astro-ph0608446/ms.tex
1: \documentclass[aaspp4,11pt,preprint,apjfonts]{emulateapj}
2: %\documentclass[12pt,preprint]{aastex}
3: 
4: \slugcomment{Accepted for publication in ApJL}
5: \shorttitle{Suppressed Star Formation in $z\sim2.3$ Galaxies}
6: \shortauthors{Kriek et al.}
7: 
8: \newcommand{\ha}{H$\alpha$}
9: \newcommand{\wha}{$W_{\rm H\alpha}$}
10: 
11: \begin{document}
12: 
13: \title{Spectroscopic Identification of Massive Galaxies at $z\sim2.3$
14:   with Strongly Suppressed Star Formation \altaffilmark{1}}
15: 
16: \author{Mariska Kriek\altaffilmark{2,3},
17: Pieter G. van Dokkum\altaffilmark{3,4}, 
18: Marijn Franx\altaffilmark{2},  
19: Ryan Quadri\altaffilmark{3}, 
20: Eric Gawiser\altaffilmark{3,4,5},
21: David Herrera\altaffilmark{3,4},
22: Garth D. Illingworth\altaffilmark{6}, 
23: Ivo Labb\'e\altaffilmark{7},
24: Paulina Lira\altaffilmark{8},
25: Danilo Marchesini\altaffilmark{3}, 
26: Hans-Walter Rix\altaffilmark{9},
27: Gregory Rudnick\altaffilmark{10},
28: Edward N. Taylor\altaffilmark{2},
29: Sune Toft\altaffilmark{11},
30: C. Megan Urry\altaffilmark{4},
31: and Stijn Wuyts\altaffilmark{2}}
32: 
33: %\email{mariska@strw.leidenuniv.nl}
34: 
35: \altaffiltext{1}{Based on observations obtained at the Gemini
36:   Observatory, which is operated by the Association of Universities for
37:   Research in Astronomy, Inc., under a cooperative agreement with the
38:   NSF on behalf of the Gemini partnership.}
39: 
40: \altaffiltext{2}{Leiden Observatory, PO Box 9513, 2300 RA Leiden,
41:   Netherlands}
42: 
43: \altaffiltext{3}{Department of Astronomy, Yale University, New Haven,
44:   CT 06520}
45: 
46: \altaffiltext{4}{Yale Center for Astronomy and Astrophysics, Yale
47:   University, New Haven, CT 06520}
48: 
49: \altaffiltext{5}{NSF Astronomy and Astrophysics Postdoctoral Fellow}
50: 
51: \altaffiltext{6}{UCO/Lick Observatory, University of California, Santa
52:   Cruz, CA 95064}
53: 
54: \altaffiltext{7}{Carnegie Fellow, Carnegie Observatories, 813 Santa
55:   Barbara Street, Pasadena, CA 91101}
56: 
57: \altaffiltext{8}{Departamento de Astronom{\'i}a, Universidad de Chile, 
58:   Casilla 36-D, Santiago, Chile}
59: 
60: \altaffiltext{9}{Max-Planck-Institute f\"ur Astronomie, K\"onigstuhl
61:   17, Heidelberg, Germany}
62: 
63: \altaffiltext{10}{Goldberg Fellow, National Optical Astronomy Observatory, 
64:   950 North Cherry Avenue, Tucson, AZ 85719}
65:  
66: \altaffiltext{11}{European Southern Observatory,
67:   Karl-Schwarzschild-Str. 2, 85748 Garching bei M\"unchen, Germany}
68: 
69: 
70: \begin{abstract} 
71:   We present first results of a spectroscopic survey targeting
72:   $K$-selected galaxies at $z=2.0-2.7$ using the GNIRS instrument on
73:   Gemini-South. We obtained near-infrared spectra with a wavelength
74:   coverage of 1.0--2.5~$\rm\mu$m for 26 $K$-bright galaxies ($K<19.7$)
75:   selected from the MUSYC survey using photometric redshifts. We
76:   successfully derived spectroscopic redshifts for all 26 galaxies
77:   using rest-frame optical emission lines or the redshifted
78:   Balmer/4000\AA\ break. Twenty galaxies have spectroscopic redshifts
79:   in the range $2.0<z<2.7$, for which bright emission lines like \ha\
80:   and [O\,{\sc iii}] fall in atmospheric windows. Surprisingly, we
81:   detected no emission lines for nine of these 20 galaxies. The median
82:   $2\sigma$ upper limit on the rest-frame equivalent width of \ha\ for
83:   these nine galaxies is $\sim$10\,\AA. The stellar continuum emission
84:   of these same nine galaxies is best fitted by evolved stellar
85:   population models. The best-fit star formation rate (SFR) is zero
86:   for five out of nine galaxies, and consistent with zero within
87:   $1\sigma$ for the remaining four.  Thus, both the \ha\ measurements
88:   and the independent stellar continuum modeling imply that 45\% of
89:   our $K$-selected galaxies are not forming stars intensely. This high
90:   fraction of galaxies without detected line emission and low SFRs may
91:   imply that the suppression of star formation in massive galaxies
92:   occurs at higher redshift than is predicted by current CDM galaxy
93:   formation models. However, obscured star formation may have been
94:   missed, and deep mid-infrared imaging is needed to clarify this
95:   situation.
96: \end{abstract}
97: 
98: \keywords{galaxies: high redshift --- galaxies: evolution --- 
99:   galaxies: formation}
100: 
101: \begin{figure*}
102:   \epsscale{1.15}
103:   %\epsscale{1.}
104:   \plotone{fig1.eps}
105:   \caption{NIR spectra ({\it purple}) and optical-to-NIR photometry
106:     ({\it blue}) of the nine galaxies at $2.0<z<2.7$ for which we
107:     detected no \ha\ emission. The upper panels show the 1D original
108:     spectrum. The `low resolution' binned spectra (160 \AA\ per bin)
109:     and the photometry are presented in the lower panels. All fluxes
110:     are given in $10^{-19} \rm ergs~s^{-1}~cm^{-2}~\AA^{-1}$. Regions
111:     with low or variable atmospheric transmission or with strong sky
112:     line emission are indicated in light gray. The best fit to the
113:     optical photometry and low resolution spectrum, allowing all
114:     values for $A_V$, is overplotted in dark gray. The best-fit model
115:     parameters are printed in each panel. The corresponding errors are
116:     the 68\% confidence intervals, derived from 200 Monte Carlo
117:     simulations. The $\chi^2$ values are given per degree of freedom
118:     ($N_{\rm deg}=28$). Models without dust generally provide good
119:     fits as well ({\it light gray}), and imply ages that are about a
120:     factor of 2 higher than the listed ages. All galaxies without
121:     detected \ha\ emission are best fitted by evolved stellar
122:     population models with low specific SFRs.
123:     \label{spectra}}
124: \end{figure*}
125: 
126: \begin{figure}
127:   \begin{center}
128:     \epsscale{1.15}
129:     %\epsscale{0.6}
130:     \plotone{fig2.eps}
131:     \caption{The equivalent width of \ha\ vs. the specific SFR derived
132:       from our model fits to the spectra. Filled red circles are
133:       galaxies with no detected \ha\ emission and purple squares are
134:       emission line galaxies in our sample. Blue crosses are
135:       UV-selected galaxies by \cite{er06a,er06b,er06c}. DRGs are
136:       indicated by open green diamonds. Upper limits for \ha\ and the
137:       specific SFR are 2$\sigma$. Expected relations between \wha\ and
138:       specific SFR are derived from the \cite{bc03} and \cite{ke98}
139:       models and drawn for a $\tau_{\rm 100 Myr}$ ({\it gray dotted
140:         line}) and a $\tau_{\rm 1000 Myr}$ ({\it gray dashed line})
141:       model for the first 3 Gyr. Both \ha\ measurements and the
142:       independent SED modeling demonstrate that in 9 out of 20
143:       galaxies in our sample the star formation has been strongly
144:       suppressed.\label{wha}}
145:   \end{center}
146: \end{figure}
147: 
148: 
149: \section{INTRODUCTION}
150: 
151: Observations imply that stellar populations of high-mass galaxies were
152: formed at higher redshift than those of low-mass galaxies
153: \citep[e.g.,][]{co96,ju05}. Recent hierarchical CDM models are able to
154: produce dead, massive galaxies at high redshift by incorporating
155: feed-back from active galactic nuclei \citep[AGN;
156: e.g.,][]{cr06,bo06,dl06,ho06}. In order to observationally determine
157: when and how star formation in massive galaxies was suppressed it is
158: necessary to identify and study these objects out to the highest
159: redshifts.
160: 
161: Recently, massive and apparently dead galaxies have been identified at
162: $z>1.5$ \citep[e.g.,][]{mc04,sa05,da05,la05,re05,re06,pa06}. Beyond
163: $z=2$ most studies rely on photometric redshifts and broad-band colors
164: to identify these galaxies. However, as dust and age have similar
165: effects on the broad-band spectral energy distribution (SED), the star
166: formation rates (SFRs) are often not well constrained. Furthermore,
167: their photometric redshifts are not well calibrated, as only a few at
168: $z>2$ have spectroscopic redshifts \citep{da05}. Thus spectroscopic
169: redshifts and independent stellar population diagnostics are needed to
170: determine the prevalence of ``red and dead'' galaxies beyond $z=2$.
171: 
172: Spectroscopic confirmation of non-star-forming galaxies at $z>2$ is
173: complicated due to their faint rest-frame UV emission and lack of
174: nebular emission lines. Deep near-infrared (NIR) spectroscopy provides
175: the best option to confirm such galaxies at $z>2$, as their relatively
176: bright rest-frame optical luminosity allows direct detection of the
177: stellar continuum. The optical continuum shape, and in particular the
178: Balmer/4000 \AA\ break, can be used to derive redshifts for galaxies
179: without emission lines, and provides independent constraints on
180: stellar populations \citep{kr06}.  
181: 
182: To study a high-redshift spectroscopic sample that is not biased
183: towards galaxies with bright emission lines, we are conducting a NIR
184: spectroscopic survey of $K$-selected galaxies with photometric
185: redshifts $z\sim2.3$. Here we report on a surprising result of our
186: survey: the large fraction of galaxies with no detected emission
187: lines. Throughout the letter we assume a $\Lambda$CDM cosmology with
188: $\Omega_{\rm m}=0.3$, $\Omega_{\rm \Lambda}=0.7$, and $H_{\rm
189:   0}=70$~km s$^{-1}$ Mpc$^{-1}$. All broadband magnitudes are given in
190: the Vega-based photometric system.
191: 
192: 
193: \section{GALAXY SAMPLE AND DATA}
194: 
195: The galaxies presented in this letter are drawn from the
196: Multi-wavelength Survey by Yale-Chile (MUSYC). This survey provides
197: optical and deep NIR photometry for several southern and equatorial
198: fields \citep{ga06,qu06}. The targets were selected in $K$ ($K<19.7$)
199: to reduce the dispersion in stellar mass and to ensure adequate $S/N$
200: in the NIR spectra. Additionally, we required a photometric redshift
201: in the range $2.0<z<2.7$, for which bright rest-frame optical emission
202: lines such as [O\,{\sc iii}] and \ha\ fall in the $H$ and $K$
203: atmospheric windows. The photometric redshifts are derived following
204: the procedure described in \cite{ru01,ru03}.
205: 
206: We observed 26 galaxies with the Gemini near-infrared spectrograph
207: (GNIRS) in 2004 September (GS-2004B-Q-38), 2005 May (GS-2005A-Q-20),
208: 2006 January (GS-2005B-C-12) and 2006 February (GS-2006A-C-6). The
209: spectra of two of these galaxies have already been presented by
210: \cite{vd05} and \cite{kr06}. All galaxies were observed in
211: cross-dispersed mode, in combination with the short wavelength camera,
212: the 32 line mm$^{-1}$ grating (R=1000) and the 0\farcs675 by 6\farcs2
213: slit. In this configuration we obtained a wavelength coverage of 1.0
214: -- 2.5 $\rm \mu m$. The galaxies were observed for 1-4 hours,
215: depending on the brightness of the target and the weather conditions.
216: The observational techniques and reduction of the GNIRS spectra are
217: described in detail by \cite{kr06}. For each galaxy we extract a
218: one-dimensional original and low resolution binned spectrum.
219: 
220: To derive the stellar population properties and obtain redshifts for
221: galaxies without emission lines, we fit stellar population models to
222: the low resolution continuum spectra together with the $UBVRIz$
223: fluxes, following the technique described in \cite{kr06}.  We use the
224: \cite{bc03} models with a set of exponentially declining star
225: formation histories, a \cite{sa55} initial mass function (IMF) between
226: 0.1 and 100 $M_{\odot}$, and solar metallicity, and adopt the
227: \cite{ca00} reddening law. The assumed model parameters (IMF,
228: reddening law, metallicity) are identical to those used by e.g.,
229: \cite{fo04}, \cite{sh05}, and \cite{pa06}.
230: 
231: We obtained spectroscopic redshifts for all 26 galaxies using
232: rest-frame optical emission lines or the Balmer/4000\AA\ break. The
233: `break' redshifts have a median uncertainty of $|\Delta
234: z|/(1+z)=0.017$, as determined from fitting the low-resolution
235: continua of emission line galaxies with $z$ as a free parameter. 20 of
236: 26 galaxies have spectroscopic redshifts in the range $2.0<z<2.7$, for
237: which \ha\ falls in the $K$-band. In what follows we restrict the
238: sample to the galaxies in this redshift range; the full sample will be
239: described elsewhere (M.~Kriek et al., in preparation). We note that
240: the six galaxies that fall out of this redshift range have
241: $z=1.75-1.95$.\footnote{The median of the differences between
242:   spectroscopic and photometric redshifts $(z_{\rm spec}-z_{\rm
243:     phot})/(1+z_{\rm spec})$ is only $-0.001$ for the full sample of
244:   26 galaxies.}
245: 
246: \section{SUPPRESSED STAR FORMATION}
247: 
248: Surprisingly, nine out of the 20 galaxies in the sample show no
249: emission lines in their rest-frame optical spectra. The spectra and
250: best fits of these galaxies are presented in Fig.~\ref{spectra}. For
251: these galaxies we derived upper limits on the \ha equivalent width
252: (\wha) as follows. We have drawn 200 random redshifts from the
253: redshift probability distribution, and determined the 2$\sigma$ upper
254: limit of \wha\ in each case from the measured noise properties,
255: assuming a rest-frame \ha\ FWHM of 500 km s$^{-1}$ \citep[see][]{vd04}
256: and the best-fit stellar continuum. The adopted limit is the maximum
257: value found in the simulations, excluding the highest 5\%. The median
258: \wha\ upper limit of these nine galaxies (corrected for Balmer
259: absorption) is 10\,\AA.
260: 
261: \wha\ is a measure of the ratio of current to past star formation, and
262: the limits on \wha\ may imply very low SFRs in these galaxies. We
263: investigate this in Fig.~\ref{wha}, in which we plot the rest-frame
264: \wha\ upper limits (corrected for Balmer absorption) versus the
265: specific SFRs (SFR / $M_*$) derived from our model fits to the
266: spectra. Remarkably, all nine galaxies without detected emission
267: lines ({\it filled red circles}) are best fitted by stellar population
268: models with low specific SFRs, and the data points are broadly
269: consistent with the expected relations between these properties
270: \citep{ke98,bc03}. For five galaxies the best-fit SFR is zero and the
271: four remaining galaxies have best-fit values that are consistent with
272: zero within $1\sigma$. The midmean\footnote{mean of the central two
273:   quadrants} specific and absolute SFRs are $0.004\rm\,Gyr^{-1}$ and
274: $0.9\,M_{\odot}/$yr respectively, significantly lower than the
275: $0.56\rm\,Gyr^{-1}$ and $128\,M_{\odot}/$yr found for the eleven
276: emission line galaxies.  Thus, both the \ha\ measurements and the
277: stellar continuum modeling imply that the star formation in these nine
278: galaxies has been strongly suppressed. We note that we find a similar
279: relation when we plot $L_{\rm H\alpha}$ vs.\ the modeled absolute SFR,
280: as our galaxies span only a small range in stellar mass
281: ($0.9-4.6\,\times\,10^{11}M_{\odot}$). These nine galaxies have a
282: median stellar mass of $2.6\times 10^{11} M_{\odot}$, a median $J-K$
283: color of 2.45, and have undergone a median of 21 age/$\tau$ e-folding
284: times. Six out of nine are distant red galaxies \citep[DRGs,
285: $J-K>2.3$, ][]{fr03}. As can be seen in Fig.~\ref{wha} several of the
286: emission line galaxies are also best fitted by stellar population
287: models with low specific SFRs. This may suggest that the gas in these
288: galaxies is not ionized by hot stars; we will explore this in a future
289: paper.
290: 
291: Formally, we find high best-fit values for the dust content for eight
292: out of nine galaxies without emission lines. However, $A_V$ is poorly
293: constrained for most of these galaxies, and models with zero or only
294: small amounts of dust are consistent within $1\sigma$. We refitted all
295: 9 galaxies allowing only models without dust. The best fits
296: (Fig.~\ref{spectra}, {\it light gray}) also yield low specific SFRs,
297: ranging from $0.001-0.013\rm\,Gyr^{-1}$, with a median value of
298: $0.002\rm\,Gyr^{-1}$. The median best-fit stellar age is 0.9 Gyr,
299: which is a factor of $\sim2$ larger than when allowing dust. While we
300: cannot draw firm conclusions about the dust content in these galaxies,
301: we note that high $A_V$ may indicate that they are still in the
302: process of losing their gas and dust.
303: 
304: Figure~\ref{wha} also shows the UV-selected $z\sim2$ galaxies of
305: \cite{er06a,er06b,er06c}. These galaxies have similar \wha\ and
306: specific SFRs as the $K$-selected galaxies with emission
307: lines. However, there is no overlap with the $K$-selected galaxies
308: without detected \ha\ emission; both the \wha\ and the modeled
309: specific SFR are higher for the UV-selected galaxies. Although
310: UV-selection is able to find massive galaxies
311: \citep[e.g.,][]{sh05,er06b}, it does not sample the full distribution
312: of their properties \citep[see also][]{vd06}.
313: 
314: \section{DISCUSSION}
315: 
316: We find that nine out of 20 $K$-selected galaxies at $2.0<z_{\rm
317:   spec}<2.7$ have no detected \ha\ emission (\wha\,$\lesssim$~10\AA)
318: and are best fit by stellar population models with low specific SFRs
319: ($\sim\,0.004\rm\,Gyr^{-1}$), implying a fraction of galaxies with
320: strongly suppressed star formation of $45^{+18}_{-12}$\,\%.  The
321: quoted uncertainty is derived assuming Poisson statistics, and does
322: not include the following systematic errors and caveats. First, our
323: $z_{\rm phot}$ selection criterion could introduce biases, as
324: systematic errors in photometric redshift may correlate with SED
325: type\footnote{This effect may be small, as the distribution of
326:   specific SFRs of the six objects at $z_{\rm spec}<2$ is similar to
327:   that of the 20 remaining objects}. Also, the $K$-band selection
328: criterion could bias our sample, as for starburst galaxies strong
329: emission lines can contribute significantly to the $K$-band flux
330: \citep{er06a}. We do not expect this bias to be strong, as the median
331: contribution of the emission lines to the $K$-band is only 0.04 mag
332: for the eleven emission line galaxies. Third, incompleteness may play
333: a role, as we observed only $\sim$20\% of the galaxies that meet the
334: selection criteria. We note, however, that according to RS- and
335: KS-tests our $K$-selected sample has a similar distribution of
336: rest-frame $U-V$-colors as the large mass-limited sample ($>10^{11}
337: M_{\odot}$) by \cite{vd06}, when applying the same $K$-magnitude
338: cut. Furthermore, the assumption that \ha\ emission is only due to
339: star formation may lead us to underestimate the fraction.
340: 
341: Finally, and perhaps most importantly, we may miss star formation with
342: very high rest-frame optical extinction. Although most local ULIRGs
343: have high integrated \wha\ \citep[$W_{\rm H\alpha+[N\,{\sc ii}]} \sim
344: 90$\AA,][]{li95}, for some the star-burst regions are almost
345: completely obscured (e.g., Arp 220 has $W_{\rm H\alpha+[N\,{\sc ii}]}
346: = 18$\AA) and these objects might be misinterpreted in our
347: analysis. Available 24 $\mu$m imaging on DRG samples shows that
348: 30-50\% have no MIPS counterpart \citep{we06,pa06,re06}, and
349: \citet{re06} finds that red galaxies with low MIPS fluxes typically
350: have low specific SFRs. To resolve this issue it is necessary to
351: combine our spectra with deep mid/far-infrared imaging to detect
352: hidden star formation.
353:   
354: It is interesting to compare our fraction of galaxies with strongly
355: suppressed star formation to previous studies\footnote{The fractions
356:   of galaxies with suppressed star formation among DRGs and
357:   $K$-selected galaxies are similar in our sample}. Our result is
358: consistent with the fraction found by \cite{la05}, as they identified
359: three ``red and dead'' galaxies from a sample of 11 DRGs in the HDF-S,
360: and with the study of \cite{re06}, who find that seven out of 24 DRGs
361: at $1.5<z<2.6$ are not detected at 24 $\mu$m, and have an average low
362: specific SFR of 0.05 Gyr$^{-1}$. However, our fraction is
363: significantly higher than the fractions found by \cite{da04} and
364: \cite{pa06}. Using the $BzK$ criterion, \cite{da04} identify all their
365: 11 $K$-selected galaxies at $2.0<z<2.7$ as star-forming galaxies. The
366: difference may be partly explained by different definitions, as
367: $\sim$4 of our 9 galaxies with suppressed star formation galaxies
368: would have been identified as star-forming galaxies by the $BzK$
369: criterion. \cite{pa06} finds that $\sim$10\% of 153 DRGs at $1.5\le
370: z_{\rm phot}\le 3$ show no signs of current star formation. Again, the
371: selection criteria could play a role; none of our galaxies would have
372: been classified as ``dead'' by \cite{pa06}, as they apply the
373: following criteria: age $>1 \rm Gyr$ , $E(B-V)\le0.1$, age$/\tau>3$
374: and no X-ray or 24$\mu$m detection. The sample selection could also be
375: a factor, as the contribution of dusty star-forming galaxies to the
376: DRG population is expected to be higher at $z<2$. Furthermore, the
377: CDF-S field -- in which both the studies by \cite{pa06} and
378: \cite{da04} were performed -- may be atypical \citep{vd06}. We stress
379: that our study is the first that is based on spectroscopic redshifts,
380: and that this might also account for differences in the obtained
381: fractions. This will be explored in a future paper.
382: 
383: We note that most current CDM galaxy formation models fail to produce
384: the high fraction of red galaxies at the massive end as found by
385: \cite{vd06}. One easy way to solve this is by allowing more dust in
386: the galaxies, but we have shown here that a large fraction of galaxies
387: has low specific SFRs, and these are generally absent in these models
388: at $z>2$ \citep[e.g.,][]{so04,na05,ka06}. Our results may indicate
389: that the suppression of the star formation in the most massive
390: galaxies occurs at higher redshift than has been predicted by current
391: models. In this context, it is interesting to note that not all of the
392: line-emission in our sample may be due to star formation: from our
393: low-resolution GNIRS spectra it appears that several of the
394: emission-line galaxies with low specific SFRs exhibit high [N\,{\sc
395:   ii}]/H$\alpha$ ratios, possibly indicating AGN. High resolution
396: spectra of the line-emitting objects in our sample will be discussed
397: in a future paper (M.~Kriek et al., in preparation).
398: 
399: \acknowledgments We thank the referee for constructive comments which
400: improved the manuscript. This research was supported by grants from
401: the Netherlands Foundation for Research (NWO), the Leids
402: Kerkhoven-Bosscha Fonds, National Science Foundation grant NSF CAREER
403: AST-044967, and NASA LTSA NNG04GE12G. 
404: 
405: \begin{thebibliography}{}
406: \bibitem[Bower et al.(2006)]{bo06} Bower, R.G., et al. 2006, \mnras,
407:   370, 645
408: \bibitem[Bruzual \& Charlot(2003)]{bc03} Bruzual, G. \& Charlot, S. 2003, 
409:   \mnras, 344, 1000
410: \bibitem[Calzetti et al.(2000)]{ca00} Calzetti, D., Armus, L., Bohlin, R.C., 
411:   Kinney, A.L., Koornheef, J., \& Storchi-Bergmann, T. 2000, \apj, 533, 682 
412: \bibitem[Cowie et al.(1996)]{co96} Cowie, L.L., Songaila, A., Hu, E., \&
413:   Cohen, J.G. 1996, \aj, 112, 839
414: \bibitem[Croton et al.(2006)]{cr06} Croton, D.J., et al. 2006, \mnras, 365, 11
415: \bibitem[Daddi et al.(2004)]{da04} Daddi, E., et al. 2004, \apj, 617,
416:   746
417: \bibitem[Daddi et al.(2005)]{da05} Daddi, E., et al. 2005, \apj, 626, 680
418: \bibitem[De Lucia et al.(2006)]{dl06} De Lucia, G., Springel, V.,
419:   White, S.D.M., Croton, D., \& Kauffmann, G. 2006, \mnras, 366, 499
420: \bibitem[Erb et al.(2006a)]{er06a}  Erb, D.K., Shapley, A.E., Pettini, M.,
421:   Steidel, C.C., Reddy, N.A., \& Adelberger, K.L. 2006, \apj, 644, 813
422: \bibitem[Erb et al.(2006b)]{er06b}  Erb, D.K., Steidel, C.C., Shapley, A.E., 
423:   Pettini, M., Reddy, N.A., \& Adelberger, K.L. 2006, \apj, 646, 107
424: \bibitem[Erb et al.(2006c)]{er06c}  Erb, D.K., Steidel, C.C., Shapley, A.E., 
425:   Pettini, M., Reddy, N.A., \& Adelberger, K.L. 2006, \apj\, 647, 128
426: \bibitem[F\"orster Schreiber et al.(2004)]{fo04} F\"orster Schreiber, N.M.
427:   et al. 2004, \apj, 616, 40
428: \bibitem[Franx et al.(2003)]{fr03} Franx, M. et al. 2003, \apj, 587, L79
429: \bibitem[Gawiser et al.(2006)]{ga06} Gawiser, E. et al. 2006, \apjs, 162, 1
430: \bibitem[Hopkins et al.(2006)]{ho06} Hopkins, P.F., Hernquist, L.,
431:   Cox, T.J., Robertson, B., \& Springel, V. 2006, \apj, 163, 50
432: \bibitem[Juneau et al.(2005)]{ju05} Juneau, S., et al. 2005, \apj, 619, L135
433: \bibitem[Kang et al.(2006)]{ka06} Kang, X., Jing, Y.P., \& Silk, J. 2006,
434:   \apj\ in press (astro-ph/0601685)
435: \bibitem[Kennicutt(1998)]{ke98} Kennicutt, R.C. 1998, ARA\&A, 36, 189
436: \bibitem[Kriek et al.(2006)]{kr06} Kriek, M. et al. 2006, \apj, 645, 44
437: \bibitem[Labb\'e et al.(2005)]{la05} Labb\'e, I. et al. 2005, \apj, 624, L81
438: \bibitem[Liu \& Kennicutt(1995)]{li95} Liu, C.T., \& Kennicutt, R.C. 1995, 
439:   \apj, 450, 547
440: \bibitem[McCarthy et al.(2004)]{mc04} McCarthy, P.J. et al. 2004, \apj, 614, L9
441: \bibitem[Nagamine et al.(2005)]{na05} Nagamine, K., Cen, R., Hernquist, L.,
442:   Ostriker, J.P., \& Springel, V. 2005, \apj, 627, 608
443: \bibitem[Papovich et al.(2006)]{pa06} Papovich, C., et al. 2006, \apj, 
444:   640, 92
445: \bibitem[Quadri et al.(2006)]{qu06} Quadri, R., et al. 2006, submitted
446:   to \apj\ (astro-ph/0606330)
447: \bibitem[Reddy et al.(2005)]{re05} Reddy, N.A., Erb, D.K., Steidel, C.C.,
448:   Shapley, A.E., Adelberger, K.L., \& Pettini, M. 2005, \apj, 633, 748
449: \bibitem[Reddy et al.(2006)]{re06} Reddy, N.A., et al. 2006, 644, 792
450: \bibitem[Rudnick et al.(2001)]{ru01} Rudnick, G., et al. 2001, \aj, 122, 2205
451: \bibitem[Rudnick et al.(2003)]{ru03} Rudnick, G., et al. 2003, \apj, 599, 847
452: \bibitem[Salpeter(1955)]{sa55} Salpeter, E.E. 1955, \apj, 121, 161
453: \bibitem[Saracco et al.(2005)] {sa05} Saracco, P. et al. 2005, \mnras, 357, L40
454: \bibitem[Shapley et al.(2005)]{sh05} Shapley, A.E., et al. 2005, \apj, 
455:   626, 698
456: \bibitem[Somerville(2004)]{so04} Somerville, R.S. et al. 2004, \apj,
457: 600, 135
458: \bibitem[van Dokkum et al.(2004)]{vd04} van Dokkum, P.G. et al. 2004, \apj, 
459:   611, 703 
460: \bibitem[van Dokkum et al.(2005)]{vd05} van Dokkum, P.G., Kriek, M., Rodgers,
461:  B., Franx, M., \& Puxley, P. 2005, \apj, 622, L13
462: \bibitem[van Dokkum et al.(2006)]{vd06} van Dokkum, P.G., et al. 2006, \apj,
463:   638, L59
464: \bibitem[Webb et al.(2006)]{we06} Webb, T.M.A., et al. 2006, \apj, 636, L17
465: \end{thebibliography}
466: 
467: 
468: \end{document}
469: