1: %\documentclass[12pt,preprint]{aastex}
2: \documentclass{emulateapj}
3:
4: \newcommand{\msun}{$M_{\odot}\ $}
5: \newcommand{\etal}{et al.~}
6: \newcommand{\kms}{km s$^{-1}$}
7: \newcommand{\hbetao}{H$\beta_{\rm o}$}
8: \newcommand{\hbetal}{H$\beta_{\rm LICK}$}
9: \newcommand{\hbeta}{H$\beta$}
10: \newcommand{\hgamma}{H$\gamma$}
11: \newcommand{\hdelta}{H$\delta$}
12:
13: \shorttitle {Blue stragglers rejuvenating globular clusters} \shortauthors {A.~J.~Cenarro et al.}
14:
15: \begin{document}
16:
17: \title {Evidence for blue straggler stars rejuvenating the integrated
18: spectra of globular clusters}
19:
20: \author{A. Javier Cenarro, J.L. Cervantes, Michael A. Beasley, Antonio Mar\'in--Franch, Alexandre Vazdekis}
21: \affil{Instituto de Astrof\'isica de Canarias, V\'ia L\'actea s/n, 38200 La Laguna,
22: Tenerife, Spain}
23: \email{cenarro@iac.es}
24:
25: \begin{abstract}
26: Integrated spectroscopy is the method of choice for deriving the ages
27: of unresolved stellar systems. However, hot stellar evolutionary
28: stages, such as hot horizontal branch stars and blue straggler stars
29: (BSSs), can affect the integrated ages measured using Balmer
30: lines. Such hot, ``non-canonical'' stars may lead to overestimations
31: of the temperature of the main sequence turn-off, and therefore
32: underestimations of the integrated age of a stellar population. Using
33: an optimized \hbeta\ index in conjunction with HST/WFPC2
34: color-magnitude diagrams (CMDs), we show that Galactic globular
35: clusters exhibit a large scatter in their apparent ``spectroscopic''
36: ages, which does not correspond to that in their CMD-derived ages. We
37: find for the first time that the specific frequency of BSSs, defined
38: within the same aperture as the integrated spectra, shows a clear
39: correspondence with \hbeta\ in the sense that, at fixed metallicity,
40: higher BSS ratios lead to younger $apparent$ spectroscopic ages. Thus,
41: the specific frequency of BSSs in globular clusters sets a fundamental
42: limit on the accuracy for which spectroscopic ages can be determined
43: for globular clusters, and maybe for other stellar systems like
44: galaxies. The observational implications of this result are discussed.
45: \end{abstract}
46:
47: \keywords{blue stragglers --- globular clusters: general --- galaxies:
48: star clusters --- galaxies: stellar content}
49:
50: \section{Introduction} \label{Introduction}
51:
52: A common method for estimating the ages of unresolved stellar systems
53: is to measure Balmer lines and metal lines from integrated spectra,
54: and compare them to stellar population models. The method relies on
55: the fact that Balmer lines are mostly sensitive to the effective
56: temperature ($T_{\rm eff}$) of the main sequence (MS) turn-off of a
57: stellar population (e.g.~Buzzoni \etal1994). This, in turn, provides a
58: measure of the population age.
59:
60: However, a longstanding concern over the use of Balmer lines to
61: estimate spectroscopic ages is the effect of non-canonical
62: evolutionary stages on the integrated stellar population spectra.
63: Particularly worrisome are the possible effects that bluer horizontal
64: branch (HB) stars and blue straggler stars (BSSs) may have on the
65: inferred $T_{\rm eff}$ of the MS turn-off. HB stars are helium-burning
66: stars which occupy a region in the CMD with a typical absolute $V$
67: magnitude of $\sim 0.7$\,mag (e.g.~Harris 2001). BSSs are identified
68: as blue, luminous extensions of MS stars (Sandage 1953). They are
69: thought to form due to H refuelling processes after the MS stage,
70: either from collisional stellar encounters (e.g.~Davies \etal1994) or
71: from mass-transfer binaries (McCrea 1964). Both HB stars and BSSs with
72: $T_{\rm eff} > 6500$\,K show prominent Balmer lines in their spectra,
73: which can mimic the presence of younger stellar populations (e.g.~Rose
74: 1985, Lee \etal2000, Schiavon \etal2004; Trager \etal2005).
75:
76: In a recent paper, Cervantes \& Vazdekis (2008) defined an optimized
77: line-strength index for \hbeta, called \hbetao, that minimizes the
78: metallicity dependence of the Balmer line in favor of its age
79: sensitivity. Interestingly, as pointed out in that paper, the
80: integrated spectra of Galactic globular clusters (GGCs) from Schiavon
81: \etal(2005; hereafter S05) exhibit a clear intrinsic scatter in their
82: \hbetao\ strengths, particularly with increasing metallicity.
83:
84: In order to shed light on the above issue, in this {\it letter} we
85: combine resolved and unresolved data of such GGCs to investigate
86: whether CMD-based age differences among GGCs and/or different relative
87: contributions of non-canonical stages can be responsible for the
88: distinct integrated \hbetao\ indices. In Section~\ref{TheData} we
89: discuss the data used in this study, the analysis of which is
90: described in Section~\ref{Analysis}. Finally, in
91: Section~\ref{ConclusionsandDiscussion} we present our conclusions and
92: discuss our findings within the context of estimating ages for
93: unresolved stellar populations.
94:
95: \section{The Data} \label{TheData}
96:
97:
98:
99: Integrated optical spectra for 41 GGCs were taken from S05. These data
100: were obtained by drift-scanning the core diameter of each GGC --with a
101: spectroscopic aperture equal to this diameter-- in order to construct
102: a representative integrated spectrum. The spectra cover a wavelength
103: range of $\sim3350-6430$\AA\ with a FWHM of $\sim 3.1$\AA\ and typical
104: signal-to-noise ratios (S/N) of $\sim 250$\,\AA$^{-1}$ in the
105: continuum of the \hbeta\ line. We refer the reader to S05 for more
106: details on these data.
107:
108: In Figure~\ref{fig1} we present the indices \hbetao, \hbetal\ (Worthey
109: \etal1994) and [MgFe] (Gonz\'alez 1993) measured for the 41 GGC
110: spectra at the S05 spectral resolution. Uncertainties in the index
111: measurements (1$\sigma$ error bars) account for the S/N spectra and
112: the typical radial velocity error provided by S05 for each GGC. To
113: guide the eye, based on the MILES stellar library
114: (S\'anchez-Bl\'azquez \etal2006a; Cenarro \etal2007a), an extension of
115: the simple stellar population (SSP) models in Vazdekis (1999) are
116: overplotted at 3.1\AA\ spectral resolution (hereafter V99$+$). The
117: fact that most GGCs lie below the model grids arises from the
118: well-known zero-point problem of SSP models (e.g.~Vazdekis \etal2001;
119: Schiavon \etal2002), although this does not affect our results which
120: are based on relative differences.
121:
122: As expected, in the \hbetao\ plot (Fig.~\ref{fig1}a) GGCs follow --on
123: average-- an old sequence in metallicity. However, by simple visual
124: inspection they seem to separate into two groups, particularly at the
125: high metallicity end ([MgFe] $\gtrsim 1.2$). Hence, solid and open
126: symbols are employed to indicate, respectively, GGCs with $high$ and
127: $low$ \hbetao\ values at fixed [MgFe]. Henceforth, these are referred
128: to as GGC$H$ and GGC$L$ respectively.
129:
130: When using the \hbetal\ definition (Fig.~\ref{fig1}b), both GGC groups
131: are still distinguished, although the greater age--metallicity
132: degeneracy of \hbetal\ would make their differentiation less clear to
133: detect if the distinct symbol codes were not present. Even so, Puzia
134: \etal(2002) already pointed out the existence of unexpected \hbetal\
135: differences among certain metal rich (MR) GGCs. It therefore appears
136: that some property differs between both groups leading to different
137: Balmer-line strengths at a given metallicity. In fact, since the
138: effect seems to increase with the increasing metallicity, we focus
139: our analysis on those 25 GGCs from S05 with [Fe/H] $> -1.35$ ([MgFe]
140: $\gtrsim 1.2$) among which clear differences in \hbeta\ at fixed
141: metallicity are observed.
142:
143: Together with the indices in Fig.~\ref{fig1}, the adopted CMD-derived
144: parameters for the 25 GGCs are listed in Table~\ref{table}. The HB
145: morphology, as measured by the HBR parameter (Lee \etal1994), and the
146: specific frequency of RR Lyrae variables, $S_{\rm RR Lyrae}$, are
147: taken from the Harris (1996) catalogue (Feb 2003 revision; hereafter
148: H03). We adopted the relative age estimates from De Angeli \etal(2005)
149: and Recio-Blanco \etal(2006), whose applied the so-called vertical
150: method on the GGC HST/WFPC2 snapshot catalogue of Piotto
151: \etal(2002). Also based on that catalogue, Recio-Blanco \etal(2006)
152: estimated the maximum $T_{\rm eff}$ along the HB, $T_{\rm eff}\,{\rm
153: HB}$ --considered as an HB morphology parameter--, whereas Moretti
154: \etal(2008; hereafter M08) measured the logarithm of the number of
155: BSSs inside the core radius, $r_{\rm c}$, normalised to the sampled
156: luminosity --in units of $10^4$\,L$\sun$-- in the $F555W$ HST band
157: within the same aperture. The last quantity can be considered as a
158: logarithmic specific frequency of BSSs inside the GGC $r_{\rm c}$,
159: hereafter $S_{\rm BSS}^{r_{\rm c}}$, and is representative of the
160: spectroscopic data in S05 as they are both computed within the same
161: aperture.
162:
163: \section{Analysis}
164: \label{Analysis}
165:
166:
167:
168: With the aim of constraining the origin of the intrinsic scatter in
169: the Balmer line-strengths of our GGC subsample, in Figure~\ref{fig2}
170: we show the CMD-derived parameters of Section~\ref{TheData}
171: --where available-- as a function of the GGC metallicity from H03.
172: Symbol codes are kept as in Fig~\ref{fig1} except for the MR GGC$H$s
173: NGC6388 and NGC6441, which being well-known {\it second parameter}
174: clusters (Rich \etal1997), are plotted as open stars --rather than
175: solid squares-- to facilitate further discussion.
176:
177: In Fig.~\ref{fig2}a, we see that there is no dependence of the group
178: location on the CMD-derived age. The difference in \hbeta\ strengths
179: between both groups is therefore not due to age differences among the
180: GGCs, as would normally be inferred from a classical SSP index-index
181: analysis. Note also that the typical dispersion of $\sim 1$\,Gyr
182: quoted by De Angeli \etal(2005) among the CMD-derived ages of
183: intermediate metallicity GGCs could never explain the large scatter in
184: \hbeta. We also rule out the possibility that the number of RR lyraes
185: in the instability strip is playing a significant role, as the two GGC
186: groups are well mixed (Fig.~\ref{fig2}b). In addition we find that
187: there is no obvious dependence on the HB morphology, as measured by
188: either the HBR parameter or the maximum $T_{\rm eff}$ of the HB
189: (Fig.~\ref{fig2}c and Fig.~\ref{fig2}d, respectively). As expected,
190: the second parameter clusters NGC6388 and NGC6441 (open stars) do
191: stand out of the general trends in both panels. Their high \hbeta\
192: strengths are naturally explained by the addition of hot HB stars in
193: their integrated spectra.
194:
195: Besides basic properties of the GGC CMDs, literature estimates of
196: [$\alpha$/Fe] for the GGCs in S05 (where available) show a high degree
197: of homogeneity (e.g.~Pritzl \etal2005), so differing levels of
198: $\alpha$--elements cannot account for the observed differences. We
199: have also ruled out that the Balmer lines of GGC$L$s are
200: sistematically filled-in by emission. In fact, emission was found and
201: corrected by S05 for only NGC6171 and NGC6553 (GGC$H$s), and NGC6352
202: (GGC$L$). Interestingly, the fact that NGC6352 has the weakest \hbeta\
203: line of the sample may suggest that a residual emission could still be
204: present.
205:
206: Having rejected the above mechanisms from being responsible for the
207: observed differences in \hbeta\ between GGC$H$s and GGC$L$s --except
208: in the obvious case of NGC6388 and NGC6441--, in Fig~\ref{fig2}e we
209: show the GGC metallicities versus $S_{\rm BSS}^{r_{\rm
210: c}}$. Interestingly, GGC$H$s and GGC$L$s --which were identified
211: spectroscopically-- separate cleanly into two groups in terms of their
212: BSS specific frequencies. At a given metallicity, GGCs with higher
213: $S_{\rm BSS}^{r_{\rm c}}$ values exhibit stronger \hbeta\ lines,
214: suggesting that BSSs are indeed affecting their integrated spectra.
215:
216: To reinforce this result, we quantify the impact of BSSs on the
217: integrated spectrum of NGC6342, the GGC$H$ with the highest $S_{\rm
218: BSS}^{r_{\rm c}}$. Based on photometric data from H03 and M08 for this
219: GGC and its BSS population, we obtain that $13\%$ of the GGC flux in
220: $V$ band within $r_{\rm c}$ comes from a population of 7 BSSs with
221: $0.22 \leq B-V \leq 0.52$, and a luminosity-weighted $B-V$ of $0.33$.
222: For each BSS, assuming [Fe/H] $= -0.65$ and its $B-V$, we estimate
223: $T_{\rm eff}$ using the $(B-V)-T_{\rm eff}$ relation for dwarfs from
224: Alonso \etal(1996). The $T_{\rm eff}$ values of the 7 BSSs are in the
225: range $5925 \leq T_{\rm eff} \leq 7500$\,K. We then compute spectral
226: BSS templates with the above parameters on the basis of the MILES
227: stellar library, using the interpolating algorithm in Vazdekis
228: \etal(2003; Appendix B). After the 7 BSS templates are scaled
229: according to their $V$ fluxes and subtracted from the NGC6342
230: spectrum, its integrated \hbetao\ and \hbetal\ indices decrease by
231: $0.70$\,\AA\ and $0.62$\,\AA\ respectively. Taking into account that
232: the averaged $S_{\rm BSS}^{r_{\rm c}}$ of GGC$L$s with [Fe/H]$ > -0.8$
233: (similar to that of NGC6342) is $\sim 1.4$ (Fig.~\ref{fig2}e), and
234: assuming similar $T_{\rm eff}$ values for their BSS populations, the
235: relative offsets in \hbetao\ and \hbetal\ between NGC6342 and the
236: above GGC$L$s are, respectively, $0.58$\,\AA\ and $0.51$\,\AA. This
237: agrees with the differences between GGC$H$s and GGC$L$s obtained at
238: this metallicity regime ([MgFe]$\gtrsim 2$; Fig.~\ref{fig1}),
239: supporting the idea that BSSs are responsible for the observed
240: differences in \hbeta. Even more, the effect of BSSs is also detected
241: among the subsample of GGC$L$s with [Fe/H]$ > -0.8$. NGC0104 and
242: NGC6624, with extreme values of $S_{\rm BSS}^{r_{\rm c}}$, pose the
243: lowest and largest \hbetao\ values respectively.
244:
245: \section{Discussion}
246: \label{ConclusionsandDiscussion}
247:
248: Based on the close correspondence between the specific frequency of
249: BSSs in GGCs with [Fe/H] $> -1.35$ and their integrated \hbeta\
250: strengths at fixed metallicity, we conclude that BSSs are primarily
251: responsible for the \hbeta\ variations observed in the integrated
252: spectra of GGCs of intermediate-to-high metallicity. Far from
253: discussing on the origin for the distinct $S_{\rm BSS}^{r_{\rm c}}$
254: values among GGCs (see M08 for a thorough study on this topic), we
255: here analyze the implications of the above result in the context of
256: age-dating unresolved stellar populations.
257:
258: First, caution must be employed in Balmer-line based age-metallicity
259: studies of unresolved extragalactic globular clusters (EGCs). Cenarro
260: \etal(2007b) already reported the existence of EGCs with strong Balmer
261: lines that were consistent with hosting an additional population of
262: either blue HB stars and/or BSSs. Since the BSS fraction of EGCs is
263: generally not known, the finding in this letter sets a fundamental
264: limit to the reliability with which ages may be determined for EGCs
265: using Balmer lines and SSP models. Taking the S05 data as a
266: representative old GC system, we can estimate this limit from the
267: averaged offsets in the measured \hbeta\ lines of GGC$H$s and
268: GGC$L$s. Since the offsets seem to vary with metallicity, local linear
269: fits to all GGC$H$s and GGC$L$s in Fig.~\ref{fig1} with [MgFe]$> 1.5$
270: ([Fe/H] $\gtrsim -1.0$) have been performed. For instance, at the
271: location of 47Tuc (NGC0104; [MgFe]$\sim 2.31$), we obtain
272: $\Delta$\hbetao\ $= 0.46 \pm 0.03$\AA\ and $\Delta$\hbetal\ $= 0.33
273: \pm 0.04$\AA\ (at the S05 resolution) with uncertainties accounting
274: for the standard errors of the means. Thus, assuming that GGC$L$s are
275: $\sim14$\,Gyr old --the largest SSP age in Fig.~\ref{fig1}-- the two
276: offsets can be consistently misinterpreted on the basis of SSP models
277: as GGC$H$s being $\sim 6-7$\,Gyr old, that is, as a rejuvenation of up
278: to $\sim 8$\,Gyr. Differences between GGC$H$s and GGC$L$s also exist
279: for the Lick \hgamma\ and \hdelta\ indices (Worthey \& Ottaviani
280: 1997), although they are not so apparent, probably due to their
281: limited age-disentangling power for old SSPs. For these indices, the
282: above test leads to rejuvenations of up to $\sim 4 - 5$\,Gyr.
283:
284: The role of metallicity in the present discussion is worthwhile
285: considering. Although the relation between $S_{\rm BSS}^{r_{\rm c}}$
286: and metallicity is not statistically significant --but marginally
287: positive-- over the entire GGC sample (in agreement with M08), we find
288: clear correlations for GGC$H$s and GGC$L$s separately, as illustrated
289: by the solid lines in Fig.~\ref{fig2}e. The different slopes seem to
290: indicate that, when BSSs are important, their relative contributions
291: are larger at high metallicities. Note that the fading with
292: metallicity expected in $F555W$ band for the most MR GGCs only
293: accounts for up to $\Delta S_{\rm BSS}^{r_{\rm c}} \sim 0.2$, so the
294: above trends are irrespective of this effect.
295:
296: These results may also have important consequences for EGC studies. To
297: understand the origin of color bimodality in GC systems within a
298: context of galaxy formation, age-dating GC subpopulations --through
299: the analysis of their integrated Balmer lines-- is a common practice
300: (see Brodie \& Strader 2006 and references therein). Interestingly,
301: some papers have reported that the MR GC subpopulation of certain
302: galaxies show on average a smaller mean age and a larger age scatter
303: than their metal poor (MP) counterparts (e.g.~Puzia
304: \etal2005). Although the present finding does not rule out the
305: existence of {\it true} age differences between MP and MR GC
306: subpopulations, the increasing importance of BSSs with metallicity
307: might, at least, partially affect the results of previous work.
308:
309: Whether all the above results can compromise the integrated ages of
310: other stellar systems, like galaxies, may rely on the mechanism that
311: dominates the formation of BSSs. If stellar encounters were driving
312: the BSS population, then one should not expect a major effect in
313: galaxies because of their much lower stellar densities. However, this
314: would not apply if mass-transfer binaries were the progenitors of most
315: BSSs. In fact, Momany \etal(2007) and Mapelli \etal(2007) support the
316: last scenario to explain the large BSS populations of dwarf spheroidal
317: galaxies, and Han \etal(2007) have demonstrated the importance of
318: binary interactions to understand the UV-upturn of elliptical galaxies
319: (Es). It therefore seems that BSSs could play a non-negligible role
320: in the integrated spectra of galaxies as long as they host an
321: important fraction of binary stars. If this were the case and the
322: potential increasing importance of BSSs with metallicity would still
323: hold for massive Es, then BSSs could contribute to the age scatter
324: reported for massive Es and to the fact that $younger$ Es have higher
325: metallicities than $older$ Es (e.g.~Trager \etal2000;
326: S\'anchez-Bl\'azquez \etal 2006b). This picture, however, requires of
327: further investigation which is out of the scope of this paper.
328:
329: \acknowledgments We acknowledge the anonymous referee for very useful
330: comments. A.J.C. and A.M-F are Juan de la Cierva Fellows of the
331: Spanish Ministry of Education and Science (SMES). JLC acknowledges the
332: SMES for a FPU PhD fellowship. This work has been funded by the SMES
333: through grant AYA2007-67752-C03-01.
334:
335: \begin{thebibliography}{}
336: \bibitem[Alonso \etal(1996)]{1996A&A...313..873A} Alonso, A., Arribas, S., \& Mart\'{\i}nez-Roger, C.\ 1996, \aap, 313, 873
337: \bibitem[Brodie \& Strader (2006)]{2006ARA&A..44..193B} Brodie, J.~P., \& Strader, J.\ 2006, \araa, 44, 193
338: \bibitem[Buzzoni \etal(1994)]{1994AJ....107..513B} Buzzoni, A., Mantegazza, L., \& Gariboldi, G.\ 1994, \aj, 107, 513
339: \bibitem[Cenarro \etal(2007a)]{2007MNRAS.374..664C} Cenarro, A.~J., Peletier, R.~F., S\'nchez-Blázquez, P., Selam, S.~O., Toloba, E., Cardiel, N., Falc\'n-Barroso, J., Gorgas, J., Jim\'enez-Vicente, J., \& Vazdekis, A.\ 2007a, \mnras, 374, 664
340: \bibitem[Cenarro \etal(2007b)]{2007AJ....134..391C} Cenarro, A.~J., Beasley, M.~A., Strader, J., Brodie, J.~P., \& Forbes, D.~A.\ 2007b, \aj, 134, 391
341: \bibitem[Cervantes \& Vazdekis (2008)]{356} Cervantes, J.~L., \& Vazdekis, A.\ 2008, \mnras, in press (arXiv:0810.3240)
342: \bibitem[Davies \etal(1994)]{1994ApJ...424..870D} Davies, M.~B., Benz, W., \& Hills, J.~G.\ 1994, \apj, 424, 870
343: \bibitem[De Angeli \etal(2005)]{2005AJ....130..116D} De Angeli, F., Piotto, G., Cassisi, S., Busso, G., Recio-Blanco, A., Salaris, M., Aparicio, A., \& Rosenberg, A.\ 2005, \aj, 130, 116.
344: \bibitem[Gonz\'alez (1993)]{1993PhDT.......172G} Gonz\'alez, J.~J., \ 1993, PhDT, University of California
345: \bibitem[Han \etal(2007)]{2007MNRAS.380.1098H} Han, Z., Podsiadlowski, Ph., \& Lynas-Gray, A.~E.\ 2007, \mnras, 380, 1098
346: \bibitem[Harris (1996)]{1996AJ....112.1487H} Harris, W.~E., 1996, \aj, 112, 1487
347: \bibitem[Harris (2001)]{2001stcl.conf..223H} Harris, W.~E.\ 2001, in Star Clusters, ed. L. Labhardt \& Binggeli (Berlin: Springer), 223
348: \bibitem[Lee \etal(1994)]{1994ApJ...423..248L} Lee, Y., Demarque, P., \& Zinn, R.\ 1994, \apj, 423, 248
349: \bibitem[Lee \etal(2000)]{2000AJ....120..998L} Lee, H., Yoon, S., \& Lee, Y.\ 2000, \aj, 120, 998
350: \bibitem[Mapelli (2007)]{2007MNRAS.380.1127M} Mapelli, M., Ripamonti, E., Tolstoy, E., Sigurdsson, S., Irwin, M.~J., \& Battaglia, G.\ 2007, \mnras, 380, 1127
351: \bibitem[McCrea (1964)]{1964MNRAS.128..147M} McCrea, W.~H.\ 1964, \mnras, 128, 147
352: \bibitem[Momany \etal(2007)]{2007A&A...468..973M} Momany, Y., Held, E.~V., Saviane, I., Zaggia, S., Rizzi, L., \& Gullieuszik, M. \ 2007, \aap, 468, 973
353: \bibitem[Moretti \etal(2008)]{2008A&A...483..183M} Moretti, A., De Angeli, F., \& Piotto, G.\ 2008, \aap, 483, 183 (M08)
354: \bibitem[Piotto \etal(2002)]{Pio02} Piotto et al., 2002, \aap, 391, 945
355: \bibitem[Pritzl \etal(2005)]{2005AJ....130.2140P} Pritzl, B.~J., Venn, K.~A., \& Irwin, M.\ 2005, \aj, 130, 2140
356: \bibitem[Puzia \etal(2002)]{2002A&A...395...45P} Puzia, T.~H., Saglia, R.~P., Kissler-Patig, M., Maraston, C., Greggio, L., Renzini, A., \& Ortolani, S.\ 2002, \aap, 395, 45
357: \bibitem[Puzia \etal(2005)]{2005A&A...439..997P} Puzia, T.~H., Kissler-Patig, M., Thomas, D., Maraston, C., Saglia, R.~P., Bender, R., Goudfrooij, P., \& Hempel, M.,\ 2005, \aap, 439, 997
358: \bibitem[Rich \etal(1997)]{1997ApJ...484L..25R} Rich, R.~M., Sosin, C., Djorgovski, S.~G., Piotto, G., King, I.~R., Renzini, A., Phinney, E.~S., Dorman, B., Liebert, J., \& Meylan, G.\ 1997, \apj, 484, 25
359: \bibitem[Recio-Blanco \etal(2006)]{2006A&A...452..875R} Recio-Blanco, A., Aparicio, A., Piotto, G., De Angeli, F., \& Djorgovski, S.~G.\ 2006, \aap, 452, 875
360: \bibitem[Rose (1985)]{1985AJ.....90.1927R} Rose, J.~A.\ 1985, \aj, 90, 1927
361: \bibitem[S\'anchez-Bl\'azquez \etal(2006a)]{2006MNRAS.371..703S} S\'anchez-Bl\'azquez, P., Peletier, R.~F., Jim\'nez-Vicente, J., Cardiel, N., Cenarro, A.~J., Falc\'on-Barroso, J., Gorgas, J., Selam, S., \& Vazdekis, A.\ 2006, \mnras, 371, 703
362: \bibitem[S\'anchez-Bl\'azquez \etal(2006b)]{2006A&A...457..809S} S\'anchez-Bl\'azquez, P., Gorgas, J., Cardiel, N., \& Gonz\'alez, J.~J.\ 2006, \aap, 457, 809
363: \bibitem[Sandage (1953)]{1953AJ.....58...61S} Sandage, A.~R.\ 1953, \aj, 58, 61
364: \bibitem[Schiavon \etal(2002)]{2002ApJ...580..873S} Schiavon, R.~P., Faber S.~M., Rose, J.~A., \& Castilho, B.~V.\ 2002, \apj, 580, 873
365: \bibitem[Schiavon \etal(2004)]{2004ApJ...608L..33S} Schiavon, R.~P., Rose, J.~A., Courteau, S., \& MacArthur, L.~A.\ 2004, \apj, 608, 33
366: \bibitem[Schiavon \etal(2005)]{2005ApJS..160..163S} Schiavon, R.~P., Rose, J.~A., Courteau, S., \& MacArthur, L.~A.\ 2005, \apjs, 160, 163 (S05)
367: \bibitem[Trager \etal(2000)]{2000AJ....120..165T} Trager, S.~C., Faber, S.~M., Worthey, G., \& Gonz\'alez, J.~J.\ 2000, \aj, 120, 165
368: \bibitem[Trager \etal(2005)]{2005MNRAS.362....2T} Trager, S.~C., Worthey, G., Faber, S.~M., \& Dressler, A.\ 2005, \mnras, 362, 2
369: \bibitem[Vazdekis \etal(2001)]{2001ApJ...549..274V}Vazdekis, A., Salaris, M., Arimoto, N., \& Rose, J.~A.\ 2001, \apj, 549, 274
370: \bibitem[Vazdekis (1999)]{1999ApJ...513..224V} Vazdekis, A.\ 1999, \apj, 513, 224
371: \bibitem[Vazdekis \etal(2003)]{2003MNRAS.340.1317V}Vazdekis, A., Cenarro, A.~J., Gorgas, J., Cardiel, N., \& Peletier, R.~F.\ 2003, \mnras, 340, 1317
372: \bibitem[Worthey \etal(1994)]{1994ApJS...94..687W} Worthey, G., Faber, S.~M., Gonz\'alez, J.~J, \& Burstein, D.\ 1994, \apjs, 94, 687
373: \bibitem[Worthey \& Ottaviani (1997)]{1997ApJS..111..377W} Worthey, G., \& Ottaviani, D.~L.\ 1997, \apjs, 111, 377
374: \bibitem[Zoccali \etal(2000)]{2000ApJ...538..289Z}Zoccali, M., Cassisi, S., Bono, G., Piotto, G., Rich, R.~M., \& Djorgovski, S.~G.\ 2000, \apj, 538, 289
375: \end{thebibliography}{}
376:
377: %\clearpage
378:
379: \input{tab1.tex}
380:
381: \begin{figure}
382: %\epsscale{0.65}
383: \epsscale{0.45}
384: \plotone{f1.eps}
385: \caption{\hbetao\ (a) and \hbetal\ (b) indices versus [MgFe] for the
386: 41 GGCs in S05. Solid and open symbols indicate, respectively, GGCs
387: with high (GGC$H$) and low (GGC$L$) \hbeta\ indices at fixed
388: [MgFe]. SSP models from V99$+$ at FWHM = 3.1\AA\ are overplotted,
389: with dotted and dashed lines indicating fixed ages and
390: metallicities as in the labels.}
391: \label{fig1}
392: \end{figure}
393:
394: \begin{figure*}
395: \epsscale{1.19}
396: %\epsscale{1.}
397: \plotone{f2.eps}
398: \caption{Metallicity versus CMD-derived parameters (Age$_{\rm NORM}$,
399: $S_{\rm RRLyr}$, HBR, $log(T_{\rm eff}\,{\rm HB})$, and $S_{\rm
400: BSS}^{r_{\rm c}}$; see definition in the text) for a subsample of 25
401: GGCs from S05 with [Fe/H]$ > -1.35$. Symbol codes are kept as in
402: Fig~\ref{fig1} except for the second parameter GGCs NGC6388 and
403: NGC6441 (open stars). Solid lines in panel $e$ illustrate individual
404: linear fits to solid and open squares.}
405: \label{fig2}
406: \end{figure*}
407:
408: \end{document}
409:
410: