1: % aa.dem
2: % AA vers. 6, LaTeX class for Astronomy & Astrophysics
3: % demonstration file
4: % (c) Springer-Verlag HD
5: % revised by EDP Sciences
6: %-----------------------------------------------------------------------
7: %
8: %\documentclass[referee]{aa} % for a referee version
9: %\documentclass[onecolumn]{aa} % for a paper on 1 column
10: %\documentclass[longauth]{aa} % for the long lists of affiliations
11: %\documentclass[rnote]{aa} % for the research notes
12: %
13: \documentclass{aa}
14: %
15: \usepackage{graphicx}
16: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
17: \usepackage{txfonts,natbib}
18: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
19: %
20: \begin{document}
21: %
22: \title{Dynamical evidence of the age--metallicity relation in the Milky Way disk\thanks{This
23: paper is based on data collected at the Observat\'orio do Pico dos Dias, operated by Laborat\'orio
24: Nacional de Astrof\'{\i}sica, MCT, Brazil.}}
25:
26: \author{H. J. Rocha-Pinto\inst{1}
27: \and R. H. O. Rangel\inst{1}
28: \and G. F. Porto de Mello\inst{1}
29: \and G. A. Bragan\c ca\inst{1}
30: \and W. J. Maciel\inst{2}
31: }
32:
33:
34: \offprints{H. J. Rocha-Pinto}
35:
36: \institute{Observat\'orio do Valongo, Universidade Federal do Rio de Janeiro,
37: Ladeira do Pedro Ant\^onio 43, 20080-090 Rio de Janeiro RJ, Brazil.
38: \and
39: Instituto de Astronomia, Geof\'{\i}sica e Ci\^encias Atmosf\'ericas, Universidade de S\~ao Paulo,
40: R. do Mat\~ao 1226, 05508-900 S\~ao Paulo SP, Brazil
41: }
42:
43: \titlerunning{Dynamical evidence of the age--metallicity relation}
44: \authorrunning{H. J. Rocha-Pinto et al.}
45:
46: \date{Received date; accepted date}
47:
48:
49:
50: % \abstract{}{}{}{}{}
51: % 5 {} token are mandatory
52:
53: \abstract
54: % context heading (optional)
55: % {} leave it empty if necessary
56: {}
57: % aims heading (mandatory)
58: {We studied the relationship between the average stellar abundance of several elements
59: and the orbital evolution of stars in the neighbourhood of the Sun. }
60: % methods heading (mandatory)
61: {We used both observational data for 325 late-type dwarfs in
62: a volume-complete sample and simulations of the orbital diffusion. Metallicities, ages, and initial
63: position and velocities for the simulated stars are sampled from empirical distributions of these
64: quantities in the Milky Way.}
65: % results heading (mandatory)
66: {We found that that there is a relationship between the average stellar abundance of Fe, Na, Si, Ca, Ni, and Ba
67: and the mean orbital radius of stars currently passing through the solar neighbourhood. The greater the difference
68: between the mean orbital radius and the solar Galactocentric distance, the more deficient the star is, on average,
69: in these chemical species. }
70: % conclusions heading (optional), leave it empty if necessary
71: {The stars that take a longer time to come from their birthplaces to arrive in the present solar neighbourhood are more
72: likely to be more metal-poor than those that were born here. This result is a direct, independent indication
73: that a tightly defined Galactic age-metallicity relation exists.}
74:
75: \keywords{Galaxy: evolution -- stars: late-type -- solar neighbourhood}
76:
77: \maketitle
78:
79:
80: \section{Introduction}
81:
82: Although one of the core consequences of the chemical evolution
83: theory is the gradual increase in the metal content of the
84: interstellar medium (ISM) and the
85: progressive enrichment of subsequent stellar generations, some authours
86: have found little, if any, indication that
87: an age--metallicity relation (AMR) exists amongst solar
88: neighbourhood late-type stars (\citealt{feltzing,bruxa}).
89: None of these papers question the predictions of the chemical evolution
90: theory concerning the enrichment of the ISM. The discrepancy
91: between theory and data is usually interpreted as a result of an
92: inefficient mix of the stellar ejecta into the surrounding medium, in spite of
93: the growing evidence that the present dispersion in the ISM over distances of 1
94: kpc is very small
95: \citep{scaloelmo}. However, other authors have found a much tighter
96: relationship between age and metallicity (see \citealt{r00}).
97:
98: A way out of
99: this deadlock was recently proposed by \citet{pont}, who showed
100: that the non-linear separation of the isochrones in the HR diagram
101: can introduce a statistical bias in the isochrone age measurement,
102: artificially assigning a greater isochrone age to the stars.
103: \citet{pont} showed that if a proper statistical method is used to
104: correct for this bias, a tight AMR can be found from the same data
105: used by some of the `loose' AMR proponents. On the other hand, it
106: is true that chromospheric ages also suffer from their own biases
107: and cannot be considered perfect age estimators
108: \citep{donahue,pace}. Since much of the problem stems from the
109: drawbacks of age determination methods, the existence of an AMR
110: can be tested with the use of an age-related quantity that is
111: independent of any age determination method. In this Letter, we
112: show that the mean orbital radius of stars presently passing
113: through the solar neighbourhood, $R_m$, is mildly related to the
114: stellar age and that it can provide this independent test of the
115: existence of the AMR.
116:
117: \section{Sample selection and observations}
118:
119: Our sample consists of stars from the Hipparcos catalogue
120: \citep{hipparcos} that also have {\it uvby} indices measured by
121: \citet{olsen83,olsen93,olsen94} or compiled by \citet{HM98} and
122: which could be observed from the southern hemisphere. A
123: heliocentric distance cut of 25 pc was used in order to have a
124: sample of nearly 300 stars, so it would be possible to complete
125: the programme within a few observation nights at a small telescope.
126: Our final sample has 325 stars.
127:
128: The observations were carried out at the Laborat\'orio Nacional de
129: Astrof\'{\i}sica, as part of a survey of the chemical abundances of
130: late-type dwarfs, during 15 nights between July 1999 and October
131: 2001. The coud\'e spectrograph of the 1.60m telescope was used
132: with the 1800 l/mm grating and a 1024-pixel CCD, pixel size being
133: 24$\mu$. The spectra were centred at $\lambda 6145\AA$, covered about
134: 150 $\AA$, and had a nominal resolution of 0.3 $\AA$, with an
135: average S/N ratio of 100 to 200. All spectra were reduced in the
136: standard way; continuum normalization was performed homogeneously
137: for all spectra. The spectral range contains a number of
138: moderately strong, unblended transitions: 2 for Na I, 3 for Si I,
139: 3 for Ca I, 10 for Fe I, 2 for Fe II, 5 for Ni I, and one for Ba
140: II. The equivalent widths (EWs) of the metal lines were converted
141: to elemental abundances using the MARCS model atmospheres
142: described by \citet{edv}, in a differential analysis with the Sun
143: as standard star. Solar $gf$ values were derived by forcing EWs
144: measured off Moon spectra, obtained in the same conditions as
145: the stellar spectra, to reproduce the standard solar abundances.
146: Effective temperatures for the program stars were derived from
147: $(b-y)$, $(B-V)$ and $(B_{\rm T}-V_{\rm T})$ calibrations described
148: by \citet{delpelosoetal2005}. Surface gravities were obtained from
149: the ionization equilibrium of the Fe I and Fe II lines. A full
150: description of the abundance analysis will be presented elsewhere.
151: The internal uncertainty of the [X/H] abundances is $\sim$~0.1 dex,
152: except for Ba II, which has more uncertain abundances, with only
153: one spectral line measured.
154:
155: Radial velocities ($v_{\rm r}$) for most of the 325 observed stars
156: were calculated from Th-Ar comparison spectra. For a few of them,
157: comparison spectra could not be obtained, and their $v_{\rm r}$
158: were taken from the SIMBAD database. We calculated the $U$,
159: $V$, and $W$ components of the spatial velocity around the Galactic
160: centre, from the $v_{\rm r}$, Hipparcos parallaxes and proper motions.
161: Thereafter, the orbital parameters (mean orbital
162: galactocentric radius, eccentricity, maximum height above the
163: plane) for each star were calculated as described in
164: \citet[][hereafter, R04]{r04}. In this paper, we have adopted the
165: present solar Galactocentric radius, $R_\odot$, as 8.0 kpc.
166:
167: A full chemokinematical analysis of this data sample will be
168: presented elsewhere, but here, we are interested in an apparent trend
169: between $R_m$ and the stellar abundances. Figure \ref{ztrend}
170: shows plots of average elemental abundance, [X/H], averaged over
171: several $R_m$ bins, for Fe, Na, Si, Ca, Ni, and Ba. According to
172: \citet{grenon} and \citet{edv}, $R_m$ can be taken as an indicator
173: of the stellar Galactocentric birth radius, $R^\ast$. \citet{wfd}
174: have shown that this relation between age and $R_m$ can only be
175: considered in a statistical sense. However, \citeauthor*{r04}
176: found a correlation between the age and $R_m$ range for the stars
177: presently crossing the solar neighbourhood: younger stellar
178: generations have $R_m$ near $R_\odot$, while the older generations
179: have the broader $R_m$ distributions typical of stars coming from
180: different parts of the Galaxy into the solar neighbourhood. This
181: finding is consistent with Grenon's proposition that $R_m$ stays
182: more or less close to $R^\ast$ and reinforces previous claims that
183: $R_m$ can be used to estimate $R^\ast$.
184:
185: In Fig. \ref{ztrend}, we show that the average abundances for
186: stars presently crossing the solar neighbourhood also differ as a
187: function of $R_m$. The stars with $R_m \approx R_\odot$ have
188: near-solar $\langle[{\rm X/H}]\rangle$, while stars supposedly
189: born farther and farther from the solar neighbourhood (i.e., those
190: with larger $|R_m-R_\odot|$) have progressively decreasing
191: $\langle[{\rm X/H}]\rangle$. This decrease can be as high as 0.3
192: dex for Fe and 0.4 dex for Ba, between $R_m=8$ and $R_m=6$ kpc.
193:
194: An anonymous referee has pointed out that the same trend can be seen
195: in seen in Table 14 of \citet{edv}; we have also found it for [Me/H] amongst
196: late-type dwarfs from the Geneva-Copenhagen survey (see
197: Fig. \ref{result}).
198:
199:
200:
201: \begin{figure}
202: \resizebox{\hsize}{!}{\includegraphics{f1.eps}}
203: \caption[]{Average abundance of Fe, Na, Si, Ca, Ni, and Ba as a function of the mean orbital Galactocentric radius for stars
204: presently crossing the solar neighbourhood.
205: The average abundance is larger for stars having $R_m$ around the present solar Galactocentric radius, and it decreases by
206: more than 0.2 dex for stars at extreme $R_m$. The vertical and horizontal error bars are standard
207: deviations around $\langle[{\rm X/H}]\rangle$ and $\langle R_m\rangle$, respectively, at each bin.}
208: \label{ztrend}
209: \end{figure}
210:
211: Disk stellar velocities peak near $\Omega(R^\ast)$, the average
212: rotational velocity at the stellar birth radius. Most of this
213: motion goes in the azimuthal direction, not in the radial
214: direction. Radial and vertical movement is mostly driven by the
215: orbital diffusion \citep{wielen}, on account of stellar encounters
216: with massive objects like giant molecular clouds. As
217: time goes on, the distance between the apogalactic and perigalactic
218: radius increases due to the random kicks that
219: stars get from collisions, and the chance the star could be caught
220: at a different galactocentric distance from its birth radius
221: increases as well. Thus, since $R_m$ can be seen as a stellar
222: birth radius indicator, we expect that for our local sample
223: $|R_m - R_\odot|$ reflects the time the star spent to come from
224: $R^\ast$ to $R_\odot$. In other words, $|R_m - R_\odot|$ can be
225: taken as an independent age estimator. It has an obvious advantage
226: of neither depending on stellar evolution models nor being based on
227: chromospheric properties. However, like other kinematical age
228: estimators, it does not allow us to assign ages for individual
229: stars, but to a stellar group. This is easily seen from Figure 7
230: by \citeauthor*{r04}: The larger the age spectrum, and consequently the age
231: uncertainty, the smaller $|R_m - R_\odot|$. Hence, while
232: stars with $R_m\approx R_\odot$ have a wide age range, stars
233: that came into the solar neighbourhood from other radii are, on
234: average, older than the stars born near $R_\odot$.
235:
236: If this is the case, the relation between average abundance and
237: $R_m$ seen in Fig. \ref{ztrend} reflects the existence of a
238: relation between age and elemental abundances that is similar to the AMR;
239: otherwise, if there were no AMR in the Galactic disk, as hinted
240: by \citet{feltzing}, solar neighbourhood stars with large $|R_m -
241: R_\odot|$ (old stars according to the reasoning presented in
242: the previous paragraph) could have any abundance within the typical
243: abundance range of disk stars, and their average would not be dissimilar
244: to the average abundance of stars having $|R_m - R_\odot|\approx 0$.
245:
246: \section{Simulation of the $R_m$--abundance plane}
247:
248: To test this hypothesis, we ran chemodynamical
249: simulations of a stellar sample, using a reasonable Galactic model
250: constrained by empirical data. Because we are interested in thin
251: disk stars, we did not address the simulation of the thick disk or
252: halo, although the effect of these Galactic components on the
253: Galactic potential is taken into account.
254:
255: The starting position of a simulated star in $X$, $Y$, and $Z$
256: coordinates with respect to the Galactic centre is randomly chosen
257: following the thin disk density law of \citet{siegel}. The star
258: formation rate was considered to be constant, and a randomly uniform age
259: from 0 to 15 Gyr is assigned to each star. The simulated stellar
260: metallicity follows the distribution
261: %
262: \begin{equation}
263: {\rm [Fe/H]} \sim N\left(z(\tau) + (R-R_\odot)\partial_{R}{\rm [Fe/H]},\sigma^2\right),
264: \label{metsim}
265: \end{equation}
266: %
267: where $z(\tau)$ represents the AMR at age
268: $\tau$, $\sigma$ is the cosmic scatter, and $\partial_{R}{\rm
269: [Fe/H]}$ the radial metallicity gradient at the galactocentric
270: distance $R$. Finally, the starting velocities $V_r$, $V_\phi$, and
271: $V_z$ for each star, in the radial, azimuthal, and vertical
272: direction, respectively, with respect to the Galactic centre, were
273: sampled from a normal distribution $\sim N\left(\langle
274: V_i\rangle, \sigma^2_{V_i}\right)$, where $\langle V_r\rangle =
275: \langle V_z\rangle = 0$ and $\langle V_\phi \rangle = \Theta(R)$,
276: i.e., the rotation velocity at $R$, taken from \citet{law}. We took
277: $[\sigma_{V_r},\sigma_{V_\phi},\sigma_{V_z}]=[23,12,9]$ km/s from
278: \citeauthor*{r04}.
279:
280: The orbit of each star was then followed from its birth ---
281: given by the randomly assigned age --- to the present time, using
282: a program developed by Chris Flynn (details in \citeauthor*{r04}).
283: To simulate stellar encounters with massive objects, a random push
284: in each velocity component was given at each integration step. The
285: magnitude of the push in each velocity component was set by
286: requiring that the increase in the velocity dispersion with age of
287: the stellar group for the simulated stars in the solar vicinity
288: fitted the observed age--velocity dispersion relation (given by \citeauthor*{r04}).
289: After integrating the orbit, stars whose present-time galactocentric
290: radius differed by less than 0.5 kpc from $R_\odot$ were selected
291: as ``solar neighbourhood'' stars. It is from this simulated
292: subsample of stars passing through the solar neighbourhood that we
293: take the predicted $R_m$--abundance relation to compare with the
294: results from Fig. \ref{ztrend}.
295:
296: \begin{figure}
297: \resizebox{\hsize}{!}{\includegraphics{f2.eps}}
298: \caption[]{Comparison between the observed and predicted
299: $R_m$--[Fe/H] trend for stars passing through
300: the solar neighbourhood. The hatched grey area and
301: white squares correspond to standard deviations around the
302: average [Fe/H] for the same $R_m$ intervals for 13017
303: late-type dwarfs in the Geneva-Copenhagen survey \citep{bruxa}.
304: The lines correspond to
305: the results of a Galactic model with no AMR (dashed line),
306: loose AMR (dotted line), and
307: tight AMR (solid line), as detailed by the parametrizations
308: in Table \ref{paramet}.}
309: \label{result}
310: \end{figure}
311:
312: Since our goal was simply to show that the observed trend in the
313: $R_m$--abundance plane is determined by the existence of a
314: well-defined AMR, we used three single parametrizations for the
315: AMR: i) pure scatter relation, for which no change in the
316: average metallicity or abundance scatter occurs with time; ii) a loose
317: AMR, corresponding to a polinomial fit to the data in \citet{edv};
318: iii) a relatively tight relation according to which the average
319: metallicity of younger stars is higher than that for older stars.
320: The adopted parametrizations for these two cases are given in
321: Table \ref{paramet}.
322:
323: \begin{table}
324: \caption[]{Parametrizations used in the simulations}
325: \label{paramet}
326: \begin{flushleft}
327: \begin{tabular}{lc}
328: \hline\hline
329: function & parametrization \\
330: \hline\hline
331: no AMR & $z(\tau)=-0.15$; $\sigma=0.3\,{\rm dex}$ \\
332: loose AMR & $z(\tau)=0.0239 - 0.0355\tau - 0.0007\tau^2$ \\
333: & $\sigma=0.22\,{\rm dex}$ \\
334: % \hline
335: \strut tight AMR & $z(\tau)= 1.51 - 44.25(34.23-\tau)^{-1}\strut$ \\
336: & $\sigma=0.1\,{\rm dex}$ \\
337: \hline
338: single-slope $\partial_{R}$[Fe/H] & $-$0.07 dex kpc$^{-1}\phantom{, \quad R < R_\odot}$ \cr
339: % \hline
340: two-slope $\partial_{R}$[Fe/H] & $-$0.05 dex kpc$^{-1}, \quad R \le R_\odot $\cr
341: & $-$0.03 dex kpc$^{-1}, \quad R > R_\odot$ \cr
342: \hline\hline
343: \end{tabular}
344: \end{flushleft}
345: \end{table}
346:
347: Our simulations consider an initial number of 6000 stars, in order
348: to have nearly 300 stars within 0.5 kpc from $R_\odot$ at the end
349: of the integration to compare with the observational sample.
350: Fig. \ref{result} summarises our simulations of the
351: $R_m$--abundance plane.
352: %
353: Three curves are shown, which correspond to
354: the $R_m$--$\langle{\rm [Fe/H]}\rangle$ trend found using the tight AMR, loose AMR,
355: and no--AMR case, each with the two-slope metallicity
356: gradient given in Table \ref{paramet}.
357: %
358: Clearly the no--AMR
359: simulation failed to produce the arc-shaped $R_m$--$\langle{\rm [Fe/H]}\rangle$ trend
360: seen in the observational data. This failure is not difficult to
361: understand, because if there is no AMR in the Galactic disk, there is no
362: relation between how much time a star takes to come from $R^\ast$
363: to $R_\odot$ and its metallicity. The $R_m$--$\langle{\rm [Fe/H]}\rangle$ trend for this
364: case is close to the adopted average metallicity for this
365: parametrization ([Fe/H] $\approx -0.15$ dex). The trend also
366: decreases very slowly with $R_m$ on account of the adopted
367: $\partial_{R}{\rm [Fe/H]}$.
368:
369: The solid curve in Fig. \ref{result} is very similar to the
370: observed trend. It is the resulting $R_m$--$\langle{\rm [Fe/H]}\rangle$ trend for the
371: case of a tight AMR with very small intrinsic metallicity
372: dispersion. We have made no attempt to find a best-fit
373: curve, since the goal of the simulations is to show
374: that the observational trend between $R_m$ and [Fe/H] can be
375: explained with reasonable assumptions for the disk parameters,
376: provided that there is a tight AMR. Several other parametrizations
377: for the AMR and metallicity gradient were used, in order to verify
378: the effect of the parametrization on the $R_m$--$\langle{\rm [Fe/H]}\rangle$ trend. We
379: verified that the shape of the resulting curve is
380: most sensitive to the overall metallicity growth from the
381: beginning of the Galactic disk to the present time, as well as to
382: $\sigma$. A lower overall metallicity growth and/or larger
383: $\sigma$ flatten the simulated $R_m$--$\langle{\rm [Fe/H]}\rangle$ relation, making it look
384: closer to the no--AMR case. On the other hand, the effect of
385: $\partial_{R}{\rm [Fe/H]}$ is more pronounced on the tails of the
386: $R_m$--$\langle{\rm [Fe/H]}\rangle$ curve. We verified that a two-slope metallicity
387: gradient \citep{maciel} yields a better eye-ball fit to the observed trend than
388: a single-slope gradient. However, the effect is small and should
389: be investigated more before any conclusions can be made.
390:
391: Our simulations have only considered the $R_m$--$\langle{\rm [Fe/H]}\rangle$ trend
392: because there is much more information available in the literature
393: for the observed relations between [Fe/H], $\tau$, and $R$. However,
394: the same exercise could be done for Ca, Na, Si, Ni, and Ba, as long
395: as their abundances with respect to H have grown since the
396: beginning of the Galactic disk and the intrinsic cosmic scatter at
397: any age was low. It is interesting to verify in Fig.
398: \ref{ztrend} that the biggest difference in the abundance at
399: $R_m\approx R_\odot$ and large $|R_m-R_\odot|$ is found for Ba.
400: Barium is thought to present a larger overall growth with age than
401: iron \citep{edv}, since the bulk of its nuclesynthesis is due
402: to long-lived AGB stars, which make their presence felt in the
403: chemical evolution of the Galaxy with a lag with respect to iron
404: sources. From our
405: simulations, a larger growth in abundance with age increases the
406: arc-like shape of the $R_m$--$\langle{\rm [Fe/H]}\rangle$ relation, on account of the
407: difference in abundance from the older stars with larger
408: $|R_m-R_\odot|$ to the average abundance stars with $R_m\approx
409: R_\odot$. It is also interesting to note that Ca and Si (and
410: perhaps also Na) seem to have the flatter profiles with $R_m$ in
411: Fig. 1. That they are $\alpha$-elements, almost exclusively
412: products of type II supernovae, could account for this, since a less
413: well-defined age--abundance relation is expected in this case.
414:
415: \section{Conclusions}
416:
417: Spectroscopic abundances for
418: 6 elements in 325 stars show that, on average, solar neighbourhood
419: stars with large $|R_m - R_\odot|$ have lower abundances,
420: regardless of the chemical element considered. This relation is
421: interpreted as evidence of a well-defined age--abundance relation,
422: since stars with large $|R_m - R_\odot|$ would mainly be old. A
423: number of simulations of the orbits of disk stars born with
424: chemokinematical properties that are reasonable for the Galactic disk
425: confirm this interpretation.
426:
427:
428: The constraint provided by our data on the tightness of the AMR is
429: relatively strong. Figure \ref{result} shows that a loose AMR, with
430: $\sigma\approx 0.20$ dex, is still compatible with an
431: arc-shaped $R_m$--$\langle{\rm [Fe/H]}\rangle$, albeit much less pronounced
432: than what is observed.
433:
434: Our method highlights the possibility of studying the AMR and its dispersion through a
435: technique that does not depend on age determination drawbacks. This could be achieved
436: through a larger stellar sample and accurate (preferably, spectroscopic) abundances in order
437: to better define the empirical $R_m$--$\langle{\rm [X/H]}\rangle$ relation.
438:
439:
440:
441: \begin{acknowledgements}
442: %GFPM
443: We thank the suggestions by an anonymous referee that substantially improved this Letter.
444: The authors acknowledge financial support by FAPERJ, FAPESP, CNPq,
445: PRONEX/FINEP, FUJB/UFRJ, and PIBIC/UFRJ. This research made use of the
446: SIMBAD database, operated by the CDS, Strasbourg, France.
447: \end{acknowledgements}
448: \begin{thebibliography}{}
449: %
450: \bibitem[del Peloso et al. (2005)]{delpelosoetal2005} del Peloso, E. F., da Silva, L. \& Porto de Mello, G.
451: F. 2005, A\&A, 434, 275
452: \bibitem[Donahue(1998)]{donahue} Donahue R.A. 1998, in Stellar Systems and the Sun, Donahue R.A. and Bookbinder J.A., eds.,
453: ASP Conf. Ser. 154, 1235
454: \bibitem[Edvardsson et al.(1993)]{edv} Edvardsson, B., Anderson, J., Gustafsson, B., Lambert, D. L.,
455: Nissen, P. E., \& Tomkin, J. 1993, A\&A, 275, 101
456: \bibitem[ESA(1997)]{hipparcos} ESA 1997, The Hipparcos and Tycho Catalogues, ESA SP-1200
457: \bibitem[Feltzing et al.(2001)Feltzing, Holmberg \& Hurley]{feltzing} Feltzing, S., Holmberg, J.,
458: \& Hurley, J. R. 2001, A\&A, 377, 911
459: \bibitem[Grenon(1987)]{grenon} Grenon, M. 1987, JA\&A, 123
460: \bibitem[Hauck \& Mermilliod(1998)]{HM98} Hauck, B., \& Mermilliod, M., 1998, A\&AS, 129, 431
461: \bibitem[Law et al.(2005)]{law} Law, D. R., Johnston, K. V., \& Majewski, S. R. 2005, ApJ, 619, 807
462: \bibitem[Maciel et al.(2006)]{maciel} Maciel, W. J., Lago, L. G., \& Costa, R. D. D. 2006, A\&A, in press (astro-ph/0603419)
463: \bibitem[Nordstr\"om et al.(2004)]{bruxa} Nordstr\"om, B., Mayor, M., Andersen, J., Holmberg, J.,
464: Pont, F., J\o rgensen, B. R., Olsen, E. H., Udry, S., \& Mowlavi, N. 2004, A\&A, 418, 989
465: \bibitem[Olsen(1983)]{olsen83} Olsen, E. H. 1983, A\&AS, 54, 55
466: \bibitem[Olsen(1993)]{olsen93} Olsen, E. H. 1993, A\&AS, 102, 89
467: \bibitem[Olsen(1994)]{olsen94} Olsen, E. H. 1994, A\&AS, 104, 429
468: \bibitem[Pace \& Paquini(2004)]{pace} Pace, G., \& Pasquini, L. 2004, A\&A, 426, 1021
469: \bibitem[Pont \& Eyer(2004)]{pont} Pont, F., \& Eyer, L. 2004, MNRAS, 351, 487
470: \bibitem[Rocha-Pinto et al.(2000)]{r00} Rocha-Pinto, H. J., Maciel, W. J., Scalo, J., \& Flynn, C. 2000,
471: A\&A, 358, 850
472: \bibitem[Rocha-Pinto et al.(2004)R04]{r04} Rocha-Pinto, H. J., Flynn, C., Scalo, J., H\"anninen, J., Maciel, W. J.,
473: \& Hensler, G. 2004, A\&A, 423, 517 (R04)
474: \bibitem[Scalo \& Elmegreen(2004)]{scaloelmo} Scalo, J., \& Elmegreen, B. G. 2004, ARA\&A, 42, 275
475: \bibitem[Siegel et al.(2002)]{siegel} Siegel M. H., Majewski, S. R., Reid, I. N., Thompson, I. B. 2002, ApJ, 578, 151
476: \bibitem[Wielen(1977)]{wielen} Wielen, R. 1977, A\&A, 60, 263
477: \bibitem[Wielen et al.(1996)]{wfd} Wielen, R., Fuchs, B., \& Dettbarn, C. 1996, A\&A, 314, 438
478:
479: \end{thebibliography}
480:
481:
482: % \bibitem[Ibukiyama \& Arimoto(2002)]{ibukiyama} Ibukiyama, A., \& Arimoto, N. 2002, A\&A 394, 927
483: % \bibitem[Johnson \& Soderblom(1987)]{johnson} Johnson, D. R. H., \& Soderblom, D. R. 1987, AJ, 864
484: % \bibitem[Marsakov, Shevelev, \& Suchkov(1990)]{marsakov} Marsakov, V. A., Shevelev, Iu. G., \&
485: % Suchkov, A. A. 1990, Ap\&SS, 172, 51
486: % \bibitem[Meusinger, Reimann, \& Stecklum(1991)]{meusinger} Meusinger H., Reimann H.-G., \& Stecklum B., 1991, A\&A, 245, 57
487: % \bibitem[Ng \& Bertelli(1998)]{ng} Ng, Y. K., \& Bertelli, G. 1998, A\&A, 329, 943
488:
489: \end{document}
490: