1: \documentclass[12pt,preprint]{aastex}
2: \begin{document}
3: \title{Beryllium in the Ultra-Lithium-Deficient, \\
4: Metal-Poor Halo Dwarf, G186-26\altaffilmark{1}}
5:
6: \author{Ann Merchant Boesgaard\altaffilmark{2} and Megan C. Novicki\altaffilmark{2}}
7:
8: \altaffiltext{1}{Based on data obtained with the Subaru Telescope, which is
9: operated by the National Astronomical Observatory of Japan.}
10:
11: \altaffiltext{2}{Institute for Astronomy, University of Hawai`i at M\-anoa,
12: 2680 Woodlawn Drive, Honolulu, HI {\ \ }96822, boes@ifa.hawaii.edu,
13: mnovicki@ifa.hawaii.edu}
14:
15: \begin{abstract}
16: The vast majority of low-metal halo dwarfs show a similar amount of Li; this
17: has been attributed to the Li that was produced in the Big Bang. However,
18: there are nine known halo stars with T $>$ 5900 K and [Fe/H] $<$ $-$1.0 that
19: are ultra-Li-deficient. We have looked for Be in the very low metallicity
20: star, G 186-26 at [Fe/H] = $-$2.71, which is one of the ultra-Li-deficient
21: stars. This star is also ultra-Be deficient. Relative to Be in the Li-normal
22: stars at [Fe/H] = $-$2.7, G 182-26 is down in Be by more than 0.8 dex. Of two
23: potential causes for the Li-deficiency -- mass-transfer in a pre-blue
24: straggler or extra rotationally-induced mixing in a star that was initially a
25: very rapid rotator -- the absence of Be favors the blue-straggler hypothesis,
26: but the rotation model cannot be ruled-out completely.
27: \end{abstract}
28:
29: \keywords{stars: abundances; stars: evolution; stars: late-type;
30: subdwarfs; stars: individual (G 186-26); stars: Population II; Galaxy: halo}
31:
32:
33: \section{Introduction}
34:
35: Although nearly all metal-poor dwarf and turn-off stars with T$_{\rm eff}$
36: $>$5600 K and [Fe/H] $\lesssim$$-$1.3 show similar Li abundances, there are a
37: few such stars which seem to be ultra-deficient in Li. The discovery by Spite
38: \& Spite (1982) of a plateau in the Li abundances in metal-poor stars has been
39: followed by much research to determine Li in many additional ancient halo
40: stars and to derive the primordial Li abundance produced by Big Bang
41: nucleosynthesis (BBN) (e.g. Thorburn 1994, Bonifacio \& Molaro 1997, Ryan et
42: al.~1999, 2000, Mel\'endez \& Ram\'\i rez 2004). In fact, in the follow-up
43: work to the initial paper by Spite \& Spite (1982), Spite et al.~(1984)
44: discovered a Li-deficient star, HD 97916, with a Li abundance at least an
45: order of magnitude below the plateau. The plateau value of A(Li) ( = log
46: N(Li/H) + 12.00) is near 2.2, but ``the plateau'' seems to have both a
47: temperature and a metallicity dependence, e.g. Novicki (2005); those results
48: are consistent with the results of Thorburn (1994) and Ryan et al.~(2001a) in
49: terms of both the coefficients of the dependencies and the sample size. For
50: HD 97916 at 6124 K the upper limit on A(Li) is $<$1.2. The extreme halo star,
51: G 186-26, was found to be Li deficient by Hobbs, Welty, \& Thorburn (1991).
52: Additional Li-deficient metal-poor stars have been discovered by Hobbs \&
53: Mathieu (1991), Thorburn (1994), and Ryan et al.~(2001b). There are now nine
54: known halo dwarfs with [Fe/H] $<$$-$1.1 all having T$_{\rm eff}$ $>$ 5980 K.
55:
56: These ``ultra-Li-deficient stars'' are important in estimating the primordial
57: Li value, A(Li)$_p$. Do they represent a true dispersion in the plateau? Do
58: they indicate greater Li depletion and are thus sign-posts of general (more
59: mild) Li depletion in all or most of the plateau stars? If all plateau stars
60: have undergone some depletion, then A(Li)$_p$ is higher than the currently
61: measured value, with concomitant implications for cosmology, such as the
62: baryon-to-photon ratio, $\eta$.
63:
64: There are at least two possible origins for the Li deficiencies. Ryan et
65: al.~(2001a, 2002) argue that the original Li is reduced by the mechanism that
66: operates in blue-straggler stars, primarily a mass transfer event or a stellar
67: merger, but in ``blue-stragglers-to-be.'' Pinsonneault et al.~(1999, 2002)
68: propose that stellar mixing caused by rotation has lowered Li in the plateau
69: and has produced larger Li depletions in some fraction of the stars that were
70: originally the most rapid rotators. (Ryan et al.~2001a rule out diffusion and
71: a Hyades-like Li-dip as explanations for the severely Li-deficient halo
72: stars.) These two hypotheses have different implications for the Be content
73: of the ultra-Li-deficient stars. In the blue-straggler model the star would
74: undergo substantial internal mixing and/or mass transfer which would destroy
75: both Li and Be completely as atoms of both elements would be in environments
76: where the temperatures are high enough to destroy them: $\sim$2.5 x 10$^6$ K
77: for Li and $\sim$3.5 x 10$^6$ K for Be. (Mass transferred from an evolved
78: giant would also be diluted in Li and Be.) In the rotationally-induced mixing
79: model Li may be partly or completely destroyed, while Be may be totally or
80: partly preserved (Pinsonneault, Deliyannis \& Demarque 1992, hereafter PDD92).
81:
82: The nine ultra-Li-deficient are not a uniform group. Four of them are
83: single-line spectroscopic binaries (Carney et al.~2001); three of those four
84: plus one other have measureable rotation velocities, while two (including G
85: 186-26) have very sharp lines with only upper limits on v sin i (Elliott \&
86: Ryan 2005); three or four (including G 186-26) are single stars (Ryan et
87: al.~2001a). There seem to be no peculiarities in composition that
88: characterize all of them (e.g. Spite et al.~1993, Norris et al.~1997; Elliott
89: \& Ryan 2005).
90:
91: In this letter we report on the Be abundance of one of the warmest and most
92: metal-poor of the ultra-Li deficient stars, G 186-26. The value of A(Li) for
93: this star is $<$1.1 and [Fe/H] = $-$2.8 (Thorburn 1994).
94:
95: \section{Observations and Analysis}
96:
97: The spectra for G 186-26 were obtained on one night in 2003 May 27 (UT) with
98: the high-dispersion spectrograph (HDS) at the Subaru 8.2 m telescope on Mauna
99: Kea (Noguchi et al.~2002). We used the blue collimator and blue
100: cross-disperser with the standard HDS setup: StdUb. Our slit was 0.7 x 4.4
101: arcsec and the binning was 2 x 2. There are two EEV-CCDs covering 2048 x 4100
102: pixels (pixel size = 13.5 $\mu$) corresponding to a wavelength coverage of
103: 2970 - 4640 \AA. Calibration exposures were taken of the bias, halogen lamps
104: for flat-fielding, and Th-Ar comparison spectra. We obtained two integrations
105: of 50 minutes each of this V = 10.82 star for a total signal-to-noise (per
106: pixel) of 98 at 3130 \AA. The spectral resolution is $\sim$50,000 or 0.062
107: \AA. The dispersion is 0.0187 \AA\ pix$^{-1}$ and the resolution element is
108: 3.3 pix. Both resonance lines of Be II at 3130.416 and 3131.064 \AA\ appear
109: in two different orders of the spectrum and we analyzed both separately, but
110: the flux in the order where the Be II lines are centered is 8.5 times that in
111: the lower order. Standard data reduction procedures were used. After
112: division by the normalized flat field, scattered light and cosmic ray events
113: were removed. The spectra were traced and extracted, and wavelength
114: calibrated from the Th-Ar spectra. The two exposures were combined and the
115: continuum was determined. Since G 186-26 has a metallicity of 500 times below
116: solar, there is little blending and the continuum placement was
117: straight-forward.
118:
119: The stellar parameters for G 186-26 have been determined by Novicki (2005).
120: We used the photometric indices of $(b-y)$$_0$, $(V-K)$$_0$, and $(R-I)$$_0$
121: and the calibrations of Carney (1983) to determine the effective temperature,
122: T$_{\rm eff}$. (The $(b-y)$ values are from Shuster \& Nissen (1988), $(V-K)$
123: from Alonso et al.~(1994) and Carney (1983), the $(R-I)$ from Eggen (1979).)
124: A weighting of 4:2:1 was used for T$_{\rm eff}$($b-y)_0$: T$_{\rm
125: eff}$($V-K)_0$: T$_{\rm eff}$($R-I)_0$. (For G 186-26 the reddening is
126: negligible and was ignored: E($b-y$) = 0.011.) The three temperatures are
127: 6196 K, 6181 K, and 6256 K with a weighted mean of 6200 K $\pm$25 K, but we
128: use $\pm$40 K as a more realistic estimate of the uncertainty. For comparison
129: Alonso et al.~(1996) find a weighted mean of 6428 K from H, J, K photometry
130: for this star. We have found six values for [Fe/H] in the literature from
131: high-resolution, high S/N determinations. These have been normalized to the
132: same scale (e.g. to solar log N(Fe/H) = 7.51) and averaged. Two lower
133: resolution results have been included also. The final weighted mean for
134: [Fe/H] is $-$2.71 (Novicki 2005). We used Str\"omgren photometry, ($b-y$)$_0$
135: and c$_0$, to estimate log g via the Yi, Demarque \& Kim (2004) 10 Gyr
136: isochrone for Y = 0.23. This gives a value of 4.48 and we estimate the
137: uncertainty as 0.20 dex. (The 13 Gyr isochrone gives log g of 4.44 which
138: would result in a smaller Be abundance by 0.018 dex, so our value of 4.48 is a
139: conservative choice.) The microturbulent velocity was taken to be 1.5 km
140: s$^{-1}$, the typical value for low-metal stars according to Magain (1987).
141: Our parameters for G 186-26 are T$_{\rm eff}$ = 6200 $\pm$40 K, log g = 4.48
142: $\pm$ 0.20, [Fe/H] = $-$2.71 $\pm$0.12, which are in good agreement with those
143: of Thorburn (1994), Norris et al.~(1997), and Akerman et al.~(2004).
144:
145: The Be abundance was determined using spectrum synthesis with the program
146: MOOG, version 2002, (Sneden 1973; http://verdi.as.utexas.edu/moog.html). The
147: stellar parameters were used to generate a model atmosphere using the Kurucz
148: (1993) grid of models. All elements except Be and O are reduced by the same
149: amount as [Fe/H]. The O abundance matters as there are many blending OH lines
150: in the Be II spectral region. We have used the value of [O/Fe] = +0.49
151: derived for this star by Akerman et al.~(2004). Figure 1 shows the spectrum
152: synthesis fit for Be in our combined spectrum of G 186-26. Neither of the two
153: Be II resonance lines is present. Shown for comparison in Figure 1 is the
154: Li-normal star, BD +3 740, which has normal Be for its temperature and [Fe/H],
155: both of which are similar to those parameters for G 186-26; this spectrum is
156: from Boesgaard et al.~(1999) but has been reanalyzed with the latest version
157: of MOOG. Figure 2 shows an enlarged view of the Be II lines in G 186-26 with
158: the same synthesis as in Figure 1. The value for A(Be) of $-$1.58 was
159: selected as the expected initial Be abundance (see $\S$3.2). The uncertainty
160: in the Be abundances that is due to the uncertainties in the stellar
161: parameters is $\pm$0.09.
162:
163: \section{Results and Interpretation}
164:
165: It can be seen from Figures 1 and 2 that the spectrum of G 186-26 is
166: consistent with no Be present in the stellar atmosphere, but we suggest an
167: upper limit on A(Be) of $-$2.00 $\pm$0.09 dex. As Figure 1 shows, we have
168: detected Be II lines in BD +3 740 at 6227 K and [Fe/H] = $-$2.81 and derive
169: A(Be) = $-$1.37. Furthermore, Primas et al.~(2000a) show the presence of Be
170: in G 64-12, a star with a similar temperature (6400 K) and even lower
171: metallicity, [Fe/H] = $-$3.28, with A(Be) = $-$1.15.
172:
173: We compare our upper limit on Be for G 186-26 with Be abundances for other
174: stars at its metallicity. Our Figure 3 shows the Be abundance limit of G
175: 186-26 in the context of Be detections in other metal-poor, Li-normal stars as
176: a function of metallicity. The stars below [Fe/H] of $-$0.9 appear to fall on
177: the same Fe-Be trend considering the scatter in the Be abundances at a given
178: Fe, i.e.~they have normal Be abundances. They also all have normal Li plateau
179: abundances. Our A(Be) limit, $<$$-$2.0, is less than other such stars at this
180: metallicity by 0.8 dex. Allowing for the $\pm$0.12 uncertainty in [Fe/H], it
181: is lower by 0.7 - 0.9 dex.
182:
183: \subsection{Blue Straggler}
184:
185: Ryan et al.~(2001) pose a question about the blue straggler phenomenon which
186: essentially is: why should the processes that form blue stragglers be limited
187: only to those stars that have masses {\it above} the turn-off for main
188: sequence stars? Those higher mass stars are apparent because of their unusual
189: colors -- too bright and too blue for their ages. The phenomenon involving
190: mass transfer and mergers could be at work in lower-mass, sub-turnoff stars as
191: well. Peculiar colors would not be the signature of this behavior, but Li
192: depletion would result. For example, blue stragglers examined for Li in M 67
193: by Pritchet \& Glaspey (1991) have no detectable Li. Hobbs \& Mathieu (1991)
194: found no Li line in blue stragglers in the two halo field stars they studied.
195: Those two stars were ``known'' blue stragglers before they were on the list
196: of the nine ultra-Li deficient stars. Carney et al.~(2005) find that 4 of the
197: 5 Li-deficient field blue stragglers that they studied are single-lined
198: spectroscopic binaries.
199:
200: Pritchet \& Glaspey (1991) conclude that the blue straggler formation must be
201: due to binary coalescence, binary mass transfer, or other deep mixing. Such
202: stars would be thoroughly mixed. All three light elements, Li, Be, and B,
203: would be destroyed in this scenario as all the atoms of these elements would
204: be subjected to temperatures where they would be destroyed by nuclear
205: reactions. Preston \& Sneden (2000) argue that most field blue stragglers are
206: produced by mass transfer events because collisions are less common in the
207: field than in clusters. The former-giant primary that transfers the mass
208: would deposit Li-depleted deep envelope material onto the current primary.
209: Our measurement of A(Be) and Thorburn's (1994) measurement of A(Li) show that
210: both Li and Be are severely depleted -- if present at all. We may be
211: observing a star that is a blue-straggler-to-be. It does not have the mass to
212: have evolved beyond the main sequence, yet it has the signature of a coalesced
213: binary or mass transfer in its Li and Be deficiencies.
214:
215: Latham et al.~(2002) did not find a period for G 186-26. The Carney et
216: al.~(1996) radial velocity measurements show a variation from $-$317.78 to
217: $-$323.19 km s$^{-1}$ for a range of 5.41 km s$^{-1}$ with measurement errors
218: of $\pm$0.71 to $\pm$1.89 (or $\pm$0.98, the mean internal error). Their 43
219: measurements with error bars over 3220 days are shown in Figure 4. Four
220: additional measurements by Latham et al.~(2002) some 3300 days later spanning
221: 139 days center around $-$322.6 km s$^{-1}$ with errors of $\sim$0.9 km
222: s$^{-1}$. There are no signs of double lines in our spectra.
223:
224: If there has been mass transfer from a former AGB star onto G 186-26, it would
225: have left composition signatures. In particular, Sneden, Preston \& Cowan
226: (2003) have shown that blue stragglers would have overabundances of the
227: s-process elements, Sr, Ba, and Pb. Norris et al.~(1997) found an
228: overabundance of Ba in G 186-26, with [Ba/Fe] = 0.35. Abundances of many
229: elements in six ultra-Li-deficient halo stars were reported by Elliott \& Ryan
230: (2005); they find that Sr is overabundant by +0.5 dex and Ba by +1.0 dex in G
231: 186-26. Norris et al.~(2001), Stephens \& Boesgaard (2002) and Fulbright
232: (2002) all find that such an overabundance of [Ba/Fe] is extremely rare at
233: this low [Fe/H].
234:
235: Support for the blue straggler model for the ultra Li-deficient stars comes
236: from the apparent lack of Be in G 186-26. Additional support for this comes
237: from the super-solar content of neutron-capture elements. However, there is
238: no evidence that G 186-26 is a binary star, but it may be a coalesced binary.
239:
240: \subsection{Rotation}
241:
242: Although all the Li abundances in the ultra Li-deficient stars are upper
243: limits, the presence of Be could indicate that rotationally-induced mixing has
244: occurred, but not down to internal temperatures as high as $\sim$3.5 x 10$^6$
245: K. Therefore, abundances of Be in the ultra-deficient Li halo stars could
246: provide a clue about the efficiency of internal mixing. That is, if Be is
247: found to be present in the ultra-Li deficient stars, that would imply that the
248: {\it Li depletion} is likely to be due to mixing caused by high initial
249: angular momentum.
250:
251: The effect of rotationally-induced mixing on the Li and Be abundances in halo
252: stars was first addressed by PDD92. According to the standard models, there
253: is no Be depletion (their Tables 3A and 3B). The models which track the
254: instabilities due to spin-down and angular momentum loss do predict depletion
255: of both Li and Be. These authors address whether there is a range in initial
256: rotation rates (i.e. in initial angular momentum) that results in a range in
257: Li at the present epoch. Since Li is destroyed at shallower layers than Be,
258: this mixing mechanism should produce less Be destruction than Li destruction,
259: unless mixing is deep enough to deplete both elements substantially.
260:
261: We have tried to determine the amount of Li and Be depletion that is predicted
262: for our star by the PDD92 models. Therefore, we have examined the latest
263: ``Yale'' isochrones (Yi et al.~2004) appropriate for our star's [Fe/H], log g,
264: and effective temperature to find the mass for G 186-26. The 13 Gyr isochrone
265: gives a mass of 0.735 M$_{\odot}$. Tables 5A, 5B, and 5C in PDD92 show the
266: expected depletion for Li and Be for Z = 0.0001 for three values of the
267: initial angular momentum. We looked at Table 5C primarily, in order to find
268: the largest Li and Be depletions. From our interpolation we find Li depleted
269: from the initial value by $-$1.21 dex and Be by $-$0.40 dex, compared with at
270: least $-$0.8 dex in G186-26. If the initial Li is 2.2 dex (like the current
271: plateau value), the present value would be 1.0 dex; the measured limit on
272: A(Li) by Thorburn (1994) is consistent at $<$1.1 dex. The initial value for
273: Be can be found from the relationship between A(Be) and [Fe/H] for the iron
274: abundance in G 186-26 of [Fe/H] = $-$2.71. From Figure 5 of Boesgaard et
275: al.~(1999) and Figure 10.3 of Boesgaard (2004) we find the initial Be was
276: about $-$1.18 dex and thus the current value would be $-$1.18 + $-$0.40 =
277: $-$1.58 for A(Be). Figures 1 and 2 show this value in the spectrum synthesis
278: as well as 2 and 4 times more Be than that, $-$2.00 (2.6 times less Be), and
279: zero Be. Although no Be gives the best fit, an upper limit of $-$2.0 dex is
280: consistent with the noise seen in the continuum and the overall fit of the
281: synthesis. We note there are uncertainties in the initial Be assumed for this
282: star and for the amount of Be depletion from the models.
283:
284: Newer models and observations on this issue have been discussed by
285: Pinsonneault et al.~(1999, 2002). Pinsonneault et al.~(2002) explain that the
286: early models of PDD92 did not take into account that there could be a
287: saturation effect in the loss of angular momentum for the rapidly rotating
288: stars. They state that the amount of Li depletion could not be {\it as large
289: as} the PDD92 rates. This would likely reduce the amount of Be depletion
290: also. The hypothesis that the ultra Li-deficient stars were originally part
291: of a group of very rapidly rotating stars which have now spun down, along with
292: some destruction of Li and Be, is not easy to reconcile with such a low upper
293: limit on A(Be). This issue of Be abundances should be readdressed with modern
294: models.
295:
296: \section{Summary and Conclusions}
297:
298: The nature of the ultra Li-deficient metal-poor stars has been the subject of
299: several papers and is relevant for the determination of the value of
300: primordial Li. Two leading hypotheses for the Li deficiencies are: (1) They
301: are analogs to blue stragglers which have lost surface Li through binary
302: mass-transfer or mergers. (2) They are the descendents of a subset of stars
303: with initial rapid rotation that have depleted Li by rotationally-induced
304: mixing during spin-down through their evolution. In the blue straggler idea
305: there would be complete destruction of both Li and Be. If extra mixing
306: resulted from rotation, much of the Li would be destroyed, but little or no
307: Be. Thus Be acts as a discriminator between these two hypotheses because it
308: is less fragile than Li.
309:
310: We have made observations of Be in a Li-deficient, very metal-poor halo star,
311: G 186-26. This star has [Fe/H] = $-$2.71, T$_{\rm eff}$ = 6200, and A(Li)
312: $<-$1.1. Our Subaru/HDS spectrum has a spectral resolution of $\sim$50,000
313: and S/N of 98 in the Be II spectral region near 3130 \AA. An upper limit for
314: A(Be) is $<-$2.00. The data support the blue straggler hypothesis over
315: rotation. A coalesced binary and mass transfer episodes could result in the
316: observed of deficits of Li and Be.
317:
318: For G 186-26 the blue straggler analog fits well, but it is necessary to look
319: at Be in other ultra Li-deficient, low-metal stars. To more fully evaluate
320: the rotation hypothesis, new model predictions should be done for Be. Elliott
321: \& Ryan (2005) have shown that there are no commonalities in other abundances,
322: but four of their six stars have measureable rotational velocities; perhaps
323: rotation plays a role in these stars.
324:
325: \acknowledgements We thank Dr. Sean Ryan for his suggestions on an earlier
326: version of this paper. We are grateful for the assistance at the telescope by
327: the enthusiastic crew at Subaru. This work was supported by NSF grants
328: AST-0097945 and AST-0505899 to A.M.B.
329:
330: \begin{references}
331:
332: \reference{}
333: Akerman, C.J., Carigi, L., Nissen, P.E., Pettini, M. \& Asplund, M. 2004,
334: \aap, 414, 931
335:
336: \reference{}
337: Alonso, A., Arribas, S. \& Mart\'\i nez-Roger, C. 1994, \aaps, 107, 365
338:
339: \reference{}
340: Alonso, A., Arribas, S. \& Mart\'\i nez-Roger, C. 1996, \aaps, 117, 227
341:
342: \reference{} Boesgaard, A.M. 2000, in ``The Light Elements and their
343: Evolution'' IAU Sym. \# 198, ed. L. Da Silva, M. Spite, J.R. Medeiros, p. 389
344:
345: \reference{}
346: Boesgaard, A.M. 2004, in ``The Origin and Evolution of the Elements''
347: ed. A. McWilliam \& M. Rauch, (Cambridge: Cambridge University Press), p. 117
348:
349: \reference{}
350: Boesgaard, A.M., Deliyannis, C.P., King, J.R., Ryan, S.G., Vogt, S.S. \&
351: Beers, T.C. 1999, \aj, 117, 1549
352:
353: \reference{}
354: Bonifacio, P. \& Molaro, P. 1997, \mnras, 285, 847
355:
356: \reference{}
357: Carney, B.W. 1983, \aj, 88, 623
358:
359: \reference{}
360: Carney, B.W., Latham, D.W. \& Laird, J.B. 2005, \aj, 129, 466
361:
362: \reference{}
363: Carney, B.W., Latham, D.W., Laird, J.B. \& Aguilar, L.A. 1994, \aj, 107, 2240
364:
365: \reference{}
366: Carney, B.W., Latham, D.W., Laird, J.B., Grant, C.E. \& Morse, J.A. 2001. \aj,
367: 122, 3419
368:
369: \reference{}
370: Eggen, O.J. 1979, \apj, 229, 158
371:
372: \reference{}
373: Elliott, L. \& Ryan, S.G. 2005, IAU Symposium 228, in press
374:
375: \reference{}
376: Fulbright, J. 2002, \aj, 123, 404
377:
378: \reference{}
379: Hobbs, L.M. \& Mathieu, R.D. 1991, \pasp, 103, 431
380:
381: \reference{}
382: Hobbs, L.M., Welty, D.E. \& Thorburn, J.A. 1991, \apj, 373, L47
383:
384: \reference{}
385: Kurucz, R. 1993, CD-ROM 1, Atomic Data for Opacity Calculations
386: (Cambridge: SAO)
387:
388: \reference{}
389: Latham, D.W., Stefanik, R.P., Torres, G., Davis, R.J., Mazeh, T., Carney,
390: B.W., Laird, J.B. \& Morse, J.A. 2002, \aj, 124, 1144
391:
392: \reference{}
393: Mel\'endez, J. \& Ram\'\i rez, I. 2004, \apj, 615, L33
394:
395: \reference{}
396: Nissen, P.E., Gustafsson, B., Edvardsson, B. \& Gilmore, G. 1994, \aap, 285,
397: 440
398:
399: \reference{}
400: Noguchi, K., et al. 2002, \pasj, 54, 855
401:
402: \reference{}
403: Norris, J.E., Ryan, S.G., Beers, T.C. \& Deliyannis, C.P. 1997, \apj, 485, 370
404:
405: \reference{}
406: Norris, J.E., Ryan, S.G., \& Beers, T.C. 2001, \apj, 561, 1034
407:
408: \reference{}
409: Novicki, M. 2005, Ph.D. Thesis, Univ. of Hawaii at Manoa
410:
411: \reference{}
412: Pinsoneault, M. H., Deliyannis, C.P. \& Demarque, P., 1992, \apjs, 78, 179
413:
414: \reference{}
415: Pinsoneault, M. H., Steigman, G., Walker, T.P. \& Narayanan, V.K. 2002, \apj,
416: 574, 398
417:
418: \reference{}
419: Pinsoneault, M. H., Walker, T.P , Steigman, G. \& Narayanan, V.K. 1999, \apj,
420: 527, 180
421:
422: \reference{}
423: Preston, G.W. \& Sneden, C. 2000, \aj, 120, 1014
424:
425: \reference{}
426: Primas, F., Asplund, M, Nissen, P.E. \& Hill, V. 2000a, \aap, 364, L42
427:
428: \reference{}
429: Primas, F., Asplund, M, Bonifacio, P. \& Hill, V. 2000b, \aap, 362, 666
430:
431: \reference{}
432: Pritchet, C.J. \& Glaspey, J.W. 1991, \apj, 373, 105
433:
434: \reference{}
435: Ryan, S.G., Norris, J.E. \& Beers, T.C. 1999, \apj, 523, 654
436:
437: \reference{}
438: Ryan, S.G., Beers, T.C., Olive, K.A., Fields, B.D. \& Norris, J.E. 2000, \apj,
439: 530, L57
440:
441: \reference{}
442: Ryan, S.G., Beers, T.C., Kajino, T. \& Rosolankova, K 2001a, \apj, 547, 231
443:
444: \reference{}
445: Ryan, S.G., Kajino, T., Beers, T.C., Suzuki, T., Romano, D., Matteucci, F. \&
446: Rosolankova, K 2001b, \apj,
447:
448: \reference{}
449: Ryan, S.G., Gregory, S.G., Kolb, U., Beers, T.C. \& Kajino, T. 2002, \apj,
450: 571, 501
451:
452: \reference{}
453: Shuster, W.J. \& Nissen, P.E. 1988, \aaps, 73, 225
454:
455: \reference{}
456: Sneden, C. 1973, Ph.D. thesis, Univ. of Texas at Austin
457:
458: \reference{}
459: Spite, F. \& Spite, M. 1982, \aap, 115, 357
460:
461: \reference{}
462: Spite, M., Maillard, J.P. \& Spite, F. 1984, \aap, 141, 56
463:
464: \reference{}
465: Spite, M., Molaro, P., Francois, P. \& Spite, F. 1993, \aap, 271, L1
466:
467: \reference{}
468: Stephens, A. \& Boesgaard, A.M. 2002, \aj, 123, 1647
469:
470: \reference{}
471: Stephens, A., Boesgaard, A.M., King, J.R. \& Deliyannis, C.P. 1997, \apj, 491,
472: 339
473:
474: \reference{}
475: Thorburn, J.A. 1992, \apj, 399, L83
476:
477: \reference{}
478: Thorburn, J.A. 1994, \apj, 421, 318
479:
480: \reference{}
481: Yi, S.K., Demarque, P. \& Kim, Y. 2004, \apss, 291, 261
482:
483: \end{references}
484:
485: \begin{figure}
486: \plotone{f1.ps}
487: \caption{The synthetic spectrum in the Be region for G 186-26. The small
488: squares are the data. The solid line corresponds to our line list predictions
489: for the Kurucz model for this star for no Be. The short dashed line
490: corresponds to the Be abundance expected from the PDD92 calculations ($-$1.58)
491: and the dotted line is a factor of 2 more Be and the dashed dotted is a factor
492: of 4 more Be. The upper dotted line is our upper limit, A(Be) = $-$2.00, but
493: see Figure 2 for more clarity. The lower part of the figure is a spectrum of
494: a star with normal Li and normal Be, BD +3 740 (displaced, but on the same
495: scale). G 186-26 and BD +3 740 have similar temperatures and metallicities.
496: For such metal-poor stars the stronger Be II line $\lambda$3130 is a better
497: abundance indicator than $\lambda$3131.}
498: \end{figure}
499:
500:
501: \begin{figure}
502: \plotone{f2.ps}
503: \caption{Like Figure 1 of G 186-26, but zoomed in on the Be II lines.}
504: \end{figure}
505:
506: \begin{figure}
507: \plotone{f3.ps}
508: \caption{The trend of Fe and Be showing the Be upper limit for G 186-26 as the
509: solid triangle with Li-normal stars. Open squares are from Primas et
510: al.~(2000a, 2000b); open circles are from Boesgaard et al.~(1999), Boesgaard
511: (2000), Stephens et al.~(1997).}
512: \end{figure}
513:
514: \begin{figure}
515: \plotone{f4.ps}
516: \caption{Radial velocities measured for G 186-26 by Carney et al.~(1994).
517: Latham et al.~2002 find no evidence for binarity in the star. The line is
518: simply the least squares fit through the data and is not meant to imply that
519: there is an increase in the radial velocity over time.}
520: \end{figure}
521:
522:
523: \end{document}
524:
525: