1: %%
2: %% Beginning of file 'ms.tex'
3: %%
4: %% Multiple stellar populations in LMC star clusters.
5: %% Second revised version, 22/05/2008.
6: %%
7:
8: %% The command below calls the preprint style
9: %% which will produce a one-column, single-spaced document.
10: %% Examples of commands for other substyles follow. Use
11: %% whichever is most appropriate for your purposes.
12: %%
13: \documentclass[12pt,preprint]{aastex}
14:
15: %% manuscript produces a one-column, double-spaced document:
16: %% \documentclass[manuscript]{aastex}
17:
18: %% These commands will produce file in emulated formatted
19: %% ApJL style (two columns, fonts, etc)
20: %% \documentclass[apjl]{emulateapj}
21: %% \usepackage{apjfonts}
22:
23: %% preprint2 produces a double-column, single-spaced document:
24: %% \documentclass[preprint2]{aastex}
25:
26: %% Sometimes a paper's abstract is too long to fit on the
27: %% title page in preprint2 mode. When that is the case,
28: %% use the longabstract style option.
29: %% \documentclass[preprint2,longabstract]{aastex}
30:
31: %% If you want to create your own macros, you can do so
32: %% using \newcommand. Your macros should appear before
33: %% the \begin{document} command.
34:
35: \newcommand{\myemail}{dmy@roe.ac.uk}
36:
37: %% You can insert a short comment on the title page using the command below.
38: %% \slugcomment{Draft 21/09/06}
39:
40: %% If you wish, you may supply running head information, although
41: %% this information may be modified by the editorial offices.
42: %% The left head contains a list of authors,
43: %% usually a maximum of three (otherwise use et al.). The right
44: %% head is a modified title of up to roughly 44 characters.
45: %% Running heads will not print in the manuscript style.
46:
47: \shorttitle{Multiple stellar populations in LMC clusters}
48: \shortauthors{Mackey et al.}
49:
50: %% This is the end of the preamble. Indicate the beginning of the
51: %% paper itself with \begin{document}.
52:
53: \begin{document}
54:
55: \title{Multiple stellar populations in three rich Large Magellanic Cloud star clusters\altaffilmark{1}}
56:
57: %% Use \author, \affil, and the \and command to format
58: %% author and affiliation information.
59: %% Note that \email has replaced the old \authoremail command
60: %% from AASTeX v4.0. You can use \email to mark an email address
61: %% anywhere in the paper, not just in the front matter.
62: %% As in the title, use \\ to force line breaks.
63:
64: \author{A.D. Mackey\altaffilmark{2}, P. Broby Nielsen\altaffilmark{2}, A.M.N. Ferguson\altaffilmark{2}, J.C. Richardson\altaffilmark{2}}
65:
66: %% Notice that each of these authors has alternate affiliations, which
67: %% are identified by the \altaffilmark after each name. Specify alternate
68: %% affiliation information with \altaffiltext, with one command per each
69: %% affiliation.
70:
71: \altaffiltext{1}{Based on observations made with the NASA/ESA Hubble Space
72: Telescope, obtained at the Space Telescope Science Institute, which is operated
73: by the Association of Universities for Research in Astronomy, Inc., under NASA
74: contract NAS 5-26555.}
75: \altaffiltext{2}{Institute for Astronomy, University of Edinburgh, Royal Observatory,
76: Blackford Hill, Edinburgh, EH9 3HJ, UK}
77:
78: %% Mark off your abstract in the ``abstract'' environment. In the manuscript
79: %% style, abstract will output a Received/Accepted line after the
80: %% title and affiliation information. No date will appear since the author
81: %% does not have this information. The dates will be filled in by the
82: %% editorial office after submission.
83:
84: \begin{abstract}
85: We present deep color-magnitude diagrams for three rich intermediate-age star
86: clusters in the LMC, constructed from archival ACS F435W and F814W
87: imaging. All three clusters exhibit clear evidence for peculiar main-sequence turn-offs.
88: NGC 1846 and 1806 each possess two distinct turn-off branches, while the turn-off for
89: NGC 1783 shows a much larger spread in color than can be explained by the photometric
90: uncertainties. We demonstrate that although all three clusters contain significant
91: populations of unresolved binary stars, these cannot be the underlying cause of
92: the observed turn-off morphologies. The simplest explanation is that each cluster
93: is composed of at least two different stellar populations with very similar metal
94: abundances but ages separated by up to $\sim 300$ Myr. The origin of these unusual
95: properties remains unidentified; however, the fact that at least three massive clusters
96: containing multiple stellar populations are now known in the LMC suggests a potentially
97: significant formation channel.
98: \end{abstract}
99:
100: %% Keywords should appear after the \end{abstract} command. The uncommented
101: %% example has been keyed in ApJ style. See the instructions to authors
102: %% for the journal to which you are submitting your paper to determine
103: %% what keyword punctuation is appropriate.
104:
105: \keywords{galaxies: star clusters --- globular clusters: general --- Magellanic Clouds}
106:
107: %% From the front matter, we move on to the body of the paper.
108: %% In the first two sections, notice the use of the natbib \citep
109: %% and \citet commands to identify citations.
110:
111: \section{Introduction}
112: It is a long-held notion that rich star clusters, including globular
113: clusters, are composed of single stellar populations -- that is, that each
114: such object is made up of stars of a uniform age and chemical composition.
115: New observations are challenging this accepted picture, however.
116: It has recently been discovered that several of the most
117: massive Galactic globular clusters harbor multiple stellar populations
118: with a wide variety of unexpected characteristics.
119: The most striking example is $\omega$ Centauri, which exhibits
120: at least four different populations covering a wide spread in metal abundance
121: and age across the sub-giant branch (SGB) \citep{villanova:07}, as well
122: as a main sequence bifurcation which implies the
123: presence of a vastly helium-enriched population \citep{piotto:05}.
124: NGC 2808 possesses a triple main sequence split, suggesting two
125: helium-enriched populations \citep{piotto:07}; while NGC 1851
126: and NGC 6388 show clear splits in their SGBs, possibly indicative of
127: populations with ages $\sim 1$ Gyr apart, or of sizeable intracluster
128: variations in chemical composition \citep{milone:08,piotto:08,salaris:08}.
129: In addition, several massive globular clusters in M31, including G1,
130: apparently possess significant internal spreads in metal abundance
131: \citep{meylan:01,fuentescarrera:08}.
132: Overall, these observed properties pose serious challenges
133: for conventional models of globular cluster formation and evolution.
134:
135: Rich intermediate-age star clusters in the Magellanic Clouds have the potential
136: to open a new angle on this problem. These objects have masses comparable to many
137: Galactic globulars (although are an order of magnitude less massive than
138: clusters like NGC 2808), but possess color-magnitude diagrams (CMDs) far more
139: sensitive to internal dispersions in age, for example. We recently discovered
140: that the massive LMC cluster NGC 1846 possesses a CMD exhibiting two clearly distinct
141: main-sequence turn-offs (MSTOs) \citep{mackey:07}, which we interpreted as belonging
142: to two stellar populations with similar metal abundance but ages $\approx 300$
143: Myr apart. Motivated by this discovery we have searched the HST/ACS archive for
144: additional observations of both NGC 1846 and other rich intermediate-age LMC
145: clusters. In this Letter we report on the first results from this search, presenting
146: a new deep CMD for NGC 1846 as well as CMDs for two additional objects -- NGC 1783
147: and 1806 -- that exhibit peculiar MSTOs. Similar results for these clusters have
148: simultaneously and independently been obtained by another group \citep{goudfrooij:08}.
149:
150: \section{Observations and data reduction}
151: ACS/WFC imaging of the three clusters was taken through the F435W and F814W filters
152: as part of HST program GO 10595 (P.I. Goudfrooij), between 2005 September 29 -- 2006
153: January 14. Each cluster was imaged three times per filter -- two long exposures of
154: duration 340s each, and one short exposure of duration 90s in F435W and 8s in F814W.
155:
156: We used the ACS module of the {\sc dolphot} photometry software \citep{dolphin:00}
157: to photometer the images. {\sc dolphot} performs point-spread function (PSF) fitting
158: using PSFs especially tailored to the ACS cameras, and works directly on
159: flat-fielded images from the STScI archive relative to some deep reference image
160: (we used the drizzled combination of the two long-exposure F814W images for each
161: cluster). {\sc dolphot} accounts for the hot-pixel and cosmic-ray masking
162: information provided with each flat-fielded image, fits the sky locally around each
163: detected source, and provides magnitudes corrected for charge-transfer efficiency
164: (CTE) effects on the calibrated VEGAMAG scale of \citet{sirianni:05}.
165:
166: For a given cluster we found that the narrowest CMD sequences were obtained
167: by treating the four long-exposure images together and the two short-exposure images
168: together in two separate {\sc dolphot} runs. We used the photometric quality
169: information provided by {\sc dolphot} to clean the two resulting detection lists before
170: combining them. In each list we selected objects of stellar type, with valid measurements
171: on all input images, global sharpness parameter between $\pm 0.3$ in each filter, and
172: ``crowding'' parameter less than $0.5$ mag in each filter. We combined the
173: cleaned lists by selecting all long-exposure detections fainter than, and all short-exposure
174: detections brighter than $0.5$ mag below the long-exposure saturation level at
175: $m_{{\rm F435W}} \approx 19$. A small region of overlap between the lists was used to
176: calculate and eliminate any small global offsets (typically\ $\la 0.02$ mag).
177:
178: \section{Analysis}
179: Our CMDs for NGC 1846, 1783, and 1806 are presented in Fig.~\ref{f:cmds}. In these plots,
180: we have isolated the cluster sequences from the surrounding low-level field populations by
181: imposing radial cuts of $30\arcsec$ and $60\arcsec$ as marked. Full CMDs are shown in the
182: left-hand column, while the second column shows close-ups of the cluster MSTOs and the
183: third column presents Hess diagrams of these MSTO regions. In the right-hand column are
184: CMDs constructed by taking stars as far as possible from each cluster center over areas
185: equivalent to those used for the central MSTO extractions. These outer CMDs likely still
186: contain a few cluster members, and are hence not completely pure field samples. Nonetheless,
187: they clearly demonstrate that the unusual CMD features described below are not the result
188: of field star contamination.
189:
190: All three clusters exhibit striking, unusual MSTOs. That for NGC 1846 possesses two clear
191: branches, as demonstrated previously by \citet{mackey:07} from shallower ACS observations
192: (S$/$N$\,\sim 150$ at the MSTO, compared with S$/$N$\,\sim 250$ here). The
193: MSTO for NGC 1806 also possesses two clear branches, spaced similarly to those of NGC 1846.
194: In contrast, the MSTO for NGC 1783 is not evidently split into branches; instead, the spread
195: in color of the MSTO for this cluster is much larger than can be explained by the
196: photometric uncertainties (marked in Fig.~\ref{f:cmds}; note also the narrowness of the
197: main sequence below the turn-off). This may represent individual MSTO branches
198: which are closely spaced and unresolved by the observations, or a smoothly spread distribution
199: of stars. We find no evidence in the three clusters that stars belonging to different
200: MSTO sections (e.g., upper/lower branches) possess different spatial distributions.
201: \citet{mucciarelli:07} previously noted a spread in color about the MSTO of NGC
202: 1783; however, their CMD was from shallower data such that the observed spread was
203: consistent with that expected from the photometric uncertainties.
204:
205: Apart from the peculiar MSTO morphologies, the CMDs in Fig.~\ref{f:cmds} are as expected
206: for intermediate-age clusters. The main sequences, sub-giant branches (SGBs) and red giant
207: branches (RGBs) are narrow and the red clumps are well defined, implying that the MSTO
208: structures are not due to significant line-of-sight depth or differential reddening
209: in these clusters. The narrow CMD sequences further suggest minimal internal dispersions
210: in $[$Fe$/$H$]$; however we note they place no constraints on the possibility of internal
211: variations in other chemical abundances -- for example, CN, O, Na, or $[\alpha/$Fe$]$ --
212: as are observed for several of the peculiar massive Galactic globular clusters
213: \citep[e.g.,][]{piotto:08}.
214:
215: Also evident on each cluster CMD is a clear spread of stars above and to the red of
216: the main sequence, implying a non-negligible population of unresolved binary systems
217: \citep[e.g.,][]{hurley:98}. To quantify this in a given cluster, we
218: selected all stars within $30\arcsec$ of the center
219: and in the range $22 \le m_{{\rm F435W}} \le 23.2$ (i.e., along the upper main sequence).
220: We fit a ridgeline to the main sequence, and constructed a histogram of the spread in
221: stellar color about this peak. The dispersion to the blue represents photometric
222: uncertainties only; to the red, there is an additional component due to binaries.
223: Reflecting the blue-side dispersion to the red and subtracting this from the histogram
224: thus allows the number of binary stars to be approximately determined.
225: Because of the presence of photometric uncertainties, only binaries with mass ratios
226: sufficiently large that they lie more than a few$\,\times\,0.01$ mag in color from
227: the main sequence peak remain after this subtraction -- experiments with synthetic CMDs
228: (see below) suggest a limiting mass ratio of $q \sim 0.5$. We measured binary fractions
229: of $\approx 40\%$ for each of our three clusters, from $\sim1500$ stars per
230: histogram. Unresolved binaries can strongly influence a cluster's MSTO morphology and
231: it is therefore important to assess whether these significant inferred populations
232: can explain the peculiar MSTOs. This is unlikely to be the case, since
233: inspection of Fig. \ref{f:cmds} reveals that each peculiar MSTO clearly branches from
234: the {\it single star} main sequence rather than from the binary sequence.
235: We explore this question in more detail below with the aid of synthetic CMDs.
236:
237: The most straightforward interpretation of our CMDs is that each cluster possesses at
238: least two stellar populations with very similar $[$Fe$/$H$]$ but differing ages.
239: Since the possibility of variations in other chemical abundances is as-yet unconstrained,
240: in what follows we adopt the simplest assumption of complete chemical homogeneity in each system.
241:
242: To quantify the properties of each cluster's stellar populations in this scenario, we
243: generated a grid of isochrones in the ACS/WFC (F435W, F814W) plane spanning certain
244: ranges of age and $[$M$/$H$]$, and searched this grid to find the isochrone with a shape
245: most closely matching that of the relevant CMD. We adopted recent isochrones from the
246: Padova group \citep{marigo:08}, using their interactive web form
247: ({\it http://stev.oapd.inaf.it/cmd}) to generate a fine grid sampling an age
248: range $1.0 \le \tau \le 3.0$ Gyr at intervals of $0.1$ Gyr, and a metal abundance range
249: $0.0025 \le Z \le 0.0095$ at intervals of $\Delta Z = 0.0005$. These values are
250: appropriate for NGC 1846 and 1783 which have $\tau \sim 1.5-2.0$ Gyr and $[$M$/$H$]\sim -0.4$
251: \citep[e.g.,][]{mackey:07,mucciarelli:07}, as well as for NGC 1806 which is poorly studied
252: but possesses a very similar CMD to that of NGC 1846.
253:
254: Full details of the isochrone fitting procedure are presented by \citet{mackey:07}.
255: For a given cluster we first matched isochrones to the upper MSTO branch; for NGC 1783 we
256: used the upper envelope of the MSTO with the aim of determining the maximum age spread
257: allowed by the width of this feature. Having calculated the necessary shifts in color
258: and magnitude required to overlay the best-fitting isochrone to the upper MSTO branch,
259: we next applied our technique to the lower MSTO branch (the lower envelope of the MSTO
260: for NGC 1783) with the added constraint that the best-fitting isochrone must have
261: equivalent such shifts. This ensured a consistent solution for each cluster in the
262: absence of any significant line-of-sight depth or differential reddening.
263:
264: Our results are presented in Fig.~\ref{f:isofit}. In all three cases the highest
265: quality solutions were for $Z=0.0075$ isochrones. This corresponds to
266: $[$M$/$H$]\approx -0.40$, which, as noted above, matches previous spectroscopic and
267: photometric estimates for both NGC 1846 and 1783. In NGC 1846 and 1806,
268: the difference in age between the two MSTO branches is $\approx 300$ Myr, with the
269: oldest branch representing a population of age $\sim 1.9$ Gyr and the youngest
270: branch a population of age $\sim 1.6$ Gyr in both clusters. In NGC 1783, the upper
271: and lower envelopes of the MSTO imply a maximum age spread of $\approx 400$ Myr, with
272: mean age $\sim 1.8$ Gyr. \citet{mucciarelli:07} previously found
273: $\tau\approx 1.4\pm 0.2$ Gyr for this cluster using
274: isochrones from the Pisa Evolutionary Library with mild overshooting.
275:
276: Our best-fitting isochrones match the cluster main-sequences, MSTOs, SGBs, lower RGBs,
277: and red clumps well. The primary discrepancy between the isochrones and CMDs occurs
278: in each case on the upper RGB where the isochrones appear bluer than the observed stars.
279: This may represent some unidentified limitation in either the stellar models or our
280: photometry. Selecting isochrones sufficiently metal-rich to closely trace the upper
281: RGB sequences results in badly degraded fits to the other CMD regions, particularly
282: the main-sequences and MSTO shapes. Our fitting technique allows rough
283: uncertainties to be placed on our best values for age and metal abundance. Moving
284: to isochrones adjacent on the grid to the best models already leads to noticeably
285: poorer fits -- hence we conservatively estimate random uncertainties of $\pm 0.001$
286: in $Z$ and $\pm 0.1$ Gyr in relative age.
287:
288: We next constructed synthetic CMDs to investigate the role played by unresolved
289: binary stars around the MSTO regions, and in particular investigate whether these
290: objects might reproduce the observed MSTO structures. To do this we generated $5000$
291: stellar detections of which $2000$ ($40\%$) were unresolved binaries. These numbers
292: approximately match those for the CMD of NGC 1846. We drew single star and
293: primary masses (and hence luminosities) randomly from a uniform distribution along
294: the $Z = 0.0075$, $\tau = 1.9$ Gyr isochrone, which best-fit the lower MSTO branch
295: in NGC 1846. In the scenario where the MSTO structure is due to binaries, the lower
296: MSTO branch must represent single stars and the upper branch the binary sequence. To
297: produce the binary sequence we generated a secondary star for each primary by randomly
298: selecting a mass ratio from a uniform distribution in the range $0.5 - 1.0$.
299: This synthesis procedure is clearly simplistic -- more realistic would be to
300: generate stellar masses along the isochrone by randomly selecting from an appropriate
301: mass function, as well as allowing a non-uniform distribution of mass ratios.
302: We emphasize, however, that our aim here is to observe the area on the CMD
303: occupied by the binary sequence. Varying the generation algorithms presently represents
304: an unnecessary level of complexity.
305:
306: The upper row in Fig.~\ref{f:synth} shows our results.
307: The left panel is the synthetic CMD generated as described above. In the right
308: panel we added random photometric uncertainties to the F435W and F814W
309: magnitudes of each star in this synthetic CMD by measuring the uncertainties
310: determined by {\sc dolphot} for detected stars in NGC 1846 of equivalent
311: brightness and color, and then generating errors from Gaussian distributions
312: with these widths. These plots clearly demonstrate that unresolved binaries
313: cannot alone reproduce the peculiar MSTO structures.
314: Specifically, the binary sequence does not result in a distinct branch above and
315: to the blue of the lower MSTO branch, as is observed for NGC 1846 (Fig. \ref{f:cmds}).
316:
317: We next altered our synthesis procedure to use both best-fitting isochrones
318: for NGC 1846 (Fig.~\ref{f:synth}, lower panels).
319: This new CMD closely resembles the observed CMD, demonstrating that
320: two stellar populations in NGC 1846 (and by extension NGC 1806) are necessary
321: to explain the observed MSTO structure. At least some of the spread of stars
322: between the two branches in NGC 1846 must be due to unresolved binaries; however,
323: it is intriguing that the two branches each appear to have a larger spread than
324: those in the synthetic CMD, which are quite well defined even with the addition of
325: photometric uncertainties. We do not assert here that this spread is intrinsic,
326: merely that it is worthy of more detailed investigation.
327:
328: \section{Concluding remarks}
329: In this Letter we have demonstrated that three rich intermediate-age LMC star
330: clusters are comprised of multiple stellar populations. Several other LMC
331: objects with similar properties have previously been tentatively identified.
332: The MSTO in NGC 2173 ($\tau \sim 1.5$ Gyr) suggests a possible $\sim 300$ Myr
333: internal age dispersion \citep{bertelli:03}, while NGC 1868 may exhibit a weak
334: secondary MSTO \citep{santiago:02}. The young cluster NGC 2011 apparently
335: possesses twin main sequences \citep{gouliermis:06}.
336:
337: It is difficult to imagine how our three LMC clusters, each an order of magnitude
338: less massive than the peculiar Galactic globulars, might have retained sufficient
339: gas to undergo multiple widely-separated episodes of star formation. Previous
340: work has shown that the time-dependent potential of a forming star cluster may
341: allow a population of older field stars to be captured \citep[e.g.,][]{pflamm:07};
342: however, such models cannot convincingly account for the large fraction of captured
343: stars ($\sim 50\%$) required for our clusters, nor the apparent homogeneity
344: in age and metal abundance of the older population within each cluster.
345: A plausible alternative to these suggestions is that each of our three objects is
346: the merger product of two or more clusters formed separately within a single giant
347: molecular cloud \citep{mackey:07}. The LMC possesses numerous ``double'' star
348: clusters (e.g., NGC 1850) and its low tidal field could facilitate such mergers.
349: Large-scale realistic $N$-body modelling, along with measurements of the structures
350: and internal kinematics of the three present clusters should allow strong constraints
351: to be placed on this possibility.
352:
353: \acknowledgments
354: ADM and AMNF are supported by a Marie Curie Excellence Grant from the European
355: Commission under contract MCEXT-CT-2005-025869.
356: %We are grateful to the referee,
357: %Giampaolo Piotto, for constructive comments which helped improve this paper.
358:
359: %% To help institutions obtain information on the effectiveness of their
360: %% telescopes, the AAS Journals has created a group of keywords for telescope
361: %% facilities. A common set of keywords will make these types of searches
362: %% significantly easier and more accurate. In addition, they will also be
363: %% useful in linking papers together which utilize the same telescopes
364: %% within the framework of the National Virtual Observatory.
365: %% See the AASTeX Web site at http://www.journals.uchicago.edu/AAS/AASTeX
366: %% for information on obtaining the facility keywords.
367:
368: %% After the acknowledgments section, use the following syntax and the
369: %% \facility{} macro to list the keywords of facilities used in the research
370: %% for the paper. Each keyword will be checked against the master list during
371: %% copy editing. Individual instruments or configurations can be provided
372: %% in parentheses, after the keyword, but they will not be verified.
373:
374: {\it Facilities:} \facility{HST (ACS)}.
375:
376: %% The reference list follows the main body and any appendices.
377: %% Use LaTeX's thebibliography environment to mark up your reference list.
378: %% Note \begin{thebibliography} is followed by an empty set of
379: %% curly braces. If you forget this, LaTeX will generate the error
380: %% "Perhaps a missing \item?".
381:
382: \begin{thebibliography}{}
383: \bibitem[Bertelli et al.(2003)]{bertelli:03} Bertelli, G., Nasi, E., Girardi, L.,
384: Chiosi, C., Zoccali, M., \& Gallart, C. 2003, AJ, 125, 770
385: \bibitem[Dolphin(2000)]{dolphin:00} Dolphin, A. E. 2000, PASP, 112, 1383
386: \bibitem[Fuentes-Carrera et al.(2008)]{fuentescarrera:08} Fuentes-Carrera, I.,
387: Jablonka, P., Sarajedini, A., Bridges, T., Djorgovski, G., \& Meylan, G. 2008, A\&A, in press
388: \bibitem[Goudfrooij et al.(2008)]{goudfrooij:08} Goudfrooij, P., Puzia, T. H.,
389: Kozhurina-Platais, V., \& Chandar, R. 2008, in preparation
390: \bibitem[Gouliermis et al.(2006)]{gouliermis:06} Gouliermis, D. A., Lianou, S.,
391: Kontizas, M., Kontizas, E., \& Dapergolas, A. 2006, ApJ, 652, L93
392: \bibitem[Hurley \& Tout(1998)]{hurley:98} Hurley, J. R., \& Tout, C. A. 1998, MNRAS, 300, 977
393: \bibitem[Mackey \& Broby Nielsen(2007)]{mackey:07} Mackey, A. D., \& Broby Nielsen P.
394: 2007, MNRAS, 379, 151
395: \bibitem[Marigo et al.(2008)]{marigo:08} Marigo, P., Girardi, L., Bressan, A.,
396: Groenewegen, M. A. T., Silva, L., \& Granato, G. L. 2008, A\&A, in press
397: \bibitem[Meylan et al.(2001)]{meylan:01} Meylan, G., Sarajedini, A., Jablonka, P.,
398: Djorgovski, S., Bridges, T., \& Rich, R. 2001, AJ, 122, 830
399: \bibitem[Milone et al.(2008)]{milone:08} Milone, A. P., et al. 2008, ApJ, 673, 241
400: \bibitem[Mucciarelli et al.(2007)]{mucciarelli:07} Mucciarelli, A., Origlia L., \&
401: Ferraro, F. R. 2007, AJ, 134, 1813
402: \bibitem[Pflamm-Altenburg \& Kroupa(2007)]{pflamm:07} Pflamm-Altenburg, J., \&
403: Kroupa, P. 2007, MNRAS, 375, 855
404: \bibitem[Piotto et al.(2005)]{piotto:05} Piotto, G., et al. 2005, ApJ, 621, 777
405: \bibitem[Piotto et al.(2007)]{piotto:07} Piotto, G., et al. 2007, ApJ, 661, L53
406: \bibitem[Piotto(2008)]{piotto:08} Piotto, G. 2008, Mem.S.A.It., 79, 3
407: \bibitem[Salaris et al.(2008)]{salaris:08} Salaris, M., Cassisi, S., \& Pietrinferni, A.
408: 2008, ApJ, 678, L25
409: \bibitem[Santiago et al.(2002)]{santiago:02} Santiago, B., Kerber, L., Castro, R.,
410: \& de Grijs, R. 2002, MNRAS, 336, 139
411: \bibitem[Sirianni et al.(2005)]{sirianni:05} Sirianni, M., et al. 2005, PASP, 117, 1049
412: \bibitem[Villanova et al.(2007)]{villanova:07} Villanova, S., et al. 2007, ApJ, 663, 296
413: \end{thebibliography}
414:
415: \clearpage
416:
417: \begin{figure}
418: \begin{center}
419: \epsscale{1.0}
420: \plotone{f1.eps}
421: \end{center}
422: \caption{ACS/WFC CMDs and Hess diagrams for NGC 1846, 1783, and 1806. Radial cuts
423: of $30\arcsec$ or $60\arcsec$ have been imposed as marked. In the second column the
424: sizes of typical photometric uncertainties are indicated. Each Hess diagram has bin
425: sizes of $0.01$ mag in color and $0.05$ mag in $m_{{\rm F435W}}$. The field CMDs
426: consist of stars taken as far as possible from each cluster center (typically beyond
427: $\sim 120\arcsec$) over areas equivalent to those used for the central MSTO extractions.
428: \label{f:cmds}}
429: \end{figure}
430:
431: \begin{figure}
432: \begin{center}
433: \epsscale{1.0}
434: \plotone{f2.eps}
435: \end{center}
436: \caption{Best-fitting isochrones from \citet{marigo:08} overlaid on the three cluster
437: CMDs.\label{f:isofit}}
438: \end{figure}
439:
440: \begin{figure}
441: \begin{center}
442: \epsscale{0.7}
443: \plotone{f3.eps}
444: \end{center}
445: \caption{Synthetic CMDs calculated as described in the text. A single
446: stellar population cannot reproduce the MSTO structure of, e.g.,
447: NGC1846, even with a significant fraction of unresolved binaries; however
448: including two separate populations results in a close match.\label{f:synth}}
449: \end{figure}
450:
451: \end{document}
452:
453: