{}
813: Baars, J.W.M., Carrasco, L., Schloerb, F.P., 1998, Bulletin of the
814: American Astronomical Society, vol. 31, p. 840
815: %\bibitem[\protect\citename{Dickinson et al. }2001]{dickinson01}
816: %Dickinson M., et al., 2001, BAAS, 198, 2501
817: %\bibitem{} Elbaz, D., et al., 1999, A\&A 351, L37
818: \bibitem{} Barger, A.J., Cowie, L.L., Richards, E.A., 2000, AJ 119, 2092
819: \bibitem{} Barger, A., Cowie, L.L., Bradt, W.N., Capak, P., Garmire,
820: G.P., Hornschemeier, A.E., Steffen, A.T., Wehner, E.H., 2002, AJ 124, 1839
821: \bibitem{} Broadhurst, T., Lehar, J., 1995, ApJ 450, L41
822: %\bibitem{} Carilli, C.L., \& Yun, M.S., 2000a, ApJ 530, 618
823: %\bibitem{} Carilli, C.L., \& Yun, M.S., 2000b, ApJ 539, 1024
824: \bibitem{} Chapman, S.C., Smail, I., Ivison, R.J., Blain, A.W., 2002,
825: MNRAS 335, L17
826: %\bibitem{} Chapman, S.C., Smail, I., Ivison, R.J., Helou, G., Dale,
827: %D.A., Lagache, G., 2002, ApJ 573, 66 %%% FIRBACK
828: \bibitem{} Chapman, S.C., Blain, A.W., Ivison, R.J., Smail, I.R.,
829: 2003, Nature 422, 695
830: \bibitem{} Clements, D.L., Rowan-Robinson, M., Lawrence, A.,
831: Broadhurst, T., McMahon, R., 1992, MNRAS 256, 35P
832: \bibitem{} Cutri, R.M., et al., 2003, Explanatory Supplement to the
833: 2MASS All Sky Data Release, IPAC
834: \bibitem{} Dannerbauer, H., Lehnert, M.D., Lutz, D., Tacconi, L.,
835: Bertoldi, F., Carilli, C., Genzel, R., Menten, K., 2002, ApJ 573, 473
836: \bibitem{} Downes, D., Solomon, P.M., Radford, S.J.E., 1993, ApJ 414, L13
837: \bibitem{} Downes D., Solomon P. M., Sanders D. B., Evans A. S., 1996,
838: A\&A, 313, 91
839: \bibitem{}
840: Dunlop, J.S., 2002, in
841: %{\it A New Era In Cosmology},
842: ASP conference
843: proceedings vol. 283, eds. T. Shanks and N. Metcalfe, San Fransisco:
844: Astronomical Society of the Pacific (astro-ph/0203183)
845: \bibitem{} Dunlop, J.S., \& Peacock, J.A., 1990, MNRAS 247, 19
846: \bibitem{} Dunlop, J.S., McLure, R.J., Kukula, M.J., Baum, S.A.,
847: O'Dea, C.P., Hughes, D.H., 2003a, MNRAS 340, 1095
848: \bibitem{} Dunlop, J.S., et al., 2003b, MNRAS submitted (astro-ph/0205480)
849: \bibitem{} Dunne, L., Clements, D.L., Eales S.A., 2000, MNRAS 319, 813
850: \bibitem{} Eales, S.A., \& Rawlings, S., 1993, ApJ 411, 67
851: \bibitem{} Eales, S.A., Rawlings, S., Law-Green, J.D.B., Cotter, G.,
852: Lacy, M., 1997, MNRAS 291, 593
853: \bibitem{} Efstathiou, A., \& Rowan-Robinson, M., 2003, MNRAS 343, 322
854: \bibitem{} Eisenhardt, P., \& Dickinson, M., 1992, ApJ 399, L47
855: \bibitem{} Eisenhardt, P., Armus, L., Hogg, D.W., Soifer, B.T.,
856: Neugebauer, G., Werner, M.W., 1996, ApJ 461, 72
857: \bibitem{}
858: Farrah, D., Verma, A., Oliver, S., Rowan-Robinson, M., McMahon, R.,
859: 2002a, MNRAS 329, 605
860: \bibitem{} Farrah, D., Serjeant, S., Efstathiou, A., Rowan-Robinson,
861: M., Verma, A., 2002b, MNRAS 335, 1163
862: %\bibitem{} Farrah, D., Afonso, J., Efstathiou, A., Rowan-Robinson, M.,
863: %Fox, M., Clements, D., 2003a, MNRAS 343, 585
864: \bibitem{}
865: Farrah, D., Geach, J., Fox, M., Serjeant, S., Oliver, S., Verma, A.,
866: Kaviani, A., Rowan-Robinson, M., 2003, MNRAS submitted
867: \bibitem{} Fernandez-Soto, A., Lanzetta, K.M., Yahil, A., 1999, ApJ
868: 513, 34
869: \bibitem{} Fox, M.J., et al., 2002, MNRAS 331, 817
870: \bibitem{} Gear, W.K., Lilly, S.J., Stevens, J.A., Clements, D.L.,
871: Webb, T.M., Eales, S.A., Dunne, L., 2003, MNRAS submitted
872: (astro-ph/0007054)
873: %\bibitem{} Graham, J.R., \& Dey, A., 1996, ApJ 471, 720
874: \bibitem{} Graham, J.R., \& Liu, M.C., 1995, ApJ 449, L29
875: \bibitem{} Graham, J.R., et al., 1994, ApJ, 420, L5
876: \bibitem{}
877: Ho, L.C., Filippenko, A.V., Sargent, W.L.W., 1997, ApJS, 112, 315
878: \bibitem{}
879: Holland, W.S., Duncan, W., Kelly, B.D., Irwin, K.D., Walton, A.J.,
880: Ade, P.A.R., Robson, E.I., 2003, in Millimeter and Submillimeter
881: Detectors for Astronomy, eds. T.G. Phillips and J. Zmuidzinas,
882: Proceedings of the SPIE, vol. 4855, pp. 1-18
883: \bibitem{} Huang, J.S., Glazebrook, K., Cowie, L.L., Tinney, C., 2003,
884: ApJ 584, 203
885: \bibitem{} Hughes, D.H., et al., 1998, Nature 394, 241
886: %\bibitem{} Ivison, R.J., Smail, I., Barger, A.J., Kneib, J.-P., Blain,
887: %A.W., Owen, F.N., Kerr, T.H., Cowie, L.L., 2000, MNRAS 315, 209
888: \bibitem{} Ivison, R.J., et al., 2002, MNRAS 337, 1
889: \bibitem{} Jarvis, M.J., Rawlings, S., Eales, S.A., Blundell, K.M.,
890: Croft, S., McLure, R.J., Willot, C.J., 2001, MNRAS 326, 1585
891: %\bibitem{} Jarvis, M.J., et al., 2001b, MNRAS 326, 1563
892: \bibitem{} Kaviani, A., Haehnelt, M.G., Kauffmann, G., 2003, MNRAS
893: 340, 739
894: \bibitem{} Keel, W.C., Wentao, W., Waddington, I., Windhorst, I.,
895: Pascarelle, S.M., 2002, AJ 123, 3041
896: \bibitem{} Kim, D.-C., Veilleux, S., Sanders, D.B., 2002, ApJS 143, 277
897: \bibitem{} Kleinmann, S.G., Hamilton, D., Keel, W.C., Wynn-Williams,
898: C.G., Eales, S.A., Becklin, E.E., Kuntz, K.D., 1988, ApJ 328, 161
899: \bibitem{} Lacy, M., Rawlings, S., Serjeant, S., 1998, MNRAS 299, 1220
900: %\bibitem{} Lacy, M., Bunker, A.J., Ridgway, S.E., 2000, AJ 120, 68
901: \bibitem{} Laing, R.A., Riley, J.M., Longair, M.S., 1983, MNRAS 204,
902: 151
903: %\bibitem{} Le Floch, E., et al 2003, A\&A 400, 499
904: \bibitem{} Lilly, S.J., Longair, M.S., 1984, MNRAS 211, 833
905: \bibitem{} Lilly, S.J., Eales, S.A., Gear, W.K.P., Hammer, F., Le
906: F\'{e}vre, O., Crampton, D., Bond, J.R., Dunne, L., 1999 ApJ 518, 641
907: \bibitem{} Liu, M.C., Graham, J.R., Wright, G.S., 1996, ApJ 470, 771
908: %\bibitem{} Lutz, D., et al., 2001, A\&A 378, L70
909: \bibitem{} Lonsdale, C., et al., 2003, PASP 115, 897
910: \bibitem[\protect\citename{Pearson }2001]{cpp01a}Pearson, C.P., 2001,
911: MNRAS 325, 1511
912: \bibitem{} Priddey, R.S., Isaak, K.G., McMahon, R.G., Omont, A., 2003,
913: MNRAS 339, 1183
914: \bibitem{} Rowan-Robinson, M., et al., 1991, Nature 351, 719
915: \bibitem{} Rowan-Robinson, M., 2000, MNRAS 316, 885
916: \bibitem{} Rowan-Robinson, M., 2001, ApJ 549, 745
917: \bibitem{} Scott, S., et al., 2002 MNRAS 331, 817
918: \bibitem{} Serjeant, S., Lacy, M., Rawlings, S., King, L.J., Clements,
919: D.L., 1995, MNRAS 276, L31
920: \bibitem{} Serjeant, S., Rawlings, S., Lacy, M., Maddox, S.J., Baker,
921: J.C., Clements, D., Lilje, P., 1998, MNRAS 294, 494
922: \bibitem{} Serjeant, S., et al., 2003, MNRAS in press
923: \bibitem{} Smail, I., Ivison, R.J., Kneib, J.-P., Cowie, L.L., Blain,
924: A.W., Barger, A.J., Owen, F.N., Morrison, G., 1999, MNRAS 308, 1061
925: \bibitem{} Smail, I., Owen, F., Morrison, G.E., Keel, W.C., Ivison,
926: R.J., Ledlow, M.J., 2002a, ApJ 581, 844
927: \bibitem{} Smail, I., Ivison, R.J., Blain, A.W., Kneib, J.-P.,
928: 2002b, MNRAS 331, 495
929: \bibitem{} Smail, I., Ivison, R.J., Gilbank, D.G., Dunlop, J.S., Keel,
930: W.C., Motohara, K., Stevens, J.A., 2002, ApJ 583, 551
931: \bibitem{} Stanford, S.A., Stern, D., van Breugel, W., De Breuk, C.,
932: 2000, ApJS 131, 185
933: \bibitem{} Takagi, T., Arimoto, N., Hanami, H., 2003, MNRAS 340, 813
934: \bibitem{} Veilleux, S., Kim, D.-C., Sanders, D.B., 2002, ApJS 143, 315
935: \bibitem{} Verma, A., Rowan-Robinson, M., McMahon, R., Efstathiou, A.,
936: 2002, MNRAS 335, 574
937: \bibitem{}
938: Willott, C.J., et al., 2001, in Recent results of XMM-Newton and
939: Chandra, XXXVIth Recontres de Moriond, XXIst
940: Moriond Astrophysics Meeting, eds. D.M. Neumann \& J.T.T. Van.
941: (astro-ph/0105560)
942: \bibitem{}
943: Willott, C.J., Rawlings, S., Archibald, E.N., Dunlop, J.S., 2002,
944: MNRAS 331, 435
945: \bibitem{}
946: Willott, C.J., Rawlings, S., Jarvis, M.J., Blundell, K.M., 2003, MNRAS
947: 339, 173
948:
949: \end{thebibliography}
950:
951:
952:
953: \end{document}
954:
955:
956:
957:
958: