1: % mn2esample.tex
2: %
3: % v2.1 released 22nd May 2002 (G. Hutton)
4: %
5: % The mnsample.tex file has been amended to highlight
6: % the proper use of LaTeX2e code with the class file
7: % and using natbib cross-referencing. These changes
8: % do not reflect the original paper by A. V. Raveendran.
9: %
10: % Previous versions of this sample document were
11: % compatible with the LaTeX 2.09 style file mn.sty
12: % v1.2 released 5th September 1994 (M. Reed)
13: % v1.1 released 18th July 1994
14: % v1.0 released 28th January 1994
15:
16: \documentclass[useAMS,usenatbib]{mn2e}
17: \usepackage{times}
18: \usepackage{epsfig}
19:
20: % If your system does not have the AMS fonts version 2.0 installed, then
21: % remove the useAMS option.
22: %
23: % useAMS allows you to obtain upright Greek characters.
24: % e.g. \umu, \upi etc. See the section on "Upright Greek characters" in
25: % this guide for further information.
26: %
27: % If you are using AMS 2.0 fonts, bold math letters/symbols are available
28: % at a larger range of sizes for NFSS release 1 and 2 (using \boldmath or
29: % preferably \bmath).
30: %
31: % The usenatbib command allows the use of Patrick Daly's natbib.sty for
32: % cross-referencing.
33: %
34: % If you wish to typeset the paper in Times font (if you do not have the
35: % PostScript Type 1 Computer Modern fonts you will need to do this to get
36: % smoother fonts in a PDF file) then uncomment the next line
37: % \usepackage{Times}
38:
39: %%%%% AUTHORS - PLACE YOUR OWN MACROS HERE %%%%%
40:
41: \newcommand{\ebv}{E(B$-$V)}
42:
43: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
44:
45: \title[An Asssessment of Broadband Optical Colours as Age Indicators for Star Clusters]{An Asssessment of Broadband Optical Colours as Age Indicators for Star Clusters}
46:
47:
48: \author[M. Hancock, B. J. Smith, M. L. Giroux, and C. Struck]{M. Hancock$^{1}$\thanks{E-mail:
49: hancockm@etu.edu (MH); smithbj@etsu.edu (BJS); girouxm@etsu.edu (MLG); curt@iastate.edu (CS)}, B. J. Smith$^{1}$, M. L. Giroux$^{1}$, and C. Struck$^{2}$\\
50: $^{1}$Department of Physics, Astronomy, and Geology, East Tennessee State University, Box 70652, Johnson City, TN 37614\\
51: $^{2}$Department of Physics and Astronomy, Iowa State University, Ames IA 50011}
52:
53:
54: \begin{document}
55:
56: %\date{Accepted 1988 December 15. Received 1988 December 14; in original form 1988 October 11}
57:
58: \pagerange{\pageref{firstpage}--\pageref{lastpage}} \pubyear{2008}
59:
60: \maketitle
61:
62: \label{firstpage}
63:
64: \begin{abstract}
65: We present an empirical assessment of the use of broadband optical
66: colours as age indicators for unresolved extragalactic clusters and
67: investigate stochastic sampling effects on integrated colours. We
68: use the integrated properties of Galactic open clusters as models for
69: unresolved extragalactic clusters. The population synthesis code {\it
70: Starburst99} \citep{lei99} and four optical colours were used to
71: estimate how well we can recover the ages of 62 well-studied Galactic
72: open clusters with published ages. We provide a method for
73: estimating the ages of unresolved clusters and for reliably
74: determining the uncertainties in the age estimates. Our results
75: support earlier conclusions based on comparisons to synthetic
76: clusters, namely the (U$-$B) colour is critical to the estimation of
77: the ages of star forming regions. We compare the observed optical
78: colours with those obtained from {\it Starburst99} using the published
79: ages and get good agreement. The scatter in the
80: (B$-$V)$_{observed}-$(B$-$V)$_{model}$ is larger for lower luminosity
81: clusters, perhaps due to stochastic effects.
82: \end{abstract}
83:
84: \begin{keywords}
85: open clusters and associations: general --- galaxies: star clusters --- galaxies: stellar content --- methods: data analysis
86: \end{keywords}
87:
88:
89: \section{INTRODUCTION}
90:
91: Thousands of luminous young star clusters have been discovered in
92: external galaxies by the Hubble Space Telescope (HST) (e.g.,
93: \citealp{hol92,hol96,whi93,whi95,meu95,joh99,ben02,kee03,han03,whi03,wei04},
94: and references therein). Population synthesis analyses show that
95: these clusters are sometimes very young, with ages of a few to a few
96: tens of Myr. Such age information can be extremely valuable, because
97: it provides clues to cluster formation processes as well as cluster
98: destruction mechanisms. A radial gradient in the ages of clusters
99: within a galaxy can tell us about gas inflow driven by a bar or an
100: interaction. For interacting galaxies, comparison of cluster ages
101: with dynamical models can provide information about star forming
102: mechanisms. For example, gas compression along the Arp 107 tidal arm
103: may have caused a gradient of ages along the arm (e.g.,
104: \citealp{smi05a}), while in M51, a burst of cluster formation occurred
105: at around the time of the latest passage of the companion
106: (\citealp{bas05,lee05}). Cluster synthesis studies have led to the
107: suggestion that there is a high `infant mortality' in star clusters,
108: with many dissolving within 10 Myr (e.g., \citealp{bas05,fal05}).
109: Star clusters are themselves sometimes clustered into complexes with
110: characteristic radii of $\sim$1 kpc (e.g., \citealp{zha01,lar04}), and
111: the more massive clusters tend to be located near the centre of these
112: complexes, suggesting cluster merging. Accurate age dating of young
113: clusters associated with `ultraluminous X-ray sources' (ULXs) can help
114: distinguish between stellar-mass and intermediate-mass (100$-$1000
115: M$_{\sun}$) black holes for the origin of the X-ray emission (e.g.,
116: \citealp{smi05b}).
117:
118:
119: Stellar population synthesis models have evolved dramatically in the
120: last several decades, since the pioneering works where data on
121: globular clusters and individual giant stars were used to fill gaps in
122: the evolutionary tracks to make early colour models (e.g.\
123: \citealp{tin68,tin72,str78}). Recently, several groups have
124: introduced evolutionary synthesis codes, e.g., \citet{bru93} (B\&C),
125: \citet{fri94} (GALEV), \citet{fio97} (PEGASE), and \citet{lei99}
126: (Starburst99). These new models vary in terms of their input physics,
127: stellar spectral libraries and extinction laws. In general, however,
128: there is good agreement among these models (e.g., \citealp{cha96,vaz05}).
129:
130:
131: Unfortunately, however, at the present time, it is not clear how
132: accurate age estimates based on these models are for unresolved
133: extragalactic clusters. In most cases spectra of the targets are not
134: available, so the amount of extinction and metal abundance are
135: unknown. In these cases, one typically compares some combination of
136: UV, optical, and/or near-IR broadband colours to model cluster spectral
137: energy distributions (SED) to simultaneously estimate age and
138: extinction assuming some metallicity (e.g. \citealp{pas03,han03,han07,smi08}).
139:
140: There are several parameters that can affect the colour of a cluster
141: other than age. Reddening plays a very important role, as does the
142: chemical composition. However, in some age ranges and colours,
143: reddening and chemical composition are degenerate with age.
144: Furthermore, for low mass clusters ($\la10^{5}$ M$_{\odot}$), the
145: observed integrated colours can be affected by stochastic sampling of
146: the initial mass function (IMF) (see for example,
147: \citealp{cer06,cer04,cer03} and references therein). For example,
148: the random addition of a small number of high mass stars will affect
149: the integrated colour of a cluster. This suggests that the observed
150: integrated colours of low mass clusters may not be good indicators of
151: age when compared to the integrated colours of a stellar population
152: model with a fully sampled IMF.
153:
154:
155: It is not clear how well a particular set of colours
156: can predict the age of a cluster. Are some colour sets better suited
157: than others? What uncertainties can be expected because of the choice
158: of colours in the comparison? What uncertainties can be expected
159: because of the assumptions made in generating the model SEDs?
160:
161:
162: Several authors have investigated the use of broadband colours as age
163: indicators by comparison of model colours to the colours of synthetic
164: clusters (e.g. \citealp{gil02,and04,deg05}). We use an alternative
165: method by comparing model colours to the integrated colours of
166: resolved Galactic open clusters (OCs). We compare the published
167: integrated colours of well-studied OCs to a set of population
168: synthesis models. These clusters have published ages previously
169: determined by the turn-off of the zero-age main sequence on the H-R
170: diagram. This work parallels that of \citet{pes08}, who do a similar
171: analysis of Magellanic Cloud clusters using both optical and near-IR
172: data.
173:
174:
175: The present paper is organized as follows. In \S2 we describe our
176: sample of OCs. We describe the population synthesis models used in
177: this study in \S3 and the data analysis in \S4. Predicting the ages
178: of the OCs, predicting the amount of extinction and the effects of
179: metallicity, stochastic sampling effects and cluster dissolution
180: are discussed in \S5. Finally, we summarize in \S6.
181:
182: \section{THE DATA SAMPLE}
183:
184: We started with the set of Galactic open clusters
185: from the WEBDA\footnote{http://www.univie.ac.at/webda/}(Web Base
186: Donn\'{e}es Amas) database operated at the Institute for Astronomy of
187: the University of Vienna \citep{mer95}. The WEBDA database contains
188: 379 OCs, with well-determined integrated (U$-$B), (B$-$V), (V$-$R),
189: and (V$-$I) colours, and colour excesses, as found by several authors
190: (\citealp{lat02, bat94, pan89, spa85, sag83, gra65}). We then culled the
191: WEBDA sample to include only the sample of well-studied OCs in
192: \citet{pau06}, who established a list of 72 open clusters with
193: the most accurate known parameters to serve as a standard table for
194: testing isochrones and stellar models. The age uncertainties in
195: \citet{pau06} were determined by measuring the standard deviation of
196: all the published ages in the literature for each of the OCs.
197:
198:
199: Not all of the 4 integrated colours were determined for each of the
200: OCs in the \citet{pau06} standard set. To model observations with
201: unknown dust extinction, we reversed the extinction corrections to the
202: published colours using the published values of \ebv\ and the
203: conversions from \ebv\ to the other colour excesses. We used the
204: conversions in \citet{lat02}, namely E(U$-$B)=0.72 \ebv$+0.05$
205: \ebv$^2$, E(V$-$R)=0.6 \ebv, and E(V$-$I)=1.25 \ebv.
206:
207:
208: From the standard set we created sub-samples of OCs for each of the 4
209: optical colours and 10 different combinations of colours. When
210: multiple integrated colour measurements were available, we adopted
211: the most recently determined values. Our final data sets include OCs
212: with mean age uncertainties of 19\% and ages ranging from 8 Myr to 8.8
213: Gyr. The metallicities associated with this sample ranges from [Fe/H]
214: $\sim$1/7 solar to $\sim$2.3 solar.
215:
216:
217: Unfortunately, the WEBDA database does not give the uncertainties on
218: the total colours for individual OCs. According to \citet{sag83},
219: the maximum uncertainty in the published integrated colours for the
220: WEBDA data set is $\pm0.2$ mag. One of the sources of uncertainty
221: listed is the error in the reddening. Because we have used the
222: published colour excesses to reverse the extinction corrections, we
223: can neglect the extinction uncertainty in the colours. Removing this
224: from the total uncertainty, assuming it was originally added in
225: quadrature, the maximum uncertainty in colour is $\sim$0.14. We
226: assume that all the measured colours have this maximum uncertainty.
227: This assumption further allows us to make fair comparisons of both
228: the accuracy and precision afforded by each colour in age estimation.
229:
230:
231: \section{THE MODEL CLUSTERS}
232:
233:
234: We used a set of evolutionary synthesis models from the {\it
235: Starburst99} (SB99) code \citep{lei99}. We used the new v5.1 code,
236: which includes the Padova asymptotic giant branch (AGB) stellar models
237: \citep{vaz05}. The new version accounts for all stellar phases that
238: contribute to the integrated light of a stellar population with
239: arbitrary age from extreme UV to NIR. Strictly speaking, the Geneva
240: tracks are more appropriate for modeling young clusters, less than 10
241: Myr, when O stars are present. Most of our sample OCs have ages
242: greater than 100 Myr so for simplicity we only consider models with
243: the Padova tracks.
244:
245:
246: Our SB99 model spectral energy distributions (SEDs) were generated
247: assuming a Kroupa initial mass function (IMF) (favors high mass stars)
248: \citep{kro02} with exponents of 1.3 and 2.3 and mass ranges from
249: $0.1-0.5$ M$_{\odot}$ and $0.5-100$ M$_{\odot}$ respectively.
250: We have also assumed instantaneous (single burst) star formation, and solar
251: abundances. It has been demonstrated that adopting different forms of
252: the IMF has a minor impact on optical colours \citep{mac04}, so we
253: have not explored various IMFs. We also generated models with
254: abundances less than (0.2$\times$) and greater than (2.5$\times$)
255: solar. The model SEDs were reddened from 0.0 mag to 2.0 mag in 0.02
256: mag increments using the \citet{car89} reddening law. Finally, the
257: model SEDs were convolved with the Johnson and Kron-Cousins {\it
258: UBVRI} filter bandpasses and the broadband optical colours were
259: determined. The models span a range of ages from 1 Myr to 20 Gyr.
260: From 1 Myr to 1 Gyr we used a step size of 1 Myr; from 1.1 to 20 Gyr,
261: the step size was 100 Myr. Because the models will be compared to the
262: integrated colours of resolved stellar populations, only the stellar
263: contributions were included in the SEDs. Nebular emission lines have
264: been shown to be important in the first $10^7$ yr \citep{and03} and
265: should be considered when studying unresolved stellar populations (see
266: \S5.3).
267:
268:
269: Figure 1 plots the (B$-$V) vs.\ (U$-$B) colour-colour diagram.
270: The black curves are our SB99 models with solar abundances and varying
271: E(B$-$V). From the bottom to the top, E(B$-$V)=0.0, 0.5, 1.0, and 1.5
272: mags. The filled squares are the observed colours of the standard set
273: of open clusters in the WEBDA sample. For each curve, age increases
274: to the right from 1 Myr to 20 Gyr. Note that we have assumed the
275: maximum uncertainty in colour for each OC.
276:
277: \begin{figure}
278: \centering{\epsfig{figure=f1.eps, height=7.0cm}}
279: \caption[f1.eps]{(B$-$V) vs.\ (U$-$B) colour-colour diagram. The black curves are our SB99 models with solar abundances and varying E(B$-$V). From the bottom to the top, E(B$-$V)=0.0 (thick solid curve), 0.5 (dashed curve), 1.0 (dashed curve), and 1.5 (dotted curve) mags. The filled squares are the observed colours of the standard set of open clusters in the WEBDA sample. For each curve, age increases to the right from 1 Myr to 20 Gyr. Note that we have assumed the maximum uncertainty in colour for each OC.} \label{f1}
280: \end{figure}
281:
282:
283: \section{DATA ANALYSIS}
284:
285: This study is not intended as a test of the SB99 model's ability to
286: reproduce the colours of young clusters. This has already been
287: demonstrated (see for e.g., \citealp{vaz05}). We intend to test how
288: well various colours and initial model assumptions can recover the
289: ages of well studied Galactic open clusters. In addition to using the
290: observed colours to estimate ages, we compared the published colours
291: with the model colours calculated assuming the published ages to look
292: for evidence of stochastic sampling effects.
293:
294:
295: To determine the model predicted ages of the OCs we compared the
296: observed integrated colours to each of the reddened (and unreddened)
297: SB99 model colours, for a single assumed metallicity. We used a
298: $\chi^2$ minimization calculation (e.g., \citealp{pas03,deg05}) to determine
299: the best match of the observed colours to the models and hence the
300: ages of the open clusters:
301: \[
302: \chi^2 = \sum^{N}_{i=1}\left(\frac{obs_{i} - model_{i}}{\sigma_{i}}\right)^2
303: \]
304:
305: \noindent where N is the number of colours (1-4) used in the analysis,
306: obs$_i$ is the observed colour, model$_i$ is the corresponding model
307: colour, and $\sigma_i$ is the uncertainty in the obs$_i$ colour. All
308: ages with a fit of $\chi^2\leq$ N were considered good fits. For
309: each colour set there was therefore a range of predicted ages for each
310: OC. The age associated with the minimum $\chi^2$ is taken as the
311: best-fit age. To determine the uncertainties in the predicted age we
312: find the minimum and maximum ages within a $\Delta\chi^2$ defined to
313: give a 68\% confidence level (e.g., \citealp{pre92}). Additionally,
314: we add the model step size to the age uncertainty. We have also
315: predicted the amount of extinction in each case. The amount of
316: extinction applied to the model associated with the best-fit age is
317: the best-fit \ebv. The uncertainty in the predicted \ebv\ is
318: determined from the minimum and maximum \ebv\ within the same
319: $\Delta\chi^2$ mentioned above. Additionally, we add the model step
320: size to the \ebv\ uncertainty.
321:
322: \begin{table*}
323: \centering
324: \begin{minipage}{140mm}
325: \caption[ages colour differences]{Ages and Colour Differences}
326: %\vspace{.2in}
327: \begin{tabular}{lccrrrrrrr}
328: \hline \hline
329: Name & Age$^1$ & Age$^2$ & M$_{V}$ & (U$-$B) & (U$-$B) & diff. & (B$-$V) & (B$-$V) & diff. \\
330: & (pub) & (pred) & & measured & model & & measured & model & \\
331: \hline
332: Bochum 10& 8$\pm$2 & 1$_{-0}^{+31}$ & -6.52 & -1.05 & -0.67 & -0.38 & -0.24 & 0.27 & -0.51 \\
333: NGC 6871& 9$\pm$2 & 23$_{-22}^{+60}$ & -7.28 & -0.89 & -0.80 & -0.09 & -0.24 & 0.17 & -0.41 \\
334: Stock 14& 10$\pm$2 & 31$_{-30}^{+96}$ & -8.25 & -0.39 & -0.76 & 0.37 & 0.34 & 0.17 & 0.17 \\
335: King 12& 11$\pm$1 & 13$_{-12}^{+67}$ & -4.90 & -0.88 & -0.72 & -0.16 & -0.20 & 0.14 & -0.34 \\
336: IC 2581& 13$\pm$3 & 120$_{-114}^{+980}$ & -8.11 & -0.33 & -0.67 & 0.34 & 0.07 & 0.05 & 0.02 \\
337: NGC 3105& 21$\pm$3 & 4$_{-3}^{+49}$ & -6.95 & -0.69 & -0.67 & -0.02 & 0.11 & 0.06 & 0.05 \\
338: NGC 6250& 22$\pm$5 & 66$_{-61}^{+157}$ & -4.69 & -0.59 & -0.66 & 0.07 & -0.16 & 0.06 & -0.22 \\
339: NGC 6396& 30$\pm$8 & 38$_{-37}^{+2862}$ & -5.03 & -0.64 & -0.58 & -0.06 & -0.05 & 0.07 & -0.12 \\
340: Trumpler 1& 30$\pm$6 & 11$_{-10}^{+29}$ & -5.65 & -0.88 & -0.58 & -0.30 & -0.04 & 0.07 & -0.11 \\
341: Lynga 6& 35$\pm$9 & 30$_{-29}^{+103}$ & -4.38 & -0.65 & -0.54 & -0.11 & 0.01 & 0.07 & -0.06 \\
342: Melotte 20& 43$\pm$18 & 83$_{-77}^{+114}$ & -5.71 & -0.42 & -0.48 & 0.06 & 0.02 & 0.07 & -0.05 \\
343: King 10& 45$\pm$11 & 106$_{-100}^{+10594}$ & -5.34 & -0.58 & -0.47 & -0.11 & -0.19 & 0.07 & -0.26 \\
344: NGC 129& 62$\pm$15 & 9$_{-8}^{+20}$ & -6.03 & -0.64 & -0.40 & -0.24 & 0.39 & 0.09 & 0.30 \\
345: NGC 6834& 65$\pm$18 & 131$_{-125}^{+1269}$ & -6.20 & -0.45 & -0.39 & -0.06 & -0.10 & 0.10 & -0.20 \\
346: NGC 6405& 71$\pm$21 & 181$_{-134}^{+146}$ & -3.65 & -0.27 & -0.38 & 0.11 & -0.03 & 0.11 & -0.14 \\
347: Collinder 394& 74$\pm$6 & 97$_{-91}^{+178}$ & -4.74 & -0.26 & -0.38 & 0.12 & 0.23 & 0.11 & 0.12 \\
348: NGC 6025& 74$\pm$22 & 82$_{-76}^{+121}$ & -4.38 & -0.46 & -0.38 & -0.08 & -0.12 & 0.11 & -0.23 \\
349: NGC 5662& 77$\pm$20 & 78$_{-73}^{+1222}$ & -4.38 & -0.24 & -0.37 & 0.13 & 0.30 & 0.11 & 0.19 \\
350: NGC 6087& 78$\pm$19 & 143$_{-137}^{+309}$ & -5.09 & -0.21 & -0.37 & 0.16 & 0.18 & 0.12 & 0.06 \\
351: NGC 1778& 129$\pm$29 & 149$_{-143}^{+659}$ & -4.16 & -0.27 & -0.25 & -0.02 & 0.09 & 0.14 & -0.05 \\
352: NGC 1647& 130$\pm$25 & 111$_{-105}^{+197}$ & -2.35 & -0.35 & -0.25 & -0.10 & 0.07 & 0.14 & -0.07 \\
353: NGC 5316& 166$\pm$33 & 119$_{-113}^{+12681}$ & -4.91 & 0.13 & -0.19 & 0.32 & 0.67 & 0.15 & 0.52 \\
354: NGC 744& 184$\pm$49 & 256$_{-170}^{+1044}$ & -5.17 & -0.14 & -0.16 & 0.02 & 0.06 & 0.15 & -0.09 \\
355: Melotte 105& 224$\pm$53 & 281$_{-183}^{+1119}$ & -4.27 & -0.11 & -0.10 & -0.01 & 0.07 & 0.16 & -0.09 \\
356: NGC 3532& 262$\pm$46 & 243$_{-163}^{+363}$ & -5.50 & -0.03 & -0.04 & 0.01 & 0.22 & 0.18 & 0.04 \\
357: NGC 2548& 364$\pm$102 & 486$_{-341}^{+411}$ & -3.23 & 0.12 & 0.07 & 0.05 & 0.26 & 0.23 & 0.03 \\
358: Melotte 111& 522$\pm$82 & 384$_{-217}^{+273}$ & -1.95 & 0.12 & 0.16 & -0.04 & 0.20 & 0.33 & -0.13 \\
359: NGC 2527& 619$\pm$163 & 317$_{-204}^{+783}$ & -1.90 & 0.09 & 0.18 & -0.09 & 0.30 & 0.39 & -0.09 \\
360: NGC 2266& 736$\pm$77 & 498$_{-348}^{+1302}$ & -3.83 & 0.40 & 0.18 & 0.22 & 0.62 & 0.46 & 0.16 \\
361: Berkeley 2& 794$\pm$1 & 999$_{-953}^{+19101}$ & -4.18 & 0.14 & 0.17 & -0.03 & 0.46 & 0.49 & -0.03 \\
362: NGC 2355& 833$\pm$137 & 238$_{-187}^{+3562}$ & -3.67 & 0.22 & 0.17 & 0.05 & 0.59 & 0.51 & 0.08 \\
363: Berkeley 69& 867$\pm$48 & 506$_{-373}^{+19594}$ & -3.16 & 0.21 & 0.17 & 0.04 & 0.38 & 0.52 & -0.14 \\
364: NGC 2477& 875$\pm$238 & 250$_{-180}^{+10350}$ & -5.70 & 0.17 & 0.17 & 0.00 & 0.49 & 0.53 & -0.04 \\
365: NGC 2192& 1072$\pm$48 & 1000$_{-956}^{+2200}$ & -3.53 & 0.10 & 0.17 & -0.07 & 0.49 & 0.62 & -0.13 \\
366: NGC 2660& 1351$\pm$291 & 694$_{-585}^{+19406}$ & -4.08 & 0.34 & 0.22 & 0.12 & 0.61 & 0.72 & -0.11 \\
367: NGC 1798& 1421$\pm$16 & 47$_{-44}^{+9653}$ & -4.86 & 0.15 & 0.25 & -0.10 & 0.82 & 0.77 & 0.05 \\
368: NGC 2506& 1648$\pm$485 & 259$_{-175}^{+2941}$ & -4.31 & 0.22 & 0.28 & -0.06 & 0.54 & 0.82 & -0.28 \\
369: NGC 7044& 1824$\pm$361 & 19999$_{-19810}^{+102}$ & -3.93 & 0.48 & 0.26 & 0.22 & 0.58 & 0.81 & -0.23 \\
370: Berkeley 32& 3477$\pm$698 & 4100$_{-4094}^{+8600}$ & -2.97 & 0.28 & 0.36 & -0.08 & 0.78 & 0.91 & -0.13 \\
371: NGC 6253& 3949$\pm$1086 & 6800$_{-6794}^{+10399}$ & -2.68 & 0.34 & 0.38 & -0.04 & 0.81 & 0.92 & -0.11 \\
372: NGC 2682& 4093$\pm$958 & 1300$_{-1294}^{+8200}$ & -3.16 & 0.28 & 0.38 & -0.10 & 0.78 & 0.92 & -0.14 \\
373: NGC 6791& 7850$\pm$2026 & 19999$_{-19805}^{+102}$ & -4.14 & 0.82 & 0.46 & 0.36 & 1.02 & 0.97 & 0.05 \\
374: \hline \\
375: \multicolumn{10}{l}{\footnotesize{$^1$ Ages in Myr from \citet{pau06}}}\\
376: \multicolumn{10}{l}{\footnotesize{$^2$ Ages in Myr predicted from (U$-$B) and (B$-$V)}}\\
377: \end{tabular}
378: \end{minipage}
379: \end{table*}
380:
381:
382: In Table 1, we list the subset of our sample that has the {\it U},
383: {\it B}, and {\it V} magnitudes available. Column one is the name of
384: the OC, column two is the published age, column three is the predicted
385: age determined from the (U$-$B) and (B$-$V), column four is M$_{V}$,
386: columns five and six are the reddening-corrected measured and zero
387: reddened model (U$-$B) colours respectively, column seven is the
388: difference of the measured and model (U$-$B), and columns eight, nine
389: and ten are the same but for the (B$-$V) colour. It can be seen from
390: Table 1 that for 74\% of the OCs, the (U$-$B) differences are
391: $\la0.14$, the assumed maximum measurement uncertainty in measured
392: colour, while 60\% of the (B$-$V) differences are $\la0.14$.
393:
394:
395: \section{RESULTS}
396:
397: \subsection{Predicting Ages}
398:
399: In Figure 2, we plot the predicted ages against the published ages, if
400: only (U$-$B) and (B$-$V) are used to estimate the ages. We also provide
401: a histogram showing the distribution of the differences between the
402: predicted and published ages, a plot of the predicted E(B$-$V)s against the
403: published E(B$-$V)s, and a histogram showing the distribution of the
404: differences between the predicted and published E(B$-$V)s.
405: In this Figure, we used the solar abundance models. Plots of the
406: predictions determined with the other colour combinations and at other
407: assumed metallicities are similar so are not shown.
408: We see from this figure that with the combination of (U$-$B) and (B$-$V)
409: the ages of the OCs can be predicted with reasonable accuracy.
410:
411: \begin{figure*}
412: \centering{\epsfig{figure=f2a.eps, height=7.0cm}}
413: \centering{\epsfig{figure=f2b.eps, height=7.0cm}}
414: \centering{\epsfig{figure=f2c.eps, height=7.0cm}}
415: \centering{\epsfig{figure=f2d.eps, height=7.0cm}}
416: \caption[f2a,b,c,d]{ Top Left: Model predicted age versus the published age of the OCs in the WEBDA sample, as determined by the (U$-$B) vs.\ (B$-$V) colours. The solid black line is y=x, the inner dashed lines represent a factor of 3 difference in age, and the outer long-dashed lines represent a factor of 10 difference in age. Top Right: Histogram showing the distribution of the difference between the predicted and published ages. The inner and outer dashed lines represent a factor of 3 and 10 change in age respectively. Bottom Left: Model predicted extinction vs.\ published extinction. The solid black line is y=x, the inner dashed lines represent a change of 0.2 mag, and the outer long-dashed line represents a change of 0.5 mag. Bottom Right: Histogram showing the distribution of the difference between the predicted and published E(B$-$V). The inner and outer dashed lines represent a change of 0.2 and 0.5 mag respectively. }
417: \label{f2}
418: \end{figure*}
419:
420:
421: \begin{table*}
422: %\setlength{\tabcolsep}{0.04in}
423: \centering
424: \begin{minipage}{140mm}
425: \caption[fit-quality]{Prediction Quality}
426: %\vspace{.2in}
427: \begin{tabular}{lrcrlllrrr}
428: \hline \hline
429: Colours & OCs & Good & Rec. & Unc & Unc & Unc & w/in 3x & w/in 3x & w/in 3x \\
430: Used & & Fits$^1$ & Ages$^2$ & down$^2$ & up$^2$ & mean$^2$ & $\odot $ & 0.2 $\odot $ & 2.5 $\odot $ \\
431: \hline
432: U$-$B & 44 & 44/44/44 & 98\% & 1.03 & 2.08 & 1.56 & 45\% & 57\% & 59\% \\
433: B$-$V & 62 & 62/62/62 & 100\% & 1.18 & 2.42 & 1.80 & 45\% & 45\% & 42\% \\
434: V$-$R & 6 & 6/6/6 & 100\% & 1.37 & 2.87 & 2.12 & 33\% & 17\% & 50\% \\
435: V$-$I & 15 & 15/15/15 & 100\% & 1.40 & 2.61 & 2.01 & 47\% & 27\% & 40\% \\
436: U$-$B B$-$V & 42 & 42/42/42 & 98\% & 0.84 & 1.14 & 0.99 & 83\% & 69\% & 74\% \\
437: U$-$B V$-$R & 5 & 5/4/5 & 100\% & 1.33 & 1.48 & 1.40 & 20\% & 25\% & 20\% \\
438: U$-$B V$-$I & 11 & 11/8/11 & 100\% & 0.75 & 1.07 & 0.91 & 45\% & 62\% & 45\% \\
439: B$-$V V$-$R & 6 & 6/6/6 & 100\% & 1.26 & 2.89 & 2.07 & 17\% & 17\% & 17\% \\
440: B$-$V V$-$I & 15 & 15/15/15 & 87\% & 0.69 & 2.89 & 1.79 & 67\% & 13\% & 60\% \\
441: V$-$R V$-$I & 5 & 4/4/5 & 100\% & 0.59 & 3.50 & 2.05 & 75\% & 25\% & 60\% \\
442: U$-$B B$-$V V$-$R & 5 & 5/5/5 & 100\% & 1.16 & 1.65 & 1.40 & 0\% & 80\% & 20\% \\
443: U$-$B B$-$V V$-$I & 11 & 10/7/11 & 100\% & 0.63 & 1.22 & 0.93 & 60\% & 57\% & 55\% \\
444: B$-$V V$-$R V$-$I & 5 & 4/4/5 & 100\% & 0.36 & 3.72 & 2.04 & 50\% & 0\% & 40\% \\
445: U$-$B B$-$V V$-$R V$-$I & 4 & 3/2/3 & 100\% & 0.39 & 1.46 & 0.93 & 67\% & 50\% & 67\% \\
446: \hline \\
447: %\footnotesize{$^1$ solar/0.2 solar/2.5 solar}\\
448: \multicolumn{10}{l}{\footnotesize{$^1$ solar/0.2 solar/2.5 solar}}\\
449: \multicolumn{10}{l}{\footnotesize{$^2$ solar metallicity assumed}}\\
450: \end{tabular}
451: \end{minipage}
452: \end{table*}
453:
454:
455: We present Table 2 to discuss both the accuracy and precision of
456: various colours as age indicators. Column one is the colour
457: combination, column two is the number of OCs in the sample with those
458: colours available, and column three is the number of good fits for the
459: solar, 0.2 solar and 2.5 solar models, where the number of good fits
460: is the number of models in our grid that give $\chi^2\leq$ N. Column
461: four is the percentage of recovered ages, which describes how often
462: the predicted age uncertainties overlapped with the uncertainties in
463: the measured ages when solar metallicity is assumed. The results
464: from comparisons to models of 0.2 solar and 2.5 solar abundances are
465: similar so are not shown. Column five lists the average uncertainty
466: in predicted log(t) in the negative direction for all the OCs, column
467: six lists the average uncertainty in predicted log(t) in the positive
468: direction for all the OCs, and column seven lists the average of the
469: positive and negative uncertainties for all the OCs. It should be
470: noted that the model ages range from a log(t) of 6 to 10.3, therefore
471: 2.15 is the maximum mean positive and negative uncertainty in log(t).
472: Columns eight, nine, and ten are the percentages of predicted ages
473: that are within a factor of 3 of the published ages for the solar
474: abundance model, the 0.2 solar abundance model, and the 2.5 solar
475: abundance model, respectively. This percentage is determined for the
476: best-fit ages without consideration of the predicted or measured age
477: uncertainties. Unfortunately, the only samples of OCs large enough to
478: draw any significant interpretations are the (U$-$B), (B$-$V) and
479: (U$-$B) and (B$-$V). Results from the other combinations of colours
480: are merely suggestive.
481:
482: We see from Table 2 that for several colour combinations the
483: uncertainties in predicted age are relatively large. This is because
484: we have assumed the maximum uncertainty in measured colours for the
485: OCs. Using the actual uncertainties in integrated colours for this
486: sample would not change the predicted ages but would result in better
487: age constraints. Unfortunately the uncertainties in integrated
488: colours of each of these OCs are not available. However, the relative
489: predicted age uncertainties provide a good measure of the precision
490: afforded by each colour and colour combination for age-dating.
491: Looking at the mean uncertainties (column 7 in Table 2) we see that
492: the precision is improved when the (U$-$B) colour is included.
493:
494: We see from this table that by using the (U$-$B) and (B$-$V) colours
495: up to 83\% of the predictions are within a factor of 3 of the
496: published values. For the best-fit (U$-$B), (B$-$V), and (U$-$B) and
497: (B$-$V) samples, we have determined the differences between the
498: best-fit predicted and measured log ages. We find that 68\% of the
499: differences are less than 0.77 dex, 0.86 dex, and 0.44 dex for the
500: (U$-$B), (B$-$V), and (U$-$B) and (B$-$V) samples respectively. This
501: indicates that (U$-$B) alone gives more accurate ages than (B$-$V)
502: alone. We also see that combining (U$-$B) with (B$-$V) further
503: improves the accuracy of the predicted ages.
504:
505: Table 2 shows the importance of the (U$-$B) colour in measuring the
506: age of a cluster. To the extent that our sample size allows such
507: comparisons, combinations of colours that include (U$-$B) tend to
508: have more accurate and more precise age predictions. The fact that
509: single colours do not do a good job of constraining the ages is likely
510: because there are three degenerate unknowns, age, reddening, and
511: metallicity, and two filters are not enough to break these
512: degeneracies. This is consistent with results from \citet{and04}.
513: However, the (U$-$B) colour alone is a better predictor of age than
514: any combination of colours that do not include (U$-$B), regardless of
515: the assumptions about metallicity, consistent with results from
516: \citet{gil02} and \citet{and04}. The (U$-$B) colour reflects the
517: contributions from the young and intermediate age stars. Tracing the
518: younger stars in a population is critical to constrain the age.
519: Colours that do not include {\it U} only follow the slower-evolving
520: stars and are not as sensitive to changes in age.
521:
522:
523: \subsection{Extinction and Metallicity}
524:
525: The extinction is predicted well by the same colours that predict the
526: age well. For example, using the (U$-$B) and (B$-$V) colours
527: we find that 62\% of the \ebv\ estimates are within 0.2 mag of the
528: published values. The colour that is the most sensitive to age,
529: (U$-$B), is once again critical.
530:
531: We get similar numbers of good fits to the cluster ages whether we
532: assume 0.2 solar, solar, or 2.5 solar abundances (see Table 2). This
533: implies that, at least within the metallicity range of our sample OCs,
534: knowing the metallicity is not critical for getting a reasonable fit
535: to the cluster age. Note that we are not fitting for the
536: metallicities of these clusters, but instead are testing to see how
537: well we can recover the correct ages and extinctions with standard
538: solar and near solar metallicity models.
539:
540: \begin{figure}
541: \centering{\epsfig{figure=f3.eps, height=7.0cm}}
542: \caption[f3.eps]{Published M$_{V}$ plotted against the published ages of the open clusters in our sample. The curve approximates the lowest luminosity limit curve in \citet{cer04}. The filled black squares are the OCs fainter than the curve while the filled black triangles are brighter. \label{f3}}
543: \end{figure}
544:
545:
546: \subsection{The Effect of H$\alpha$}
547:
548: The SB99 model colours do not include nebular H$\alpha$ emission.
549: This does not impact our analysis of the integrated colours of the OCs
550: because the stellar components are resolved and no H$\alpha$ is
551: included in either the models or the data. However, this should be
552: considered when models are compared to young unresolved clusters
553: in other galaxies.
554:
555:
556: For clusters younger than $\sim$10 Myr, the H$\alpha$ emission
557: contributes significantly to the Johnson {\it R} band and minimally to the
558: Johnson {\it V} band. From SB99, it is seen that a 1 Myr model has an
559: EW(H$\alpha$) of $\sim$2500 \AA, a 5 Myr model has an EW(H$\alpha$)
560: of $\sim$400 \AA, and by 10 Myr the EW(H$\alpha$) is $\sim$10 \AA.
561: For comparisons to young unresolved clusters, H$\alpha$ emission
562: should be included in the model SEDs. For a discussion about adding
563: H$\alpha$ emission to models to compare to the clusters in Arp 285,
564: see \citet{smi08}.
565:
566:
567:
568: \subsection{The Effect of Stochastic Sampling of the IMF}
569:
570: It has been shown that traditional synthesis models usually return the
571: mean value of the integrated stellar population distribution
572: (e.g. \citealp{cer06,lur06}). While this is correct on average, it
573: is not necessarily correct in individual cases because the mean may not be
574: representative of the actual values. Galactic open clusters are
575: typically low mass so will be subject to the effects of stochastic
576: sampling of the IMF (e.g. \citealp{cer03}).
577:
578: \begin{figure}
579: \centering{\epsfig{figure=f4a.eps, height=7.0cm}}
580: \centering{\epsfig{figure=f4b.eps, height=7.0cm}}
581: \caption[f4a,b]{Top: Distribution of the difference between the published (U$-$B)$_{pub}$ and the model (U$-$B)$_{mod}$ for the open clusters in our sample. The model colours assume the published ages. The colours are reddening-corrected. The histograms are divided into two groups; fainter than and brighter than the lowest luminosity limit curve, as shown in Figure 3. Bottom: Same as above but for the (B$-$V) colours. \label{f4}}
582: \end{figure}
583:
584:
585: \begin{figure}
586: \centering{\epsfig{figure=f5a.eps, height=7.0cm}}
587: \centering{\epsfig{figure=f5b.eps, height=7.0cm}}
588: \caption[f5a,b]{Top: Difference between the published and model (U$-$B) colours against the published ages of the clusters. Bottom: Difference between the published and model (B$-$V) colours against the published ages of the clusters. The two panels describe the same two groups as above. The dashed lines show zero colour difference. \label{f5}}
589: \end{figure}
590:
591:
592: \citet{cer04} define the lowest luminosity limit as a criterion that
593: must be met in order to compare an observed cluster to a model
594: cluster. The total luminosity of the observed cluster must be larger
595: than the individual contribution of any of the stars in the model.
596: According to \citet{cer04}, clusters brighter than this limit may
597: or may not have a well-sampled IMF, but clusters fainter than this
598: limit will be misrepresented by synthesis models.
599:
600: Figure 3 is a plot of the published M$_{V}$ plotted against the
601: published ages of the OCs in our sample. The curve approximates the
602: lowest luminosity limit curve in \citet{cer04}. The filled squares
603: are the OCs fainter than the curve, while the filled triangles are OCs
604: that are brighter. There are 15 OCs fainter and 27 OCs brighter
605: than the curve. From this figure it can be seen that several of
606: the OCs in our sample are brighter than the lowest luminosity limit,
607: but the majority are fainter. This suggests that many clusters in
608: our sample (those fainter) should show these stochastic effects, while
609: several might not (those brighter).
610:
611: Figure 4 shows the distributions of the differences between the
612: published and the model (U$-$B) colours and the published and model
613: (B$-$V) colours for the OCs in our sample. The measured colours are
614: reddening-corrected. The model colours assume the published ages. To
615: make comparisons, only OCs with both integrated (U$-$B) and (B$-$V)
616: colours are included in these distributions. The histograms are
617: divided into 2 groups, those brighter and those fainter than the
618: lowest luminosity curve, as in Figure 3.
619:
620:
621: For the (U$-$B) colour, a Kolmogorov-Smirnov (K-S) test cannot rule out
622: that the brighter and fainter samples originated from the same parent
623: distributions. The rms for the differences in (U$-$B) are 0.16 and
624: 0.17 for the brighter and fainter samples respectively. However, for
625: (B$-$V), a K-S test suggests a probability of only 0.012 that the two
626: samples (brighter and fainter) originate from the same parent
627: distribution. The rms values are 0.15 and 0.25 for the (B$-$V)
628: distributions respectively.
629: The scatter appears slightly larger for the fainter samples,
630: with the (B$-$V) difference being slightly skewed to negative
631: numbers (see Figure 4).
632:
633:
634: Figure 5 shows the differences between the published and model (U$-$B)
635: and (B$-$V) colours plotted against the published ages using the same
636: sample as in Figure 4. The measured colours are reddening-corrected.
637: For each colour difference the OCs are divided into the same bins as
638: described above. From Figure 5 it can be seen that the scatter in the
639: (U$-$B) colour differences do not appear to be correlated with age.
640: The fainter (B$-$V) difference sample appears to have a slight
641: dependence on age in that the published (B$-$V) colours are bluer than
642: the model colours for the younger OCs, consistent with results in
643: \citet{bru02} (see below). Given this, the K-S results, and the
644: larger rms, we might be seeing stochastic sampling effects in the
645: differences between the measured and model (B$-$V) colours.
646:
647:
648:
649: In Figure 6, we plot the published reddening-corrected (U$-$B) colour
650: against the published reddening-corrected (B$-$V) colour for our OCs.
651: This figure is divided into the same two groups described above. Also
652: on this figure are our SB99 models with E(B$-$V) of 0.0 mag (solid
653: curve) and 0.5 mag (dashed curve). This figure shows that the scatter
654: of these colours around the models is not a function of M$_{V}$.
655: Note that relatively young clusters (that is, blue clusters)
656: tend to lie above and to the left of the zero extinction model curve.
657: Figures 4 and 5 show that this is predominately due to the fainter clusters
658: being bluer in (B$-$V) than the models.
659:
660:
661: \begin{figure}
662: \centering{\epsfig{figure=f6.eps, height=7.0cm}}
663: \caption[f6]{Published (U$-$B) against published (B$-$V) for the OCs. The figures are divided into 2 groups as described above. The curves are our SB99 models with E(B$-$V) of 0.0 mag (solid curve) and 0.5 mag (dashed curve). \label{f6}}
664: \end{figure}
665:
666: Our results can be compared to population synthesis studies that
667: explicitly investigate stochastic effects. For example, \citet{bru02}
668: and \citet{bru03} have run Monte Carlo simulations of low mass star
669: clusters, to investigate how stochastic effects can cause observed
670: colours to differ from models with full-sampled IMFs. In their
671: simulations, young model clusters with masses of 10$^3$ M$_{\sun}$
672: also lie above models with fully-sampled IMFs on a (U$-$B) vs. (B$-$V)
673: plot \citep{bru02}, consistent with what we see in Figure 6. Thus at
674: least some of the observed scatter may be due to stochastic effects.
675:
676:
677: From these figures and tables we see suggestive evidence of stochastic
678: sampling of the cluster IMF in this sample of OCs. It is possible
679: that the uncertainties in the measured colours and/or ages are large
680: enough (0.14 mag and 19\% respectively) that the effects of stochastic
681: sampling are washed out. However, figure 1 in \citet{cer03} shows
682: that the expected scatter in the integrated (B$-$V) of a 10 Myr
683: cluster with $10^{3}$ or less stars is much larger than our assumed
684: colour uncertainties. Most of the OCs in our sample are much older
685: than 10 Myr. It is possible that stochastic sampling effects are
686: smaller in (B$-$V) for older clusters. After $\sim10$ Myr the most
687: luminous stars will have evolved off of the main sequence and will
688: contribute more strongly to the red end of the integrated cluster SED.
689: \citet{deg05} remark that stochastic sampling effects affect broadband
690: photometry to a smaller extent than spectroscopy. In a follow up
691: paper (Hancock et al., in prep.) we will explore a more detailed
692: analysis of the stellar content of these resolved OCs to further
693: investigate the extent to which stochastic sampling of the IMF affects
694: the integrated broadband colours. If stochastic sampling effects are
695: observationally characterized in this sample of OCs, they will
696: provide a good bench mark for testing future population synthesis
697: models.
698:
699: \subsection{The Effect of Cluster Dissolution}
700:
701: Current stellar population synthesis models do not take into account
702: the effects of cluster dissolution. Mass segregation causes the
703: massive stars to concentrate towards the centre of a cluster while the
704: low-mass stars tend to populate the outer regions. A consequence of
705: mass segregation is that low-mass stars tend to be preferentially
706: ejected from star clusters. The mass function (MF) therefore changes
707: drastically during the evolution of a cluster, as a consequence,
708: becoming less steep than the IMF (e.g. \citealp{bau03}). The
709: preferential loss of low-mass stars will change the integrated colours
710: of a cluster (e.g. \citealp{lam06a}). Such a time evolution of the
711: integrated colours resulting from cluster dissolution can be difficult
712: to disentangle from stellar evolutionary effects.
713:
714: The colour change of a star cluster due to dissolution depends
715: upon its age, its mass, and its dissolution time scale (t$_{dis}$),
716: with the effect being larger for later ages and longer t$_{dis}$.
717: \citet{lam06a} show that, for a $10^5$ M$_{\odot}$ cluster, the
718: (B$-$V), (V$-$I) and (V$-$K) colours are not significantly affected by
719: the preferential loss of low-mass stars during the first 80\% of
720: their $t_{dis}$, with $\Delta$(V$-$I)$=0.03$ and 0.1 mag for a
721: t$_{dis}$ of 1 Gyr and 10 Gyr respectively. However, in the
722: last 20\% of a cluster's life, the colour effect can be quite
723: large, up to $\Delta$(V$-$I)$=0.3$ mag.
724:
725: Our sample clusters likely have masses between $10^3$ and $10^4$
726: M$_{\odot}$. This implies t$_{dis}$ between $\sim$100 Myr and 2 Gyr
727: \citep{lam06b}. Given that 45\% of our sample have ages less than 80
728: Myr and 86\% have ages less than 1.6 Gyr (see Table 1), it is likely
729: that most of our clusters are not in the last 20\% of their lifetimes,
730: but we can not rule out the possibility that some might be near the end
731: of their lifetimes. Thus the effect of dissolution on their colours
732: is likely small, typically smaller than our assumed uncertainties of
733: 0.14 magnitudes.
734:
735: Using the stellar data available on WEBDA for a sample of our OCs
736: covering a representative range of distances, we found that $\sim85\%$
737: contained stars $\la$0.5 M$_{\odot}$. To determine the effects that
738: low mass stars have on the model integrated colours we compared
739: Kroupa models with lower mass cutoff of 0.1 M$_{\odot}$ with those
740: with 0.5 M$_{\odot}$. We found no significant differences in
741: the (U$-$B), (B$-$V), (V$-$R), or (V$-$I) colours.
742:
743:
744: \section{SUMMARY}
745:
746: We present an empirical assessment of the use of broadband colours as
747: age indicators for unresolved extragalactic clusters and investigate
748: stochastic sampling effects on integrated colours. The population
749: synthesis code {\it Starburst99} \citep{lei99} and four optical
750: colours were used to estimate how well we can recover the ages of 62
751: well-studied Galactic open clusters with published ages. We conclude
752: the following:
753:
754: 1) Galactic open clusters can be used for testing the integrated
755: properties from population synthesis and serve as reasonable
756: benchmarks for future assessments of age-dating methods.
757:
758: 2) The (U$-$B) colour is critical. Only colour combinations that
759: included (U$-$B) resulted in good age and extinction predictions,
760: consistent with previous results.
761:
762: 3) Only with (U$-$B) included were the predicted age uncertainties
763: reasonably constrained.
764:
765: 4) Changes of a factor of $\sim$2 in assumed metal abundance do not
766: result in significantly different predictions of cluster age. This
767: indicates that the uncertainties in predicting the age of a cluster
768: resulting from the age-reddening degeneracy dominate over the other
769: sources of degeneracy in these optical bands, over our age range
770: and metallicity range. Another possibility is that uncertainties in
771: the measured colours (and hence the predicted ages) dwarf the
772: metallicity-age degeneracy.
773:
774: 5) A $\chi^2$ minimization and a $\Delta\chi^2$ defined to give 68\%
775: confidence levels provide reliable age estimates, and more
776: importantly, reliable age uncertainties. The difference in the
777: photometric ages and the ages derived from the HR-diagrams of our
778: selected cluster sample with both the (U$-$B) and (B$-$V) colours are
779: smaller than 0.5 dex for 79\% of the clusters, with a maximum
780: difference of 1.5 dex.
781:
782: 6) It is likely that the uncertainties in the measured colours and/or
783: ages are large enough (0.14 mag and $\sim20\%$ respectively) that the
784: effects of stochastic sampling are washed out. A more detailed
785: analysis of the stellar content of these resolved OCs will be required
786: to further investigate the extent to which stochastic sampling of the
787: IMF affects the integrated broadband colours.
788:
789: 7) If stochastic sampling effects are observationally characterized
790: in this sample of OCs, they will provide a good bench mark for testing
791: future population synthesis models.
792:
793:
794: \section*{Acknowledgments}
795:
796: We thank the anonymous referee for his/her helpful comments and
797: suggestions. This research has made use of the WEBDA database,
798: operated at the Institute for Astronomy of the University of Vienna.
799: The authors would like to thank Ernst Paunzen for providing ASCII
800: tables of the WEBDA data. The authors also thank Alessandra Stone for
801: help with the data acquisition. This work has been supported by the
802: NASA LTSA grant NAG5-13079.
803:
804:
805: \begin{thebibliography}{99}
806:
807: \bibitem[Anders et al.(2004)]{and04}Anders P., Bissantz N., Fritze-v. Alvensblen U., de Grijs R. 2004, MNRAS, 347,196
808:
809: \bibitem[Anders \& Fritze-v. Alvensleben(2003)]{and03}Anders P., Fritze-v. Alvensblen U. 2003, A\&A, 2003, 401, 1063
810:
811: \bibitem[Bastian et al.(2005)]{bas05}Bastian N., Gieles M., Efremov Y. N., Lamers H. J. G. L. M. 2005, A\&A, 443, 79
812:
813: \bibitem[Battinelli, Brandimarti, \& Capuzzo-Dolcetta(1994)]{bat94}Battinelli P., Brandimarti A., Capuzzo-Dolcetta R. 1994, A\&AS, 104, 379
814:
815: \bibitem[Baumgardt \& Makino(2003)]{bau03}Baumgardt H., Makino J. 2003, MNRAS, 340, 227
816:
817: \bibitem[Benedict et al.(2002)]{ben02}Benedict G. F., Howell D. A., J\o rgensen I., Kenney J. D. P., Smith B. J. 2002, AJ, 123, 1411
818:
819: \bibitem[Bruzual \& Charlot(2003)]{bru03}Bruzual A. G., Charlot S. 2003, MNRAS, 344, 1000
820:
821: \bibitem[Bruzual(2002)]{bru02}Bruzual A. G. 2002, in IAU Symp. Ser. 207, Extragalactic Star Clusters, ed. D. Geisler, E. K. Grebel, \& D. Minniti, 616
822:
823: \bibitem[Bruzual \& Charlot(1993)]{bru93}Bruzual A. G., Charlot S. 1993, ApJ, 405, 538
824:
825: \bibitem[Cardelli, Clayton, \& Mathis(1989)]{car89}Cardelli J. A., Clayton G. C., Mathis J. S. 1989, ApJ, 345, 245
826:
827: \bibitem[Cervi\~{n}o \& Luridiana(2006)]{cer06}Cervi\~{n}o M., Luridiana V. 2006, A\&A, 451, 475
828:
829: \bibitem[Cervi\~{n}o \& Luridiana(2004)]{cer04}Cervi\~{n}o M., Luridiana V. 2004, A\&A, 413, 145
830:
831: \bibitem[Cervi\~{n}o \& Valls-Gabaud(2003)]{cer03}Cervi\~{n}o M., Valls-Gabaud D. 2003, MNRAS, 338, 481
832:
833: \bibitem[Charlot, Worthey, \& Bressan(1996)]{cha96}Charlot S., Worthey G., Bressan A. 1996, ApJ, 457, 625
834:
835: \bibitem[de Grijs et al.(2005)]{deg05}de Grijs R., Anders P., Lamers H. J. G. L. M., Bastian N., Fritze-v. Alvensleben U., Parmentier G., Sharina M. E., Yi S. 2005, MNRAS, 359, 874
836:
837: \bibitem[Fall, Chandar, \& Whitmore(2005)]{fal05}Fall S. M., Chandar R., Whitmore B. C. 2005, ApJ, 631, L133
838:
839: \bibitem[Fioc \& Rocca-Volmerange(1997)]{fio97}Fioc M., Rocca-Volmerange B. 1997, A\&A, 326, 950
840:
841: \bibitem[Fritze-v.\ Alvensleben \& Gerhard(1994)]{fri94}Fritze-v. Alvensleben U., Gerhard O. E. 1994, A\&A, 285, 751
842:
843: \bibitem[Gil de Paz \& Madore(2002)]{gil02}Gil de Paz A., Madore B. 2002, AJ, 123, 1864
844:
845: \bibitem[Gray(1965)]{gra65}Gray D.F. 1965, AJ, 70, 362
846:
847:
848: \bibitem[Hancock et al.(2007)]{han07}Hancock M., Smith B. J., Struck C., Giroux M. L., Appleton P. N., Charmandaris V., Reach W. T. 2007, AJ, 133, 676
849:
850: \bibitem[Hancock et al.(2003)]{han03}Hancock M., Weistrop D., Eggers D., Nelson C. H. 2003, AJ, 125, 1696
851:
852:
853: \bibitem[Holtzman et al.(1992)]{hol92}Holtzman J. A. et al. 1992, AJ, 103, 691
854:
855: \bibitem[Holtzman et al.(1996)]{hol96}Holtzman J. A. et al. 1996, AJ, 112, 416
856:
857: \bibitem[Johnson et al.(1999)]{joh99}Johnson K. E., Vacca W. D., Leitherer C., Conti P., Lipscy S. J. 1999, AJ, 117, 1708
858:
859: \bibitem[Keel \& Borne(2003)]{kee03}Keel W. C., Borne K. D. 2003, AJ, 126, 1227
860:
861: \bibitem[Kroupa(2002)]{kro02}Kroupa P. 2002, in ASP Conf. Ser. 285, Modes of Star Formation and the Origin of Field Populations, ed. E. K. Grebel \& W. Brandner (San Francisco: ASP), 86
862:
863: \bibitem[Lamers, Anders, \& de Grijs(2006)]{lam06a}Lamers H. J. G. L. M., Anders P., de Grijs R. 2006, A\&A, 452, 131
864:
865: \bibitem[Lamers \& Gieles(2006)]{lam06b}Lamers H. J. G. L. M., Gieles M. 2006, A\&A, 455, L17
866:
867: \bibitem[Larsen(2004)]{lar04}Larsen S. S. 2004, A\&A, 416, 537
868:
869: \bibitem[Lata et al.(2002)]{lat02}Lata S., Pandey A.K., Sagar R., Mohan V. 2002, A\&A, 388, 158
870:
871: \bibitem[Lee, Chandar, \& Whitmore(2005)]{lee05}Lee M. G., Chandar R., Whitmore B. C. 2005, AJ, 130, 2128
872:
873: \bibitem[Leitherer et al.(1999)]{lei99}Leitherer C. et al. 1999, ApJS, 123, 3
874:
875: \bibitem[Luridiana \& Cervi\~{n}o(2007)]{lur06}Luridiana V., Cervi\~{n}o M. 2007, ASPC, 374, 393
876:
877:
878: \bibitem[MacArthur et al.(2004)]{mac04}MacArthur L. A., Courteau S., Bell E. F., Holtzman J. A. 2004, ApJS, 152, 175
879:
880: \bibitem[Mermilliod(1995)]{mer95}Mermilliod J.-C. 1995, in ``Information and On-Line Data in Astronomy'', Eds D. Egret \& M.A. Albrecht (Kluwer Academic Press, Dordrecht), p. 127-138
881:
882: \bibitem[Meurer et al.(1995)]{meu95}Meurer G. R., Heckman T. M., Leitherer C., Kinney A., Robert C., Garnett D. R. 1995, AJ, 110, 2665
883:
884: \bibitem[Pandey et al.(1989)]{pan89}Pandey A.K., Bhatt B.C., Mahra H.S., Sagar R. 1989, MNRAS, 236, 263
885:
886: \bibitem[Pasquali, de Grijs, \& Gallagher(2003)]{pas03}Pasquali A., de Grijs R., Gallagher J. S., 2003, MNRAS, 345, 161
887:
888: \bibitem[Paunzen \& Netopil(2006)]{pau06}Paunzen E., Netopil M. 2006, MNRAS, 371, 1641
889:
890: \bibitem[Pessev et al.(2008)]{pes08}Pessev P. M., Goudfrooij P., Puzia T. H., Chandar R. 2008, MNRAS, 385, 1535
891:
892: \bibitem[Press et al.(1992)]{pre92}Press W. H., Teukolsky S. A., Vetterling W. T., Flannery B. P. 2002, Numerical Recipes in Fortran: the art of scientific computing, 2nd ed. (Cambridge University Press)
893:
894: \bibitem[Sagar, Joshi, \& Sinvhal(1983)]{sag83}Sagar R., Joshi U.C., Sinvhal S.D. 1983, Bull. Astron. Soc. India, 11, 44
895:
896: \bibitem[Smith et al.(2005a)]{smi05a}Smith B. J., Struck C., Appleton P. N., Charmandaris V., Reach W., Eitter J. J. 2005, AJ, 130, 2117
897:
898: \bibitem[Smith et al.(2005b)]{smi05b}Smith B. J., Struck C., Nowak M. A. 2005b, AJ, 129, 1350
899:
900: \bibitem[Smith et al.(2008)]{smi08}Smith B. J. et al. 2008, AJ, 135, 2406
901:
902: \bibitem[Spassova \& Beav(1985)]{spa85}Spassova N.M., Beav P.V. 1985, Astrophys. Space Sci., 112, 111
903:
904: \bibitem[Struck-Marcell \& Tinsley(1978)]{str78}Struck-Marcell C., Tinsley B. M. 1978, ApJ, 221, 562
905:
906: \bibitem[Tinsley(1968)]{tin68}Tinsley B. M. 1968, ApJ, 151, 547
907:
908: \bibitem[Tinsley(1972)]{tin72}Tinsley B. M. 1972, A\&A, 20, 383
909:
910: \bibitem[V\'{a}zquez \& Leitherer(2005)]{vaz05}V\'{a}zquez G. A., Leitherer C. 2005, ApJ, 621, 695
911:
912: \bibitem[Weistrop et al.(2004)]{wei04}Weistrop D., Eggers D., Hancock M., Nelson C. H., Bachilla R., Kaiser M. E. 2004, AJ, 127, 1360
913:
914: \bibitem[Whitmore(2003)]{whi03}Whitmore B. C. 2003, in Space Telescope Science Inst. Symp. Series 14, A Decade of Hubble Space Telescope Science, ed. M. Livio, K. Noll, \& M. Stiavelli (Cambridge: Cambridge Univ. Press), 153
915:
916: \bibitem[Whitmore et al.(1993)]{whi93}Whitmore B. C. et al. 1993, AJ, 106, 1365
917:
918: \bibitem[Whitmore \& Schweizer(1995)]{whi95}Whitmore B. C., Schweizer F. 1995, AJ, 109, 960
919:
920:
921: \bibitem[Zhang, Fall, \& Whitmore(2001)]{zha01}Zhang Q., Fall S. M., Whitmore B. C. 2001, ApJ, 561, 727
922:
923:
924:
925: \end{thebibliography}
926:
927: \bsp
928:
929: \label{lastpage}
930:
931: \end{document}
932: