1: \documentclass[12pt,preprint]{aastex}
2:
3: \received{}
4: \revised{}
5: \accepted{}
6: \ccc{}
7: \cpright{}{}
8:
9: \slugcomment{}
10: \shorttitle{M13 Subgiants}
11: \shortauthors{Briley, Cohen, \& Stetson}
12:
13: \newcommand{\subsun}{\mbox{$_{\odot}$}}
14: \newcommand{\teff}{$T_{eff}$}
15: \newcommand{\grav}{log($g$)}
16: \newcommand{\kms}{km~s$^{-1}$}
17: \newcommand{\fe}{[Fe/H]}
18: \newcommand{\cabund}{[C/Fe]}
19: \newcommand{\ciso}{$^{12}$C/$^{13}$C}
20:
21:
22: \begin{document}
23:
24: \title{Carbon Abundances of Faint Stars in M13 - Evidence for
25: Two Abundance Altering Mechanisms
26: \altaffilmark{1}}
27:
28: \author{Michael M. Briley}
29: \affil{Department of Physics and Astronomy, University of Wisconsin
30: Oshkosh, 800 Algoma Boulevard, Oshkosh, WI 54901}
31:
32: \author{Judith G. Cohen}
33: \affil{Palomar Observatory, MS 105-24, California Institute of Technology,
34: Pasadena, CA 91125}
35:
36: \and
37:
38: \author{Peter B. Stetson}
39: \affil{Dominion Astrophysical Observatory, 5071 West Saanich Road,
40: Victoria, BC V9E 2E7, Canada}
41:
42: \altaffiltext{1}{Based on observations obtained at the
43: W.M. Keck Observatory, which is operated jointly by the California
44: Institute of Technology, the University of California, and the
45: National Aeronautics and Space Administration.}
46:
47: \begin{abstract}
48: We present an analysis of CH band strengths in Keck LRIS spectra of a
49: sample of 81 stars in M13 within 2 magnitudes of the main-sequence turnoff.
50: The subgiants clearly exhibit a substantial (a factor of $\sim$6) spread in \cabund.
51: Moreover, the bulk of the subgiants possess C abundances larger than those
52: found among their more luminous counterparts.
53: The turnoff stars themselves are too warm for appreciable CH formation, but the
54: relatively small range in the observed CH band strength for stars just below
55: the turnoff nevertheless translates into this same spread in \cabund.
56: Still fainter, the sample size is small, but the same range in \cabund\ appears to
57: be present.
58: On the basis of these observations we suggest that a process external to
59: the present stars has resulted in a substantial star-to-star dispersion in
60: \cabund\ (and possibly other light elements) among all stars in M13.
61: In addition, the surface C abundances among the more
62: luminous stars have been further modified by the operation of an internal
63: deep-mixing mechanism during red giant branch ascent.
64: The amplitude of the scatter we find in \cabund\ at all luminosities may
65: prove difficult to explain via accretion from intermediate mass
66: AGB stars as the external ``polluting'' mechanism.
67:
68: \end{abstract}
69:
70: \keywords{globular clusters: general ---
71: globular clusters: individual (M13) --- stars: evolution -- stars:abundances}
72:
73:
74: \section{INTRODUCTION}
75:
76: A detailed history of the thirty years of progress in our understanding
77: and interpretation of the compositions of globular cluster stars is far too long and
78: complex to offer in the present work.
79: Hence we present below only the shortest of summaries.
80: For a more proper treatment, we recommend to the reader the more comprehensive
81: reviews of \citet{Kraft94} and \citet{DaCosta98}.
82:
83: Contrary to the canonical picture of globular clusters as chemically homogeneous
84: populations of stars, significant star-to-star differences in the abundances of C,
85: N, and often other light elements associated with proton capture reactions, have
86: been found in every Galactic globular cluster studied to date.
87: Locked within these inhomogeneities appears to be at least
88: a partial history of processes at work in the early cluster environment.
89: However, these patterns are also likely subject to further modification
90: driven by a less than well-understood mixing mechanism at work within the
91: stars themselves.
92: With this in mind, intra-cluster abundance inhomogeneities are usually
93: discussed in terms of these two distinct origins: internal
94: post-first dredge-up ``deep mixing'' or a process external to the present
95: day cluster stars (e.g., the accretion of ejecta from more massive stars).
96:
97: In the case of deep mixing, material exposed to the CN (and possibly ON) cycle
98: regions above the H-burning shell of an evolving low-mass star is
99: circulated into the outer convective envelope during red giant branch (RGB) ascent
100: resulting in increased surface N abundances at the expense of C (and possibly O).
101: Observational evidence that such a process operates can be found
102: among low-mass metal-poor ($-2 \le$ \fe $\le -1$) field giants where
103: \citet{Gratton00} note lowered
104: C abundances, enhanced N, depleted Li, and lower \ciso ~ ratios than
105: expected from first dredge-up \citep*[see also][]{Charbonnel98,Keller01}.
106: Mixing may be even more efficient in the metal-poor ($-2 \approx$ \fe) globulars where
107: substantial decreases in \cabund ~ ($\approx$ 1 dex) with increasing luminosity
108: have been observed along the RGBs of M92 \citep{Bellman01}, M15
109: \citep{Trefzger83}, and NGC 6397 \citep{Briley90}.
110: Mixing has also been suggested as an explanation for the detections
111: of O underabundances and O-Na, Al-Mg anticorrelations among the bright
112: giants of numerous globular clusters, e.g. M5, M3, M10, M13
113: \citep[see the composite figure for O-Na relations assembled by][their Fig. 12,
114: and references therein]{Ramirez02} - trends notably absent from the field
115: population \citep{Gratton00}.
116: If proton capture reactions on Ne and Mg resulting in Na and Al are
117: also operating at ON-cycle temperatures \citep[][and references
118: therein]{Cavallo00} one could then expect the
119: abundance patterns observed in most cluster giants to be at least qualitatively
120: reproduced under the conditions of sufficiently deep mixing.
121:
122: But the underlying physical process connecting the shell region with the
123: envelope beyond first dredge-up is not included in standard models of low
124: mass stellar evolution and remains poorly understood.
125: Some possible mechanisms explored include meridional circulation
126: \citep{Sweigart79}, various treatments of diffusion \citep[e.g.][]{Charbonnel95,
127: Denissenkov00}, and shell flashes \citep*{Aikawa01} - see the
128: review of the theory of red giant branch stars
129: by \citet*{Salaris02} for a more through discussion.
130:
131: As early as 24 years ago it was known that the subgiant branch (SGB) stars and
132: likely the main-sequence (MS) stars of the metal-rich (\fe $= -0.8$)
133: cluster 47 Tuc exhibit C and
134: N differences similar to those found on its RGB \citep{Hesser78}.
135: These results have been confirmed by \citet[and references therein]{Cannon98},
136: and significant C/N inhomogeneities among SGB and MS stars have since been observed
137: in NGC 6752 and M5 \citep*{Suntzeff91,Cohen02},
138: and M71 \citep{Cohen99a}.
139: Variations in other light elements have also been noted among the SGB/MS stars
140: of these very clusters: Na-CN correlated on the MS of 47 Tuc \citep{Briley96},
141: O-Na and Mg-Al anticorrelated among the MS stars of NGC 6752
142: \citep{Gratton01}, and O-Na anticorrelated in M71 SGB stars \citep{Ramirez02}.
143: Most recently, \citet{Ramirez03} have shown a significant dispersion in Na extending
144: from the M5 RGB to its MS turnoff.
145: These inhomogeneities are of course difficult to explain within the framework
146: of mixing theories and imply that some component of the
147: abundance differences is in place well before appreciable RGB ascent,
148: possibly embedded in the present-day stars much earlier in the cluster
149: history.
150: Indeed, the nuclei observed to vary among the SGB/MS stars, i.e., C, N, O, Na,
151: Al, and Mg, suggest as a source the hot bottom burning regions of intermediate mass
152: asymptotic giant branch (AGB) stars \citep{Ventura01}.
153: As put forth by \citet{Cottrell81}, the proton-exposed ejecta from such
154: stars may then have been accreted by and thus ``polluted'' the present day cluster
155: stars \citep[see][for a more thorough discussion]{Cannon98}.
156:
157: It has become increasingly clear that a full description of cluster abundance
158: inhomogeneities requires an understanding of the extent of the contributions
159: of {\it both} deep internal mixing and external sources of chemical differences.
160: Unfortunately, the same reaction chains (proton captures on C, N, O, Ne and Mg)
161: appear to be associated with both mechanisms.
162: Perhaps the only method of disentangling the role of each process
163: is to compare the abundances of the less luminous cluster stars with their
164: more evolved counterparts.
165: Because of the strength of the CH band, for technical reasons this is easier to accomplish
166: at present for C than for the other light elements listed above.
167:
168: In this regard, the globular cluster M13 provides us with a particularly interesting
169: target.
170: With an overall metallicity of \fe = $-1.51$ dex
171: \citep{Kraft92}, it is perhaps
172: one of the most thoroughly studied ``intermediate metallicity'' clusters.
173: It has long been known to contain giants with a large and bimodal spread in
174: CN band strengths, anticorrelated with CH, whose origin lies in an
175: apparent C-N anticorrelation \citep[e.g.][]{Smith96}.
176: In addition, among its RGB stars, large\footnote{As commented by \citet{Kraft97},
177: the M13 RGB stars exhibit some of the largest inhomogeneities known.}
178: O differences are present (anticorrelated with N), as well as significant star-to-star
179: scatter in Na and Al \citep[Na anticorrelated with Mg and correlated with O, see][]
180: {Kraft97, Cavallo00}.
181: Such observations are suggestive of the mixing of proton exposed
182: material to the surfaces of some cluster members and there is evidence
183: of shifts towards lower O and higher Na and Al abundances among the
184: most luminous RGB stars ($M_V < -1.7$) \citep{Pilachowski96, Kraft97,
185: Cavallo00}.
186: Yet, at the same time, a large dispersion in these elements is present in
187: the data of Kraft et al. to at least within a 0.25 mag of the luminosity
188: function (LF) bump at $V \approx 14.75$ \citep{Paltrinieri98}.
189: A large scatter in C and N abundances was also followed to a
190: similar luminosity by \citet{Suntzeff81}.
191: \citet*{Grundahl98} found scatter in Str\"{o}mgren
192: photometry among the RGB and SGB stars in M13 which they attributed to a
193: spread in CNO abundances probably extending at least to the main-sequence
194: turnoff.
195:
196: We report here on recent observations and derived C abundances of a sample
197: of SGB/MS stars in the globular cluster M13 which appear to support the
198: hypothesis of a substantial intrinsic spread in the distribution of light elements
199: within the cluster which is further modified during RGB ascent.
200:
201:
202: \section{OBSERVATIONS}
203:
204: The construction of
205: the photometric catalog used here for M13 is described in
206: \citet*{Stetson98} and \citet{Stetson00}
207: and is similar to that built for M5 and described in detail
208: in our earlier work \citep{Cohen02}. Slitmasks
209: were designed focusing on the subgiants in M13 and on the region just
210: below the turnoff. We note that the base of the RGB in M13
211: is at about V = 17.7, I = 16.9, while the MSTO is at about
212: V = 18.2, I = 17.8.
213:
214: These slitmasks were used on the blue side
215: \citep{McCarthy98} of LRIS \citep{Oke95}
216: at the Keck Telescope in May 2001 and in May 2002. As the
217: weather was sub-optimal, four 1000 sec exposures were obtained with
218: the slitmask intended for the fainter stars.
219: The 600 g/mm grism was used with a 0.7 arcsec wide slit, yielding
220: a dispersion of 1.0 \AA/pixel, and a spectral resolution of 4\AA.
221: The spectral coverage was from 3500 to 5500~\AA\ for essentially
222: all the stars.
223: There are 81 stars with spectra good enough for present purposes,
224: namely more than 450 ADU/pixel (with 2 $e^-$/ADU) in the continuum at
225: 4360~\AA.
226:
227:
228: Non-members are very easily recognized due to the low metallicity
229: of M13 and its large negative radial velocity. Since the stars
230: were selected to lie on the cluster isochrone and M13 is at
231: high galactic latitude, one would expect very few non-members.
232: All of the stars in the sample presented here are members of M13.
233: Full details of the sample of stars and the observations
234: will be given in a forthcoming paper.
235:
236:
237: \section{ANALYSIS}
238:
239: Our analysis essentially parallels that of \citet[hereafter BC01]{Briley01}
240: and the reader is referred to that work for details.
241:
242: The strengths of the 4350\AA ~ CH (G) bands of our program stars were
243: measured via the I(CH) index of \citet{Cohen99a, Cohen99b} - a double sided index
244: which compares the flux removed by the G-band to the adjacent continuum.
245: The resulting indices (corrected for the radial velocity of M13) are plotted as a
246: function of I magnitude in Figure \ref{fig_allch}, where a significant spread
247: among both the SGB and MS stars can be seen.
248: As is discussed in BC01, the decrease in the spread in I(CH) near $I \approx
249: 18$ is the result of the high temperatures and low metallicities of the M13 MS
250: turnoff stars (i.e., there simply is little CH formation in such stars).
251:
252: In order to relate the observed indices to the underlying \cabund, we
253: employ a series of synthetic spectra based on MARCS \citep{Gustafsson75}
254: model atmospheres.
255: We assume that the iron abundance is constant within M13 as
256: suggested by the high dispersion spectroscopic studies
257: of low luminosity globular cluster samples by
258: \citet{Gratton01} and by \citet{Ramirez02, Ramirez03}.
259: Our MS turnoff models are those used in BC01 to which we have added an
260: additional 3 luminous SGB models (at \teff/\grav = 5168/2.96, 5224/3.15,
261: 5281/3.33) and 5 fainter MS points (at 6118/4.47, 6034/4.53, 5833/4.59, 5601/4.66,
262: 5351/4.70) taken from the 16 Gyr \fe = $-$1.48 O-enhanced isochrone
263: grid of \citet{Bergbusch92}.
264: As in BC01, we adopt a geometric distance of 7.2 kpc (plus an
265: additional 0.11 magnitudes in
266: $(m-M)_V$) and a reddening of E($B-V$) = 0.02 magnitudes
267: following \citet{Harris96}.
268:
269: From each model and a given set of C/N/O abundances, synthetic spectra
270: were computed using the SSG program \citep*[][and references therein]
271: {Bell94} and the linelist of \citet{Tripicco95} (see BC01
272: for further details).
273: The result was then convolved with a Gaussian to match
274: the resolution of the observed spectra and I(CH) indices measured.
275: Four compositions are illustrated here ([C/Fe]/[N/Fe]/[O/Fe]):
276: $-0.85/+0.7/+0.4$ and $-1.1/+1.2/-0.5$ which (as in BC01) roughly match
277: the observed compositions of M13's CN-weak and strong bright giants
278: respectively, $0.0/+0.4/+0.4,$ an arbitrary ``pre-mixing''
279: composition, and $-0.5/+1.4/0.0,$ a redistribution of the ``pre-mixing'' composition
280: holding the sum of C+N+O constant.
281: The curves plotted in Figure \ref{fig_allch}
282: are I(CH) indices calculated from these synthetic spectra
283: where it can be seen that the small scatter in I(CH) among the MS turnoff
284: stars is due to the intrinsic weakness of the G-band, rather than an absence
285: of inhomogeneities (as pointed out by BC01).
286:
287: Note that among such warm relatively metal-poor stars the resulting
288: indices are remarkably insensitive to all but our choice of \cabund.
289: This is the result of the low metallicity; CN and CO play a less
290: important role in the molecular equilibrium for carbon.
291: For example, using the 5281/3.33 model ($M_I = +1.89$) with
292: $-0.5/+1.4/0.0$ as a baseline, increasing [O/Fe] by
293: 0.4 dex decreases I(CH) by 0.006, decreasing [N/Fe] by 1.4 dex
294: increases I(CH) by 0.006, and increasing \ciso\ to 90 has essentially no effect.
295: The greatest sensitivity appears to be to microturbulence ($v_t$), as increasing
296: $v_t$ from 1.5 to 2.5 \kms increases I(CH) by 0.010 (an 8\% increase),
297: which is negligible compared to the changes resulting from different
298: values of \cabund.
299:
300: We have applied the method of \citet{Briley90} to convert the observed
301: I(CH) indices to corresponding C abundances: the model isoabundance curves
302: were interpolated to the $M_I$ of each program star, and the observed I(CH) index
303: converted into \cabund\ based on the synthetic indices at that $M_I$.
304: As such, we are assuming that a given star's $M_I$ maps to the \teff, \grav, and
305: mass specified by the \citet{Bergbusch92} isochrone based on our calculation of
306: model luminosity/color and choice of distance modulus and reddening.
307: Returning to the 5281/3.33 model (the faintest model used in deriving
308: SGB \cabund), we note that an underestimation in \teff\ of 100K
309: would correspond to an error of +0.03 in $V-I$ and a derived
310: \cabund\ some 0.12 dex too large at [C/Fe] $=-0.5$.
311: Likewise, an underestimation of \grav\ by 0.1 dex would be equivalent to a
312: shift of $-0.25$ in $M_I$ and $-0.08$ dex in \cabund.
313: Essentially identical changes result when considering our most luminous
314: SGB model (5168/2.96).
315: Thus if one takes 100K and 0.1 dex as reasonable errors in \teff\ and \grav,
316: we estimate our uncertainties in \cabund\ to be roughly 0.1 dex.
317:
318: \section{RESULTS}
319:
320: The resulting \cabund\ values for the subgiants only
321: are plotted in Figure \ref{fig_chbright}, as well as the
322: \cabund\ values derived by \citet{Suntzeff81} and \citet{Smith96} for
323: more luminous cluster members.
324: Immediately obvious from Figure \ref{fig_chbright}
325: is a marked decline in \cabund\ with
326: increasing luminosity among the M13 giants - the mean
327: \cabund\ of Sunzteff's ``RGB'' sample is $-0.9 \pm 0.3$ dex, while for
328: the present SGB stars the value is $-0.4 \pm -0.2$ dex.
329: This pattern in \cabund\ was noted by Suntzeff, but the depletions
330: appear to set in at a luminosity below the limit of his sample.
331: Such as result can most easily be understood in the context of the deep-mixing
332: framework where C-depleted material from the CN-burning region above
333: the H-burning shell is progressively circulated into the stellar envelope
334: as discussed above.
335: We note however that changes in the distribution of O, Na and Al in M13 are
336: suggested by \citet{Pilachowski96}, \citet{Kraft97}, and \citet{Cavallo00}
337: to occur only among its most luminous giants, i.e., $V < 13$ which is also the
338: point at which the C abundances drop precipitously in Figure~\ref{fig_chbright}.
339: It would thus appear that during the bulk of RGB ascent, mixing only into CN-cycle
340: regions is taking place in M13.
341:
342: At the same time, a large scatter in \cabund ~ (roughly 0.8 dex), comparable
343: to that found among the more luminous stars, is present in the LRIS SGB sample
344: in Figure~\ref{fig_chbright}.
345: Figure~\ref{fig_allch} further suggests a large dispersion in C abundance is also
346: present at luminosities below the MS turnoff.
347: (We have no information on N abundances at present.)
348: These inhomogeneities are in place well below the LF bump
349: and the operation of proposed deep mixing mechanisms.
350: Thus it appears that the action of an external process has shaped much of
351: the star-to-star scatter present in M13, i.e., the large range of C, and likely
352: other light abundances found among the giants (at least fainter than $V = 13$)
353: is to a great extent {\it not} the result of varying mixing efficiencies
354: within the stars themselves.
355:
356: If we assume the initial \cabund\ values to be roughly solar in the
proto-M13 gas cloud, the C-poor stars
357: in our sample might be explained by the incorporation of C-poor, N-rich AGB
358: ejecta.
359: However, the large spread in C abundances seen here requires far more than
360: the ``pollution'' of the stellar surfaces by simple accretion of AGB ejecta - such
361: surface contaminations would be quickly erased by the deepening convective
362: envelope following the MS phase.
363: This is clearly not observed and a substantial fraction of the stars' total mass
364: must therefore be homogeneous in these elements.
365: This in turn requires the accretion of not just of a sprinkling but rather of a
366: significant amount of material: if we consider a
367: typical C-poor M13 star (\cabund\ $= -0.6$), some 70\% of the star's mass
368: must be captured ejecta (assuming the accreted matter to be completely
369: free of C).
370: While this initially seems implausible, we note the recent calculations of
371: \citet{Thoul02} which demonstrate that large accumulations
372: may be possible, particularly in clusters with small core radii, e.g. 47 Tuc
373: where they estimate as much as 80\% of the mass of a 1 M\subsun ~ star
374: could be accreted material.
375: Unfortunately, the core radius of M13 is considerably larger, and similar
376: calculations yield an expected accretion of only 8\%.
377: Whether this is a problem that can be overcome with more detailed accretion
378: models (e.g., Thoul et al. included several assumptions such as chaotic orbits
379: with stars spending 20\% of their time in the cluster core, which was also the
380: spatial extent of the central gas reservoir), a reflection of changes in the
381: structure/dynamics of M13 since its formation, or can be better explained by
382: an alternative mechanism remains to be seen.
383:
384: The work presented here implies we are seeing the effects of a significant
385: external ``contamination'' of unexplained origin
386: upon which a poorly understood mixing mechanism acts during the
387: later stages of RGB ascent.
388: However, there remain a number of issues to address:
389: The critical region just fainter than the M13 LF bump at
390: $V \approx 14.75$ has yet to be explored, leaving
391: the exact luminosity of C-depletion onset uncertain.
392: Also, in comparing the present results with those of \citet{Suntzeff81} and
393: \citet{Smith96}, we are comparing abundances based on different instruments,
394: indices, and analysis tools - note the apparent 0.3 dex offset between the C
395: abundances of Suntzeff and Smith in Figure~\ref{fig_chbright}.
396: Furthermore, we urgently need a larger sample of stars below the MSTO in M13.
397: To address these items, we will be returning to M13 in the near future.
398:
399: \acknowledgements
400: The entire Keck/LRIS user communities owes a huge debt to
401: Jerry Nelson, Gerry Smith, Bev Oke, and many other
402: people who have worked to make the Keck Telescope and LRIS
403: a reality and to operate and maintain the Keck Observatory.
404: We are grateful to the W. M. Keck Foundation for the vision to fund
405: the construction of the W. M. Keck Observatory.
406: We wish to express our thanks to Roger Bell whose SSG code was
407: instrumental in this project.
408: Partial support was provided by the National Science Foundation under
409: grant AST-0098489 to MMB and grant AST-9819614 and AST-0205951 to JGC and by
410: the F. John Barlow professorship (MMB).
411:
412: \begin{thebibliography}{}
413:
414: \bibitem[Aikawa et al.(2001)Aikawa, Fujimoto, \& Kato]{Aikawa01}
415: Aikawa,ÊM., Fujimoto,ÊM.ÊY., \& Kato,ÊK. 2001, \apj, 560, 937
416:
417: \bibitem[Bell et al.(1994)Bell, Paltoglou, \& Tripicco]{Bell94} Bell, R. A., Paltoglou,
418: G., \& Tripicco, M. J. 1994, \mnras, 268, 771
419:
420: \bibitem[Bellman et al.(2001)]{Bellman01} Bellman, S., Briley, M. M., Smith, G. H.
421: \& Claver, C. F., 2001, \pasp, 113, 326
422:
423: \bibitem[Bergbusch \& Vandenberg(1992)]{Bergbusch92} Bergbusch, P. A \&
424: Vandenberg, D. A., 1992, \apjs, 81, 163
425:
426: \bibitem[Briley et al.(1990)]{Briley90} Briley,ÊM. M., Bell,ÊR. A., Hoban,ÊS.,
427: Dickens,ÊR. J. 1990, \apj, 359, 307
428:
429: \bibitem[Briley \& Cohen(2001)]{Briley01} Briley, M. M., \& Cohen, J. G.
430: 2001, \aj, 122, 1429
431:
432: \bibitem[Briley et al.(1996)]{Briley96} Briley, M. M., Smith, V. V., Suntzeff, N. B.,
433: Lambert, D. L., Bell, R. A. \& Hesser, J. E., 1996, Nature, 383, 604
434:
435: \bibitem[Cannon et al.(1998)]{Cannon98} Cannon, R. D., Croke, B. F. W., Bell,
436: R. A., Hesser, J. E. \& Stathakis, R. A., 1998, \mnras, 298, 601
437:
438: \bibitem[Cavallo \& Nagar(2000)]{Cavallo00} Cavallo,ÊR.ÊM., \& Nagar,ÊN.ÊM.
439: 2000, \aj, 120, 1364
440:
441: \bibitem[Charbonnel(1995)]{Charbonnel95} Charbonnel, C., 1995, \apjl, 453, L4
442:
443: \bibitem[Charbonnel \& do Nascimento(1998)]{Charbonnel98} Charbonnel,ÊC.,
444: \& doÊNascimento,ÊJ. D.Ê1998, \aap, 336, 915
445:
446: \bibitem[Cohen(1999a)]{Cohen99a} Cohen, J. G., 1999a, \aj, 117, 2434
447:
448: \bibitem[Cohen(1999b)]{Cohen99b} Cohen, J.~G., 1999b \aj, 117, 2428
449:
450: \bibitem[Cohen et al.(2002)Cohen, Briley, \& Stetson]{Cohen02} Cohen,ÊJ.ÊG.,
451: Briley,ÊM.ÊM., \& Stetson, P.ÊB. 2002, \aj, 123, 2525
452:
453: \bibitem[Cottrell \& DaÊCosta(1981)]{Cottrell81} Cottrell,ÊP.ÊL., \& DaÊCosta,
454: G.ÊS. 1981, \apj, 245, L79
455:
456: \bibitem[DaÊCosta(1998)]{DaCosta98} DaÊCosta,ÊG. S. 1998, in Proceedings
457: of IAU Symposium 189, On Fundamental Stellar Properties: The Interaction
458: between Observation and Theory, eds. T. R. Bedding, A. J. Booth \& J. Davis
459: (Kluwer:Dordrecht), p. 193
460:
461: \bibitem[Denissenkov \& Tout(2000)]{Denissenkov00} Denissenkov,ÊP.ÊA., \& Tout,
462: C.ÊA. 2000, \mnras, 316, 395
463:
464: \bibitem[Gratton et al.(2001)]{Gratton01} Gratton, R. G. et al., 2001, \aap, 369, 87
465:
466: \bibitem[Gratton et al.(2000)]{Gratton00} Gratton, R. G., Sneden, C.,
467: Carretta, E. \& Bragaglia, A. 2000, \aap, 354, 169
468:
469: \bibitem[Grundahl et al.(1998)Grundahl,VandenBerg, \& Andersen]{Grundahl98}
470: Grundahl, F. VandenBerg, D. A., \& Andersen, M. L. 1998, \apj, 500, L179
471:
472: \bibitem[Gustafsson et al.(1975)]{Gustafsson75} Gustafsson, B., Bell, R. A.,
473: Eriksson, K., \& Nordlund, \AA\ 1975, \aap, 42, 407
474:
475: \bibitem[Harris(1996)]{Harris96} Harris, W. E. 1996, \aj, 112, 1487
476:
477: \bibitem[Hesser(1978)]{Hesser78} Hesser, J. E., 1978, \apj, 223, L117
478:
479: \bibitem[Keller et al.(2001)Keller, Pilachowski, \& Sneden]{Keller01}
480: Keller,ÊL. D., Pilachowski,ÊC. A., \& Sneden, C. 2001, \aj, 122, 2554
481:
482: \bibitem[Kraft(1994)]{Kraft94} Kraft, R. P. 1994, \pasp , 106, 553
483:
484: \bibitem[Kraft et al.(1992)]{Kraft92} Kraft, R. P., Sneden, C., Langer, G. E., \& Prosser,
485: C. F. 1992, \aj, 104, 645
486:
487: \bibitem[Kraft et al.(1997)]{Kraft97} Kraft,ÊR.ÊP., Sneden,ÊC., Smith,ÊG.ÊH., Shetrone,
488: M.ÊD., Langer,ÊG.ÊE., \& Pilachowski,ÊC.ÊA. 1997, \aj, 113, 279
489:
490: \bibitem[McCarthy et al.(1998)]{McCarthy98} McCarthy, J. K., et al. 1998, SPIE, 3355, 81
491:
492: \bibitem[Oke et al.(1995)]{Oke95} Oke, J. B., et al. 1995, \pasp, 107, 375
493:
494: \bibitem[Paltrinieri et al.(1998)]{Paltrinieri98} Paltrinieri,ÊB., Ferraro,ÊF.ÊR.,
495: Carretta,ÊE., FusiÊPecci,ÊF. 1998, \mnras, 293, 434
496:
497: \bibitem[Pilachowski et al.(1996)]{Pilachowski96} Pilachowski,ÊC.ÊA., Sneden,
498: C., Kraft,ÊR.ÊP., \& Langer,ÊG.ÊE. 1996, \aj, 112, 545
499:
500: \bibitem[Ram\'{\i}rez \& Cohen(2002)]{Ramirez02} Ram\'{\i}rez, S. V. \&
501: Cohen, J. G., 2002 \aj, 123, 3277
502:
503: \bibitem[Ram\'{\i}rez \& Cohen(2003)]{Ramirez03} Ram\'{\i}rez, S. V. \&
504: Cohen, J. G., 2003 \aj, submitted
505:
506: \bibitem[Salaris et al.(2002)Salaris, Cassisi, \& Weiss]{Salaris02}
507: Salaris,ÊM., Cassisi,ÊS., \& Weiss,ÊA. 2002, \pasp, 114, 388
508:
509: \bibitem[Smith et al.(1996)]{Smith96} Smith, G. H., Shetrone, M. D., Bell,
510: R. A., Churchill, C. W., \& Briley, M. M. 1996, \aj, 112, 1511
511:
512: \bibitem[Stetson(2000)]{Stetson00} Stetson, P.B., 2000, \pasp, 112, 925
513:
514: \bibitem[Stetson et al.(1998)Stetson, Hesser, \& Smecker-Hane]{Stetson98}
515: Stetson, P.B., Hesser, J.E. \& Smecker-Hane, T., 1998, \pasp, 110, 533
516:
517: \bibitem[Suntzeff(1981)]{Suntzeff81} Suntzeff, N. B., 1981, \apjs, 47, 1
518:
519: \bibitem[Suntzeff \& Smith(1991)]{Suntzeff91} Suntzeff, N. B. \& Smith, V. V.
520: 1991, \apj, 381, 160
521:
522: \bibitem[Sweigart \& Mengel(1979)]{Sweigart79} Sweigart, A. V. \& Mengel,
523: J. G., 1979, \apj, 229, 624
524:
525: \bibitem[Thoul et al.(2002)]{Thoul02} Thoul,ÊA., Jorissen,ÊA., Goriely,ÊS., Jehin,
526: E., Magain,ÊP., Noels,ÊA., \& Parmentier,ÊG. 2002, \aap, 383, 491
527:
528: \bibitem[Trefzger et al.(1983)]{Trefzger83} Trefzger, C. F., Carbon, D. F., Langer.
529: G. E., Suntzeff, N. B. \& Kraft, R. P., 1983, \apj, 266, 144
530:
531: \bibitem[Tripicco \& Bell(1995)]{Tripicco95} Tripicco, M. J., \& Bell, R. A. 1995,
532: \aj, 110, 3035
533:
534: \bibitem[Ventura et al.(2001)]{Ventura01} Ventura, P., D'Antona, F., Mazzitelli, I.,
535: \& Gratton, R.\ 2001, \apjl, 550, L65
536:
537: \end{thebibliography}
538:
539: \clearpage
540:
541: \begin{figure}
542: \plotone{f1.eps}
543: \caption{Observed and synthetic I(CH) indices
544: for stars in M13 are plotted as a function
545: of luminosity. Also plotted are the indices from \citet{Cohen99b}. One sigma error bars
546: based on Poisson statistics are included. A large spread in I(CH) is
547: present for both the SGB and pre-MS turnoff stars, as well as among the stars at the
548: base of the turnoff. Using the model isoabundance lines as a guide,
549: the scatter in I(CH) appears consistent with a constant range in \cabund\
550: over the entire luminosity range.
551: \label{fig_allch}}
552: \end{figure}
553:
554: \begin{figure}
555: \plotone{f2.eps}
556: \caption{Carbon abundances derived from the present SGB stars
557: in M13 are plotted with the results of \citet{Suntzeff81} and \citet{Smith96}.
558: Although a large and consistent scatter in \cabund\ apparently exists from the SGB
559: to the tip of the RGB, the decline in \cabund\ with luminosity
560: near the RGB tip is striking.
561: \label{fig_chbright}}
562: \end{figure}
563:
564: \end{document}
565: