1: %\documentclass[12pt,preprint]{aastex}
2:
3: \documentclass[apjl]{emulateapj}
4: %\usepackage{natbib}
5: %\citestyle{aa}
6:
7: \shorttitle{A genuine intermediate-Age Globular Cluster in M33}
8:
9: \shortauthors{Chandar et al.}
10:
11: \begin{document}
12: \title{A genuine intermediate-Age Globular Cluster in M33\altaffilmark{*}}
13:
14: \author{Rupali Chandar\altaffilmark{1}, Thomas H. Puzia\altaffilmark{2}, Ata
15: Sarajedini\altaffilmark{3}, \& Paul Goudfrooij\altaffilmark{2}}
16: \altaffiltext{*}{Based on observations obtained with the Hectospec instrument
17: at the MMT Observatory. The MMT Observatory is a joint venture of the
18: Smithsonian Institute and the University of Arizona. Based on observations
19: made with the NASA/ESA Hubble Space Telescope, obtained from the data
20: archive at the Space Telescope Science Institute. STScI is operated by the
21: Association of Universities for Research in Astronomy, Inc. under NASA
22: contract NAS 5-26555.}
23: \altaffiltext{1}{Department of Physics and Astronomy, Johns Hopkins
24: University, 3400 North Charles Street, Baltimore, MD 21218, USA.
25: {\tt rupali@pha.jhu.edu}}
26: \altaffiltext{2}{Space Telescope Science Institute, 3700 San Martin Drive,
27: Baltimore, MD 21218, USA. {\tt tpuzia@stsci.edu, goudfroo@stsci.edu}}
28: \altaffiltext{3}{Department of Astronomy, University of Florida, 211 Bryant
29: Space Science Center, P.O. Box 112055, Gainesville, FL 32611, USA.
30: {\tt ata@astro.ufl.edu}}
31:
32:
33: \begin{abstract}
34: We present deep integrated-light spectroscopy of nine M33 globular
35: clusters taken with the Hectospec instrument at the MMT Observatory.
36: Based on our spectroscopy and previous deep color-magnitude diagrams
37: obtained with HST/WFPC2, we present evidence for the presence of a genuine
38: intermediate-age globular cluster in M33. The analysis of Lick line
39: indices indicates that all globular clusters are metal-poor
40: ([Z/H]~$\la-1.0$) and that cluster M33-C38 is $\sim\!5\!-\!8$ Gyr younger
41: than the rest of the sample M33 star clusters. We find no evidence for a
42: population of blue horizontal branch stars in the CMD of M33-C38, which
43: rules out the possibility of an artificially young spectroscopic age due
44: to the presence of hot stars. We infer a total mass of
45: $5-9\times10^{4}~M_{\odot}$ for M33-C38, which implies that M33-C38 has
46: survived $\sim2-3$ times longer than some dynamical evolution model
47: predictions for star clusters in M33, although it is not yet clear to
48: which dynamical component of M33 -- thin disk, thick disk, halo -- the
49: cluster is associated.
50: \end{abstract}
51:
52: \keywords{globular clusters: general --- galaxies: star clusters ---
53: galaxies: evolution --- galaxies: formation --- galaxies: structure}
54:
55: \section{Introduction}
56:
57: Star clusters which have survived for several ($\sim2-8$) billion years
58: (intermediate-age clusters) can provide important clues concerning both
59: the evolution history of their parent galaxies and for understanding the
60: destruction processes which erode star cluster systems \citep[e.g.,][and
61: references therein]{goudfrooij04}. In the Galaxy, clusters of this age all
62: fall in the category of ``old open clusters'' \citep[see compilation
63: in][]{friel95} -- they all reside in the Galactic thin disk, and have
64: lower masses ($\sim$few $10^3$~$M_{\odot}$) than typical ancient globular
65: clusters ($\ga10^{4.5}~M_{\odot}$). The oldest of these old open clusters
66: help to pinpoint the age of the thin disk, while the ages and masses of
67: the entire population help constrain the survival rates of clusters in the
68: Galactic disk. Intermediate-age clusters are also known to exist in the
69: Large and Small Magellanic Clouds (LMC and SMC). The LMC contains a
70: population of clusters which are $1-3$~Gyr, while the SMC formed numerous
71: clusters $4-8$~Gyr ago.
72:
73: In more distant galaxies, it becomes increasingly difficult to establish
74: the presence of intermediate-age clusters, primarily because the
75: techniques available to study compact clusters become more limited. For
76: example, the integrated optical colors of intermediate-age clusters are
77: degenerate with those of ancient ($\geq8$~Gyr) clusters. Although a number
78: of clusters have absorption line index strengths (measured from
79: integrated-light spectroscopy) which suggest that these objects are of
80: intermediate age, their absolute ages cannot be definitively established
81: due to the possible presence of hot stars in the core helium burning stage
82: (hot ``horizontal branch'' or HB stars). These hot HB stars could
83: potentially boost the Balmer absorption line strengths sufficiently to
84: mimic younger ages for old globular clusters \citep[e.g.,][]{lee05,
85: maraston00}. Ideally, cluster ages should be determined from deep
86: color-magnitude diagrams (CMDs) which reveal the main sequence turnoff; to
87: date however, this has only been accomplished for a single cluster beyond
88: the Magellanic Clouds \citep[M31-SKHB 312;][]{brown04}, using 129 orbits
89: of {\it Hubble Space Telescope} (HST) time.
90:
91: In this Letter, we present integrated-light (optical) spectroscopy for
92: nine ancient star cluster candidates in the nearby spiral galaxy M33.
93: These clusters have deep CMDs available from HST/WFPC2 observations, which
94: clearly reveal the HB morphology. The combination of HB morphology and
95: absorption line measurements result in robust relative age estimates, and
96: are ideal to search for intermediate-age clusters in M33.
97:
98: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
99:
100: \section{Data and Analysis}
101:
102: \subsection{Integrated-Light Spectroscopy}
103:
104: We obtained integrated spectroscopy for $\sim\!150$ star clusters and
105: cluster candidates in M33 using the MMT/Hectospec instrument
106: \citep{fabricant05} at a dispersion of 1.2~\AA~pix$^{-1}$. In addition to
107: clusters, a significant number of background regions were chosen, in order
108: to sample the range of background light provided by the galaxy itself. The
109: clusters were observed on the night of October 27/28, 2005, with four 1100
110: second exposures (for a total of 73.3 minutes). Each of the 300 Hectospec
111: fibers subtends $1.5\arcsec$ on the sky. The data were reduced using the
112: HSRED reduction pipeline developed by R. Cool\footnote{Available from
113: http://mizar.as.arizona.edu/\~{}rcool/hsred.}. All observations were bias
114: subtracted, overscan corrected, and trimmed. The science exposures were
115: flat-fielded and combined together to eliminate cosmic rays, and one
116: dimensional spectra were extracted and wavelength calibrated. The final
117: resolution measured from the FWHM of HeNeAr comparison spectra was
118: $\sim4.7$~\AA. Spectra for the entire M33 cluster sample will be presented
119: in a future work. Here, we focus on a subset of nine clusters which have
120: available CMDs reaching below the HB \citep[hereafter S00]{sarajedini98, sarajedini00}.
121:
122: Because the underlying galaxy light within each object fiber can affect
123: the absorption line strengths, we created a mean background spectrum for
124: each cluster. This was accomplished by averaging several background
125: spectra chosen to match the galaxy background level measured in an annulus
126: around each cluster from reduced M33 images provided by the Local Group
127: Survey\footnote{http://www.lowell.edu/\~{}masey/lgsurvey.}). These
128: customized background spectra were then subtracted from each object. For
129: the nine clusters presented here, the cluster spectra have such high
130: signal-to-noise, that the background subtraction has little impact. In
131: Figure~\ref{fig1} we present integrated, background-subtracted spectra of
132: the globular clusters discussed in this work.
133:
134: \begin{figure}
135: \centering
136: \includegraphics[bb= 0 0 400 390, width=8cm]{f1.eps}
137: \caption{The blue portion of the background subtracted spectra of nine
138: M33 globular clusters, taken with MMT/Hectospec are shown. The spectra
139: have been normalized by a spline fit to their continua.}
140: \label{fig1}
141: \end{figure}
142:
143: The relative ages and abundances of stellar populations can be estimated
144: from integrated-light spectra, by comparing the absorption line strengths
145: for age and metallicity-sensitive lines. The measurements and accompanying
146: uncertainties are computed using the techniques described in
147: \cite{puzia02}, for line indices defined in \cite{worthey94} and
148: \cite{worthey97}. The passband definitions were shifted to account for the
149: radial velocity of each cluster, assuming the measurements given in
150: \cite{chandar02}. Because no observations of Lick standard stars have been
151: made to date with MMT/Hectospec, we do not calibrate the index
152: measurements to the Lick/IDS system \citep{burstein84}, and instead focus
153: on {\it relative} cluster ages. A comparison of the line indices measured
154: for the nine clusters presented in this work with predictions of the
155: \cite{thomas03, thomas04} models is illustrated in Figure~\ref{ps:am}.
156:
157: \subsection{Color-Magnitude Diagrams}
158: The CMDs for the nine clusters analyzed in this work were originally
159: presented in Sarajedini et al. (2000; hereafter S00). These are based on
160: $V_{\rm F555W}$ and $I_{\rm F814W}$ observations taken with the HST/WFPC2
161: instrument. We refer the reader to S00 for details of data acquisition,
162: reduction, and analysis. The CMDs show that only M9 and U77 have blue HB
163: stars; the rest of the M33 clusters in the S00 sample have HBs which lie
164: completely redward of the RR Lyrae instability strip, in the so-called
165: ``red clump''.
166:
167: Here, we present new CMDs based on the S00 data for two clusters of
168: particular interest -- M9 and C38. These have been radially cleaned, so
169: that blends and other contaminants from the crowded central regions are
170: not included in the CMD. In Figure~\ref{fig:cmd} we show the new CMDs for
171: these two clusters which are further discussed below.
172:
173:
174: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
175: \section{Discussion}
176:
177: \subsection{Evidence for an Intermediate-Age Cluster}
178: In Figure~\ref{ps:am}, we show the measurements for the metal-sensitive
179: absorption line index [MgFe]\arcmin\ versus the age-sensitive Balmer line
180: indices H$\delta_{\rm A}$, H$\gamma_{\rm A}$, and H$\beta$, as is
181: typically done to estimate the ages and chemical compositions of stellar
182: populations from integrated-light spectroscopy \citep[e.g.,][]{puzia05}.
183: For comparison, we show simple stellar population (SSP) model predictions
184: from \cite{thomas04} for six different ages (solid lines) and six
185: different values of [Z/H] (dashed lines). We also compared the M33 cluster
186: measurements with the solar-scaled SSP model predictions of \cite{bc03} at
187: 4.7~\AA\ resolution (not shown), which yield results consistent with the
188: \cite{thomas04} models.~All panels in Figure~\ref{ps:am} show a ``cloud''
189: of points located towards low metallicities and old ages, although we
190: cannot rule out an age spread of the order of $\Delta t/t\la 0.3$, which
191: translates into a $\sim\!4$~Gyr spread at 12 Gyr absolute age. Cluster C38
192: is clearly offset from the ``cloud" in every panel, and clearly appears to
193: have a younger age than the other clusters -- recall that our measurements
194: have not been absolutely calibrated to the models, but should be robust in
195: a relative sense. Therefore, Figure~\ref{ps:am} implies that C38 is
196: $\sim\!5-8$ Gyr younger than the other M33 clusters plotted here. Clusters
197: with H$\alpha$ seen in emission (R14 and U77) are plotted with different
198: symbols in Figure~\ref{ps:am}. The Balmer absorption-line strengths for
199: these clusters thus represent lower limits.
200:
201: \begin{figure*}
202: \centering
203: \includegraphics[bb= 0 0 400 400, width=5.5cm]{f2a.eps}
204: \includegraphics[bb= 0 0 400 400, width=5.5cm]{f2b.eps}
205: \includegraphics[bb= 0 0 400 400, width=5.5cm]{f2c.eps}
206: \caption{Age-metallicity diagnostic plots for the sample M33 globular
207: clusters, constructed from three different Balmer line indices:
208: H$\delta_{\rm A}$ ({\it left panel}), H$\gamma_{\rm A}$ ({\it middle
209: panel}), and H$\beta$ ({\it right panel}) vs.~[MgFe]\arcmin\
210: \citep[see also Fig.~9 of][for details]{puzia05}. Over-plotted are SSP models of
211: \cite{thomas04} for [$\alpha$/Fe]~$=0.3$, metallicities [Z/H]=$-2.25,
212: -1.35, -0.33, 0.00$, 0.35, and 0.67 dex ({\it dashed lines}), and ages
213: 15, 10, 5, 3, 2, 1 ({\it solid lines}). The thick solid line is the 5
214: Gyr isochrone, and is used to split between old and intermediate age
215: globular clusters. Note that some clusters are not plotted in the
216: individual panels due to contamination of their index passband
217: definitions by bad pixels. Clusters with H$\alpha$ observed in
218: emission are indicated by open circles. All clusters are labeled as in
219: Fig.~\ref{fig1}. We point out that the Lick index measurements are not
220: calibrated to the Lick system so that only relative measures should be
221: derived from the shown diagrams.}
222: \label{ps:am}
223: \end{figure*}
224:
225: Could the stronger Balmer line indices for C38, which suggest a younger
226: age, be due to the presence of blue/hot HB stars? In Figure~\ref{fig:cmd},
227: we show the resolved CMD of C38 based on deep HST/WFPC2 observations for
228: this cluster, and also for M9 (one of the older clusters) for comparison.
229: C38 shows no blue HB stars, while M9 does. Note that the HST observations
230: have sufficient depth and resolution to detect blue HBs if they are
231: present, but they clearly do not exist in C38. Therefore, the presence of
232: hot HB stars, which has been suggested as a plausible explanation for
233: apparent intermediate ages of clusters in different galaxies \citep[such
234: as in M31;][]{puzia05b}, cannot be the correct explanation in this case.
235: The line indices and the knowledge of the HB morphology taken together,
236: indicate that it is the {\it age\/} of the cluster which drives the
237: observed Balmer line strengths.
238:
239: There are two features from the resolved CMD of C38 shown in
240: Figure~\ref{fig:cmd} which add weight to this conclusion. First, S00
241: found that the absolute I band magnitude of the red clump for this cluster
242: is more luminous than all of the others in their sample, and that this
243: behavior is reasonably well fit by the $\sim2$~Gyr theoretical HB models
244: of \cite[see also Figure~26 in S00]{girardi99}.
245: Secondly, when the CMD for C38 is cleaned to exclude the most crowded
246: central regions, there appears to be a hint of a main sequence turnoff
247: (MSTO) in the CMD, as illustrated in the left panel of
248: Figure~\ref{fig:cmd}. In contrast, M9 does not show evidence for a
249: MSTO.~Also in this figure, we show a Z = 0.004 isochrone at an age of
250: 2~Gyr from Girardi et al. (2000) for comparison. Taken together with the
251: spectroscopic results and the abnormally bright red clump, the presence of
252: a MSTO in the CMD of C38 provides strong evidence that C38 is a genuine
253: intermediate-age ($\sim2-5$~Gyr) star cluster in M33.
254:
255: \begin{figure}[!t]
256: \centering
257: \includegraphics[bb=46 291 585 752,width=8.5cm]{f3.eps}
258: \caption{$V-I$ versus $I$ CMD for M33 clusters C38 and M9. The
259: CMDs have been restricted to include stars in an annulus between 1 and
260: 2\arcsec\ away from the center of each cluster, in order to minimize
261: blending and contamination. A Girardi isochrone for an age of 2 Gyr and
262: $Z=0.004$ is overplotted onto the C38 CMD, which shows the red clump stars
263: and a hint of a main sequence turnoff. A fiducial HB for the Galactic
264: globular cluster M3 \citep{johnson98} is overplotted onto the M33-M9 CMD,
265: which clearly reveals the blue horizontal branch stars in this cluser.}
266: \label{fig:cmd}
267: \end{figure}
268:
269:
270: \subsection{Properties of M33-C38}
271:
272: Here we compile known properties of M33-C38. This cluster lies at a
273: projected distance of $\sim\!4.5$ kpc from the center of M33, has a total
274: $V$-band magnitude of 18.01 and $B-V=$ 0.73 and $V-I=0.89$ colors
275: \citep{christian88, chandar99, chandar02}. S00 estimated
276: that it has an [Fe/H] of $-0.65\pm0.16$ from the slope of the red giant
277: branch, which is consistent with our spectroscopy given the uncertainties
278: ([Fe/H]$=-1.1\pm0.6$).
279:
280: \cite{chandar02} measure a radial velocity of $-145\pm30$ km/s for this
281: cluster. The local H\,{\sc i} (disk) velocity at the location of C38 is
282: found to be $\sim-100$~km/s from the radio maps of \cite{warner73}.
283: Although the cluster velocity measurement deviates from the local disk
284: motion in Figure~\ref{fig:cmd} of \cite{chandar02}, (showing a comparison
285: of cluster velocities versus the local disk motions relative to cluster
286: age, C38 falls in a region which is ambiguous as to which M33 component
287: C38 belongs, clusters with the most deviant velocities are almost certainly
288: in a halo/thick disk component). Therefore, this object could belong to
289: either the thin disk or a halo/thick disk component. Implications for
290: these possibilities are discussed in the next section.
291:
292: Unfortunately M33-C38 does not have any published $M/L$ ratios, which
293: would provide a direct estimate of its total mass. There are, however,
294: published $M/L_V$ values for three of the clusters presented in this work
295: \citep[H38, M9, and R12;][]{larsen02}. They have an average
296: $M/L_V=1.53\pm0.18$, which is consistent with $M/L_V$ measurements for
297: Galactic GCs \citep{mclaughlin00}. The integrated spectra presented here
298: corroborate that these objects are ancient, even though their exact ages
299: are not known. Given the fact that $M/L_V$ {\it decreases} with decreasing
300: age, as an upper limit for C38 we take the highest value found empirically
301: by \cite{larsen02} of $M/L_V\!=\!1.87$. As a lower limit, we assume the
302: $M/L_V$ value predicted by the \cite{bc03} models at an age of 2~Gyr and
303: for a low metallicity system ($M/L_V\!=\!1.125$). Assuming a foreground
304: reddening $E_{B-V}=0.1$ mag towards M33 and a distance modulus of 24.64
305: mag, we estimate that C38 has a mass in the range $5-9\times10^{4}\,
306: M_{\odot}$. This is about an order of magnitude higher than the masses for
307: old Galactic open clusters, and within the range of ancient globular
308: clusters.
309:
310: \subsection{Implications of Intermediate-Age Clusters in M33}
311:
312: The confirmed presence of an intermediate-age cluster in M33 has important
313: implications for star/cluster formation processes in spiral galaxies,
314: regardless of which structural component C38 belongs to. Below we explore
315: the implications of the presence of intermediate-age clusters in both a
316: thick disk versus halo component, and also in the thin disk.
317:
318: If C38 belongs to either a thick disk or halo component, it is likely that
319: at least a few of the other clusters in our sample belong to the same
320: structural component, so that a relatively large age spread among M33
321: clusters would be present in that component. In the Galaxy, halo GCs are
322: known to be old and have an age spread of only $\sim\!3$~Gyr, whereas the
323: GCs associated with the thick disk are essentially coeval
324: \citep{deangeli05}. We suggest that a large age spread for M33 clusters
325: would favor a halo over a thick disk origin, for the following reasons.
326: The thick disk in the Milky Way is relatively metal-rich, and believed to
327: have formed during a single event, likely the accretion of a relatively
328: massive satellite galaxy which puffed up the disk
329: \citep[e.g.][]{morrison93, abadi03, martin04}. Such an origin for a thick
330: disk in M33 would imply that clusters in M33 formed for many Gyr
331: relatively undisturbed in the disk, until several Gyr ago when an
332: accretion event took place, resulting in a thick disk with a cluster
333: population spanning a large range of ages. Given that to date there is no
334: evidence for such an event in M33, and that this type of scenario would
335: require that no significant merging occurred before or since, we suggest
336: that it is unlikely that clusters residing in a thick disk would have an
337: age spread of several Gyr.
338:
339: Therefore, the presence of C38 seems to favor a halo component over a
340: thick disk component. This would point to a longer timescale for the
341: buildup of the M33 halo relative to that found for the Milky Way, as
342: suggested previously by S00 and \cite{chandar02}.
343:
344: If instead of the halo/thick disk, C38 resides in the thin disk of M33 (the
345: velocity measurements are ambiguous in discriminating between these two
346: possibilities), the intermediate age of this object also has important
347: implications. First, this would confirm the presence of relatively
348: massive, intermediate-age disk clusters, which are not widely observed in
349: our own Galaxy (but might be hidden behind large columns of dust). Second,
350: the ability of massive clusters $(\sim\!10^4~M_{\odot}$) to survive in the
351: disk of M33 has been suggested to be limited to $<1$~Gyr. \cite{lamers05}
352: used cluster samples in M33 and their theory of cluster dissolution
353: \citep[e.g.][]{boutloukos03} to estimate directly from the cluster age and
354: mass distributions the ability of clusters to survive in different
355: environments. For M33, they find that $10^4~M_{\odot}$ clusters will
356: disrupt in $\log({\rm age})=8.8$ years ($\sim\!600$~Myr), and that a
357: $5\times10^{4}\, M_{\odot}$ cluster will disrupt in $\sim1.5$~Gyr.~Mass
358: estimates for C38 put it at $5-9\times10^{4}\, M_\odot$. Although we do
359: not know its precise age, it appears that C38 has likely survived $2-3$
360: times longer than this prediction. Regardless, if C38 is part of M33's
361: thin disk, this points to the formation of a star cluster which is more
362: massive than known counterparts of a similar age in the Local Group, and
363: also to the ability of such an object to survive disruption forces in the
364: environment of M33 for several billion years.
365:
366: \vspace{-0.5cm}
367:
368: \acknowledgments We thank D. Fabricant and N. Caldwell
369: for making the wonderful Hectospec instrument available to the community.
370: R.C. and T.H.P. acknowledge support by NASA through grants GO-10402-A
371: and GO-10129 from the Space Telescope Science Institute, which is operated
372: by AURA, Inc.,~under NASA Contract NAS5-26555. A.S. gratefully
373: acknowledges support from NSF CAREER grant AST-0094048.
374:
375:
376: \begin{thebibliography}{}
377:
378: \bibitem[Abadi et al.(2003)]{abadi03} Abadi, M.~G., Navarro,
379: J.~F., Steinmetz, M., \& Eke, V.~R.\ 2003, \apj, 597, 21
380: \bibitem[Boutloukos \& Lamers(2003)]{boutloukos03} Boutloukos,
381: S.~G., \& Lamers, H.~J.~G.~L.~M.\ 2003, \mnras, 338, 717
382: \bibitem[Brown et al.(2004)]{brown04} Brown, T.~M., Ferguson,
383: H.~C., Smith, E., Kimble, R.~A., Sweigart, A.~V., Renzini, A., Rich, R.~M.,
384: \& VandenBerg, D.~A.\ 2004, \apjl, 613, L125
385: \bibitem[Bruzual \& Charlot(2003)]{bc03} Bruzual, G., \&
386: Charlot, S.\ 2003, \mnras, 344, 1000
387: \bibitem[Burstein et al.(1984)]{burstein84} Burstein, D., Faber, S.\ M.,
388: Gaskell, C.\ M., \& Krumm, N.\ 1984, ApJ, 287, 586
389: \bibitem[Chandar et al.(1999)]{chandar99} Chandar, R., Bianchi,
390: L., \& Ford, H.~C.\ 1999, \apj, 517, 668
391: \bibitem[Chandar et al.(2002)]{chandar02} Chandar, R., Bianchi,
392: L., Ford, H.~C., \& Sarajedini, A.\ 2002, \apj, 564, 712
393: \bibitem[Christian \& Schommer(1988)]{christian88} Christian,
394: C.~A., \& Schommer, R.~A.\ 1988, \aj, 95, 704
395: \bibitem[De Angeli et al.(2005)]{deangeli05} De Angeli, F.,
396: Piotto, G., Cassisi, S., Busso, G., Recio-Blanco, A., Salaris, M.,
397: Aparicio, A., \& Rosenberg, A.\ 2005, \aj, 130, 116
398: \bibitem[Fabricant et al.(2005)]{fabricant05} Fabricant, D., et
399: al.\ 2005, \pasp, 117, 1411
400: \bibitem[Friel(1995)]{friel95} Friel, E.~D.\ 1995, \araa, 33,
401: 381
402: \bibitem[Girardi(1999)]{girardi99} Girardi, L.\ 1999, \mnras,
403: 308, 818
404: \bibitem[Goudfrooij et al.(2004)]{goudfrooij04} Goudfrooij, P.,
405: Gilmore, D., Whitmore, B. C., \& Schweizer, F. 2004, \apjl, 613, L121
406: \bibitem[Johnson \& Bolte(1998)]{johnson98} Johnson, J.~A., \&
407: Bolte, M.\ 1998, \aj, 115, 693
408: \bibitem[Lamers et al.(2005)]{lamers05} Lamers, H.~J.~G.~L.~M.,
409: Gieles, M., \& Portegies Zwart, S.~F.\ 2005, \aap, 429, 173
410: \bibitem[Larsen et al.(2002)]{larsen02} Larsen, S.~S., Brodie,
411: J.~P., Sarajedini, A., \& Huchra, J.~P.\ 2002, \aj, 124, 2615
412: \bibitem[Lee \& Worthey(2005)]{lee05} Lee, H.-c., \& Worthey,
413: G.\ 2005, \apjs, 160, 176
414: %\bibitem[Lee et al. (2005)]{lee05} Lee, et al. 2005.
415: \bibitem[Maraston \& Thomas (2000)]{maraston00} Maraston, C. \& Thomas, D.
416: 2000, \apj, 541, 126
417: \bibitem[Martin et al.(2004)]{martin04} Martin, N.~F., Ibata,
418: R.~A., Bellazzini, M., Irwin, M.~J., Lewis, G.~F., \& Dehnen, W.\ 2004,
419: \mnras, 348, 12
420: \bibitem[McLaughlin(2000)]{mclaughlin00} McLaughlin, D.~E.\ 2000,
421: \apj, 539, 618
422: \bibitem[Morrison(1993)]{morrison93} Morrison, H.~L.\ 1993, \aj,
423: 105, 539
424: \bibitem[Puzia et al.(2002)]{puzia02} Puzia, T.~H., Saglia,
425: R.~P., Kissler-Patig, M., Maraston, C., Greggio, L., Renzini, A., \&
426: Ortolani, S.\ 2002, \aap, 395, 45
427: \bibitem[Puzia et al.(2005a)]{puzia05} Puzia, T.~H.,
428: Kissler-Patig, M., Thomas, D., Maraston, C., Saglia, R.~P., Bender, R.,
429: Goudfrooij, P., \& Hempel, M.\ 2005a, \aap, 439, 997
430: \bibitem[Puzia et al.(2005b)]{puzia05b} Puzia, T.~H., Perrett,
431: K.~M., \& Bridges, T.~J.\ 2005b, \aap, 434, 909
432: \bibitem[Sarajedini et al.(1998)]{sarajedini98} Sarajedini, A.,
433: Geisler, D., Harding, P., \& Schommer, R.\ 1998, \apjl, 508, L37
434: \bibitem[Sarajedini et al.(2000)]{sarajedini00} Sarajedini, A.,
435: Geisler, D., Schommer, R., \& Harding, P.\ 2000, \aj, 120, 2437
436: \bibitem[Thomas et al.(2003)]{thomas03} Thomas, D., Maraston, C., \&
437: Bender, R.\ 2003, MNRAS, 339, 897
438: \bibitem[Thomas et al.(2004)]{thomas04} Thomas, D.,
439: Maraston, C., \& Korn, A.\ 2004, MNRAS, 351, L19
440: \bibitem[Warner et al.(1973)]{warner73} Warner, P.~J., Wright,
441: M.~C.~H., \& Baldwin, J.~E.\ 1973, \mnras, 163, 163
442: \bibitem[Worthey(1994)]{worthey94} Worthey, G.\ 1994, ApJS, 95, 107
443: \bibitem[Worthey \& Ottaviani(1997)]{worthey97} Worthey, G., \& Ottaviani
444: D.L. 1997, ApJS, 111, 377
445:
446: \end{thebibliography}
447:
448: %\clearpage
449:
450: %\begin{figure}
451: %\centering
452: %\includegraphics[bb= 0 0 400 390, width=8cm]{f1.eps}
453: %\caption{The blue portion of the background subtracted spectra of nine
454: %M33 globular clusters, taken with MMT/Hectospec are shown. The spectra
455: %have been normalized by a spline fit to their continua.}
456: %\label{fig1}
457: %\end{figure}
458:
459: %\clearpage
460:
461: %\begin{figure}
462: %\includegraphics[bb= 0 0 400 400, width=5.0cm]{f2a.eps}
463: %\includegraphics[bb= 0 0 400 400, width=5.0cm]{f2b.eps}
464: %\includegraphics[bb= 0 0 400 400, width=5.0cm]{f2c.eps}
465: %\caption{Age-metallicity diagnostic plots for the sample M33 globular
466: %clusters, constructed from three different Balmer line indices:
467: %H$\delta_{\rm A}$ ({\it left panel}), H$\gamma_{\rm A}$ ({\it middle
468: %panel}), and H$\beta$ ({\it right panel}) vs.~[MgFe]\arcmin\
469: %\citep[see also Fig.~9 of][for details]{puzia05}. Over-plotted are SSP models of
470: %\cite{thomas04} for [$\alpha$/Fe]~$=0.3$, metallicities [Z/H]=$-2.25,
471: %-1.35, -0.33, 0.00$, 0.35, and 0.67 dex ({\it dashed lines}), and ages
472: %15, 10, 5, 3, 2, 1 ({\it solid lines}). The thick solid line is the 5
473: %Gyr isochrone, and is used to split between old and intermediate age
474: %globular clusters. Note that some clusters are not plotted in the
475: %individual panels due to contamination of their index passband
476: %definitions by bad pixels. Clusters with H$\alpha$ observed in
477: %emission are indicated by open circles. All clusters are labeled as in
478: %Fig.~\ref{fig1}. We point out that the Lick index measurements are not
479: %calibrated to the Lick system so that only relative measures should be
480: %derived from the shown diagrams.}
481: %\label{ps:am}
482: %\end{figure}
483:
484: %\clearpage
485:
486: %\begin{figure}
487: %\centering
488: %\includegraphics[bb=46 291 585 751,width=8.5cm]{f3.eps}
489: %\caption{$V-I$ versus $I$ CMD for M33 clusters C38 and M9. The
490: %CMDs have been restricted to include stars in an annulus between 1 and
491: %2\arcsec\ away from the center of each cluster, in order to minimize
492: %blending and contamination. A Girardi isochrone for an age of 2 Gyr and
493: %$Z=0.004$ is overplotted onto the C38 CMD, which shows the red clump stars
494: %and a hint of a main sequence turnoff. A fiducial HB for the Galactic
495: %globular cluster M3 \citep{johnson98} is overplotted onto the M33-M9 CMD,
496: %which clearly reveals the blue horizontal branch stars in this cluser.}
497: %\label{fig:cmd}
498: %\end{figure}
499:
500:
501:
502:
503:
504:
505:
506:
507:
508:
509:
510:
511:
512:
513: \end{document}
514: