1: \documentclass[12pt,preprint]{aastex}
2: \usepackage{epsfig}
3:
4:
5: \shorttitle{CLUSTER GALAXY ASSEMBLY TIMESCALE}
6: \shortauthors{CARRETERO ET AL.}
7:
8: \begin{document}
9:
10: \title{On the environmental dependence of cluster galaxy assembly timescale}
11:
12: \author{C. Carretero\altaffilmark{1}, A. Vazdekis\altaffilmark{1},
13: J. E. Beckman\altaffilmark{1,}\altaffilmark{2}, P.
14: S\'anchez-Bl\'azquez\altaffilmark{3} and J.
15: Gorgas\altaffilmark{3}}
16: \altaffiltext{1}{Instituto de
17: Astrof\'{\i}sica de Canarias, V\'{\i}a L\'actea s/n, 38200 La
18: Laguna, Tenerife, Spain; cch@iac.es}
19: \altaffiltext{2}{Consejo
20: Superior de Investigaciones Cient\'{\i}ficas, Spain.}
21: \altaffiltext{3}{Departamento de Astrof\'{\i}sica, Facultad de
22: F\'{\i}sicas, Universidad Complutense de Madrid, 28040 Madrid, Spain.}
23:
24:
25: \begin{abstract}
26: We present estimates of CN and Mg overabundances with respect to
27: Fe for early-type galaxies in 8 clusters over a range of richness
28: and morphology. Spectra were taken from the Sloan Digital Sky
29: Survey (SDSS) DR1, and from WHT and CAHA
30: observations. Abundances were derived from absorption lines and
31: single burst population models, by comparing galaxy spectra with
32: appropriately broadened synthetic model spectra. We detect
33: correlations between [Mg/CN] and [CN/Fe] and
34: cluster X-ray luminosity. No correlation is observed for [Mg/Fe].
35: We also see a clear trend with the richness and morphology of the
36: clusters. This is interpreted given varying formation
37: timescales for CN, Mg and Fe, and a varying star
38: formation history in early-type galaxies as a function of their
39: environment: intermediate-mass early-type galaxies in more massive clusters are assembled on shorter timescales than in less massive clusters, with an upper limit of $\sim1$~Gyr.
40: \end{abstract}
41:
42: \keywords{cosmology: observations --- galaxies: abundances --- galaxies: clusters:
43: general --- galaxies: formation --- galaxies: stellar content --- X-rays: galaxies:
44: clusters}
45:
46: \section{Introduction}
47:
48:
49: A key question for scenarios of galaxy formation is whether galaxies formed in
50: single structural ``monolithic'' events \citep{lar74} or by a series of
51: ``hierarchical'' processes \citep{pre74,whi91} in which large galaxies built up from smaller ones. Many structural and dynamical properties of
52: galaxies in clusters are explained in this scenario, though problems remain: the
53: absence of the predicted mass cusps in the centres of ellipticals and bulges, and the prediction of far more satellite galaxies than those observed.
54:
55: Stellar populations offer a fossil record of the formation and evolution of
56: galaxies, most clearly in elliptical galaxies, and stellar
57: population studies provide very strong
58: constraints on the principal galaxy formation scenarios. It is
59: hard to reconcile the hierarchical models with the result that massive galaxies show significantly larger mean luminosity weighted ages than their smaller counterparts \citep{kau03}.
60:
61: Understanding stellar populations in early-type
62: galaxies as a function of the environment can provide answers
63: to the puzzle. The present observational base is small. Only
64: three clusters have been observed for a detailed stellar
65: populations analysis: Virgo \citep[e.g.][]{vaz01a}, Coma
66: \citep[e.g.][]{jor99} and Fornax \citep{kun98}. Extending the
67: study to a large number of clusters covering a range of richness
68: and morphology is mandatory.
69:
70: Past studies of clusters used the original Lick/IDS spectral
71: indices \citep{wor94b} whose reliability is limited by their
72: resolution dependence, as uncertain corrections for broadening and
73: instrumental effects are needed. It is better to use modelled
74: integrated spectra, which can be broadened to match galaxy
75: velocity dispersion, $\sigma$. Such models have been developed by
76: \citet{vaz99} and allow an accurate separation of age and
77: metallicity, and subsequent individual abundance derivations.
78:
79: The study of the element abundance ratios in elliptical galaxies
80: within distinct clusters should be a powerful discriminant between
81: different star formation histories \citep[e.g.][]{wor98}. In
82: particular, overabundances of [Mg/Fe] compared with the solar
83: ratio have been found in massive elliptical galaxies
84: \citep{pel89,wor92,vaz97}. These have been interpreted via several possible scenarios based on the fact that Mg is
85: mainly produced in Type II supernovae \citep{fab92,mat94}, and include different star formation rates (SFR) and a time dependent IMF.
86:
87: Differences in the abundances of C and N as a function of the
88: environment have been recently suggested by \citet{pat03}, who found striking spectral differences between field elliptical
89: galaxies and their counterparts in the central region of the Coma
90: cluster. Galaxies in the denser environment showed
91: significantly lower CN$_2$ and C4668 absorption strength. Here we
92: explore these differences by extending the
93: study to a larger number of clusters over a range of richness and
94: morphology, applying the new analysis techniques to derive
95: abundance ratios. We have assumed a flat Universe with H$_0 =
96: 75$~km~s$^{-1}$~Mpc$^{-1}$ and $q_0 = 0.5$.
97:
98:
99: \section{Data}
100:
101: Sloan Digital Sky Survey \citep{sto02} spectra were obtained using
102: a multiobject, $3^{\prime\prime}$ diameter fiber spectrograph. Exposures ranged from 45 to 75 minutes. All the data processing
103: was performed with automated SDSS software. Redshifts were
104: measured on the reduced spectra by an automated system, which
105: models each galaxy spectrum as a linear combination of stellar
106: populations. We measured independently the redshifts of the
107: galaxies used in this study by crosscorrelating each galaxy
108: spectrum with our SSP synthetic model spectra. We found no
109: significant differences between SDSS redshift values and ours.
110:
111: From the SDSS Data Release 1 database
112: we selected galaxy spectra according to the following criteria. They must:
113:
114: {\it i) Belong to an Abell cluster.} We included this criterion
115: because the richness and the morphology of Abell clusters are
116: uniformly defined and described in the literature. Also, X-ray
117: luminosity values are available.
118:
119: {\it ii) Belong to the early-type galaxies catalogue of
120: \citet{ber03}.} This sample has $\sim9000$ early-type galaxies,
121: in the redshift range $0.01\le z\le 0.3$, selected from the SDSS
122: spectroscopic database using morphological and spectral criteria.
123: The mean spectrum signal-to-noise per pixel is 16.
124:
125: {\it iii) Have $S/N$ per pixel greater than 15.}
126:
127: {\it iv) Have velocity dispersion in the range 150~km~s$^{-1}$
128: $\le\sigma\le$ 250~km~s$^{-1}$.} Galaxies outside this range of
129: $\sigma$ were rejected because of the completeness of the sample:
130: not all clusters had spectra of dwarf and/or giant elliptical
131: galaxies because of the inner limitation of SDSS data (for dwarfs)
132: and the morphology of the clusters (for giants). Also, the quality of the spectra of the faintest galaxies was too low for our analysis requirements.
133: $\sigma$ values were obtained from \citet{ber03}.
134:
135: Using these criteria, we obtained a total of 55
136: galaxies distributed in 6 clusters. The clusters are: A257, A279,
137: A655, A1238, A1650 and A2050. Their redshift values vary in the
138: range $0.07 < z < 0.13$ and they cover a range of richness and morphology.
139: See details in Table~\ref{table1}.
140:
141:
142:
143:
144: For comparison and completeness, we added to our SDSS data
145: high-quality long-slit spectra of early-type galaxies in Coma and
146: Virgo clusters \citep[for details, see][]{pat03}. To compare them with those of the SDSS sample, we extracted spectra along the
147: slit simulating a circular aperture (distance weighted coadded
148: spectra) of radius $1.5^{\prime\prime}$ at $z = 0.1$.
149: This aperture translates into apertures of radius
150: $6^{\prime\prime}$ for Coma and $37^{\prime\prime}$ for Virgo.
151:
152:
153: \section{Galaxy measurements and results}
154:
155: To derive mean luminosity-weighted ages and metallicities, we
156: compared selected absorption line strengths with those predicted
157: by the model of \citet{vaz99}. This model provides flux-calibrated
158: spectra in the optical range at a resolution of 1.8~\AA\ (FWHM)
159: for single-burst stellar populations. This way, we can transform synthetic spectra to the resolution and
160: dispersion of the galaxy spectra instead of the opposite, as
161: required while working in the Lick system. Selected absorption
162: indices were CN$_2$, Mg$_2$ \citep{wor94a} and Fe2 (defined as Fe2 = $\frac{{\rm Fe}4383+{\rm Fe}5270}{2}$). We used
163: these features because of their low sensitivity to variations in
164: $S/N$ \citep{car03} and velocity dispersion (we have estimated
165: $\Delta(index)/index < 0.15$, for $\Delta(\sigma)$ =
166: 300~km~s$^{-1}$). This way we avoid possible variations in the
167: index value, as $\sigma$ may vary as a function of $r$, due to the
168: fact that SDSS spectra provide the light integrated within the
169: fibers of the spectrograph.
170:
171: Plots of the strengths of the selected indices versus H$\beta$
172: provide close to orthogonal model grids, allowing us to estimate accurately
173: galaxy mean ages as well as relative abundances of the
174: different elements. Figure~\ref{fig1} illustrates this method for
175: the galaxies of clusters A1238 and A655 (two clusters with extreme
176: values of X-ray luminosity). We will refer to the metallicities
177: derived in the diagrams CN$_2$--H$\beta$, Mg$_2$--H$\beta$ and
178: Fe2--H$\beta$ as $Z_{\rm CN}$, $Z_{\rm Mg}$ and $Z_{\rm Fe}$,
179: respectively. Since the CN$_2$ index is strongly dominated by C
180: and N, the Mg$_2$ index is governed by Mg and the Fe2 index by Fe
181: \citep{tri95}, these metallicities must be close to the [CN/H],
182: [Mg/H] and [Fe/H] abundances, and $[Z_{\rm CN}/Z_{\rm Fe}]$,
183: $[Z_{\rm Mg}/Z_{\rm Fe}]$ and $[Z_{\rm Mg}/Z_{\rm CN}]$ are then
184: estimates of the abundance ratios [CN/Fe], [Mg/Fe] and [Mg/CN] for
185: each galaxy. Note that an extrapolation of the model grids is
186: required for some galaxies to obtain the abundances of CN and Mg,
187: since the models extend only to [M/H] = 0.2. It is worth noting
188: that certain galaxies fall below the model grids, which can be
189: attributed to the fact that absolute age determination is subject
190: to model uncertainties \citep[see][]{vaz01b,sch02}. We neglect the
191: possible effect of nebular emission on H$_\beta$ since no
192: [\ion{O}{3}]$\lambda$5007 emission is detected in our galaxy
193: spectra. Nevertheless, assuming an upper H$_\beta$ emission
194: correction of $\sim0.5$~\AA\ \citep{kun02}, although this
195: would give rise to a significant reduction in the mean age of the
196: stellar populations of the oldest galaxies, the net effect on the
197: abundance ratios would be no more than $\sim0.05$ dex, for the
198: most affected galaxies in a cluster. Note that this correction for the
199: H$\beta$ index is larger than the one obtained by
200: varying the model prescriptions, e.g. with $\alpha$-enhanced isochrones and
201: atomic diffusion included \citep{vaz01b}, for inferring ages in
202: better agreement with the current age of the Universe, for the oldest galaxies
203: in our sample. However we do recognize that the ages quoted here, and in current articles dealing with stellar population modelling, may well span an older range than the true age range of the populations observed. This would indicate some flaw or flaws in the present quantitaive understanding of stellar evolution, and is a problem recognized by the community working in population synthesis.
204:
205:
206:
207:
208: Figure~\ref{fig2} shows the measured abundance ratios for the
209: galaxies in the whole sample of clusters as a function of
210: the velocity dispersion. To derive representative
211: relative abundances ratios for each cluster, and since relative
212: abundances correlate with $\sigma$, we have fitted to each cluster
213: a straight line with a fixed slope and a varying intercept. We
214: have assumed that this slope corresponds to that found for the
215: Coma cluster in the considered range of $\sigma$. Note that we do
216: not have any, statistically significant, evidence of a variation
217: of the slope between the different clusters. In Table~\ref{table1}
218: we list the derived relative abundances at a fixed velocity
219: dispersion of $\sigma=200$~km~s$^{-1}$ for each cluster.
220:
221:
222: As a quantitative indicator of the masses of the clusters, we have
223: used X-ray luminosities. $L_{\rm X}$ values were taken from
224: \citet{ebe98} and \citet{led03}, adapted to the cosmological
225: parameters used by the latter.
226:
227: Figure~\ref{fig3} shows the values of $[Z_{\rm CN}/Z_{\rm Fe}]$,
228: $[Z_{\rm Mg}/Z_{\rm Fe}]$ and $[Z_{\rm Mg}/Z_{\rm CN}]$ versus
229: X-ray luminosity for each cluster. We find clear correlations
230: between [CN/Fe] and [Mg/CN] values and X-ray luminosity, with probabilities of
231: no correlation of 0.046 and 0.002, respectively. On the
232: contrary, no correlation is found for [Mg/Fe]. It is worth noting
233: that the significant point in these relations is the relative
234: differences in abundance ratios, not their absolute values. It is
235: noticeable that the abundance ratio values also correlate with the
236: richness class and the morphological type (see Table~\ref{table1}), with
237: probabilities of no correlation of 0.05 and 0.02, respectively.
238:
239: \section{Discussion}
240:
241: The correlations can be interpreted in terms of the
242: different formation timescales for each element, and the different
243: star formation histories of early-type galaxies as a function of
244: their environment.
245:
246: Magnesium is ejected into the interstellar medium (ISM) by
247: Type II supernovae (SNe II) on short timescales ($<10$~Myr). On
248: the other hand, the iron-peak elements are the products of SNe~Ia,
249: which occur on timescales of $\sim1$~Gyr. Between the two
250: extremes, although there are recent suggestions that most of the C
251: come from massive stars \citep{ake04}, C and N are mainly ejected
252: into the ISM by low- and intermediate-mass stars
253: \citep{ren81,chi03}, leading to CN formation on timescales longer
254: than for Mg but shorter than for Fe. Furthermore, several authors
255: \citep[e.g.][]{ell97,sta98} argue that early-type galaxies are old
256: and passively evolving systems. In any case, the
257: luminosity-weighted ages derived from our model grids confirm that
258: the galaxies are significantly older than the formation timescales
259: of the different species. So, if we find substantial differences
260: in the abundance ratios of these elements which depend on the
261: physical properties of the environment, these must be due to the
262: fact that galaxies are assembled on different timescales as a
263: function of their environment.
264:
265: In this framework, the constancy of the [Mg/Fe]
266: values is explained in terms of the great difference in the
267: formation timescales of the two elements: the galaxies are fully
268: assembled before Type Ia SNe can significantly pollute with Fe the
269: ISM of the smaller galaxies before merging, and right after Mg is
270: fully ejected. Since [Mg/Fe] is found to be constant with the
271: X-ray luminosity of the clusters (see Fig.~\ref{fig3}), which is
272: an indicator of their mass, we conclude that this ratio is
273: independent of the environment. Similar results for the [Mg/Fe]
274: ratio have been obtained by other authors
275: \citep{jor99,kun02,pat03}, by studying field and Coma cluster
276: elliptical galaxies.
277:
278: However, when considering species with less disparate formation
279: timescales, such as CN and Fe, or CN and Mg, clear correlations
280: are found between abundance ratios and the environment, as shown
281: in Fig.~\ref{fig3}. The fact that [CN/Fe] decreases with the
282: cluster X-ray luminosity, and that [Mg/CN] increases with it,
283: suggests that galaxies in more massive clusters are fully
284: assembled on shorter timescales than those in less massive
285: clusters. We show that this difference is large enough to produce measurable
286: variations of the abundance ratios of galaxies in more or less
287: massive clusters.
288:
289: The result that there exist relative differences in the assembly timescales of
290: the galaxies due to the properties of the environment is qualitatively in agreement with the
291: hierarchical models. Discrepancies appear, however, when considering the absolute values
292: of such timescales. The fact that [CN/Fe] abundance ratio is not constant implies
293: that early-type galaxies are fully assembled on timescales around the massive
294: release of CN into the ISM. Hierarchical models, on the
295: contrary, predict longer assembly timescales.
296:
297: Other scenarios have been explored in order to explain the
298: differences in abundance ratio values as functions of the
299: environment \citep[see][]{pat03}. These include
300: a decrease in the stellar giant/dwarf ratio in high-density
301: environments, with respect to low-density ones, which would lead
302: to lower index values in the latter. But model calculations have
303: shown that the differences due to a variation in the IMF are too small to produce the observed variations. Also, a
304: difference in the luminosity-weighted mean age between high- and
305: low-density environments has been proposed. Models
306: show that, to account for the differences, the galaxies
307: in high-density environments must be $\sim8$~Gyr younger than in
308: low-density ones. This contradicts previous studies that
309: suggest that galaxies in high-density environments are, in any
310: case, older than those in low-density environments \citep{kun02}. So our interpretation here seems the most consistent.
311:
312: It is noteworthy that we have found a dependence of the
313: abundance ratios with several properties of the environment, both
314: quantitative (i.e. X-ray luminosity values) and qualitative (i.e.
315: richness and morphological types) as shown in Table~\ref{table1}.
316: This lends strong support to the basic hypothesis that the
317: characteristics of the environment affect the evolution of the
318: galaxies. The relations we have found set clear constraints on
319: models of chemical evolution and galaxy formation.
320:
321: \acknowledgements
322: The authors thank A. Aguerri, X. Barcons, L. Carigi, C. Guti\'errez, R. Peletier and the anonymous referee for useful comments. We acknowledge grant AYA2001-0435 from the Spanish Ministry of Science and Technology. The SDSS is managed by the ARC for the Participating Institutions.
323:
324:
325:
326: \begin{thebibliography}{}
327:
328: \bibitem[Akerman et al.(2004)]{ake04}Akerman, C.J., Carigi, L., Nissen, P.E.,
329: Pettini, M., \& Asplund, M. 2004, \aap, 414, 931
330: \bibitem[Bernardi et al.(2003)]{ber03}Bernardi, M. et al. 2003, \aj, 125, 1817
331: \bibitem[Cardiel et al.(2003)]{car03}Cardiel, N., Gorgas, J., S\'anchez-Bl\'azquez,
332: P., Cenarro, A.J., Pedraz, S., Bruzual, G., \& Klement, J. 2003, \aap, 409, 511
333: \bibitem[Chiappini, Romano \& Matteucci(2003)]{chi03}Chiappini, C., Romano, D., \&
334: Matteucci, F. 2003, \mnras, 339, 63
335: \bibitem[Ebeling et al.(1998)]{ebe98}Ebeling, H., Edge, A.C., B\"{o}hringer, H., Allen,
336: S.W., Crawford, C.S., Fabian, A.C., Voges, W., \& Huchra, J.P. 1998, \mnras, 301,
337: 881
338: \bibitem[Ellis et al.(1997)]{ell97}Ellis, R.S., Smail, I., Dressler, A., Couch,
339: W.J., Oemler, A.J., Butcher, H., \& Sharples, R.M. 1997, \apj, 483, 582
340: \bibitem[Faber, Worthey, \& Gonz\'alez(1992)]{fab92}Faber, S.M., Worthey, G., \& Gonz\'alez, J.J.
341: 1992, in IAU Symp. 149, Stellar Populations of Galaxies, ed. B. Barbury \& A.
342: Renzini (Dordrecht: Kluwer), 255
343: \bibitem[J{\o}rgensen(1999)]{jor99}J{\o}rgensen, I. 1999, \mnras, 306, 607
344: \bibitem[Kauffmann et al.(2003)]{kau03}Kauffmann, G. et al. 2003, \mnras, 341, 54
345: \bibitem[Kuntschner \& Davies(1998)]{kun98}Kuntschner, H., \& Davies, R. \mnras,
346: 295, 29
347: \bibitem[Kuntschner et al.(2002)]{kun02}Kuntschner, H., Smith, R., Colles, M.,
348: Davies, R., Kaldare, R., \& Vazdekis, A. 2002, \mnras, 337, 172
349: \bibitem[Larson(1974)]{lar74}Larson, R.B. 1974, \mnras, 166, 585
350: \bibitem[Ledlow et al.(2003)]{led03}Ledlow, M.J., Voges, W., Owen, F.N., \& Burns,
351: J.O. 2003, \aj, 126, 2740
352: \bibitem[Matteucci(1994)]{mat94}Matteucci, F., 1994, \aap, 288, 57
353: \bibitem[Peletier(1989)]{pel89}Peletier, R.F. 1989, Ph.D. thesis, Univ. Groningen
354: \bibitem[Press \& Schechter(1974)]{pre74}Press, W.H., \& Schechter, P. 1974,
355: \apj, 187, 425
356: \bibitem[Renzini \& Voli(1981)]{ren81}Renzini, A., \& Voli, M. 1981, \aap, 94, 175
357: \bibitem[S\'anchez-Bl\'azquez et al.(2003)]{pat03}S\'anchez-Bl\'azquez, P., Gorgas,
358: J., Cardiel, N., Cenarro, J., \& Gonz\'alez, J.J. 2003, \apjl,
359: 590, L91
360: \bibitem[Schiavon et al.(2002)]{sch02}Schiavon, R.P., Faber, S.M., Castilho,
361: B.V., \& Rose, J.A. 2002, \apj, 580, 850
362: \bibitem[Stanford, Eisenhardt, \& Dickinson(1998)]{sta98}Stanford, S.A.,
363: Eisenhardt, P., \& Dickinson, M. 1998, \apj, 492, 461
364: \bibitem[Stoughton et al.(2002)]{sto02}Stoughton, C. et al. 2002, \aj, 123, 485
365: \bibitem[Tripicco \& Bell(1995)]{tri95}Tripicco, M., \& Bell, R.A. 1995, \aj,
366: 110, 3035
367: \bibitem[Vazdekis(1999)]{vaz99}Vazdekis, A. 1999, \apj, 513, 224
368: \bibitem[Vazdekis et al.(2001b)]{vaz01b}Vazdekis, A., Kuntschner, H., Davies, R.,
369: Arimoto, N., Nakamura, O., \& Peletier, R. 2001b, \apjl, 551, L127
370: \bibitem[Vazdekis et al.(1997)]{vaz97}Vazdekis, A., Peletier,R.F., Beckman,J.E.,
371: \& Casuso,E. 1997, \apjs, 111, 203
372: \bibitem[Vazdekis et al.(2001a)]{vaz01a}Vazdekis, A., Salaris, M., Arimoto, N.,
373: \& Rose, J.A. 2001a, \apj, 549, 274
374: \bibitem[White \& Frenk(1991)]{whi91}White, S.D.M., \& Frenk, C.S. 1991, \apj,
375: 379, 521
376: \bibitem[Worthey(1994)]{wor94b}Worthey, G. 1994, \apjs, 95, 107
377: \bibitem[Worthey(1998)]{wor98}Worthey, G. 1998, \pasp, 110, 888
378: \bibitem[Worthey et al.(1992)]{wor92}Worthey, G., Faber, S.M., \& Gonz\'alez, J.J.
379: 1992, \apj, 398, 69
380: \bibitem[Worthey et al.(1994)]{wor94a}Worthey, G., Faber, S.M., Gonz\'alez, J.J., \&
381: Burstein, D. 1994 \apjs, 94, 687
382:
383: \end{thebibliography}
384:
385:
386:
387: \begin{deluxetable}{ccccccc}
388: \tablecolumns{7}
389: \tablecaption{Properties and measurements of
390: analyzed clusters}
391: \tablehead{ \colhead{Cluster} & \colhead{$z$} &
392: \colhead{Rich.\tablenotemark{a}} &
393: \colhead{Morph. $^{\rm b}$} & \colhead{$L_{\rm X}$ $^{\rm c}$
394: } & \colhead{$[Z_{\rm CN}/Z_{\rm Fe}]$} &
395: \colhead{$[Z_{\rm Mg}/Z_{\rm Fe}]$}}
396: \startdata A0279 & 0.080 & 1
397: & II--III & 0.08 & 0.43 & 0.33\\ A1238 & 0.073 & 1 & III &
398: 0.15 & 0.61 & 0.44\\ Virgo & 0.004 & 1 & III & 0.30 & 0.48 &
399: 0.36\\ A0257 & 0.070 & 1 & II--III & 0.31 & 0.45 & 0.35\\ A2050 &
400: 0.118 & 1 & II--III & 1.22 & 0.40 & 0.31\\ Coma & 0.023 & 2 & II
401: & 1.80 & 0.39 & 0.38\\ A0655 & 0.127 & 3 & I--II & 1.97 & 0.30
402: & 0.43\\ A1650 & 0.085 & 2 & I--II & 2.01 & 0.40 &~~0.39
403: \enddata
404: \tablenotetext{a}{Abell richness class. $^{\rm b}$~Bautz-Morgan morphological type. $^{\rm c}$~X-ray luminosity in units of 10$^{44}$ erg~s$^{-1}$.}
405: \label{table1}
406: \end{deluxetable}
407:
408:
409:
410: \begin{figure}
411: \figurenum{1}
412: \epsscale{0.8}
413: \plotone{f1.eps}
414: \caption{Model grids
415: for the galaxies of clusters A1238 (filled circles) and A655 (open
416: circles). These two clusters have extreme values of X-ray
417: luminosity. {\it From left to right:} Plots of H$\beta$ vs. the
418: metallicity indices CN$_2$, Mg$_2$ and Fe2 (defined in the text).
419: The galaxy velocity dispersions increase from top to bottom, in
420: the three bins of $\sigma$ quoted in the boxes, to which galaxy
421: and model spectra have been broadened. Overplotted are the models
422: by \citet{vaz99}. Lines of constant [Fe/H] $=-0.7,-0.4,0.0$ and
423: $0.2$ are shown by dot-dashed, dotted, solid and dashed lines,
424: respectively. Thin dotted horizontal lines represent models of
425: constant age, quoted in gigayears. The ages become increasingly uncertain for
426: values greater than 8 Gyr \citep[see][]{wor94b,vaz99} but these uncertainties have no
427: significant effects on relative abundance determination.}
428: \label{fig1}
429: \end{figure}
430:
431:
432: \begin{figure}
433: \figurenum{2}
434: \epsscale{0.8}
435: \plotone{f2.eps}
436: \caption{Abundance
437: ratios $[Z_{\rm CN}/Z_{\rm Fe}]$ {\it (left)}, $[Z_{\rm Mg}/Z_{\rm
438: Fe}]$ {\it (centre)} and $[Z_{\rm Mg}/Z_{\rm CN}]$ {\it (right)}
439: versus galaxy velocity dispersion, $\sigma$. Each point
440: corresponds to one galaxy within a cluster. Errors are
441: computed from the measured index error bars shown in
442: Fig.~\ref{fig1}. Galaxy clusters are ordered by decreasing X-ray
443: luminosity, from top to bottom. The dotted straight lines show the
444: level of relative abundances for each cluster. See the text for
445: more details.}
446: \label{fig2}
447: \end{figure}
448:
449:
450: \begin{figure}
451: \figurenum{3}
452: \epsscale{0.8}
453: \plotone{f3.eps}
454: \caption{Cluster
455: X-ray luminosity versus overabundance values of $[Z_{\rm
456: CN}/Z_{\rm Fe}]$ {\it(top)}, $[Z_{\rm Mg}/Z_{\rm Fe}]$
457: {\it(middle)} and $[Z_{\rm Mg}/Z_{\rm CN}]$ {\it(bottom)}.
458: Circles, squares and triangles indicate clusters with richness
459: classes of 1, 2 and 3, respectively. Each point corresponds to one
460: individual cluster and represents the interpolated abundance ratio
461: for $\sigma=200$~km~s$^{-1}$. The Spearman rank-order correlation
462: coefficients and its significance values are written in top and
463: bottom panels.}
464: \label{fig3}
465: \end{figure}
466:
467:
468:
469:
470: \end{document}
471: