0807.4558/ms.tex
1: \documentclass[12pt,preprint]{aastex}
2: 
3: \bibliographystyle{apj}
4: \newcommand{\teff}{$T_{\rm eff}$} 
5: \newcommand{\kms}{km s$^{-1}$}
6: \begin{document}
7: 
8: \title{Chemical abundances in giants stars of the tidally disrupted 
9: globular cluster NGC 6712 from high-resolution infrared 
10: spectroscopy}
11: 
12: \author{David Yong}
13: \affil{Research School of Astronomy and Astrophysics, Australian National 
14: University, Mount Stromlo Observatory, Cotter Road, Weston Creek, ACT 2611, 
15: Australia}
16: \email{yong@mso.anu.edu.au}
17: 
18: \author{Jorge Mel{\' e}ndez}
19: \affil{Centro de Astrof{\' i}sica da Universidade do Porto, Rua das Estrelas, 
20: 4150-762 Porto, Portugal}
21: \email{jorge@astro.up.pt}
22: 
23: \author{Katia Cunha\altaffilmark{1}}
24: \affil{National Optical Astronomy Observatory, Casilla 603, La Serena, Chile} 
25: \email{kcunha@noao.edu}
26: 
27: \author{Amanda I.\ Karakas}
28: \affil{Research School of Astronomy and Astrophysics, Australian National 
29: University, Mount Stromlo Observatory, Cotter Road, Weston Creek, ACT 2611, 
30: Australia}
31: \email{akarakas@mso.anu.edu.au}
32: 
33: \author{John E.\ Norris}
34: \affil{Research School of Astronomy and Astrophysics, Australian National 
35: University, Mount Stromlo Observatory, Cotter Road, Weston Creek, ACT 2611, 
36: Australia}
37: \email{jen@mso.anu.edu.au}
38: 
39: \author{Verne V.\ Smith}
40: \affil{National Optical Astronomy Observatory, Casilla 603, La Serena, Chile}
41: \email{vsmith@noao.edu}
42: 
43: \altaffiltext{1}{On leave from
44: Observat\'orio Nacional; Rio de Janeiro, Brazil}
45: 
46: \begin{abstract}
47: We present abundances of C, N, O, F, Na, and Fe in 
48: six giant stars of the tidally disrupted globular cluster NGC 6712. 
49: The abundances were derived by comparing synthetic spectra with 
50: high resolution infrared spectra obtained with the Phoenix 
51: spectrograph on the Gemini South telescope. We find large star-to-star 
52: abundance variations of the elements C, N, O, F, and Na. NGC 6712 
53: and M4 are the only globular clusters in which F has been measured in 
54: more than two stars, and both clusters reveal F abundance variations 
55: whose amplitude is comparable to, or exceeds, that of O, 
56: a pattern which may be produced in $M$ $\gtrsim$ 5$M_\odot$ 
57: AGB stars. 
58: Within the limited samples, 
59: the F abundance in globular clusters is lower than in 
60: field and bulge stars at the same metallicity. 
61: NGC 6712 and Pal 5 are tidally disrupted globular 
62: clusters whose red giant members exhibit O and Na abundance variations 
63: not seen in comparable metallicity field stars. Therefore, globular clusters 
64: like NGC 6712 and Pal 5 cannot contribute many field stars and/or 
65: field stars do not form in environments with chemical enrichment 
66: histories like that of NGC 6712 and Pal 5. Although our sample size 
67: is small, from the amplitude of the O and Na abundance 
68: variations, we infer a large 
69: initial cluster mass and tentatively confirm that NGC 6712 was once 
70: one of the most massive globular clusters in our Galaxy. 
71: \end{abstract}
72: 
73: \keywords{Galaxy: Abundances, Galaxy: Globular Clusters: Individual: Messier 
74: Number: NGC 6712, Stars: Abundances}
75: 
76: \section{Introduction}
77: \label{sec:intro}
78: 
79: The formation and evolution of our Galaxy remains one of the great 
80: unanswered questions in modern astronomy. 
81: \citet{els62} suggested formation via the monolithic collapse of a 
82: gaseous protocloud on a timescale of 10$^8$ years. 
83: \citet{sz78} challenged this notion by proposing that 
84: the halo formed through the accretion of independent fragments 
85: over a longer period, 10$^9$ years. 
86: These seminal works studied Galactic archaeology using 
87: the kinematics and metallicities of stars and globular clusters 
88: in the disk and halo. Today, Galaxy formation is discussed within 
89: the context of $\Lambda$CDM cosmology and hierarchical structure 
90: formation \citep{white78,freeman02} with the ongoing accretion of the 
91: Sagittarius dwarf galaxy being the most prominent example 
92: \citep{ibata94}. 
93: 
94: Another 
95: mechanism for populating the disk and halo is through the destruction 
96: of globular clusters via tidal shocks, two body relaxation etc 
97: \citep{gnedin97}. 
98: Although the current mass in globular clusters is small, 
99: the initial globular cluster population may have been 
100: considerably larger than the present population \citep{gnedin97}. 
101: The globular cluster Palomar 5 exhibits large tidal tails that extend 
102: over 10 degrees 
103: and contain more mass than the remaining cluster 
104: \citep{odenkirchen01,odenkirchen03}. 
105: Therefore, Pal 5 is in the process of being tidally disrupted and 
106: is currently contributing stars to the disk and halo. 
107: 
108: Chemical abundances place strong constraints upon the fraction of halo 
109: and disk stars that may come from disrupted globular clusters and/or 
110: the types of globular clusters that may populate the disk and halo. 
111: Specifically, every well studied Galactic globular cluster exhibits large 
112: star-to-star abundance variations for the light elements from C to Al 
113: \citep{smith87,kraft94,gratton04}. Although the amplitude may vary 
114: from cluster to cluster, the abundances of C and O are low when N is 
115: high, O and Na are anticorrelated as are Mg and Al. 
116: %Indeed, Str{\"o}mgren photometry 
117: %suggests that every globular cluster shows large N abundance variations 
118: %at all evolutionary stages \citep{grundahl00,6752nh}. 
119: Indeed, Str{\"o}mgren photometry reveals that every globular cluster has 
120: large star-to-star variations in the $c_1$ = $(u-v)-(v-b)$ index at all 
121: evolutionary stages \citep{grundahl00}. The Str{\"o}mgren $u$ filter includes 
122: the 3360\AA\ NH molecular lines, and \citet{6752nh} recently showed 
123: that the N abundances are directly correlated with the $c_1$ index. 
124: Therefore, it is likely that 
125: all globular clusters possess large N abundance variations at all 
126: evolutionary stages. 
127: Although hydrogen burning at 
128: high temperatures may explain the observed abundance patterns 
129: \citep{langer93,langer95,denissenkov98,karakas03}, the source of the 
130: nucleosynthesis and the nature of the 
131: pollution mechanism remain unknown. 
132: Intermediate-mass ($\sim$3 to 8$M_{\odot}$) asymptotic giant branch 
133: (AGB) stars were the assumed polluters owing to the mono-metallic nature 
134: of most GCs, even though detailed AGB models have so far mostly failed 
135: to match the observations \citep{fenner04,karakas06b}. 
136: Nevertheless, 
137: these abundance patterns seen in every cluster have rarely, if ever, been 
138: observed in field stars to date \citep{pilachowski96,gratton00b}. 
139: 
140: \citet{smith02} conducted a detailed abundance analysis of four bright 
141: giant stars in Pal 5 and found variations of O, Na, and Al. 
142: (No abundance measurements have been performed upon stars in the 
143: tidal tails of Pal 5.) 
144: While 
145: most stars lost from a tidally disrupted cluster would be main sequence
146: stars, abundance variations of O, Na, and Al have now been identified on the 
147: main sequences of globular clusters \citep{gratton01,cohen05}. 
148: Since no radial gradients are associated with the O to Al abundance 
149: variations 
150: (with the exception of 47 Tucanae [\citealt{nf79,briley97}]), 
151: observations of red giants in the cluster should be 
152: equivalent to observing red giants in the tidal tails. That 
153: abundance variations of O, Na, and Al are found in Pal 5 
154: suggests that clusters like 
155: Pal 5 cannot provide many field stars and/or field stars do not form 
156: in environments with chemical enrichment histories similar to Pal 5. 
157: Of great interest for our understanding of Galactic and globular cluster 
158: formation would be the identification of 
159: clusters undergoing tidal disruption in which no light element 
160: abundance variations are detected. 
161: 
162: Of the large sample of globular clusters studied by \citet{paresce00} 
163: using the Hubble Space Telescope, all have mass functions (as inferred 
164: from their luminosity functions) which peak at 0.25M$_\odot$. 
165: Not surprisingly, the mass function of Pal 5 
166: is flatter than other clusters revealing significant depletions of 
167: low mass stars presumably stripped by the Galactic tidal field 
168: \citep{koch04}. The globular cluster NGC 6712 is a small and sparse 
169: globular cluster whose mass function peaks at 0.75M$_\odot$ 
170: instead of 0.25M$_\odot$ \citep{demarchi99,andreuzzi01}. 
171: That is, NGC 6712 is the only cluster 
172: whose mass function decreases with decreasing mass. With an 
173: orbit penetrating deep into the bulge, $R_{\rm pericentric}$ = 0.9 kpc 
174: \citep{dinescu99}, tidal forces have stripped away a substantial 
175: fraction of NGC 6712's lower mass stellar population. 
176: Calculations suggest that NGC 6712 may have lost up to 99\% of its original 
177: mass \citep{takahashi00}. 
178: All that remains of NGC 6712 is a remnant core of a cluster that 
179: was probably once one of the most massive in the Galaxy. 
180: The presence of a high luminosity x-ray source and a surprisingly large 
181: blue straggler population reinforce the idea that NGC 6712 was once 
182: much more massive and concentrated \citep{paltrinieri01}. 
183: Therefore, NGC 6712 has almost 
184: certainly contributed stars to the disk and/or halo. 
185: Previous abundance analyses of NGC 6712 only 
186: considered one post-AGB star \citep{jasniewicz04,mooney04} 
187: whose composition may not reflect 
188: the composition of the cluster due to the rich nucleosynthesis occurring 
189: in the late phases of stellar evolution.
190: In this paper, we present the first detailed  
191: chemical abundance analysis of bright red giant stars 
192: in this tidally disrupted globular cluster. 
193: 
194: \section{Observations, data reduction, and analysis}
195: \label{sec:data}
196: 
197: NGC 6712 lies in the direction of the Galactic bulge in a region of 
198: high visual extinction. While the brightest giants are relatively faint at 
199: visual wavelengths ($V$ $\simeq$ 13.5), these giants are very bright 
200: at infrared wavelengths ($H$ $\simeq$ $K$ $\simeq$ 8.3). 
201: \citet{cudworth88} measured proper motions from which membership 
202: probabilities were determined. For bright giants with 
203: membership probabilities $>$ 90\%, optical ($V$ vs.\ $B-V$) and 
204: infrared ($K$ vs.\ $J-K$) color-magnitude
205: diagrams were constructed using the \citet{cudworth88} and 
206: 2MASS \citep{2mass} photometry. The six brightest stars were 
207: observed using the Gemini South telescope and the Phoenix 
208: spectrograph \citep{phoenix} in service mode in July and August 2007. 
209: The program stars are listed in Table \ref{tab:prog} and the log 
210: of observations is shown in Table \ref{tab:obs}. 
211: We used the 0.35\arcsec\ slit which provided a spectral resolution 
212: of $R$ = 50,000. 
213: All program stars were observed at two positions along the slit 
214: separated by 5\arcsec\ on the sky through two 
215: filters: 
216: the H6420 filter provided wavelength coverage from 15520\AA\ to 15585\AA\ 
217: and 
218: the K4308 filter provided wavelength coverage from 23300\AA\ to 23400\AA. 
219: The exposure times per star 
220: ranged from 520 seconds for the H-band observation of V10 
221: to 36 minutes for the K-band observation of LM10. 
222: The signal-to-noise ratios (S/N) 
223: exceed 150 per resolution element for each setting in each star. 
224: For each setting on each night, our observing program included 
225: a radial velocity standard, a hot star for telluric line removal, 
226: 10 flat field exposures, and 10 dark exposures. 
227: Wavelength calibrated spectra were produced using 
228: standard reduction 
229: procedures for infrared data described by \citet{smith02b} and 
230: \citet{melendez03} 
231: with the IRAF\footnote{IRAF (Image Reduction and Analysis
232: Facility) is distributed by the National Optical Astronomy
233: Observatory, which is operated by the Association of Universities
234: for Research in Astronomy, Inc., under contract with the National
235: Science Foundation.} package of programs. 
236: Examples of reduced spectra 
237: are shown in Figure \ref{fig:spectra}. 
238: 
239: Radial velocities were measured by cross correlating the cluster 
240: spectra against the radial velocity standards. For each star, we 
241: obtained a radial velocity measure from the H-band and the K-band and 
242: the velocities measured from each region were in good agreement for a 
243: given star. In Table \ref{tab:rv}, we report the radial velocities and 
244: for our six stars we find a mean cluster radial velocity 
245: $V_{\rm rad}$ = $-$109.0 \kms\ ($\sigma$ = 5.0 \kms) 
246: which is in good agreement 
247: with the value in the \citet{harris96} catalog, 
248: $V_{\rm rad}$ = $-$107.5 \kms\ 
249: as well as the value measured by 
250: \citet{jasniewicz04} for their post-AGB star, $V_{\rm rad}$ = $-$116.4 \kms. 
251: 
252: The stellar parameters were derived in the following way. The 
253: effective temperature, \teff, was calculated using the \citet{ramirez05} 
254: \teff:color:[Fe/H] calibrations for giant 
255: stars. We used the ($B-V$), ($V-J$), ($V-H$), and 
256: ($V-K$) colors from the \citet{cudworth88} and 2MASS \citep{2mass} photometry, 
257: $E(B-V)$ = 0.43 \citep{cudworth88}, and [Fe/H] = $-$1.01 from the 
258: \citet{harris96} catalog. The final \teff\ was the mean of the individual 
259: \teff\ values from each color weighted by the uncertainties for each 
260: color calibration. The surface gravity, $\log g$, was determined using 
261: \teff, a distance modulus of $(m-M)_V$ = 15.6 \citep{harris96}, 
262: bolometric corrections BC(V) from \citet{alonso99b}, and assuming a mass 
263: of 0.8M$_\odot$. The microturbulent velocity was determined using 
264: the following relation, $\xi_t$ = 4.2 $-$ 6$\times$10$^{-4}$ \teff, 
265: adopted from the optical analysis 
266: by \citet{melendez08} of thick disk and bulge stars with 
267: comparable stellar parameters. The stellar parameters 
268: are given in Table \ref{tab:param}. 
269: 
270: We found that errors in the distance modulus of $\pm$0.2 led 
271: to changes of 0.08 dex in $\log g$ and that changes of $\pm$0.02 mag 
272: in reddening resulted in \teff\ errors of 20K. 
273: Had we adopted the \citet{alonso99b} \teff:color:[Fe/H] calibration for giant 
274: stars, our values for \teff\ would be 93K ($\sigma$ = 29K) hotter and $\log g$ 
275: would be 0.07 dex ($\sigma$ = 0.05 dex) higher. 
276: We note that the \citet{schlegel98} dust maps give a reddening $E(B-V)$ = 0.39 
277: and that \citet{paltrinieri01} find a very low value of $E(B-V)$ = 0.33. 
278: Had we adopted the lowest published value 
279: $E(B-V)$ = 0.33, our \teff\ would be 108K 
280: ($\sigma$ = 20K) cooler and $\log g$ 
281: would be 0.03 dex ($\sigma$ = 0.04 dex) lower. 
282: We estimate that internal uncertainties in the stellar parameters are 
283: \teff\ $\pm$ 50K, $\log g$ $\pm$ 0.2 dex, and $\xi_t$ $\pm$ 0.2 \kms. 
284: While the zero-point of our derived 
285: abundances would shift, the amplitude of the 
286: star-to-star abundance variation for C, N, O, F, and Na would remain 
287: similar regardless of the adopted stellar parameters provided they 
288: were homogeneously applied. Therefore, our 
289: conclusions do not depend upon the adopted stellar parameters, within 
290: a reasonable error range. 
291:  
292: Abundances for a given line were derived by comparing synthetic spectra 
293: with observed spectra. The synthetic spectra were generated using 
294: the local thermodynamic equilibrium (LTE) 
295: stellar line analysis program 
296: MOOG \citep{moog} and LTE model atmospheres from 
297: the \citet{kurucz93} grid. 
298: First we derived abundances for O from the OH molecular lines at 15535.462\AA, 
299: 15536.705\AA, and 15565.880\AA. 
300: Next, abundances for C were obtained from the CO molecular lines near 
301: 15576\AA\ as well as 
302: from the large number of CO lines in the K-band spectra. Finally, N 
303: abundances were derived from the CN molecular lines at 15552.695\AA, 
304: 15553.642\AA, and 15563.355\AA. Since the abundances of C, N, and O 
305: are coupled, we iterated until self consistent abundances were obtained, 
306: which always occurred within one iteration. 
307: Abundances for F were obtained from the HF 
308: molecular line at 23358.311\AA. Na abundances 
309: were derived from the Na\,{\sc i} line at 23379.140\AA. Fe abundances 
310: were obtained from the Fe\,{\sc i} lines at 15534.260\AA\ and 15537.690\AA\ 
311: as well as the Fe blend near 15551\AA. 
312: In Figures \ref{fig:cn}, \ref{fig:na}, and \ref{fig:f}, 
313: we show examples of synthetic spectra fits to derive 
314: abundances in our sample and in Table \ref{tab:param}, we present the 
315: final abundances. 
316: The full line list used in the generation of synthetic spectra 
317: was taken from \citet{jorissen92}, \citet{melendez99}, and 
318: \citet{melendez01,melendez03}. 
319: 
320: The model atmosphere grid does not extend 
321: below $\log g$ = 0.0. For the four stars with surface gravities 
322: $\log g <$ 0.0, abundances were extrapolated from nearby models, e.g., 
323: for V10 with $\log g$ = $-$0.22, 
324: we determined abundances for $\log g$ = +0.22 and $\log g$ = 0.00 
325: and adopted $A$(X) = $A$(X)$_{{\rm log~}g=0.00}$ + 
326: ($A$(X)$_{{\rm log}~g=0.00}$ $-$ 
327:  $A$(X)$_{{\rm log}~g=0.22}$). 
328: We checked the extrapolated results by measuring abundances at an 
329: additional value of $\log g$. 
330: For the example above our additional measurement was at 
331: $\log g$ = +0.44. We note that the derived abundances at 
332: $\log g$ = +0.44, +0.22, and 0.00 were essentially linear such that 
333: an extrapolation within the grid provides accurate results. 
334: Although we are extrapolating beyond the grid, we regard the steps in 
335: surface gravity as small, $\Delta\log g$ $\le$ 
336: 0.26, and so we anticipate that our results should be reliable. 
337: The abundance dependences upon the stellar 
338: parameters are shown in Table \ref{tab:parvar}. 
339: 
340: \section{Results}
341: \label{sec:results}
342: 
343: %Based on optical and infrared color-magnitude diagrams, V8 
344: Based on optical ($V$ vs.\ $B-V$) and infrared ($K$ vs.\ $J-K$)
345: color-magnitude diagrams, V8 
346: is a likely asymptotic giant branch (AGB) star. The radial 
347: velocity and line strengths are consistent with cluster membership, 
348: as expected given the proper-motion selection criterion. For this star, the 
349: lines are considerably broader than in the rest of the sample. 
350: To match the observed spectra, 
351: the synthetic spectra for V8 were convolved with a 
352: Gaussian of width 16 \kms\ which represents the combined effect of 
353: the instrumental profile (6 \kms), 
354: atmospheric turbulence, and stellar rotation. 
355: For the remaining stars, the synthetic spectra were convolved with a 
356: Gaussian of typical width 10 \kms\ to match the observed spectra. 
357: 
358: For the elements C, N, O, Na, and F, we find large star-to-star 
359: abundance variations ($\sim$0.6 dex) even within our small sample. 
360: For these elements, the amplitude of the abundance variation 
361: far exceeds the measurement uncertainties. 
362: In this respect, NGC 6712 behaves like all other well studied 
363: Galactic globular clusters. We also find that the Fe abundance 
364: does not show any star-to-star abundance variation, although we note 
365: that the number of Fe lines available in our wavelength regions 
366: is very small. The dispersion in Fe abundances within our sample 
367: ($\sigma$ = 0.04 dex) 
368: can be attributed entirely to the measurement uncertainties. 
369: 
370: In Figure \ref{fig:cnofna}, we plot the abundances of N, O, F, Na, and 
371: Fe against C as well as O vs.\ Na. 
372: As seen in all globular clusters, the abundances of 
373: C and N are anticorrelated and the abundances of C and O are 
374: correlated. In this figure, we fit a straight line to the data 
375: taking into account both the $x$ and $y$ errors. We show the formal slope 
376: of the fitted line as well as the 1-$\sigma$ uncertainty in the slope. 
377: The C-N anticorrelation is significant at the 3-$\sigma$ level and 
378: the C-O correlation is significant at the 4-$\sigma$ level. 
379: We also find that the Na abundances are anticorrelated with C 
380: at the 6-$\sigma$ level and that 
381: Na is anticorrelated with O at the 4-$\sigma$ level. 
382: 
383: The F abundance shows a large star-to-star variation. In Figure 
384: \ref{fig:cnofna}, the F abundances are correlated with C 
385: at the 3-$\sigma$ level. Therefore, F is also correlated with O and 
386: anticorrelated with N and Na. The amplitude of the F abundance 
387: variation ($\Delta$$A$(F) = 0.80 dex) 
388: exceeds the amplitude of the O variation ($\Delta$$A$(O) = 0.64 dex). 
389: (Given the measurement uncertainties $\sigma$A(O) = 0.11 dex and 
390: $\sigma$A(F) = 0.14 dex, O and F may have comparable abundances.) 
391: Indeed, of the 
392: elements measured in our sample, F exhibits the largest amplitude 
393: abundance variation. 
394: 
395: We find that the sum of C+N+O is constant 
396: in NGC 6712 within the measurement uncertainties 
397: (C+N+O is not correlated with the C abundance). 
398: Finally, we note that the Fe abundances are not correlated with C. 
399: Adopting a solar abundance $A$(Fe)$_\odot$ = 7.48, we find a mean 
400: cluster abundance [Fe/H] = $-$0.96 $\pm$ 0.02 ($\sigma$ = 0.04) 
401: which is in good agreement with previous estimates for this cluster 
402: by \citet{zinn84}, [Fe/H] = $-$1.01 and \citet{jasniewicz04}, 
403: [Fe/H] = $-$1.2. 
404: 
405: \section{Discussion}
406: \label{sec:discussion}
407: 
408: \subsection{Abundance comparison between NGC 6712 and M4}
409: 
410: M4 is an ideal globular cluster with which 
411: to compare the chemical abundances in NGC 6712. M4 is the 
412: only other cluster in which F abundances 
413: have been measured in more than two stars 
414: \citep{smith05}, it has a comparable metallicity 
415: ([Fe/H]$_{\rm M4}$ = $-$1.20 and [Fe/H]$_{\rm NGC~6712}$ = $-$1.01 
416: [\citealt{harris96}]), and the orbital parameters are very similar 
417: ($R_{\rm apocentric}^{\rm M4}$ = 5.9 $\pm$ 0.3 kpc, 
418: $R_{\rm pericentric}^{\rm M4}$ = 0.6 $\pm$ 0.1 kpc, and 
419: $Z_{\rm max}^{\rm M4}$ = 1.5 $\pm$ 0.4 kpc and 
420: $R_{\rm apocentric}^{\rm NGC~6712}$ = 6.2  $\pm$ 0.3kpc, 
421: $R_{\rm pericentric}^{\rm NGC~6712}$ = 0.9 $\pm$ 0.1 kpc, and 
422: $Z_{\rm max}^{\rm NGC~6712}$ = 0.9 $\pm$ 0.2 kpc [\citealt{dinescu99}]). 
423: (In the globular cluster $\omega$ Cen, F has been measured in one star 
424: and an upper limit measured in another star [\citealt{cunha03}].) 
425: However, we note that M4 may be uniquely enriched in $s$-process 
426: elements among the Galactic globular clusters \citep{M4,pritzl05,rbpbm4m5}, 
427: with the usual exception of $\omega$ Cen \citep{norris95b,smith00}. 
428: 
429: In Figure \ref{fig:cno} we show the 
430: abundance ranges for C, N, and O for our six stars in NGC 6712 and the seven 
431: stars in M4 \citep{smith05}. In both clusters, 
432: the targets are located near the tip of the red giant branch. 
433: Within the small samples, NGC 6712 may have slightly larger 
434: abundance amplitudes for C, O, and C+N+O, the abundance amplitude 
435: for C+N is very similar for these clusters,  
436: and M4 has a larger abundance amplitude for N. 
437: The mean cluster 
438: abundances are very similar for O. However, NGC 6712 may 
439: have higher mean abundances of N, C+N, and C+N+O along with a lower  
440: mean C abundance than M4. 
441: The higher N abundances and lower C abundances in NGC 6712 
442: relative to M4 suggests that the extent of CN-cycling may have been 
443: greater in NGC 6712. However, 
444: NGC 6712 was formed from gas with higher amounts of C+N and C+N+O  than 
445: M4. 
446: 
447: In Figure \ref{fig:fnaona} we show the abundance ranges for F, Na, 
448: and the ratio [O/Na] for NGC 6712 and M4. 
449: Within the small samples, NGC 6712 may have  
450: larger abundance amplitudes for F, Na, and [O/Na] than M4. The mean cluster  
451: abundances of F are in agreement, however NGC 6712 has a higher mean 
452: Na abundance and a slightly lower [O/Na] ratio than M4, which is consistent 
453: with a higher degree of hydrogen burning via ($p$,$\gamma$) reactions. 
454: Larger samples are required to fully appreciate the abundance 
455: differences between these two clusters whose orbital parameters 
456: are very similar. 
457: 
458: \subsection{F destruction and constraints on AGB nucleosynthesis}
459: 
460: The general behavior of the abundances of F 
461: with respect to C, N, O, and Na in NGC 6712 is identical to that seen 
462: in M4 \citep{smith05}. 
463: We reiterate that in both NGC 6712 and M4, the amplitude of the F 
464: abundance variation is comparable to, or exceeds, 
465: the amplitude of the O variation. 
466: Therefore, any scenario invoked to explain the light element abundance 
467: variations in globular clusters must account for these large F variations. 
468: 
469: In sufficiently massive AGB stars, the 
470: base of the convective envelope can reach temperatures that permit 
471: hydrogen burning, a process called hot-bottom burning (HBB) 
472: \citep{scalo75}. 
473: Hot-bottom burning can qualitatively produce the required 
474: C, N, O, Na, Mg, and Al abundance patterns observed in globular 
475: clusters,  
476: and intermediate-mass AGB stars have long been suspected of 
477: producing the light element abundance variations \citep{cottrell81}, 
478: although AGB models have thus far failed to match the observations
479: \citep{fenner04,karakas06b}. 
480: An additional signature of 
481: HBB is F destruction via $^{19}$F($p$,$\alpha$)$^{16}$O 
482: \citep{mowlavi96,lugaro04,smith05}. 
483: In contrast, low mass AGB stars produce F \citep{jorissen92,forestini92}. 
484: 
485: If AGB stars are solely responsible for the F and O abundance 
486: variations in NGC 6712, theoretical yields 
487: \citep{karakas03,karakas07,karakas08b}
488: offer insight into the range of possible masses of these stars. 
489: These models indicate that F may be destroyed by up to 1 dex 
490: while O is destroyed by up to 0.5 dex 
491: during HBB in 5$M_\odot$ and 6$M_\odot$ $Z$ = 0.004 AGB stars, in 
492: general agreement with the observations. 
493: However, F destruction ceases and indeed F production begins 
494: to occur again when HBB is terminated, 
495: and it is during this phase when much of the mass loss occurs. 
496: Therefore, even for the most massive stars, the AGB winds contain 
497: material with O and F depleted by similar amounts. 
498: The main uncertainties in these models are convection and mass loss. 
499: Convection determines the efficiency of HBB as well 
500: as the HBB lifetime \citep{ventura05a,ventura05b}. 
501: The mass-loss rate determines when the mass is lost from the star. 
502: For example, a stronger mass-loss rate may result in more mass 
503: lost when the star was O and F poor but with a larger 
504: degree of F depletion. 
505: From 5$M_{\odot}$ models of [Fe/H] $\sim -2.3$ computed for 
506: \citet{karakas06a}, we estimate that the fluorine yields can vary by 
507: up to a factor of $\sim$3 by changing the mass-loss rate. 
508: If massive metal-poor 
509: AGB stars are responsible for the F and O variations in NGC 6712 and M4, and 
510: if the F variation exceeds that of O, 
511: then the models would require both stronger mass loss and more 
512: efficient convection. 
513: 
514: Results presented by \citet{izzard07}, that use updated NeNa and MgAl 
515: hydrogen burning rates, gave fluorine abundances increased by a factor of 
516: 72. This last result is net production of $^{19}$F as opposed to 
517: destruction by HBB in the former models, and serves to illustrate 
518: just how uncertain the AGB models are to variations in the input physics 
519: and to the nuclear uncertainties, especially at the range of temperatures 
520: found in the H and He-burning shells of AGB stars. 
521: (We refer the reader to the discussions in \citet{lugaro04,lugaro08} and 
522: \citet{izzard07} and references therein 
523: for an overview of the current uncertainties regarding AGB model yields 
524: for fluorine and other light elements.) 
525: Additional observations of F in AGB stars, such as 
526: those presented by \citet{uttenthaler08} 
527: as well as measurements 
528: in higher mass AGB stars are critical to constrain the AGB models. 
529: 
530: In addition to HBB in intermediate-mass AGB stars, 
531: another possible source of these abundance anomalies is massive 
532: stars \citep{prantzos06,smith06,decressin06}. 
533: While massive stars will also destroy F \citep{prantzos07}, quantitative 
534: yields for F and O would be of interest to constrain all currently 
535: proposed sources  
536: of the globular cluster abundance variations. At present, 
537: the star-to-star F abundance variations in NGC 6712 and M4 
538: could be explained by pollution from either a generation of massive 
539: stars or intermediate-mass AGB stars that underwent hot-bottom burning. 
540: 
541: \subsection{A comparison of [O/Fe] in NGC 6712 with 
542: the general bulge trend}
543: 
544: The O abundances vary from star-to-star in globular clusters 
545: (e.g., \citealt{kraft94}). Stars with high O abundances also show high Mg 
546: abundances along with low Na and Al. We refer to these stars 
547: as ``normal'' because comparisons have shown that 
548: the abundance patterns of these 
549: cluster stars are in accord with field stars at the same metallicity.
550: At the opposite end of the abundance distribution in globular clusters 
551: lie the O-poor, Mg-poor, Na-rich, 
552: and Al-rich stars which we refer to as ``polluted''. No field stars 
553: have been observed with compositions matching these ``polluted'' cluster 
554: stars. 
555: In globular clusters like NGC 6752 \citep{6752} and M13 \citep{sneden04a}, 
556: the ``normal'' stars with 
557: the highest O abundances have [O/Fe] ratios in 
558: agreement with field stars at the same metallicity.  
559: 
560: Following \citet{melendez08}, we adopt solar abundances of 
561: $A$(O)$_\odot$ = 8.72 and $A$(Fe)$_\odot$ = 7.48 
562: which are similar to the \citet{asplund05} values based on 
563: 3D hydrodynamical model atmospheres. 
564: For NGC 6712, our highest relative abundance is [O/Fe] = 0.59. 
565: We assume that this star is ``normal'' and that this O abundance 
566: is representative of the initial cluster value prior to the 
567: processes which produced star-to-star variations in the light element 
568: abundances. Within the measurement uncertainties, the O abundance for 
569: NGC 6712 is comparable to the values recently measured in bulge giants  
570: by \citet{melendez08}, who showed through a homogeneous 
571: differential analysis that the thick disk and bulge (and halo) had [O/Fe] 
572: ratios in agreement at a given [Fe/H]. 
573: Therefore, we tentatively conclude that the [O/Fe] 
574: ratio in ``normal'' giants in NGC 6712 is in agreement with the 
575: general bulge trend (and halo stars at the same metallicity).  
576: 
577: \subsection{A comparison of fluorine in NGC 6712 with other 
578: Galactic populations}
579: 
580: The fluorine abundances measured in NGC 6712 are now compared to those
581: from other samples of Galactic stars, which include field stars 
582: \citep{cunha03,cunha05}, along with bulge red giants
583: \citep{cunha08}, as well as the measurements for the globular
584: clusters M4 \citep{smith05} and $\omega$ Cen \citep{cunha03}. 
585: As is the case for oxygen, the
586: assumption is made that the highest fluorine abundances in NGC 6712 (as well
587: as for M4) represent the initial cluster value prior to the processes which
588: produced the globular cluster star-to-star abundance variations. 
589: In Figure \ref{fig:fvso} 
590: are plotted the abundances of A(F) versus A(O) (top panel) and
591: log[N(F)/N(O)] versus A(O) (bottom panel) for all stars from the various
592: studies.  As a first point of comparison, it is found that the most F-rich
593: stars in NGC 6712 (and in M4) are underabundant in fluorine when compared
594: to most of the bulge and field stars that have comparable values of
595: A(O) $\sim$ 8.2-8.4 (here oxygen is used as a proxy for metallicity). 
596: Most of the field stars and bulge stars fall along a similar distribution 
597: in the A(F) versus A(O) diagram, while the globular cluster stars seem to
598: define a different trend. 
599: 
600: The straight lines shown in the top panel of Figure \ref{fig:fvso} 
601: represent linear fits
602: to the globular cluster data and, in addition, fits to the field plus bulge
603: star points; extrapolations of these lines do not intersect and confirm
604: the impression that the globular clusters have a distinct mixture of F and
605: O abundances when compared to the field and bulge stars.  This observation
606: is based on the stellar samples studied to date, with a still small
607: metallicity overlap; the behavior of fluorine 
608: in the field
609: has not been probed below oxygen abundances A(O) $\sim$ 8.4.
610: However, to have a single distribution, or curve, of
611: A(F) to A(O) fit both sets of data (field and globular clusters) 
612: would require a rapid, nearly
613: discontinuous drop of about 0.8 dex in the fluorine abundance near an oxygen
614: abundance of A(O) $\sim$ 8.2-8.4.  
615: It is also noted that the rather oxygen-rich
616: bulge giant (BMB 78) studied by \citet{cunha08}, with A(O) = 9.0, but a
617: low F abundance of A(F) = 4.26, falls along the line extrapolated from the
618: globular cluster stars. 
619: 
620: Another way to compare F and O abundances is shown in the bottom panel of
621: Figure \ref{fig:fvso}, 
622: with the ratio of F/O plotted as a function of the oxygen
623: abundance.  Here again there appears to be a rather sharp, near
624: discontinuity in the values of F/O in the globular clusters in comparison
625: to the field and bulge stars near A(O) $\sim$ 8.3.  In the globular clusters,
626: the values of F/O remain nearly constant as the oxygen abundance varies
627: and this is due to the depletion of both $^{16}$O and $^{19}$F by the
628: H-burning processes that shape the peculiar chemical evolution found in
629: the globular clusters; inspection of the F and O abundances in both M4 and
630: NGC 6712 reveals nearly equal decreases in both $^{19}$F and $^{16}$O 
631: (as expected to occur in only 
632: the most massive AGB stars, as discussed above), which
633: results in nearly constant ratios of F/O within the cluster stars.  Note that
634: Figure 2 in \citet{cunha08} shows predictions for $^{19}$F production
635: via neutrino nucleosynthesis in SNe II taken from the \citet{woosley95}
636: models, as well as the approximate downward revisions to the fluorine
637: yields as suggested by \citet{heger05}.  Neutrino nucleosynthesis
638: predicts values of log[N(F)/N(O)] $\sim$ $-$5.0 
639: for models with oxygen abundances
640: of about A(O) = 8.4, which matches the envelope of values found for the stars
641: in NGC 6712, M4, as well as the two stars studied to date in $\omega$ Cen.
642: 
643: Within the uncertainties in the measurements, along with the models, it is
644: suggested that the $^{19}$F observed in the globular clusters could have been
645: created by neutrino nucleosynthesis alone (with the fluorine arising mostly
646: from core-collapse neutrinos spalling $^{20}$Ne).  The increased values of
647: F/O found in the field and bulge stars require additional sources of $^{19}$F,
648: which have been discussed and modeled by \citet{renda04} and discussed
649: in \citet{cunha08}, and consist of Wolf-Rayet winds (whose yields of fluorine
650: increase substantially with stellar envelope metallicity), along with thermally
651: pulsing AGB stars.  Such a picture would suggest that the
652: globular clusters are less polluted by Wolf Rayet winds and low-mass AGB
653: stars (AGB stars with M $>$ 2-3M$_{\odot}$ destroy $^{19}$F) than either the
654: bulge or field stars with metallicities greater than about one-third solar.
655: If the globular clusters represent the remnants of systems that formed from
656: gas that was chemically seeded by very metal-poor SN II, the low values of
657: F/O represent the ``chemical memory'' of this enrichment.
658: 
659: Although ``normal'' globular cluster stars have compositions that are 
660: indistinguishable from field stars at the same metallicity, as discussed 
661: above, $^{19}$F is an exception. Another exception 
662: is represented by the minor isotopes of Mg, whose ratios 
663: $^{25}$Mg/$^{24}$Mg and $^{26}$Mg/$^{24}$Mg in ``normal'' cluster stars 
664: exceed the values found in field stars at the same metallicity 
665: \citep{shetrone96b,6752,mghdwarf,mghsubaru}. While the contribution, 
666: or absence, of Wolf Rayet winds and/or low-mass AGB stars provides a 
667: plausible explanation for the F discrepancy 
668: as discussed above, the situation for the Mg 
669: isotopes is less clear. One explanation is that 
670: the entire globular cluster was polluted by intermediate-mass AGB 
671: stars which raised the low abundances of $^{25}$Mg and $^{26}$Mg 
672: provided by supernovae \citep{fenner03} 
673: to the high levels observed. 
674: 
675: Finally, the bulge star BMB 78, with unusually low F, may be an O-rich star 
676: whose F abundance could be attributed solely to neutrino nucleosynthesis 
677: (i.e., the F in this star 
678: has experienced little or no contribution from Wolf Rayet or AGB stars). 
679: As discussed by \citet{cunha08}, this star could therefore have important 
680: implications for the inhomogeneous chemical evolution of the bulge. 
681: An additional, and highly speculative, explanation for the unusual F and O 
682: abundances in BMB 78 is that 
683: this star was born in a globular cluster 
684: but was subsequently stripped away. 
685: 
686: \subsection{Light element abundance variations and implications for 
687: Galactic formation}
688: 
689: Star-to-star abundance variations for the light elements have been 
690: found in every well studied Galactic globular cluster. Such abundance 
691: patterns are the signature of hydrogen burning at high temperatures. 
692: The currently favored candidates are intermediate-mass AGB stars 
693: and massive stars. 
694: Our abundance measurements in NGC 6712 
695: provide new and critical information. Specifically, the F abundance 
696: is found to vary from star-to-star with an amplitude comparable to, 
697: or possibly exceeding,  
698: that of O. Therefore, the two globular clusters in which F has been 
699: measured in more than two stars both show large abundance variations. 
700: 
701: The fact that NGC 6712 exhibits large star-to-star abundance variations 
702: of the light elements has implications for Galactic 
703: formation. The current mass in globular clusters is small, but in the past 
704: there may have been many more clusters. Some fraction of field stars 
705: may have been born in globular clusters that were subsequently 
706: destroyed by the 
707: Galactic tidal field. However, the abundance signature of globular 
708: clusters, O, F, Na, Mg, and Al variations, has never been 
709: observed in field stars to date. 
710: (C and N variations are found in field halo stars
711: as well as cluster stars [e.g., \citealt{gratton00b}] 
712: and can be attributed to internal nucleosynthesis 
713: and mixing with the observed stars.)
714: Current estimates suggest that globular clusters 
715: comprise roughly 2\% of the mass of the stellar halo \citep{freeman02}. 
716: If we arbitrarily assume that one globular cluster was destroyed for every 
717: surviving cluster (i.e., the initial globular cluster population 
718: was double the current population), 
719: then for every 50 field halo stars observed, only 1 star 
720: would come from a disrupted globular cluster. However, not every 
721: star in a given globular cluster has peculiar O, F, Na, Mg, and Al abundances 
722: with respect to field stars at the same metallicity (e.g., see the 
723: earlier discussion on ``normal'' stars). 
724: If we arbitrarily assume that half the stars in globular clusters 
725: have distinct abundances of O to Al relative to field stars at the same 
726: metallicity, then  100 field halo stars need to observed 
727: to find one star whose chemical abundances indicate 
728: that it was born in a globular cluster. Nevertheless, no field halo stars 
729: have been identified that show large Na and Al enhancements along with  
730: large O depletions. \citet{gratton00b} investigated a large sample of 
731: 105 stars with $-$2 $\le$ [Fe/H] $\le$ $-$1, and so the non-detection 
732: of the globular cluster abundance anomalies 
733: in field stars is not due to a lack of effort (although larger samples 
734: may be needed). 
735: NGC 6712 and Pal 5 are tidally disrupted globular 
736: clusters which have almost certainly contributed stars to the disk and 
737: halo. Both clusters show large abundance variations for light elements 
738: which suggests that clusters like NGC 6712 and Pal 5 cannot have provided 
739: many field stars and/or field stars did not form in environments with 
740: chemical evolution histories like NGC 6712 and Pal 5. 
741: 
742: \subsection{Constraints upon the initial cluster mass from abundance
743: variations} 
744: 
745: Based on the present day luminosity function, the high luminosity 
746: x-ray source, and the large blue straggler population, it is highly 
747: likely that NGC 6712 was initially considerably more massive 
748: \citep{demarchi99,andreuzzi01,paltrinieri01}. Indeed, calculations 
749: by \citet{takahashi00} 
750: suggest that NGC 6712 may have lost 99\% of its initial mass such 
751: that it might have been one of the most massive clusters 
752: that ever formed in the Galaxy, $M_{\rm initial}$ $\sim$ 10$^7$ $M_\odot$. 
753: 
754: For eight well studied globular clusters, \citet{carretta06b} compared  
755: the interquartile range (IQR) for [O/Fe], [Na/Fe], [O/Na], and other 
756: abundance ratios with various physical parameters 
757: and 
758: found that the amplitude of the abundance variation 
759: shows a dependence upon cluster mass, as inferred from 
760: the absolute magnitude. 
761: Presumably the light element abundance variations in NGC 6712 (and in all 
762: clusters) originated early in the life of the cluster. Therefore, 
763: the currently observed abundance variations offer an independent 
764: estimate of the original mass of NGC 6712. 
765: 
766: Our measured 
767: values are IQR[O/Na] = 0.85, IQR[O/Fe] = 0.59, and IQR[Na/Fe] = 0.55 
768: (adopting solar values of 
769: $A$(O)$_\odot$ = 8.72,  
770: $A$(Na)$_\odot$ = 6.17, and
771: $A$(Fe)$_\odot$ = 7.48.) 
772: Since our sample size is 
773: small, we may be underestimating (or overestimating) the true IQRs. 
774: We fit a straight line to the \citet{carretta06b} data and find 
775: that the IQRs for [O/Na], [O/Fe], and [Na/Fe] in NGC 6712 correspond 
776: to absolute magnitudes of $-$9.6, $-$12.7, and $-$11.4 respectively. 
777: While such an analysis is far from robust, inspection of Figures 12 and 
778: 13 in \citet{carretta06b} indicate that NGC 6712 should be a 
779: very massive cluster based on the IQRs for [O/Fe], [Na/Fe], and [O/Na]. 
780: The two most massive globular clusters $\omega$ Cen and M54 have 
781: absolute magnitudes $-$10.29 and $-$10.01 respectively and both clusters 
782: are regarded as the nuclei of accreted dwarf galaxies. 
783: Despite our small sample size which may not measure the true IQRs, 
784: it is likely that 
785: NGC 6712 was initially one of the most massive clusters in our Galaxy 
786: as inferred from the large amplitude light element abundance variations. 
787: Of great 
788: interest would be the analysis of a larger number of elements in a 
789: larger sample of stars in NGC 6712 to identify abundance similarities 
790: with the massive globular cluster $\omega$ Cen. Given the narrow RGB 
791: sequence \citep{cudworth88}, a star-to-star spread in Fe seems 
792: unlikely. 
793: 
794: \section{Concluding remarks}
795: \label{sec:summary}
796: 
797: Based on high resolution infrared spectra, we derive abundances 
798: of C, N, O, F, Na, and Fe in six giant stars of the tidally 
799: disrupted globular cluster NGC 6712. For the elements C, N, O, F, 
800: and Na, we find large star-to-star abundance variations and correlations 
801: between these elements, a 
802: characteristic that NGC 6712 shares with every well studied Galactic 
803: globular cluster. This is only the second cluster in which F 
804: abundances have been measured in useful numbers of stars 
805: and both clusters show F variations whose 
806: amplitude is comparable to, or exceeds, that of O. Within the limited data, 
807: globular clusters appear to have lower F abundances than field 
808: and bulge stars at the same metallicity. 
809: Of great interest would be 
810: measurements of F in additional stars in $\omega$ Cen and 
811: other globular clusters as well as in larger samples of field stars,  
812: with both samples overlapping in metallicity. 
813: From the amplitude of the O and Na 
814: abundance variations, we tentatively confirm that NGC 6712 was 
815: once one of the most massive clusters in our Galaxy. 
816: 
817: NGC 6712 is a tidally disrupted cluster as revealed through its 
818: highly unusual luminosity function. Pal 5 is another tidally 
819: disrupted globular cluster. Both NGC 6712 and Pal 5 
820: have almost certainly contributed stars to the 
821: disk and halo. Both clusters exhibit large 
822: star-to-star abundance variations for light elements, a characteristic 
823: which has yet to be identified 
824: in field halo stars. Therefore, the light element abundance variations 
825: detected in NGC 6712 indicate that clusters like NGC 6712 and Pal 5 
826: have not provided many field stars and/or field stars did not form 
827: in environments with chemical enrichment histories like NGC 6712 and Pal 5. 
828: As pointed out by \citet{smith02}, disrupted globular clusters like 
829: Pal 5 have lost CN-strong, O-poor, Na-rich, Al-rich stars to the halo 
830: field. But where are these stars? 
831: Of great interest would be an abundance analysis of stars within the 
832: tidal tails of Pal 5 as well as a large-scale dedicated search for 
833: O, Na, and Al abundance anomalies in field halo stars. 
834: 
835: \acknowledgments
836: This paper is based on observations obtained with the Phoenix infrared
837: spectrograph, developed and operated by the National Optical Astronomy
838: Observatory. 
839: Based on observations obtained at the Gemini Observatory, which is 
840: operated by the
841: Association of Universities for Research in Astronomy, Inc., under a 
842: cooperative agreement
843: with the NSF on behalf of the Gemini partnership: the National Science 
844: Foundation (United
845: States), the Science and Technology Facilities Council (United Kingdom), the
846: National Research Council (Canada), CONICYT (Chile), the Australian Research 
847: Council
848: (Australia), Minist{\' e}rio da Ci\^{e}ncia e Tecnologia (Brazil) and SECYT 
849: (Argentina), as program GS-2007B-Q-209. 
850: This research has made use of the SIMBAD database,
851: operated at CDS, Strasbourg, France and
852: NASA's Astrophysics Data System. 
853: DY thanks Gary Da Costa and Ken Freeman for helpful discussions and 
854: the anonymous referee for helpful comments. 
855: This research was 
856: supported in part by NASA through the American Astronomical Society's Small 
857: Research Grant Program, the Australian Research Council under grants 
858: DP0663562 and DP0664105, and FCT (project PTDC/CTE-AST/65971/2006). 
859: 
860: \begin{thebibliography}{87}
861: \expandafter\ifx\csname natexlab\endcsname\relax\def\natexlab#1{#1}\fi
862: 
863: \bibitem[{{Alonso} {et~al.}(1999){Alonso}, {Arribas}, \&
864:   {Mart\'{\i}nez-Roger}}]{alonso99b}
865: {Alonso}, A., {Arribas}, S., \& {Mart\'{\i}nez-Roger}, C. 1999, \aaps, 140, 261
866: 
867: \bibitem[{{Andreuzzi} {et~al.}(2001){Andreuzzi}, {De Marchi}, {Ferraro},
868:   {Paresce}, {Pulone}, \& {Buonanno}}]{andreuzzi01}
869: {Andreuzzi}, G., {De Marchi}, G., {Ferraro}, F.~R., {Paresce}, F., {Pulone},
870:   L., \& {Buonanno}, R. 2001, \aap, 372, 851
871: 
872: \bibitem[{{Asplund} {et~al.}(2005){Asplund}, {Grevesse}, \&
873:   {Sauval}}]{asplund05}
874: {Asplund}, M., {Grevesse}, N., \& {Sauval}, A.~J. 2005, in ASP Conf. Ser. 336:
875:   Cosmic Abundances as Records of Stellar Evolution and Nucleosynthesis, ed.
876:   T.~G. {Barnes}, III \& F.~N. {Bash}, 25
877: 
878: \bibitem[{{Barbier-Brossat} \& {Figon}(2000)}]{rvs}
879: {Barbier-Brossat}, M. \& {Figon}, P. 2000, \aaps, 142, 217
880: 
881: \bibitem[{{Briley}(1997)}]{briley97}
882: {Briley}, M.~M. 1997, \aj, 114, 1051
883: 
884: \bibitem[{{Carretta}(2006)}]{carretta06b}
885: {Carretta}, E. 2006, \aj, 131, 1766
886: 
887: \bibitem[{{Cohen} \& {Mel{\' e}ndez}(2005)}]{cohen05}
888: {Cohen}, J.~G. \& {Mel{\' e}ndez}, J. 2005, \aj, 129, 303
889: 
890: \bibitem[{{Cottrell} \& {Da Costa}(1981)}]{cottrell81}
891: {Cottrell}, P.~L. \& {Da Costa}, G.~S. 1981, \apjl, 245, L79
892: 
893: \bibitem[{{Cudworth}(1988)}]{cudworth88}
894: {Cudworth}, K.~M. 1988, \aj, 96, 105
895: 
896: \bibitem[{{Cunha} \& {Smith}(2005)}]{cunha05}
897: {Cunha}, K. \& {Smith}, V.~V. 2005, \apj, 626, 425
898: 
899: \bibitem[{{Cunha} {et~al.}(2008){Cunha}, {Smith}, \& {Gibson}}]{cunha08}
900: {Cunha}, K., {Smith}, V.~V., \& {Gibson}, B.~K. 2008, \apjl, 679, L17
901: 
902: \bibitem[{{Cunha} {et~al.}(2003){Cunha}, {Smith}, {Lambert}, \&
903:   {Hinkle}}]{cunha03}
904: {Cunha}, K., {Smith}, V.~V., {Lambert}, D.~L., \& {Hinkle}, K.~H. 2003, \aj,
905:   126, 1305
906: 
907: \bibitem[{{de Marchi} {et~al.}(1999){de Marchi}, {Leibundgut}, {Paresce}, \&
908:   {Pulone}}]{demarchi99}
909: {de Marchi}, G., {Leibundgut}, B., {Paresce}, F., \& {Pulone}, L. 1999, \aap,
910:   343, L9
911: 
912: \bibitem[{{Decressin} {et~al.}(2007){Decressin}, {Meynet}, {Charbonnel},
913:   {Prantzos}, \& {Ekstr{\"o}m}}]{decressin06}
914: {Decressin}, T., {Meynet}, G., {Charbonnel}, C., {Prantzos}, N., \&
915:   {Ekstr{\"o}m}, S. 2007, \aap, 464, 1029
916: 
917: \bibitem[{{Denissenkov} {et~al.}(1998){Denissenkov}, {Da Costa}, {Norris}, \&
918:   {Weiss}}]{denissenkov98}
919: {Denissenkov}, P.~A., {Da Costa}, G.~S., {Norris}, J.~E., \& {Weiss}, A. 1998,
920:   \aap, 333, 926
921: 
922: \bibitem[{{Dinescu} {et~al.}(1999){Dinescu}, {Girard}, \& {van
923:   Altena}}]{dinescu99}
924: {Dinescu}, D.~I., {Girard}, T.~M., \& {van Altena}, W.~F. 1999, \aj, 117, 1792
925: 
926: \bibitem[{{Eggen} {et~al.}(1962){Eggen}, {Lynden-Bell}, \& {Sandage}}]{els62}
927: {Eggen}, O.~J., {Lynden-Bell}, D., \& {Sandage}, A.~R. 1962, \apj, 136, 748
928: 
929: \bibitem[{{Fenner} {et~al.}(2004){Fenner}, {Campbell}, {Karakas}, {Lattanzio},
930:   \& {Gibson}}]{fenner04}
931: {Fenner}, Y., {Campbell}, S., {Karakas}, A.~I., {Lattanzio}, J.~C., \&
932:   {Gibson}, B.~K. 2004, \mnras, 353, 789
933: 
934: \bibitem[{{Fenner} {et~al.}(2003){Fenner}, {Gibson}, {Lee}, {Karakas},
935:   {Lattanzio}, {Chieffi}, {Limongi}, \& {Yong}}]{fenner03}
936: {Fenner}, Y., {Gibson}, B.~K., {Lee}, H.-c., {Karakas}, A.~I., {Lattanzio},
937:   J.~C., {Chieffi}, A., {Limongi}, M., \& {Yong}, D. 2003, \pasa, 20, 340
938: 
939: \bibitem[{{Forestini} {et~al.}(1992){Forestini}, {Goriely}, {Jorissen}, \&
940:   {Arnould}}]{forestini92}
941: {Forestini}, M., {Goriely}, S., {Jorissen}, A., \& {Arnould}, M. 1992, \aap,
942:   261, 157
943: 
944: \bibitem[{{Freeman} \& {Bland-Hawthorn}(2002)}]{freeman02}
945: {Freeman}, K. \& {Bland-Hawthorn}, J. 2002, \araa, 40, 487
946: 
947: \bibitem[{{Gnedin} \& {Ostriker}(1997)}]{gnedin97}
948: {Gnedin}, O.~Y. \& {Ostriker}, J.~P. 1997, \apj, 474, 223
949: 
950: \bibitem[{{Gratton} {et~al.}(2004){Gratton}, {Sneden}, \&
951:   {Carretta}}]{gratton04}
952: {Gratton}, R., {Sneden}, C., \& {Carretta}, E. 2004, \araa, 42, 385
953: 
954: \bibitem[{{Gratton} {et~al.}(2001){Gratton}, {Bonifacio}, {Bragaglia},
955:   {Carretta}, {Castellani}, {Centurion}, {Chieffi}, {Claudi}, {Clementini},
956:   {D'Antona}, {Desidera}, {Fran{\c c}ois}, {Grundahl}, {Lucatello}, {Molaro},
957:   {Pasquini}, {Sneden}, {Spite}, \& {Straniero}}]{gratton01}
958: {Gratton}, R.~G., {Bonifacio}, P., {Bragaglia}, A., {Carretta}, E.,
959:   {Castellani}, V., {Centurion}, M., {Chieffi}, A., {Claudi}, R., {Clementini},
960:   G., {D'Antona}, F., {Desidera}, S., {Fran{\c c}ois}, P., {Grundahl}, F.,
961:   {Lucatello}, S., {Molaro}, P., {Pasquini}, L., {Sneden}, C., {Spite}, F., \&
962:   {Straniero}, O. 2001, \aap, 369, 87
963: 
964: \bibitem[{{Gratton} {et~al.}(2000){Gratton}, {Sneden}, {Carretta}, \&
965:   {Bragaglia}}]{gratton00b}
966: {Gratton}, R.~G., {Sneden}, C., {Carretta}, E., \& {Bragaglia}, A. 2000, \aap,
967:   354, 169
968: 
969: \bibitem[{{Grundahl} {et~al.}(2000){Grundahl}, {Vandenberg}, {Stetson},
970:   {Andersen}, \& {Briley}}]{grundahl00}
971: {Grundahl}, F., {Vandenberg}, D.~A., {Stetson}, P.~B., {Andersen}, M.~I., \&
972:   {Briley}, M. 2000, in Liege International Astrophysical Colloquia, Vol.~35,
973:   Liege International Astrophysical Colloquia, ed. A.~{Noels}, P.~{Magain},
974:   D.~{Caro}, E.~{Jehin}, G.~{Parmentier}, \& A.~A. {Thoul}, 503
975: 
976: \bibitem[{{Harris}(1996)}]{harris96}
977: {Harris}, W.~E. 1996, \aj, 112, 1487
978: 
979: \bibitem[{{Heger} {et~al.}(2005){Heger}, {Kolbe}, {Haxton}, {Langanke},
980:   {Mart{\'{\i}}nez-Pinedo}, \& {Woosley}}]{heger05}
981: {Heger}, A., {Kolbe}, E., {Haxton}, W.~C., {Langanke}, K.,
982:   {Mart{\'{\i}}nez-Pinedo}, G., \& {Woosley}, S.~E. 2005, Physics Letters B,
983:   606, 258
984: 
985: \bibitem[{{Hinkle} {et~al.}(2003){Hinkle}, {Blum}, {Joyce}, {Sharp}, {Ridgway},
986:   {Bouchet}, {van der Bliek}, {Najita}, \& {Winge}}]{phoenix}
987: {Hinkle}, K.~H., {Blum}, R.~D., {Joyce}, R.~R., {Sharp}, N., {Ridgway}, S.~T.,
988:   {Bouchet}, P., {van der Bliek}, N.~S., {Najita}, J., \& {Winge}, C. 2003, in
989:   Presented at the Society of Photo-Optical Instrumentation Engineers (SPIE)
990:   Conference, Vol. 4834, Discoveries and Research Prospects from 6- to
991:   10-Meter-Class Telescopes II. Edited by Guhathakurta, Puragra. Proceedings of
992:   the SPIE, Volume 4834, pp. 353-363 (2003)., ed. P.~{Guhathakurta}, 353--363
993: 
994: \bibitem[{{Ibata} {et~al.}(1994){Ibata}, {Gilmore}, \& {Irwin}}]{ibata94}
995: {Ibata}, R.~A., {Gilmore}, G., \& {Irwin}, M.~J. 1994, \nat, 370, 194
996: 
997: \bibitem[{{Ivans} {et~al.}(1999){Ivans}, {Sneden}, {Kraft}, {Suntzeff},
998:   {Smith}, {Langer}, \& {Fulbright}}]{M4}
999: {Ivans}, I.~I., {Sneden}, C., {Kraft}, R.~P., {Suntzeff}, N.~B., {Smith},
1000:   V.~V., {Langer}, G.~E., \& {Fulbright}, J.~P. 1999, \aj, 118, 1273
1001: 
1002: \bibitem[{{Izzard} {et~al.}(2007){Izzard}, {Lugaro}, {Karakas}, {Iliadis}, \&
1003:   {van Raai}}]{izzard07}
1004: {Izzard}, R.~G., {Lugaro}, M., {Karakas}, A.~I., {Iliadis}, C., \& {van Raai},
1005:   M. 2007, \aap, 466, 641
1006: 
1007: \bibitem[{{Jasniewicz} {et~al.}(2004){Jasniewicz}, {de Laverny},
1008:   {Parthasarathy}, {L{\`e}bre}, \& {Th{\'e}venin}}]{jasniewicz04}
1009: {Jasniewicz}, G., {de Laverny}, P., {Parthasarathy}, M., {L{\`e}bre}, A., \&
1010:   {Th{\'e}venin}, F. 2004, \aap, 423, 353
1011: 
1012: \bibitem[{{Jorissen} {et~al.}(1992){Jorissen}, {Smith}, \&
1013:   {Lambert}}]{jorissen92}
1014: {Jorissen}, A., {Smith}, V.~V., \& {Lambert}, D.~L. 1992, \aap, 261, 164
1015: 
1016: \bibitem[{{Karakas} \& {Lattanzio}(2007)}]{karakas07}
1017: {Karakas}, A. \& {Lattanzio}, J.~C. 2007, \pasa, 24, 103
1018: 
1019: \bibitem[{{Karakas} {et~al.}(2006{\natexlab{a}}){Karakas}, {Fenner}, {Sills},
1020:   {Campbell}, \& {Lattanzio}}]{karakas06b}
1021: {Karakas}, A.~I., {Fenner}, Y., {Sills}, A., {Campbell}, S.~W., \& {Lattanzio},
1022:   J.~C. 2006{\natexlab{a}}, \apj, 652, 1240
1023: 
1024: \bibitem[{{Karakas} \& {Lattanzio}(2003)}]{karakas03}
1025: {Karakas}, A.~I. \& {Lattanzio}, J.~C. 2003, \pasa, 20, 279
1026: 
1027: \bibitem[{{Karakas} {et~al.}(2008){Karakas}, {Lee}, {Lugaro}, {G{\"o}rres}, \&
1028:   {Wiescher}}]{karakas08b}
1029: {Karakas}, A.~I., {Lee}, H.~Y., {Lugaro}, M., {G{\"o}rres}, J., \& {Wiescher},
1030:   M. 2008, \apj, 676, 1254
1031: 
1032: \bibitem[{{Karakas} {et~al.}(2006{\natexlab{b}}){Karakas}, {Lugaro},
1033:   {Wiescher}, {G{\"o}rres}, \& {Ugalde}}]{karakas06a}
1034: {Karakas}, A.~I., {Lugaro}, M.~A., {Wiescher}, M., {G{\"o}rres}, J., \&
1035:   {Ugalde}, C. 2006{\natexlab{b}}, \apj, 643, 471
1036: 
1037: \bibitem[{{Koch} {et~al.}(2004){Koch}, {Grebel}, {Odenkirchen},
1038:   {Mart{\'{\i}}nez-Delgado}, \& {Caldwell}}]{koch04}
1039: {Koch}, A., {Grebel}, E.~K., {Odenkirchen}, M., {Mart{\'{\i}}nez-Delgado}, D.,
1040:   \& {Caldwell}, J.~A.~R. 2004, \aj, 128, 2274
1041: 
1042: \bibitem[{{Kraft}(1994)}]{kraft94}
1043: {Kraft}, R.~P. 1994, \pasp, 106, 553
1044: 
1045: \bibitem[{{Kurucz}(1993)}]{kurucz93}
1046: {Kurucz}, R. 1993, ATLAS9 Stellar Atmosphere Programs and 2 km/s grid.~Kurucz
1047:   CD-ROM No.~13.~ Cambridge, Mass.: Smithsonian Astrophysical Observatory,
1048:   1993., 13
1049: 
1050: \bibitem[{{Langer} {et~al.}(1993){Langer}, {Hoffman}, \& {Sneden}}]{langer93}
1051: {Langer}, G.~E., {Hoffman}, R., \& {Sneden}, C. 1993, \pasp, 105, 301
1052: 
1053: \bibitem[{{Langer} \& {Hoffman}(1995)}]{langer95}
1054: {Langer}, G.~E. \& {Hoffman}, R.~D. 1995, \pasp, 107, 1177
1055: 
1056: \bibitem[{{Lugaro} {et~al.}(2008){Lugaro}, {de Mink}, {Izzard}, {Campbell},
1057:   {Karakas}, {Cristallo}, {Pols}, {Lattanzio}, {Straniero}, {Gallino}, \&
1058:   {Beers}}]{lugaro08}
1059: {Lugaro}, M., {de Mink}, S.~E., {Izzard}, R.~G., {Campbell}, S.~W., {Karakas},
1060:   A.~I., {Cristallo}, S., {Pols}, O.~R., {Lattanzio}, J.~C., {Straniero}, O.,
1061:   {Gallino}, R., \& {Beers}, T.~C. 2008, \aap, 484, L27
1062: 
1063: \bibitem[{{Lugaro} {et~al.}(2004){Lugaro}, {Ugalde}, {Karakas}, {G{\"o}rres},
1064:   {Wiescher}, {Lattanzio}, \& {Cannon}}]{lugaro04}
1065: {Lugaro}, M., {Ugalde}, C., {Karakas}, A.~I., {G{\"o}rres}, J., {Wiescher}, M.,
1066:   {Lattanzio}, J.~C., \& {Cannon}, R.~C. 2004, \apj, 615, 934
1067: 
1068: \bibitem[{{Mel{\'e}ndez} {et~al.}(2008){Mel{\'e}ndez}, {Asplund},
1069:   {Alves-Brito}, {Cunha}, {Barbuy}, {Bessell}, {Chiappini}, {Freeman},
1070:   {Ram{\'{\i}}rez}, {Smith}, \& {Yong}}]{melendez08}
1071: {Mel{\'e}ndez}, J., {Asplund}, M., {Alves-Brito}, A., {Cunha}, K., {Barbuy},
1072:   B., {Bessell}, M.~S., {Chiappini}, C., {Freeman}, K.~C., {Ram{\'{\i}}rez},
1073:   I., {Smith}, V.~V., \& {Yong}, D. 2008, \aap, 484, L21
1074: 
1075: \bibitem[{{Mel{\'e}ndez} \& {Barbuy}(1999)}]{melendez99}
1076: {Mel{\'e}ndez}, J. \& {Barbuy}, B. 1999, \apjs, 124, 527
1077: 
1078: \bibitem[{{Mel{\'e}ndez} {et~al.}(2003){Mel{\'e}ndez}, {Barbuy}, {Bica},
1079:   {Zoccali}, {Ortolani}, {Renzini}, \& {Hill}}]{melendez03}
1080: {Mel{\'e}ndez}, J., {Barbuy}, B., {Bica}, E., {Zoccali}, M., {Ortolani}, S.,
1081:   {Renzini}, A., \& {Hill}, V. 2003, \aap, 411, 417
1082: 
1083: \bibitem[{{Mel{\'e}ndez} {et~al.}(2001){Mel{\'e}ndez}, {Barbuy}, \&
1084:   {Spite}}]{melendez01}
1085: {Mel{\'e}ndez}, J., {Barbuy}, B., \& {Spite}, F. 2001, \apj, 556, 858
1086: 
1087: \bibitem[{{Mooney} {et~al.}(2004){Mooney}, {Rolleston}, {Keenan}, {Dufton},
1088:   {Smoker}, {Ryans}, {Aller}, \& {Trundle}}]{mooney04}
1089: {Mooney}, C.~J., {Rolleston}, W.~R.~J., {Keenan}, F.~P., {Dufton}, P.~L.,
1090:   {Smoker}, J.~V., {Ryans}, R.~S.~I., {Aller}, L.~H., \& {Trundle}, C. 2004,
1091:   \aap, 419, 1123
1092: 
1093: \bibitem[{{Mowlavi} {et~al.}(1996){Mowlavi}, {Jorissen}, \&
1094:   {Arnould}}]{mowlavi96}
1095: {Mowlavi}, N., {Jorissen}, A., \& {Arnould}, M. 1996, \aap, 311, 803
1096: 
1097: \bibitem[{{Norris} \& {Da Costa}(1995)}]{norris95b}
1098: {Norris}, J.~E. \& {Da Costa}, G.~S. 1995, \apj, 447, 680
1099: 
1100: \bibitem[{{Norris} \& {Freeman}(1979)}]{nf79}
1101: {Norris}, J. \& {Freeman}, K.~C. 1979, \apjl, 230, L179
1102: 
1103: \bibitem[{{Odenkirchen} {et~al.}(2003){Odenkirchen}, {Grebel}, {Dehnen}, {Rix},
1104:   {Yanny}, {Newberg}, {Rockosi}, {Mart{\'{\i}}nez-Delgado}, {Brinkmann}, \&
1105:   {Pier}}]{odenkirchen03}
1106: {Odenkirchen}, M., {Grebel}, E.~K., {Dehnen}, W., {Rix}, H.-W., {Yanny}, B.,
1107:   {Newberg}, H.~J., {Rockosi}, C.~M., {Mart{\'{\i}}nez-Delgado}, D.,
1108:   {Brinkmann}, J., \& {Pier}, J.~R. 2003, \aj, 126, 2385
1109: 
1110: \bibitem[{{Odenkirchen} {et~al.}(2001){Odenkirchen}, {Grebel}, {Rockosi},
1111:   {Dehnen}, {Ibata}, {Rix}, {Stolte}, {Wolf}, {Anderson}, {Bahcall},
1112:   {Brinkmann}, {Csabai}, {Hennessy}, {Hindsley}, {Ivezi{\'c}}, {Lupton},
1113:   {Munn}, {Pier}, {Stoughton}, \& {York}}]{odenkirchen01}
1114: {Odenkirchen}, M., {Grebel}, E.~K., {Rockosi}, C.~M., {Dehnen}, W., {Ibata},
1115:   R., {Rix}, H.-W., {Stolte}, A., {Wolf}, C., {Anderson}, Jr., J.~E.,
1116:   {Bahcall}, N.~A., {Brinkmann}, J., {Csabai}, I., {Hennessy}, G., {Hindsley},
1117:   R.~B., {Ivezi{\'c}}, {\v Z}., {Lupton}, R.~H., {Munn}, J.~A., {Pier}, J.~R.,
1118:   {Stoughton}, C., \& {York}, D.~G. 2001, \apjl, 548, L165
1119: 
1120: \bibitem[{{Paltrinieri} {et~al.}(2001){Paltrinieri}, {Ferraro}, {Paresce}, \&
1121:   {De Marchi}}]{paltrinieri01}
1122: {Paltrinieri}, B., {Ferraro}, F.~R., {Paresce}, F., \& {De Marchi}, G. 2001,
1123:   \aj, 121, 3114
1124: 
1125: \bibitem[{{Paresce} \& {De Marchi}(2000)}]{paresce00}
1126: {Paresce}, F. \& {De Marchi}, G. 2000, \apj, 534, 870
1127: 
1128: \bibitem[{{Pilachowski} {et~al.}(1996){Pilachowski}, {Sneden}, \&
1129:   {Kraft}}]{pilachowski96}
1130: {Pilachowski}, C.~A., {Sneden}, C., \& {Kraft}, R.~P. 1996, \aj, 111, 1689
1131: 
1132: \bibitem[{{Prantzos} \& {Charbonnel}(2006)}]{prantzos06}
1133: {Prantzos}, N. \& {Charbonnel}, C. 2006, \aap, 458, 135
1134: 
1135: \bibitem[{{Prantzos} {et~al.}(2007){Prantzos}, {Charbonnel}, \&
1136:   {Iliadis}}]{prantzos07}
1137: {Prantzos}, N., {Charbonnel}, C., \& {Iliadis}, C. 2007, \aap, 470, 179
1138: 
1139: \bibitem[{{Pritzl} {et~al.}(2005){Pritzl}, {Venn}, \& {Irwin}}]{pritzl05}
1140: {Pritzl}, B.~J., {Venn}, K.~A., \& {Irwin}, M. 2005, \aj, 130, 2140
1141: 
1142: \bibitem[{{Ram{\'{\i}}rez} \& {Mel{\'e}ndez}(2005)}]{ramirez05}
1143: {Ram{\'{\i}}rez}, I. \& {Mel{\'e}ndez}, J. 2005, \apj, 626, 465
1144: 
1145: \bibitem[{{Renda} {et~al.}(2004){Renda}, {Fenner}, {Gibson}, {Karakas},
1146:   {Lattanzio}, {Campbell}, {Chieffi}, {Cunha}, \& {Smith}}]{renda04}
1147: {Renda}, A., {Fenner}, Y., {Gibson}, B.~K., {Karakas}, A.~I., {Lattanzio},
1148:   J.~C., {Campbell}, S., {Chieffi}, A., {Cunha}, K., \& {Smith}, V.~V. 2004,
1149:   \mnras, 354, 575
1150: 
1151: \bibitem[{{Scalo} {et~al.}(1975){Scalo}, {Despain}, \& {Ulrich}}]{scalo75}
1152: {Scalo}, J.~M., {Despain}, K.~H., \& {Ulrich}, R.~K. 1975, \apj, 196, 805
1153: 
1154: \bibitem[{{Schlegel} {et~al.}(1998){Schlegel}, {Finkbeiner}, \&
1155:   {Davis}}]{schlegel98}
1156: {Schlegel}, D.~J., {Finkbeiner}, D.~P., \& {Davis}, M. 1998, \apj, 500, 525
1157: 
1158: \bibitem[{{Searle} \& {Zinn}(1978)}]{sz78}
1159: {Searle}, L. \& {Zinn}, R. 1978, \apj, 225, 357
1160: 
1161: \bibitem[{{Shetrone}(1996)}]{shetrone96b}
1162: {Shetrone}, M.~D. 1996, \aj, 112, 2639
1163: 
1164: \bibitem[{{Skrutskie} {et~al.}(2006){Skrutskie}, {Cutri}, {Stiening},
1165:   {Weinberg}, {Schneider}, {Carpenter}, {Beichman}, {Capps}, {Chester},
1166:   {Elias}, {Huchra}, {Liebert}, {Lonsdale}, {Monet}, {Price}, {Seitzer},
1167:   {Jarrett}, {Kirkpatrick}, {Gizis}, {Howard}, {Evans}, {Fowler}, {Fullmer},
1168:   {Hurt}, {Light}, {Kopan}, {Marsh}, {McCallon}, {Tam}, {Van Dyk}, \&
1169:   {Wheelock}}]{2mass}
1170: {Skrutskie}, M.~F., {Cutri}, R.~M., {Stiening}, R., {Weinberg}, M.~D.,
1171:   {Schneider}, S., {Carpenter}, J.~M., {Beichman}, C., {Capps}, R., {Chester},
1172:   T., {Elias}, J., {Huchra}, J., {Liebert}, J., {Lonsdale}, C., {Monet}, D.~G.,
1173:   {Price}, S., {Seitzer}, P., {Jarrett}, T., {Kirkpatrick}, J.~D., {Gizis},
1174:   J.~E., {Howard}, E., {Evans}, T., {Fowler}, J., {Fullmer}, L., {Hurt}, R.,
1175:   {Light}, R., {Kopan}, E.~L., {Marsh}, K.~A., {McCallon}, H.~L., {Tam}, R.,
1176:   {Van Dyk}, S., \& {Wheelock}, S. 2006, \aj, 131, 1163
1177: 
1178: \bibitem[{{Smith}(1987)}]{smith87}
1179: {Smith}, G.~H. 1987, \pasp, 99, 67
1180: 
1181: \bibitem[{{Smith}(2006)}]{smith06}
1182: ---. 2006, \pasp, 118, 1225
1183: 
1184: \bibitem[{{Smith} {et~al.}(2002{\natexlab{a}}){Smith}, {Sneden}, \&
1185:   {Kraft}}]{smith02}
1186: {Smith}, G.~H., {Sneden}, C., \& {Kraft}, R.~P. 2002{\natexlab{a}}, \aj, 123,
1187:   1502
1188: 
1189: \bibitem[{{Smith} {et~al.}(2005){Smith}, {Cunha}, {Ivans}, {Lattanzio},
1190:   {Campbell}, \& {Hinkle}}]{smith05}
1191: {Smith}, V.~V., {Cunha}, K., {Ivans}, I.~I., {Lattanzio}, J.~C., {Campbell},
1192:   S., \& {Hinkle}, K.~H. 2005, \apj, 633, 392
1193: 
1194: \bibitem[{{Smith} {et~al.}(2002{\natexlab{b}}){Smith}, {Hinkle}, {Cunha},
1195:   {Plez}, {Lambert}, {Pilachowski}, {Barbuy}, {Mel{\'e}ndez}, {Balachandran},
1196:   {Bessell}, {Geisler}, {Hesser}, \& {Winge}}]{smith02b}
1197: {Smith}, V.~V., {Hinkle}, K.~H., {Cunha}, K., {Plez}, B., {Lambert}, D.~L.,
1198:   {Pilachowski}, C.~A., {Barbuy}, B., {Mel{\'e}ndez}, J., {Balachandran}, S.,
1199:   {Bessell}, M.~S., {Geisler}, D.~P., {Hesser}, J.~E., \& {Winge}, C.
1200:   2002{\natexlab{b}}, \aj, 124, 3241
1201: 
1202: \bibitem[{{Smith} {et~al.}(2000){Smith}, {Suntzeff}, {Cunha}, {Gallino},
1203:   {Busso}, {Lambert}, \& {Straniero}}]{smith00}
1204: {Smith}, V.~V., {Suntzeff}, N.~B., {Cunha}, K., {Gallino}, R., {Busso}, M.,
1205:   {Lambert}, D.~L., \& {Straniero}, O. 2000, \aj, 119, 1239
1206: 
1207: \bibitem[{{Sneden}(1973)}]{moog}
1208: {Sneden}, C. 1973, \apj, 184, 839
1209: 
1210: \bibitem[{{Sneden} {et~al.}(2004){Sneden}, {Kraft}, {Guhathakurta}, {Peterson},
1211:   \& {Fulbright}}]{sneden04a}
1212: {Sneden}, C., {Kraft}, R.~P., {Guhathakurta}, P., {Peterson}, R.~C., \&
1213:   {Fulbright}, J.~P. 2004, \aj, 127, 2162
1214: 
1215: \bibitem[{{Takahashi} \& {Portegies Zwart}(2000)}]{takahashi00}
1216: {Takahashi}, K. \& {Portegies Zwart}, S.~F. 2000, \apj, 535, 759
1217: 
1218: \bibitem[{{Uttenthaler} {et~al.}(2008){Uttenthaler}, {Lebzelter}, {Aringer},
1219:   {K{\"a}ufl}, {Siebenmorgen}, \& {Smette}}]{uttenthaler08}
1220: {Uttenthaler}, S., {Lebzelter}, T., {Aringer}, B., {K{\"a}ufl}, H.-U.,
1221:   {Siebenmorgen}, R., \& {Smette}, A. 2008, \apj\ in press (arXiv:0804.4057)
1222: 
1223: \bibitem[{{Ventura} \& {D'Antona}(2005{\natexlab{a}})}]{ventura05a}
1224: {Ventura}, P. \& {D'Antona}, F. 2005{\natexlab{a}}, \aap, 431, 279
1225: 
1226: \bibitem[{{Ventura} \& {D'Antona}(2005{\natexlab{b}})}]{ventura05b}
1227: ---. 2005{\natexlab{b}}, \aap, 439, 1075
1228: 
1229: \bibitem[{{White} \& {Rees}(1978)}]{white78}
1230: {White}, S.~D.~M. \& {Rees}, M.~J. 1978, \mnras, 183, 341
1231: 
1232: \bibitem[{{Woosley} \& {Weaver}(1995)}]{woosley95}
1233: {Woosley}, S.~E. \& {Weaver}, T.~A. 1995, \apjs, 101, 181
1234: 
1235: \bibitem[{{Yong} {et~al.}(2006){Yong}, {Aoki}, \& {Lambert}}]{mghsubaru}
1236: {Yong}, D., {Aoki}, W., \& {Lambert}, D.~L. 2006, \apj, 638, 1018
1237: 
1238: \bibitem[{{Yong} {et~al.}(2008{\natexlab{a}}){Yong}, {Grundahl}, {Johnson}, \&
1239:   {Asplund}}]{6752nh}
1240: {Yong}, D., {Grundahl}, F., {Johnson}, J.~A., \& {Asplund}, M.
1241:   2008{\natexlab{a}}, \apj\ in press (arXiv:0806.0187)
1242: 
1243: \bibitem[{{Yong} {et~al.}(2003{\natexlab{a}}){Yong}, {Grundahl}, {Lambert},
1244:   {Nissen}, \& {Shetrone}}]{6752}
1245: {Yong}, D., {Grundahl}, F., {Lambert}, D.~L., {Nissen}, P.~E., \& {Shetrone},
1246:   M.~D. 2003{\natexlab{a}}, \aap, 402, 985
1247: 
1248: \bibitem[{{Yong} {et~al.}(2003{\natexlab{b}}){Yong}, {Lambert}, \&
1249:   {Ivans}}]{mghdwarf}
1250: {Yong}, D., {Lambert}, D.~L., \& {Ivans}, I.~I. 2003{\natexlab{b}}, \apj, 599,
1251:   1357
1252: 
1253: \bibitem[{{Yong} {et~al.}(2008{\natexlab{b}}){Yong}, {Lambert}, {Paulson}, \&
1254:   {Carney}}]{rbpbm4m5}
1255: {Yong}, D., {Lambert}, D.~L., {Paulson}, D.~B., \& {Carney}, B.~W.
1256:   2008{\natexlab{b}}, \apj, 673, 854
1257: 
1258: \bibitem[{{Zinn} \& {West}(1984)}]{zinn84}
1259: {Zinn}, R. \& {West}, M.~J. 1984, \apjs, 55, 45
1260: 
1261: \end{thebibliography}
1262: 
1263: 
1264: \clearpage
1265: 
1266: \input{tab1}
1267: \input{tab2}
1268: \input{tab3}
1269: \input{tab4}
1270: \input{tab5}
1271: 
1272: \clearpage
1273: \begin{figure}
1274: \epsscale{0.8}
1275: \plotone{f1.ps}
1276: \caption{Spectra of V21 for the two wavelength regions. 
1277: Lines used in the abundance analysis are indicated. 
1278: \label{fig:spectra}}
1279: \end{figure}
1280: 
1281: \clearpage
1282: 
1283: \begin{figure}
1284: \epsscale{0.8}
1285: \plotone{f2.ps}
1286: \caption{Observed spectra (circles) and synthetic spectra 
1287: for C (upper), N (middle), and O (lower) in LM8. The synthetic 
1288: spectra show the best fit (thick black line) and unsatisfactory 
1289: fits (thin red and blue lines) 
1290: $A$(C) $\pm$ 0.15 dex, $A$(N) $\pm$ 0.20 dex, and $A$(O) $\pm$ 0.15 dex. 
1291: \label{fig:cn}}
1292: \end{figure}
1293: 
1294: \clearpage
1295: 
1296: \begin{figure}
1297: \epsscale{0.8}
1298: \plotone{f3.ps}
1299: \caption{Observed spectra (circles) and synthetic spectra 
1300: for LM10. 
1301: The lines show syntheses with different C and Na abundances. The positions of 
1302: CO and Na lines are shown. 
1303: \label{fig:na}}
1304: \end{figure}
1305: 
1306: \clearpage
1307: 
1308: \begin{figure}
1309: \epsscale{0.8}
1310: \plotone{f4.ps}
1311: \caption{Observed spectra (circles) and synthetic spectra 
1312: for F in LM10 (upper) and V10 (lower). 
1313: In the lower panel, 
1314: two sets of syntheses are plotted corresponding to 
1315: $\log g$ = +0.22 (dashed line) and 0.00 (solid line). The lines 
1316: are indistinguishable and for this element in this star, 
1317: the extrapolated abundance 
1318: for $\log g$ = $-$0.22 is $A$(F) = 2.65. 
1319: \label{fig:f}}
1320: \end{figure}
1321: 
1322: \clearpage
1323: 
1324: \begin{figure}
1325: \epsscale{0.8}
1326: \plotone{f5.ps}
1327: \caption{Elemental abundances $A$(X) vs.\ $A$(C) as well as 
1328: $A$(Na) vs.\ $A$(O) (lower right panel). A representative 
1329: error bar is shown. The dashed line is the linear least squares 
1330: fit to the data (slope and associated error are included).  
1331: \label{fig:cnofna}}
1332: \end{figure}
1333: 
1334: \clearpage
1335: 
1336: \begin{figure}
1337: \epsscale{0.8}
1338: \plotone{f6.ps}
1339: \caption{The abundance distribution of $A$(C), $A$(N), $A$(O), $A$(CN), 
1340: and $A$(CNO) for NGC 6712 (circles) and M4 (crosses). The M4 data are 
1341: from \citet{smith05}. The amplitude 
1342: of the abundance dispersion is shown along with a representative error 
1343: bar. \label{fig:cno}}
1344: \end{figure}
1345: 
1346: \clearpage
1347: 
1348: \begin{figure}
1349: \epsscale{0.8}
1350: \plotone{f7.ps}
1351: \caption{Same as Figure \ref{fig:cno} but for $A$(F), $A$(Na), 
1352: and [O/Na]. \label{fig:fnaona}}
1353: \end{figure}
1354: 
1355: \clearpage
1356: 
1357: \begin{figure}
1358: \epsscale{0.6}
1359: \plotone{f8.ps}
1360: \caption{$A$(F) vs.\ $A$(O) (upper) and $\log$[N(F)/N(O)] vs.\ 
1361: $A$(O) (lower). NGC 6712 (black circles), 
1362: M4 (black crosses: \citealt{smith05}), 
1363: $\omega$ Cen (black triangles: \citealt{cunha03}), 
1364: bulge stars (red circles: \citealt{cunha08}), and 
1365: field stars (red plus signs: \citealt{cunha03} and \citealt{cunha05}) 
1366: are shown. A representative error bar is shown. 
1367: The red and black 
1368: dashed lines are the linear least squares fits to the field+bulge 
1369: and globular cluster data respectively (excluding upper limits). 
1370: The dotted red line is the fit to the field+bulge data excluding 
1371: the upper limits and the bulge star with $A$(O) = 9.0. \label{fig:fvso}}
1372: \end{figure}
1373: 
1374: \end{document}
1375: 
1376: 
1377: