0806.0187/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{Nitrogen abundances in giant stars of the globular cluster 
9: NGC 6752\footnote{Based on observations made with ESO Telescopes 
10: at the Paranal Observatories under programme 65.L-0165(A)}}
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{Frank Grundahl}
19: \affil{Department of Physics and Astronomy, University of Aarhus, Denmark}
20: \email{fgj@phys.au.dk}
21: 
22: \author{Jennifer A.\ Johnson}
23: \affil{Department of Astronomy, Ohio State University, Columbus, OH}
24: \email{jaj@astronomy.ohio-state.edu}
25: 
26: \author{Martin Asplund}
27: \affil{Max-Planck-Institut f{\" u}r Astrophysik, Karl-Schwarzschild-Str. 1, Postfach 1317, 85741 Garching b.\ M{\" u}nchen, Germany}
28: \email{asplund@mpa-garching.mpg.de}
29: 
30: \begin{abstract}
31: 
32: We present N abundances for 21 bright giants in the globular 
33: cluster NGC 6752 based on high-resolution UVES spectra of the 3360\AA~NH 
34: lines. We confirm that the Str{\" o}mgren $c_1$ index traces the N abundance 
35: and find that the star-to-star N abundance variation is 1.95 dex, at 
36: the sample's luminosity. We find statistically significant correlations, 
37: but small amplitude variations, 
38: between the abundances of N and $\alpha$-, Fe-peak, 
39: and $s$-process elements. Analyses using model 
40: atmospheres with appropriate N, O, Na, and Al abundances would strengthen, 
41: rather than mute, these correlations. If the small variations of heavy 
42: elements are real, then the synthesis of the N anomalies must take place 
43: in stars which also synthesize $\alpha$-, Fe-peak, and $s$-process elements. 
44: These correlations 
45: offer support for contributions from both AGB and massive stars 
46: to the globular cluster abundance anomalies. 
47: 
48: \end{abstract}
49: 
50: \keywords{Galaxy: Globular Clusters: Individual: NGC 6752, Galaxy: Stars: Abundances}
51: 
52: \section{Introduction}
53: \label{sec:intro}
54: 
55: Several decades have passed since the first identifications of 
56: the star-to-star abundance variations of light elements
57: in globular 
58: clusters \citep{popper47,osborn71,nf79}. 
59: Hydrogen burning via the CNO, Ne-Na, and Mg-Al chains 
60: at high temperatures qualitatively accounts for the 
61: observed trends: the abundances of C and O are low when N is high, 
62: O and Na are anticorrelated as are Mg and Al 
63: \citep{langer95,denissenkov98,karakas03}. 
64: While the amplitude of 
65: the variation may differ from cluster to cluster, the abundance patterns 
66: have been found in every well studied Galactic globular cluster 
67: \citep{smith87,kraft94,gratton04}. Such 
68: abundance patterns have also been identified in stars in  
69: extragalactic globular clusters \citep{letarte06,johnson06}. 
70: However, the origin of these abundance anomalies remains elusive. 
71: 
72: Internal nucleosynthesis and mixing can account for C, N, and Li in 
73: giant stars whose abundances exhibit a dependence upon evolutionary 
74: status \citep{ss91,grundahl02}.  
75: These abundance patterns, C and Li destruction along with N production, 
76: are also found in field stars as the evolve up the RGB \citep{gratton00b}. 
77: However, the identification of C, N, O, Na, Mg, and Al abundance 
78: variations in unevolved cluster stars 
79: \citep{briley96,gratton01,grundahl02,cohen05} 
80: demonstrates that an external pollution mechanism 
81: must be the dominant source since unevolved stars have insufficient 
82: internal temperatures to run the necessary nuclear reactions 
83: and lack a mechanism to mix the products to surface layers. 
84: Unlike C, N, and Li, abundance 
85: variations for the elements O to Al have rarely, if ever, been 
86: identified in field stars, indicating that the abundance anomalies 
87: are likely associated with some (presently unknown) property of the cluster 
88: environment. \citet{fulbright07} found low O and high Na and Al abundances 
89: in two bulge giants. However, they suggest that these two 
90: stars may be members of the bulge globular cluster NGC 6522. 
91: Intermediate-mass asymptotic giant branch (AGB) stars 
92: \citep{cottrell81,ventura05c} and 
93: massive stars \citep{prantzos06,smith06,decressin06}  
94: are candidates for the external 
95: pollution scenario. However, neither scenario 
96: currently provides a satisfactory explanation. 
97: 
98: Nitrogen abundance variations in globular cluster giants 
99: were inferred from early 
100: observations of CH and CN variations \citep{zinn77,dacosta80}. 
101: The NH molecular lines arguably 
102: offer the best probe of N abundances, in contrast to analyses 
103: of the CN lines which require knowledge of the C and O 
104: abundances. However, few studies have measured N from NH as well 
105: as the abundances for additional elements in a large 
106: homogeneous sample. In particular, few studies have measured N from NH 
107: using high-resolution spectra for a large sample. 
108: In this paper, we perform such an analysis of bright giants 
109: in the globular cluster NGC 6752 to measure the 
110: amplitude of the N abundance variation and to identify correlations 
111: between N abundances and light element abundances as well as possible 
112: correlations between N abundances and the abundances of heavy elements. 
113: 
114: \section{Observations and data reduction}
115: \label{sec:data}
116: 
117: The targets were the same as those analyzed by \citet{grundahl02}. 
118: The sample consists of 21 stars near the RGB bump, 
119: selected from the Str{\" o}mgren $uvby$ photometry of \citet{grundahl99}, 
120: and observed in service mode in Apr-May 2000 
121: using UVES \citep{uves}. 
122: The stars were selected with the expectation that they would sample 
123: the full range of the star-to-star light element abundance variation, 
124: at this luminosity. 
125: \citet{grundahl02} suspected that the Str{\" o}mgren $c_1$ index 
126: was tracing the N abundance (the $u$ filter includes 
127: the 3360\AA~NH features), and they 
128: identified clear correlations between $c_1$ and NH and CN indices 
129: defined by \citet{bs93}. 
130: The stars were deliberately selected 
131: to span the full range of the Str{\" o}mgren $c_1$ index.  
132: In Figure \ref{fig:cmd1}, we show the $V$ versus $(v-y)$ and the 
133: $V$ versus $c_1$ color-magnitude diagrams showing the locations of 
134: the targets. 
135: 
136: The ESO pipeline reduced spectra have a resolution of R = 60,000, but 
137: were smoothed with a 5 pixel boxcar function to increase the signal-to-noise
138: ratio (S/N). The resolving power after smoothing is R $\simeq$ 30,000. 
139: Although the S/N in the smoothed spectra is difficult to 
140: estimate in the crowded region near the 3360\AA~NH lines, 
141: we estimate that all stars have a
142: typical S/N = 70 per resolution element. For more details regarding
143: target selection, observations, and data reduction see \citet{grundahl02}. 
144: 
145: \section{Stellar parameters and abundance analysis}
146: \label{sec:analysis}
147: 
148: Derivation of the stellar parameters is described in \citet{grundahl02}. 
149: Briefly, the effective temperatures (\teff) were 
150: determined from the \citet{grundahl99} $uvby$ photometry using the 
151: \citet{alonso99b} color-temperature relations. The surface gravities 
152: ($\log g$) were estimated using the stellar luminosities and \teff. 
153: The microturbulent velocities ($\xi_t$) 
154: were derived in the usual way by forcing 
155: the abundances from Fe\,{\sc i} to be independent of line strength. 
156: The stellar parameters for the program stars are presented in 
157: Table \ref{tab:param}. 
158: 
159: Visual examination of the spectra and the considerable range in 
160: NH indices measured by \citet{grundahl02} 
161: indicate a large amplitude to the star-to-star 
162: N abundance variation. Quantitative N abundances were 
163: derived by comparing the observed spectra with
164: synthetic spectra generated using the LTE 
165: spectrum synthesis and line analysis package {\sc Moog} \citep{moog}. The 
166: molecular data and subsequent line list for the 
167: (0, 0) and (1, 1) bands of the 
168: $A-X$ electronic transition of the NH molecule at 3360\AA~were taken from 
169: \citet{johnson07}. We used model atmospheres computed by \citet{kurucz93}, 
170: which are the same as those used by \citet{6752,67522}. Synthetic spectra 
171: were generated with different N abundances (we assumed that all N was in 
172: the form $^{14}$N). The difficult task of setting the continuum was 
173: achieved by selecting a handful of continuum windows and 
174: using the synthetic spectra as a guide. We adjusted the N abundances 
175: until the residuals between 
176: the synthetic and observed spectra were minimized. 
177: In Figures \ref{fig:1718fit} and \ref{fig:23fit}, we show pairs of 
178: stars with essentially identical stellar parameters. However, these pairs 
179: of stars have considerably different N abundances as can be seen 
180: qualitatively by eye or quantitatively via the $c_1$ index and final 
181: [N/Fe] abundance. 
182: The N abundances, [N/Fe], are presented in Table \ref{tab:param} 
183: assuming the newly determined 
184: solar abundance of $\log$ $\epsilon$(N) = 7.78 \citep{grevesse07}. 
185: Our conclusions would not be changed had we adopted a previous value for 
186: the solar N abundance, e.g.,  
187: $\log$ $\epsilon$(N) = 7.92 from \citet{grevesse98}. 
188:  
189: We estimate that the internal uncertainties in the measured N abundances are 
190: typically 0.2 dex, resulting from errors in the adopted 
191: stellar parameters and errors in determining the best fit. 
192: The internal errors would be slightly lower ($\sim$0.15 dex) 
193: for the most N-poor stars and 
194: slightly higher ($\sim$0.25 dex) 
195: for the most N-rich stars due to saturation concerns. 
196: For additional elements discussed in 
197: Section \ref{sec:discussion}, 
198: the derivation of abundances and their associated uncertainties 
199: were described in \citet{grundahl02} and 
200: \citet{6752,67522}. (We omit Star NGC 6752-7 from further 
201: analysis due to its deviating [Fe/H].)
202: 
203: \section{Discussion}
204: \label{sec:discussion}
205: 
206: \subsection{Nitrogen abundances and comparisons with previous studies}
207: 
208: The [N/Fe] ratios for the sample range from [N/Fe] = $-$0.43 to 
209: [N/Fe] = +1.52, almost a 2 dex variation. 
210: Recall that the 
211: sample was selected according to the Str{\" o}mgren $c_1$ index, which 
212: \citet{grundahl02} suspected was tracing the 
213: N abundance. In Figure \ref{fig:nc1}, we plot the 
214: $c_1$ index versus the N abundance. The clear correlation 
215: confirms that $c_1$ indeed traces the N abundance. Given the 
216: Str{\" o}mgren $c_1$ selection criteria, it is not surprising 
217: that the full abundance distribution is well sampled. However, 
218: our N abundances are not representative of the N distribution for 
219: the entire cluster. 
220: 
221: \citet{carretta05} measured N abundances for 9 dwarfs and 9 subgiants in 
222: NGC 6752. Their N measurements are based on analysis of 
223: the CN molecular lines (rather than the NH lines used in this analysis) 
224: which requires knowledge of the C and O abundances.  
225: %Our N abundances are not affected by uncertainties in the C and/or O 
226: %abundances and presumably the uncertainties in our N abundances will be 
227: %lower than analyses based on CN features. 
228: Our syntheses of the NH lines, and therefore our N abundances, 
229: are entirely independent of the 
230: adopted C and O abundances and presumably the uncertainties in 
231: our N abundances will be lower than analyses based on CN features. 
232: Aside from 1 subgiant with the abundance ratio [N/Fe] = 0.0, 
233: the \citet{carretta05} sample spans the range +1.0 $\le$ [N/Fe] $\le$ +1.7. 
234: Their maximum abundance [N/Fe] = +1.7 is comparable to the maximum 
235: abundance derived in this study, within the measurement uncertainties. 
236: However, their lowest abundance [N/Fe] = 0.0 is considerably higher 
237: than the minimum value found in our analysis. If CN cycling 
238: has occurred in our stars, which are more evolved than the
239: \citet{carretta05} sample, we would expect our stars to have 
240: systematically higher N abundances (assuming 
241: that the two samples were drawn from 
242: the same population and that the two samples covered the full range of the 
243: abundance variation at their respective luminosities). 
244: 
245: For the \citet{carretta05} sample, abundances for other 
246: light elements known to vary from star to star
247: have been derived by \citet{gratton01}. For convenience, we consider 
248: the [Na/Fe] ratio. In Figure \ref{fig:nan} we compare 
249: [Na/Fe] vs.\ [N/Fe] for our stars and the \citet{carretta05} stars. 
250: Their minimum value, 
251: [Na/Fe] = $-$0.29, is slightly lower than our minimum, 
252: [Na/Fe] = $-$0.10. The maximum [Na/Fe] ratio for \citet{carretta05}, 
253: [Na/Fe] = +0.62, is comparable to our maximum, 
254: [Na/Fe] = +0.67. Given that the amplitude of the [Na/Fe] variation in 
255: \citet{carretta05} exceeds our [Na/Fe] variation, 
256: we would naively expect the [N/Fe] variation in their sample 
257: to be comparable to, or larger than, our sample. 
258: Therefore, it is somewhat unusual that the amplitude of our 
259: [N/Fe] variation is considerably larger than in 
260: \citet{carretta05}. Inspection of Figure \ref{fig:nan} shows 
261: that the [N/Fe] distributions look rather different and we speculate 
262: that the difference is due in part to difficulties in deriving 
263: accurate N abundances from CN lines. Although the 
264: \citet{carretta05} star with [N/Fe] = 0 appears rather unusual compared to 
265: the bulk of their sample, it lies on the 
266: distribution defined by our stars. 
267: One possibility is that measurement uncertainties 
268: for N and/or Na have affected one or both of the analyses. 
269: We are not aware of additional analyses of N in large 
270: numbers of stars in NGC 6752. 
271: 
272: Of the well studied globular clusters, NGC 6752 exhibits one the 
273: largest amplitudes for light element abundance variations. For other 
274: well studied globular clusters that also display large abundance 
275: variations, several have measured N abundances. Work by 
276: \citet{briley02,briley04} and \citet{cohen02,cohen05b} 
277: has shown that the 
278: clusters M5, M13, M15, M71, and 47 Tuc have [N/Fe] ratios that 
279: cover 2 dex or more. These works are based on very large numbers 
280: of main sequence stars ($\sim$ 50 - 100), 
281: albeit using lower resolution spectra with abundances 
282: determined via the comparison 
283: of indices measured in observed and synthetic spectra. 
284: These indices measure the flux removed by molecular features 
285: relative to nearby continuum bandpasses \citep{briley01}. 
286: \citet{carretta05} also measured N abundances 
287: in modest numbers of dwarfs and subgiants 
288: in the globular clusters NGC 6397 and 47 Tuc. For NGC 6397, their analysis 
289: of subgiant stars shows that [N/Fe] covers a range of 2.0 dex. For 
290: 47 Tuc, their analysis of dwarfs and subgiants shows a [N/Fe] range of 
291: 1.6 dex. In summary, the amplitude of the N abundance variation in 
292: this study of NGC 6752 is comparable to that observed in other clusters. 
293: 
294: \subsection{Str{\" o}mgren photometry and N abundances}
295: 
296: To further investigate the relationship between Str{\" o}mgren 
297: photometry and N abundances, we introduce a new index $cy$ = $c_1 - (b-y)$. 
298: This index was identified through simple experiements and is a purely 
299: empirical index which appears to remove temperature from $c_1$, 
300: to first order. 
301: The $cy$ index represents $c_1$ well, but for stars on the lower RGB 
302: (fainter than the bump) $cy$ appears to show no evolutionary component. 
303: In Figure \ref{fig:cmd2}, we plot 
304: $V$ versus $cy$. While there appears to be a large scatter in $cy$ at all 
305: evolutionary phases, the width and center of the distribution are independent 
306: of the $V$ magnitude. In Figure \ref{fig:ncy}, 
307: we plot [N/Fe] versus $cy$. Remarkably, there appears 
308: to be a linear relation to which we fit 
309: [N/Fe] = 16.11$\times cy$ + 4.07. The scatter around the fit is only 0.29 dex.  
310: A linear fit between [N/Fe] and $c_1$ has a scatter of 
311: 0.31 dex and a quadratic fit has a scatter of 0.27 dex. The reduced 
312: scatter for the quadratic fit confirms the impression from 
313: Figure \ref{fig:nc1}. However, we prefer the linear relation 
314: between [N/Fe] and $cy$ which can then be used to explore in more detail  
315: the distribution of [N/Fe] 
316: for the large numbers of giant stars for which Str{\" o}mgren 
317: photometry, but not high-resolution spectra, are available. 
318: 
319: In Figure \ref{fig:ndist}, we plot the [N/Fe] distribution 
320: for all giant stars (N = 110) 
321: with $V$ magnitudes spanned by the calibration stars 
322: employing the relation given above. 
323: We find that the distribution increases towards higher [N/Fe] values. 
324: That is, the distribution is not flat, nor is it bimodal. 
325: \citet{norris81} found that the CN distribution for NGC 6752 is bimodal 
326: based on 69 giant stars. 
327: In the same Figure, we also plot the [N/Fe] distribution for all giant 
328: stars with magnitudes 13 $\le$ $V$ $\le$ 16 (N = 559). Although it is not clear 
329: whether the above relation is applicable beyond the magnitudes (i.e., 
330: stellar parameters) of the calibrating stars, we note that 
331: both distributions are very similar. These results imply that in 
332: other clusters, well calibrated $uvby$ photometry could be used
333: to probe the nitrogen abundance distribution. This may even be possible 
334: for clusters in the LMC and SMC. 
335: 
336: The $cy$-[N/Fe] relation is valid over the range of parameters covered 
337: by our sample. Having applied this relation to the entire RGB, we 
338: recover a similar [N/Fe] distribution as seen within the stars spanned 
339: by the calibrating stars. Therefore, we speculate that this relation 
340: may be applicable along the entire RGB. 
341: 
342: \subsection{Correlations between the abundances of nitrogen 
343: and other elements}
344: 
345: Statistically significant correlations, 
346: albeit of small amplitude, between Al and the
347: abundances of Si and heavier elements in NGC 6752's bright 
348: giant stars (including the RGB bump stars in this analysis) 
349: were identified by \citet{67522}. Since the N abundances 
350: exhibit a larger amplitude than the Al abundances, we seek to 
351: investigate whether there are correlations between N and 
352: heavier elements. In Figures \ref{fig:ona} to \ref{fig:baeu}, 
353: we plot the N abundances versus all measured elements in NGC 6752. 
354: In these Figures, the abundances 
355: for elements other than N are taken from \citet{grundahl02} and 
356: \citet{6752,67522}. All studies are homogeneous using the 
357: same model atmospheres, stellar parameters, and analysis techniques. 
358: 
359: For O, Na, Mg, and Al, the usual correlations with N are evident. 
360: For completeness, we note that these correlations are significant 
361: at the 10-$\sigma$ level (i.e., we fit a straight line to the 
362: data and measure the slope and uncertainty, 
363: taking into account both the x and y errors).  
364: We find that the elements Si, Ca, Sc, Ti, Ni, Cu, Y, Zr, Ba, 
365: and Nd all show correlations with N. 
366: Although the amplitude of the abundance variations for these heavier elements 
367: is small ($<$ 0.2 dex), the correlations are statistically significant
368: at the level 2-$\sigma$ or greater. 
369: (Even if we increase the x and y errors by 50\%, Si, Sc, Cu, Y, Zr, and Ba 
370: remain correlated with N at the level 2-$\sigma$ or greater.) 
371: Taken at face value, 
372: these correlations indicate that the nucleosynthetic 
373: source of the N abundance variation must also synthesize small amounts of 
374: these heavier elements. 
375: In \citet{67522}, we showed that the abundances are not correlated with 
376: \teff~and therefore any correlation between N and heavier elements 
377: is unlikely to be the result of systematic errors in the stellar parameters.
378: 
379: The model atmospheres used in the analysis were 
380: computed using scaled-solar compositions whereas the stars to which 
381: these atmospheres were applied do not have scaled-solar compositions. 
382: Therefore, we were concerned that the correlations between N and 
383: heavier elements could be due to the use of inappropriate models. 
384: That is, would the derived abundances change if we used model atmospheres
385: with the appropriately small or large N, O, Na and Al abundances? 
386: Or equivalently, does the inclusion of appropriate N, O, Na, and Al 
387: abundances change the structure of the model atmospheres to such a 
388: degree that the derived abundances of Si and heavier elements are 
389: affected? Only small changes in the heavy element abundances 
390: would be required to remove any correlations with N. 
391: 
392: In Table \ref{tab:err}, we show the abundances for a subset of 
393: lines derived using four different model atmospheres. 
394: The model atmospheres and subsequent abundances were computed using 
395: the MARCS suite of software \citep{marcs75,asplund97}. 
396: Models 1 and 2 have identical 
397: \teff, $\log g$, and [Fe/H]. However, Model 1 has a scaled-solar 
398: composition whereas Model 2 has 
399: [N/Fe] = +1.5, [O/Fe] = $-$0.1, [Na/Fe] = +0.6, [Mg/Fe] = +0.4, 
400: and [Al/Fe] = +1.2, values which are essentially identical to those measured
401: in the N-rich star NGC 6752-2. Similarly Models 3 and 4 have identical 
402: \teff, $\log g$, and [Fe/H]. However, Model 3 has a scaled-solar 
403: composition whereas Model 4 has 
404: [N/Fe] = $-$0.4, [O/Fe] = +0.6, [Na/Fe] = $-$0.1, [Mg/Fe] = +0.5, 
405: and [Al/Fe] = +0.6, values which are essentially identical to those measured 
406: in the N-poor star NGC 6752-15. For the purposes of these calculations,  
407: we assumed a strength of 10m\AA~to 
408: ensure that the lines are weak and on the linear part of the curve of growth.
409: A comparison between the abundances 
410: derived using Models 1 and 2 gauges the uncertainty 
411: when analyzing N-rich, Na-rich, Al-rich, O-poor, and Mg-poor stars. 
412: A comparison between the abundances 
413: derived using Models 3 and 4 gauges the uncertainty for N-poor, Na-poor, 
414: Al-poor, O-rich, and Mg-rich stars. And a comparison between the abundances 
415: derived using Models 2 and 4 gauges the uncertainty across the full 
416: abundance range. Given the small range in \teff~($\sim$ 250K), 
417: we assume that the calculations would be applicable to all stars in our
418: sample.  
419: 
420: While the elements considered in Table 2 were chosen somewhat arbitrarily, 
421: they are neutral and ionized species for which the correlations with N are 
422: significant (4-$\sigma$ for Y and 3-$\sigma$ for both Si and Zr) and 
423: for which any correlations, if real, would have considerable impact on our 
424: understanding of stellar and/or globular cluster evolution. For all
425: elements, the calculations indicate that adopting model atmospheres
426: with appropriate compositions would result in only small 
427: changes to the derived abundances. For the correlations that we 
428: are investigating, the N-rich stars tend to have higher abundances 
429: [X/Fe] than their N-poor counterparts. That is, the calculations indicate 
430: that the correlations between [X/Fe] and [N/Fe] 
431: would in fact be strengthened had we 
432: adopted model atmospheres with appropriate compositions. 
433: Therefore, we tentatively rule out the possibility that the correlations 
434: between N and heavier elements are due to using model atmospheres with 
435: scaled-solar compositions. Nevertheless, a more detailed analysis using model 
436: atmospheres with the appropriate, but presently unknown, abundances of He 
437: and C would be of great interest. 
438: 
439: As far as we are aware, correlations between the abundances of 
440: light elements (N and Al) and elements heavier than 
441: Si have not been noted in any other globular clusters. 
442: We attribute our detection primarily to the highly accurate and 
443: homogeneous stellar parameters derived from the \citet{grundahl99} 
444: Str{\" o}mgren photometry as well as to the high quality spectra 
445: and large sample size. 
446: 
447: \subsection{Implications for globular cluster and stellar evolution}
448: 
449: Having demonstrated that the Str{\" o}mgren $c_1$ index (or $cy$ index) 
450: directly traces 
451: the N abundance, we confirm and reiterate the claim by \citet{grundahl00}
452: that every globular cluster exhibits a considerable star-to-star 
453: variation in N abundances at all evolutionary stages as seen via 
454: the $c_1$ index. In Figure \ref{fig:cyall}
455: we plot $V$ versus $cy$ for 
456: a subset of the \citet{grundahl99} clusters. 
457: Every cluster displays a dispersion in $cy$ comparable to that 
458: seen in NGC 6752 indicating that all globular clusters exhibit a 
459: $\simeq$ 2.0 dex dispersion in [N/Fe], the value found in NGC 6752. 
460: This Figure clearly shows that 
461: all clusters exhibit a large dispersion in $cy$, i.e., 
462: N abundances, at all evolutionary phases. Note that these clusters 
463: cover a range in metallicities: 
464: [Fe/H] $\simeq$ $-$1.2 (NGC 288, NGC 362, M5),
465: [Fe/H] $\simeq$ $-$1.6 (M3, M13, NGC 6752), and 
466: [Fe/H] $\simeq$ $-$2.0 (NGC 6397, M15, M92). 
467: 
468: Pollution by ejecta from intermediate-mass AGB stars and/or 
469: massive stars is the currently favored explanation for the 
470: star-to-star abundance variation of light elements. 
471: Both scenarios are problematic, with issues relating to 
472: the initial mass function and the predicted nucleosynthesis yields 
473: \citep{fenner04,decressin06}. 
474: 
475: The handful of clusters for which N abundances have been measured in 
476: large samples of stars show a variety of N abundance distributions. 
477: M71 and 47 Tuc appear bimodal whereas M5, M13, and M15 do not appear bimodal 
478: (\citealt{cohen05b} and references therein). As discussed above, the 
479: [N/Fe] distribution for NGC 6752, using the relation between 
480: N and Str{\" o}mgren photometry, appears to increase towards increasing 
481: [N/Fe]. Such a distribution is similar to that seen in the 
482: comparably metal-poor cluster M13 (and possibly also in M5). 
483: Curiously, both NGC 6752 and M13 exhibit 
484: bimodal CN distributions, as do all clusters in which CN can be 
485: measured \citep{smith87}. 
486: The varying [N/Fe] distributions in clusters of different 
487: metallicities will provide important 
488: observational constraints upon the origin of the abundance anomalies 
489: in future globular cluster chemical evolution studies 
490: similar to that conducted by \citet{fenner04}. 
491: 
492: The correlations between N and heavy elements presented in this paper 
493: offer support for contributions from 
494: both AGB and massive stars to the observed abundance anomalies. 
495: Although quantitative yields of 
496: $s$-process elements from intermediate-mass metal-poor 
497: AGB stars are rare, 
498: $s$-process 
499: elements are expected to be synthesized, to some degree, in these stars 
500: \citep{busso01}. 
501: Fe-peak and $\alpha$ elements 
502: are synthesized by massive stars, at all metallicities \citep{chieffi04}. 
503: Given that not every element is 
504: correlated with N at a statistically significant level, and that the 
505: amplitude of the heavy element abundance variation is small, our 
506: results do not offer definitive support for either the 
507: massive star or AGB pollution scenarios. 
508: 
509: As mentioned, target selection via the Str{\" o}mgren $c_1$ index 
510: has allowed us to sample the full range of the 
511: star-to-star abundance variations, at this luminosity. Of great interest 
512: would be a further study of the full range of the abundance variations 
513: at different luminosities. An analysis of the most N-poor and N-rich stars, 
514: as selected via the $c_1$ index, would truly gauge any variation in the 
515: abundance anomalies as a function of evolutionary status. That the 
516: abundance variations for O-Al have been found in unevolved stars suggests 
517: that primordial pollution processes rather than internal evolutionary 
518: processes must be the dominant mechanism, but these 
519: observations do not entirely preclude 
520: an internal evolutionary mechanism being a second-order effect. 
521: Indeed, the brightest stars in M13 appear to exhibit a 
522: systematic abundance variation for O, Na, Mg, and Al 
523: with luminosity \citep{sneden04a}. 
524: Such trends,
525: if observed in additional clusters, 
526: would present new challenges for stellar evolution 
527: and stellar nucleosynthesis. 
528: 
529: The \citet{grundahl99} 
530: Str{\" o}mgren photometry extends to additional globular clusters. 
531: It would be of great interest to obtain high-resolution, high S/N spectra 
532: for a large sample of stars within a narrow luminosity range 
533: in other globular clusters in order to 
534: see whether the correlations between light and heavy elements are also 
535: present. The key is to again use carefully calibrated 
536: Str{\" o}mgren photometry to obtain homogeneous stellar parameters, 
537: such as in \citet{grundahl02}. 
538: 
539: Star-to-star helium abundance variations are expected when the 
540: C-Al abundance anomalies are synthesized during hydrogen burning. 
541: NGC 2808 \citep{dantona04} and 
542: $\omega$ Cen \citep{norris04} show evidence for He variations, based 
543: primarily on the interpretation of color-magnitude diagrams. 
544: Stellar models and isochrones computed by \citet{salaris06} 
545: show that the location of the RGB bump (weakly) depends upon the 
546: He abundance. Recently, \citet{carretta07} found new evidence for 
547: He variations in NGC 6218 (M12) based on RGB stars. 
548: They looked at the luminosity functions 
549: for a sample of Na-rich and a sample of Na-poor stars and found that 
550: the RGB bump differed by 0.05 mag, corresponding to a difference 
551: in $Y$ of 0.05 between the Na-poor and Na-rich populations. 
552: Although the correlation is not significant, we find 
553: marginal evidence for an increase in [Fe/H] with increasing [N/Fe]. 
554: Such a correlation may be the signature of possible He variations. 
555: 
556: Finally, one of the most stringent observational 
557: constraints upon the origin of the abundance anomalies comes from 
558: C+N+O, which is constant to within a factor of 2 \citep{smith05}. 
559: Predicted yields 
560: from AGB models are unable to satisfy this constraint 
561: \citep{lattanzio06,karakas06}. 
562: The ideal check of the constancy of C+N+O 
563: would be to measure C from CH, N from NH, and O from [OI] using 
564: high-resolution spectra for a sample that covers the full 
565: range of the abundance variations. We have measured 2 of these 3 indicators 
566: but unfortunately our spectra do not cover the 4300\AA~CH 
567: molecular lines. Measurement of C from CH for a subset of 
568: the most N-rich and N-poor stars in this sample is highly desired 
569: to examine the constancy of C+N+O and to search for correlations 
570: between C+N+O versus N.  
571: 
572: \section{Concluding remarks}
573: \label{sec:summary}
574: 
575: In this paper we present measurements of N abundances in 21 bright 
576: giant stars near the RGB bump of the globular cluster NGC 6752. 
577: The sample was chosen to span the full range of the Str{\" o}mgren $c_1$ 
578: index at this luminosity. 
579: The amplitude of the N abundance variation at the sample's luminosity 
580: is 1.95 dex. We confirm that the N abundances are 
581: correlated with the $c_1$ index. We find a linear relation between 
582: the N abundances and a new index $cy$ = $c_1$ $- (b-y)$. We apply 
583: this new relation to large numbers of 
584: RGB stars and find that the [N/Fe] distribution 
585: increases towards higher values. Broader implications are that 
586: all globular clusters show large star-to-star variations in 
587: their $cy$ indices (i.e., N abundances) at all evolutionary stages. 
588: 
589: We find that the N abundances 
590: are correlated with the light elements O, Na, Mg, and Al, a feature 
591: that is seen in every well studied globular cluster. However, we find 
592: for the first time that the N abundances are also correlated with 
593: the abundances of Si and heavier elements. While such correlations are 
594: statistically significant, the amplitude of the variation is small
595: ($<$ 0.2 dex variation in [X/Fe] as [N/Fe] varies by 2 dex). We 
596: attribute the detection to the large sample size, high quality data, 
597: and most importantly to the homogeneous and precise stellar 
598: parameters obtained from the \citet{grundahl99} Str{\" o}mgren photometry. 
599: 
600: Analysis using model atmospheres with the appropriate N, O, Na, and Al 
601: abundances gives very similar abundances to those 
602: derived using model atmospheres with scaled-solar compositions. 
603: Therefore, the correlations between N and heavier elements are unlikely 
604: to be due to using models with scaled-solar compositions. In fact, for 
605: the subset of elements investigated (Si, Y, and Zr), the correlations with N 
606: would be even more significant had we used models with appropriate 
607: compositions. 
608: 
609: Of great interest would be to search for such correlations in an 
610: additional sample of stars at a different luminosity within this 
611: cluster and in other clusters. If correlations 
612: between light and heavy elements are again 
613: identified, they would offer support to both the AGB and 
614: massive star pollution scenarios for explaining the star-to-star 
615: abundance variation of light elements. 
616: 
617: \acknowledgments
618: 
619: This research has made use of the SIMBAD database,
620: operated at CDS, Strasbourg, France and
621: NASA's Astrophysics Data System. DY thanks Mike Bessell, 
622: Amanda Karakas, John Norris for helpful discussions. We thank the 
623: anonymous referee for helpful comments. This research was 
624: supported in part by NASA through the American Astronomical Society's Small 
625: Research Grant Program. 
626: 
627: \begin{thebibliography}{58}
628: \expandafter\ifx\csname natexlab\endcsname\relax\def\natexlab#1{#1}\fi
629: 
630: \bibitem[{{Alonso} {et~al.}(1999){Alonso}, {Arribas}, \&
631:   {Mart\'{\i}nez-Roger}}]{alonso99b}
632: {Alonso}, A., {Arribas}, S., \& {Mart\'{\i}nez-Roger}, C. 1999, \aaps, 140, 261
633: 
634: \bibitem[{{Asplund} {et~al.}(1997){Asplund}, {Gustafsson}, {Kiselman}, \&
635:   {Eriksson}}]{asplund97}
636: {Asplund}, M., {Gustafsson}, B., {Kiselman}, D., \& {Eriksson}, K. 1997, \aap,
637:   318, 521
638: 
639: \bibitem[{{Briley} \& {Cohen}(2001)}]{briley01}
640: {Briley}, M.~M. \& {Cohen}, J.~G. 2001, \aj, 122, 242
641: 
642: \bibitem[{{Briley} {et~al.}(2002){Briley}, {Cohen}, \& {Stetson}}]{briley02}
643: {Briley}, M.~M., {Cohen}, J.~G., \& {Stetson}, P.~B. 2002, \apjl, 579, L17
644: 
645: \bibitem[{{Briley} {et~al.}(2004){Briley}, {Harbeck}, {Smith}, \&
646:   {Grebel}}]{briley04}
647: {Briley}, M.~M., {Harbeck}, D., {Smith}, G.~H., \& {Grebel}, E.~K. 2004, \aj,
648:   127, 1588
649: 
650: \bibitem[{{Briley} \& {Smith}(1993)}]{bs93}
651: {Briley}, M.~M. \& {Smith}, G.~H. 1993, \pasp, 105, 1260
652: 
653: \bibitem[{{Briley} {et~al.}(1996){Briley}, {Smith}, {Suntzeff}, {Lambert},
654:   {Bell}, \& {Hesser}}]{briley96}
655: {Briley}, M.~M., {Smith}, V.~V., {Suntzeff}, N.~B., {Lambert}, D.~L., {Bell},
656:   R.~A., \& {Hesser}, J.~E. 1996, \nat, 383, 604
657: 
658: \bibitem[{{Buonanno} {et~al.}(1986){Buonanno}, {Caloi}, {Castellani}, {Corsi},
659:   {Fusi Pecci}, \& {Gratton}}]{buonanno86}
660: {Buonanno}, R., {Caloi}, V., {Castellani}, V., {Corsi}, C., {Fusi Pecci}, F.,
661:   \& {Gratton}, R. 1986, \aaps, 66, 79
662: 
663: \bibitem[{{Busso} {et~al.}(2001){Busso}, {Gallino}, {Lambert}, {Travaglio}, \&
664:   {Smith}}]{busso01}
665: {Busso}, M., {Gallino}, R., {Lambert}, D.~L., {Travaglio}, C., \& {Smith},
666:   V.~V. 2001, \apj, 557, 802
667: 
668: \bibitem[{{Carretta} {et~al.}(2007){Carretta}, {Bragaglia}, {Gratton},
669:   {Catanzaro}, {Leone}, {Sabbi}, {Cassisi}, {Claudi}, {D'Antona}, {Fran{\c
670:   c}ois}, {James}, \& {Piotto}}]{carretta07}
671: {Carretta}, E., {Bragaglia}, A., {Gratton}, R.~G., {Catanzaro}, G., {Leone},
672:   F., {Sabbi}, E., {Cassisi}, S., {Claudi}, R., {D'Antona}, F., {Fran{\c
673:   c}ois}, P., {James}, G., \& {Piotto}, G. 2007, \aap, 464, 939
674: 
675: \bibitem[{{Carretta} {et~al.}(2005){Carretta}, {Gratton}, {Lucatello},
676:   {Bragaglia}, \& {Bonifacio}}]{carretta05}
677: {Carretta}, E., {Gratton}, R.~G., {Lucatello}, S., {Bragaglia}, A., \&
678:   {Bonifacio}, P. 2005, \aap, 433, 597
679: 
680: \bibitem[{{Chieffi} \& {Limongi}(2004)}]{chieffi04}
681: {Chieffi}, A. \& {Limongi}, M. 2004, \apj, 608, 405
682: 
683: \bibitem[{{Cohen} {et~al.}(2002){Cohen}, {Briley}, \& {Stetson}}]{cohen02}
684: {Cohen}, J.~G., {Briley}, M.~M., \& {Stetson}, P.~B. 2002, \aj, 123, 2525
685: 
686: \bibitem[{{Cohen} {et~al.}(2005){Cohen}, {Briley}, \& {Stetson}}]{cohen05b}
687: ---. 2005, \aj, 130, 1177
688: 
689: \bibitem[{{Cohen} \& {Mel{\' e}ndez}(2005)}]{cohen05}
690: {Cohen}, J.~G. \& {Mel{\' e}ndez}, J. 2005, \aj, 129, 303
691: 
692: \bibitem[{{Cottrell} \& {Da Costa}(1981)}]{cottrell81}
693: {Cottrell}, P.~L. \& {Da Costa}, G.~S. 1981, \apjl, 245, L79
694: 
695: \bibitem[{{Da Costa} \& {Cottrell}(1980)}]{dacosta80}
696: {Da Costa}, G.~S. \& {Cottrell}, P.~L. 1980, \apjl, 236, L83
697: 
698: \bibitem[{{D'Antona} \& {Caloi}(2004)}]{dantona04}
699: {D'Antona}, F. \& {Caloi}, V. 2004, \apj, 611, 871
700: 
701: \bibitem[{{Decressin} {et~al.}(2007){Decressin}, {Meynet}, {Charbonnel},
702:   {Prantzos}, \& {Ekstr{\"o}m}}]{decressin06}
703: {Decressin}, T., {Meynet}, G., {Charbonnel}, C., {Prantzos}, N., \&
704:   {Ekstr{\"o}m}, S. 2007, \aap, 464, 1029
705: 
706: \bibitem[{{Denissenkov} {et~al.}(1998){Denissenkov}, {Da Costa}, {Norris}, \&
707:   {Weiss}}]{denissenkov98}
708: {Denissenkov}, P.~A., {Da Costa}, G.~S., {Norris}, J.~E., \& {Weiss}, A. 1998,
709:   \aap, 333, 926
710: 
711: \bibitem[{{D'Odorico} {et~al.}(2000){D'Odorico}, {Cristiani}, {Dekker}, {Hill},
712:   {Kaufer}, {Kim}, \& {Primas}}]{uves}
713: {D'Odorico}, S., {Cristiani}, S., {Dekker}, H., {Hill}, V., {Kaufer}, A.,
714:   {Kim}, T., \& {Primas}, F. 2000, in Proc. SPIE Vol. 4005, p. 121-130,
715:   Discoveries and Research Prospects from 8- to 10-Meter-Class Telescopes,
716:   Jacqueline Bergeron; Ed., 121--130
717: 
718: \bibitem[{{Fenner} {et~al.}(2004){Fenner}, {Campbell}, {Karakas}, {Lattanzio},
719:   \& {Gibson}}]{fenner04}
720: {Fenner}, Y., {Campbell}, S., {Karakas}, A.~I., {Lattanzio}, J.~C., \&
721:   {Gibson}, B.~K. 2004, \mnras, 353, 789
722: 
723: \bibitem[{{Fulbright} {et~al.}(2007){Fulbright}, {McWilliam}, \&
724:   {Rich}}]{fulbright07}
725: {Fulbright}, J.~P., {McWilliam}, A., \& {Rich}, R.~M. 2007, \apj, 661, 1152
726: 
727: \bibitem[{{Gratton} {et~al.}(2004){Gratton}, {Sneden}, \&
728:   {Carretta}}]{gratton04}
729: {Gratton}, R., {Sneden}, C., \& {Carretta}, E. 2004, \araa, 42, 385
730: 
731: \bibitem[{{Gratton} {et~al.}(2001){Gratton}, {Bonifacio}, {Bragaglia},
732:   {Carretta}, {Castellani}, {Centurion}, {Chieffi}, {Claudi}, {Clementini},
733:   {D'Antona}, {Desidera}, {Fran{\c c}ois}, {Grundahl}, {Lucatello}, {Molaro},
734:   {Pasquini}, {Sneden}, {Spite}, \& {Straniero}}]{gratton01}
735: {Gratton}, R.~G., {Bonifacio}, P., {Bragaglia}, A., {Carretta}, E.,
736:   {Castellani}, V., {Centurion}, M., {Chieffi}, A., {Claudi}, R., {Clementini},
737:   G., {D'Antona}, F., {Desidera}, S., {Fran{\c c}ois}, P., {Grundahl}, F.,
738:   {Lucatello}, S., {Molaro}, P., {Pasquini}, L., {Sneden}, C., {Spite}, F., \&
739:   {Straniero}, O. 2001, \aap, 369, 87
740: 
741: \bibitem[{{Gratton} {et~al.}(2000){Gratton}, {Sneden}, {Carretta}, \&
742:   {Bragaglia}}]{gratton00b}
743: {Gratton}, R.~G., {Sneden}, C., {Carretta}, E., \& {Bragaglia}, A. 2000, \aap,
744:   354, 169
745: 
746: \bibitem[{{Grevesse} {et~al.}(2007){Grevesse}, {Asplund}, \&
747:   {Sauval}}]{grevesse07}
748: {Grevesse}, N., {Asplund}, M., \& {Sauval}, A.~J. 2007, Space Science Reviews,
749:   130, 105
750: 
751: \bibitem[{{Grevesse} \& {Sauval}(1998)}]{grevesse98}
752: {Grevesse}, N. \& {Sauval}, A.~J. 1998, Space Science Reviews, 85, 161
753: 
754: \bibitem[{{Grundahl} {et~al.}(2002){Grundahl}, {Briley}, {Nissen}, \&
755:   {Feltzing}}]{grundahl02}
756: {Grundahl}, F., {Briley}, M., {Nissen}, P.~E., \& {Feltzing}, S. 2002, \aap,
757:   385, L14
758: 
759: \bibitem[{{Grundahl} {et~al.}(1999){Grundahl}, {Catelan}, {Landsman},
760:   {Stetson}, \& {Andersen}}]{grundahl99}
761: {Grundahl}, F., {Catelan}, M., {Landsman}, W.~B., {Stetson}, P.~B., \&
762:   {Andersen}, M.~I. 1999, \apj, 524, 242
763: 
764: \bibitem[{{Grundahl} {et~al.}(2000){Grundahl}, {Vandenberg}, {Stetson},
765:   {Andersen}, \& {Briley}}]{grundahl00}
766: {Grundahl}, F., {Vandenberg}, D.~A., {Stetson}, P.~B., {Andersen}, M.~I., \&
767:   {Briley}, M. 2000, in Liege International Astrophysical Colloquia, Vol.~35,
768:   Liege International Astrophysical Colloquia, ed. A.~{Noels}, P.~{Magain},
769:   D.~{Caro}, E.~{Jehin}, G.~{Parmentier}, \& A.~A. {Thoul}, 503
770: 
771: \bibitem[{{Gustafsson} {et~al.}(1975){Gustafsson}, {Bell}, {Eriksson}, \&
772:   {Nordlund}}]{marcs75}
773: {Gustafsson}, B., {Bell}, R.~A., {Eriksson}, K., \& {Nordlund}, A. 1975, \aap,
774:   42, 407
775: 
776: \bibitem[{{Johnson} {et~al.}(2007){Johnson}, {Herwig}, {Beers}, \&
777:   {Christlieb}}]{johnson07}
778: {Johnson}, J.~A., {Herwig}, F., {Beers}, T.~C., \& {Christlieb}, N. 2007, \apj,
779:   658, 1203
780: 
781: \bibitem[{{Johnson} {et~al.}(2006){Johnson}, {Ivans}, \& {Stetson}}]{johnson06}
782: {Johnson}, J.~A., {Ivans}, I.~I., \& {Stetson}, P.~B. 2006, \apj, 640, 801
783: 
784: \bibitem[{{Karakas} {et~al.}(2006){Karakas}, {Fenner}, {Sills}, {Campbell}, \&
785:   {Lattanzio}}]{karakas06}
786: {Karakas}, A.~I., {Fenner}, Y., {Sills}, A., {Campbell}, S.~W., \& {Lattanzio},
787:   J.~C. 2006, \apj, 652, 1240
788: 
789: \bibitem[{{Karakas} \& {Lattanzio}(2003)}]{karakas03}
790: {Karakas}, A.~I. \& {Lattanzio}, J.~C. 2003, Publ.\ Astron.\ Soc.\ Australia,
791:   20, 279
792: 
793: \bibitem[{{Kraft}(1994)}]{kraft94}
794: {Kraft}, R.~P. 1994, \pasp, 106, 553
795: 
796: \bibitem[{{Kurucz}(1993)}]{kurucz93}
797: {Kurucz}, R. 1993, ATLAS9 Stellar Atmosphere Programs and 2 km/s grid.~Kurucz
798:   CD-ROM No.~13.~ Cambridge, Mass.: Smithsonian Astrophysical Observatory,
799:   1993., 13
800: 
801: \bibitem[{{Langer} \& {Hoffman}(1995)}]{langer95}
802: {Langer}, G.~E. \& {Hoffman}, R.~D. 1995, \pasp, 107, 1177
803: 
804: \bibitem[{{Lattanzio} \& {Tout}(2006)}]{lattanzio06}
805: {Lattanzio}, J.~C. \& {Tout}, C.~A. 2006, in EAS Publications Series, ed.
806:   T.~{Montmerle} \& C.~{Kahane}, 189--197
807: 
808: \bibitem[{{Letarte} {et~al.}(2006){Letarte}, {Hill}, {Jablonka}, {Tolstoy},
809:   {Fran{\c c}ois}, \& {Meylan}}]{letarte06}
810: {Letarte}, B., {Hill}, V., {Jablonka}, P., {Tolstoy}, E., {Fran{\c c}ois}, P.,
811:   \& {Meylan}, G. 2006, \aap, 453, 547
812: 
813: \bibitem[{{Norris} {et~al.}(1981){Norris}, {Cottrell}, {Freeman}, \& {Da
814:   Costa}}]{norris81}
815: {Norris}, J., {Cottrell}, P.~L., {Freeman}, K.~C., \& {Da Costa}, G.~S. 1981,
816:   \apj, 244, 205
817: 
818: \bibitem[{{Norris} \& {Freeman}(1979)}]{nf79}
819: {Norris}, J. \& {Freeman}, K.~C. 1979, \apjl, 230, L179
820: 
821: \bibitem[{{Norris}(2004)}]{norris04}
822: {Norris}, J.~E. 2004, \apjl, 612, L25
823: 
824: \bibitem[{{Osborn}(1971)}]{osborn71}
825: {Osborn}, W. 1971, The Observatory, 91, 223
826: 
827: \bibitem[{{Popper}(1947)}]{popper47}
828: {Popper}, D.~M. 1947, \apj, 105, 204
829: 
830: \bibitem[{{Prantzos} \& {Charbonnel}(2006)}]{prantzos06}
831: {Prantzos}, N. \& {Charbonnel}, C. 2006, \aap, 458, 135
832: 
833: \bibitem[{{Salaris} {et~al.}(2006){Salaris}, {Weiss}, {Ferguson}, \&
834:   {Fusilier}}]{salaris06}
835: {Salaris}, M., {Weiss}, A., {Ferguson}, J.~W., \& {Fusilier}, D.~J. 2006, \apj,
836:   645, 1131
837: 
838: \bibitem[{{Smith}(1987)}]{smith87}
839: {Smith}, G.~H. 1987, \pasp, 99, 67
840: 
841: \bibitem[{{Smith}(2006)}]{smith06}
842: ---. 2006, \pasp, 118, 1225
843: 
844: \bibitem[{{Smith} {et~al.}(2005){Smith}, {Cunha}, {Ivans}, {Lattanzio},
845:   {Campbell}, \& {Hinkle}}]{smith05}
846: {Smith}, V.~V., {Cunha}, K., {Ivans}, I.~I., {Lattanzio}, J.~C., {Campbell},
847:   S., \& {Hinkle}, K.~H. 2005, \apj, 633, 392
848: 
849: \bibitem[{{Sneden}(1973)}]{moog}
850: {Sneden}, C. 1973, \apj, 184, 839
851: 
852: \bibitem[{{Sneden} {et~al.}(2004){Sneden}, {Kraft}, {Guhathakurta}, {Peterson},
853:   \& {Fulbright}}]{sneden04a}
854: {Sneden}, C., {Kraft}, R.~P., {Guhathakurta}, P., {Peterson}, R.~C., \&
855:   {Fulbright}, J.~P. 2004, \aj, 127, 2162
856: 
857: \bibitem[{{Suntzeff} \& {Smith}(1991)}]{ss91}
858: {Suntzeff}, N.~B. \& {Smith}, V.~V. 1991, \apj, 381, 160
859: 
860: \bibitem[{{Ventura} \& {D'Antona}(2005)}]{ventura05c}
861: {Ventura}, P. \& {D'Antona}, F. 2005, \apjl, 635, L149
862: 
863: \bibitem[{{Yong} {et~al.}(2003){Yong}, {Grundahl}, {Lambert}, {Nissen}, \&
864:   {Shetrone}}]{6752}
865: {Yong}, D., {Grundahl}, F., {Lambert}, D.~L., {Nissen}, P.~E., \& {Shetrone},
866:   M.~D. 2003, \aap, 402, 985
867: 
868: \bibitem[{{Yong} {et~al.}(2005){Yong}, {Grundahl}, {Nissen}, {Jensen}, \&
869:   {Lambert}}]{67522}
870: {Yong}, D., {Grundahl}, F., {Nissen}, P.~E., {Jensen}, H.~R., \& {Lambert},
871:   D.~L. 2005, \aap, 438, 875
872: 
873: \bibitem[{{Zinn}(1977)}]{zinn77}
874: {Zinn}, R. 1977, \apj, 218, 96
875: 
876: \end{thebibliography}
877: 
878: 
879: \begin{figure}
880: \epsscale{0.8}
881: \plotone{fig1.ps}
882: \caption{The $V$ vs.\ $c_1$ (left) and $V$ vs.\ $(v-y)$ (right) 
883: color-magnitude diagrams using 
884: the \citet{grundahl99} photometry. The plus signs indicate the locations 
885: of our targets. (The large open circle marks NGC 6752-7.) \label{fig:cmd1}}
886: \end{figure}
887: 
888: \clearpage
889: 
890: \begin{figure}
891: \epsscale{0.8}
892: \plotone{fig2.ps}
893: \caption{Observed spectra (crosses) for NGC 6752-24 (upper panel) and 
894: NGC 6752-23 (lower panel) near the 3360\AA~NH band. 
895: These stars have very similar stellar parameters. 
896: Synthetic spectra with different N abundances are shown. 
897: The thick black line represents the best fit and unsatisfactory fits 
898: ($\pm$ 0.5 dex) are shown as thin blue and thin red lines.\label{fig:1718fit}}
899: \end{figure}
900: 
901: \clearpage
902: 
903: \begin{figure}
904: \epsscale{0.8}
905: \plotone{fig3.ps}
906: \caption{Same as Figure \ref{fig:1718fit} but for the slightly 
907: cooler pair of stars 
908: NGC 6752-3 (upper) and NGC 6752-2 (lower).\label{fig:23fit}}
909: \end{figure}
910: 
911: \clearpage
912: 
913: \begin{figure}
914: \epsscale{0.8}
915: \plotone{fig4.ps}
916: \caption{[N/Fe] vs.\ $c_1$. Star NGC 6752-7 is not included 
917: in this plot due to its deviating [Fe/H]. \label{fig:nc1}}
918: \end{figure}
919: 
920: \clearpage
921: 
922: \begin{figure}
923: \epsscale{0.8}
924: \plotone{fig5.ps}
925: \caption{[Na/Fe] vs.\ [N/Fe] for this study (closed circles) 
926: and the \citet{carretta05} sample (crosses). 
927: \label{fig:nan}}
928: \end{figure}
929: 
930: \clearpage
931: 
932: \begin{figure}
933: \epsscale{0.8}
934: \plotone{fig6.ps}
935: \caption{$V$ vs.\ $cy$ color-magnitude diagram using 
936: the \citet{grundahl99} photometry. The plus signs indicate the 
937: locations of our targets. (The large open circle marks NGC 6752-7.)
938: \label{fig:cmd2}}
939: \end{figure}
940: 
941: \clearpage
942: 
943: \begin{figure}
944: \epsscale{0.8}
945: \plotone{fig7.ps}
946: \caption{[N/Fe] vs.\ $cy$. The straight line is the best fit to the data. 
947: (The large open circle marks NGC 6752-7.)
948: \label{fig:ncy}}
949: \end{figure}
950: 
951: \clearpage
952: 
953: \begin{figure}
954: \epsscale{0.8}
955: \plotone{fig8.ps}
956: \caption{The left panel shows the 
957: [N/Fe] distribution using the relation between [N/Fe] and $cy$. The 
958: larger histogram shows all stars with 13 $\le$ $V$ $\le$ 16 while the 
959: smaller histogram 
960: shows only stars spanned by the calibrating stars. Both sets of 
961: stars are identified in the $V$ vs.\ $cy$ color-magnitude diagram in the 
962: right panel. \label{fig:ndist}}
963: \end{figure}
964: 
965: \clearpage
966: 
967: \begin{figure}
968: \epsscale{0.8}
969: \plotone{fig9.ps}
970: \caption{[X/Fe] vs.\ [N/Fe] for O, Na, Mg, and Al. The error bar shows 
971: the 1-$\sigma$ errors (see text for details). 
972: The dotted line is the mean abundance and the 
973: dashed line is the linear least squares fit to the data 
974: (slope and associated error are included). 
975: \label{fig:ona}}
976: \end{figure}
977: 
978: \clearpage
979: 
980: \begin{figure}
981: \epsscale{0.8}
982: \plotone{fig10.ps}
983: \caption{Same as Figure \ref{fig:ona} but for Si, Ca, Sc, and Ti.
984: \label{fig:siti}}
985: \end{figure}
986: 
987: \clearpage
988: 
989: \begin{figure}
990: \epsscale{0.8}
991: \plotone{fig11.ps}
992: \caption{Same as Figure \ref{fig:ona} but for V, Mn, Fe, and Co.
993: \label{fig:vca}}
994: \end{figure}
995: 
996: \clearpage
997: 
998: \begin{figure}
999: \epsscale{0.8}
1000: \plotone{fig12.ps}
1001: \caption{Same as Figure \ref{fig:ona} but for Ni, Cu, Y, and Zr.
1002: \label{fig:nizr}}
1003: \end{figure}
1004: 
1005: \clearpage
1006: 
1007: \begin{figure}
1008: \epsscale{0.8}
1009: \plotone{fig13.ps}
1010: \caption{Same as Figure \ref{fig:ona} but for Ba, Ce, Nd, and Eu.
1011: \label{fig:baeu}}
1012: \end{figure}
1013: 
1014: \clearpage
1015: 
1016: \begin{figure}
1017: \epsscale{0.8}
1018: \plotone{fig14.ps}
1019: \caption{The $V$ vs.\ $cy$ color-magnitude diagrams for a 
1020: sample of clusters taken from \citet{grundahl99}. All clusters 
1021: exhibit a large range in $cy$ at all evolutionary stages. 
1022: \label{fig:cyall}}
1023: \end{figure}
1024: 
1025: \clearpage
1026: 
1027: \input{tab1}
1028: \input{tab2}
1029: 
1030: \end{document}
1031: 
1032: 
1033: