1:
2: \documentclass[12pt,preprint]{aastex}
3: %\usepackage{graphicx}
4: %\input epsf
5:
6: \newcommand{\euve}{{\it EUVE}}
7: \newcommand{\deltE}{\Delta\kern-1ptE}
8: \newcommand{\fuse}{{\it FUSE}}
9: \newcommand{\iue}{{\it IUE}}
10: \newcommand{\hst}{{\it HST}}
11: \newcommand{\kms}{km~s$^{-1}$}
12: \newcommand{\gal}{$\alpha$}
13: \newcommand{\gb}{$\beta$}
14: \newcommand{\gla}{$\lambda$}
15: \newcommand{\etal}{et al.}
16: \newcommand{\ecs}{erg cm$^{-2}$ s$^{-1}$}
17:
18: \begin{document}
19: \slugcomment{{\bf Astronomical Journal, 2007, (October), in press}}
20: \title{{\it HST} Observations of Chromospheres in Metal Deficient
21: Field Giants}
22:
23: \author{A. K. Dupree, and Timothy Q. Li}
24: \affil{Harvard-Smithsonian Center for Astrophysics, \\
25: Cambridge MA 02138}
26: \email{adupree@cfa.harvard.edu}
27:
28: \and
29: \author{Graeme H. Smith}
30: \affil{UCO/Lick Observatory, University of California, Santa Cruz, CA 95064}
31: \email{graeme@ucolick.edu}
32:
33:
34: \begin{abstract}
35: HST high resolution spectra of metal-deficient field giants more than
36: double the stars in previous studies, span $\sim$ 3
37: magnitudes on the red giant branch, and sample an abundance range
38: [Fe/H]=$-$1 to $-3$. These stars, in spite of their age and low
39: metallicity, possess chromospheric fluxes
40: of \ion{Mg}{2} ($\lambda$2800) that are within a factor of 4
41: of Population~I stars, and give signs of
42: a dependence on the
43: metal abundance at the lowest metallicities. The \ion{Mg}{2} k-line widths
44: depend on luminosity and correlate with metallicity.
45: Line profile asymmetries
46: reveal outflows that occur at lower luminosities ($M_V$~=~$-$0.8) than detected
47: in Ca K and H-alpha lines in metal-poor giants,
48: suggesting mass outflow occurs over a larger
49: span of the red giant branch than previously thought,
50: and confirming that the \ion{Mg}{2} lines are good
51: wind diagnostics. These results do not support a
52: magnetically
53: dominated chromosphere, but appear more consistent with some sort
54: of hydrodynamic, or acoustic heating of the outer atmospheres.
55:
56: \end{abstract}
57:
58: \keywords{stars:chromospheres; stars: winds, outflows; stars:
59: Population II; ultraviolet: stars}
60:
61: \section{Introduction}
62:
63: Red giant stars of the Galactic halo have extremes in a number of properties,
64: such as age, metal abundances, and space motion, that have made them useful
65: probes of the early evolution of the Galaxy. It is well established that
66: the chromospheric activity of stars decreases with increasing age on the main
67: sequence (e.g., Soderblom 1985; Soderblom \etal\ 1991, Dravins \etal\ 1993)
68: In addition, observations show that the fluxes in
69: chromospheric and transition region emission lines decrease as stars evolve
70: into the red giant phase (e.g., Schrijver 1987; Simon \& Drake 1989;
71: B\"{o}hm-Vitense 1992). Consequently,
72: red giants of the Galactic halo present us with the opportunity to study the
73: extreme limits of another phenomenon, namely stellar activity.
74:
75: Due to the faintness of even the closest Population II red giants, investigation
76: of their chromospheres has so far centered around three spectroscopic
77: features: emission in the wings of the H\gal\ line, in the cores of
78: the \ion{Ca}{2} H and K lines, and in the cores of the \ion{Mg}{2} {\it h}
79: and {\it k} lines. Metal-poor giants brighter than
80: $\log (L/L_{\odot}) \sim 2.5$ (Cacciari \etal\ 2004) in both the
81: halo field and globular clusters frequently exhibit emission in one or both
82: of the wings of the H\gal\ absorption line (e.g., Cohen 1976; Mallia \& Pagel
83: 1978; Cacciari \& Freeman 1983, Gratton \etal\ 1984; Smith \& Dupree 1988;
84: Bates \etal\ 1993; Kemp \& Bates 1995).
85: Lyons \etal\ (1996) conclude that $\approx 80$\% of a total of 68 stars from
86: five globular clusters with $\log (L/L_{\odot}) > 2.7$ exhibit H\gal\ emission.
87: The emission profiles are more often than not asymmetric with one wing being
88: stronger than the other. However the strengths of the emission change with
89: time.
90:
91: Next to the H\gal\ emission features, the most extensively studied
92: chromospheric diagnostics among Population II giants are the central emission
93: reversals in \ion{Ca}{2} H and K lines.
94: Spectroscopic studies of these emission components
95: in both halo field and globular cluster red giants have been reported by
96: Dupree \etal\ (1990a),
97: Smith \etal\ (1992), Dupree \& Smith (1995), and Cacciari \etal\ (2004).
98: The presence of \ion{Ca}{2} H and K emission is detected further down the
99: red giant branch of Population II stars than is emission in the wings of
100: H\gal. The profiles of the Ca emission cores of metal-poor giants are often
101: asymmetric, with the sense of the asymmetry changing from blue-dominant to
102: red-dominant with increasing luminosity on the giant branch.
103: Chromospheric emission in the H and K lines has also been documented among
104: metal-poor subdwarfs (Smith \& Churchill 1998), and in the case of at least
105: one such star, HD 103095, which was included in the Mount Wilson HK survey,
106: an activity cycle of period 7.3 yr has been detected (Baliunas \etal\ 1995).
107:
108: The asymmetries and time variability of the H\gal\ and \ion{Ca}{2} emission
109: profiles are consistent with the presence of outflows within the chromospheres
110: of metal-poor stars on the upper giant branch (e.g., Dupree \etal\ 1984; Mauas
111: \etal\ 2006). This evidence is supported by the onset of
112: blueshifts in the H\gal\ at luminosities slightly lower than the onset
113: of H\gal\ emission, and blueshifts in NaD line cores at higher
114: luminosities
115: (e.g., Peterson 1981; Shetrone 1994; Kemp \& Bates 1995;
116: Shetrone \& Keane 2000; Cacciari \etal\ 2004). Furthermore, wind
117: signatures have been found in the $\lambda$10830 \ion{He}{1} line
118: of a small number of Population II field giants (Dupree \etal\ 1992;
119: Smith \etal\ 2004).
120:
121: The \ion{Mg}{2} chromospheric emission lines, analyzed for cool stars of
122: approximately solar abundance by Cardini (2005), have been less studied among
123: metal-poor giants where metallicity effects can be addressed.
124: Work based on \iue\ observations of field giants has been
125: presented by Dupree \etal\ (1990b), while a \hst\ GHRS spectrum of HD 216143
126: was discussed by Smith \& Dupree (1998). The \ion{Mg}{2} emission
127: in red giants of the globular cluster NGC 6752 have been studied using both
128: \iue\ and \hst\ GHRS (Dupree \etal\ 1990a, 1994). Despite having
129: metallicities of a factor of 100 or more less than solar, the \ion{Mg}{2}
130: emission fluxes among the Dupree \etal\ (1990b) sample were in some cases
131: not substantially less than those of Population I giants and supergiants.
132: Cuntz \etal\ (1994) found that this relative insensitivity of emission
133: flux to metallicity can be accounted for by an acoustic-heating model for
134: red giant chromospheres.
135:
136: High-resolution \ion{Mg}{2} \hst\ spectra of nine metal-poor subdwarf stars
137: with $-2.2 < {\rm [Fe/H]} < -0.6$ were obtained by Peterson \& Schrijver
138: (1997). They found that at ${\rm [Fe/H]} < -1.0$
139: the emission fluxes were no more than a factor of
140: three less than that of the solar-like dwarf \gal\ Cen A. Their observations
141: suggest that the subdwarf progenitors of metal-poor halo red giants have
142: significant levels of chromospheric activity, perhaps even comparable to
143: quiet regions on the Sun. They suggested that these subdwarfs also have
144: acoustically-heated chromospheres.
145:
146: The available evidence shows therefore that metal-poor halo stars, despite
147: being of great age, do nonetheless have chromospheres with levels of activity
148: perhaps not very dissimilar from that of the quiet Sun. The strongest of this
149: evidence comes from the fluxes of the \ion{Mg}{2} {\it h} and {\it k} emission
150: lines, which have profiles that show more contrast with the photospheric
151: spectrum than either the \ion{Ca}{2} H-K or H\gal\ lines. Given that much of
152: the \ion{Mg}{2} spectroscopy of metal-poor red giants came from the \iue\
153: spacecraft, this is an area in which \hst\ archival data can contribute
154: substantially.
155:
156: \section{Data}
157:
158: High resolution spectra of the $\lambda$2800 \ion{Mg}{2} lines in metal-poor
159: red giants were taken from the \hst\ archives for the metal-poor red giants
160: listed in Table 1. The position of the stars on a color-magnitude diagram is
161: shown in Fig. 1, where the dataset spans more than 3 magnitudes for
162: stars on the red giant branch and asymptotic giant branch, and
163: abundances
164: are less than solar by factors of 10 to 1000. The
165: spectra for all but three of these stars were obtained during
166: \hst\ Cycle 8 as part of a study of the abundances of {\it r}-process heavy
167: elements in metal-poor red giants (Proposal 8342, PI J. Cowan). The E230M echelle
168: grating was used with {\it HST}-STIS to obtain spectra covering the wavelength
169: range $\lambda\lambda$2410-3070 with a resolution of 0.09 \AA\ at the
170: $\lambda$2800 \ion{Mg}{2} lines. The STIS spectrum of HD 122563 was
171: obtained with a similar setup as part of Program 8111 (PI C. Sneden). Abundance
172: analyses based on these spectra are discussed by Cowan \etal\ (2005).
173: By contrast, GHRS spectra of HD 6833 and HD 216143 were obtained from
174: the datasets of Proposals 6010 and 6511 (PIs J. Bookbinder and G. Smith
175: respectively). The \ion{Mg}{2} spectrum of HD 216143 from these data has
176: been previously discussed by Smith \& Dupree (1998).
177: Specific details of the individual spectra are given in Table 2.
178:
179: Spectroscopic metallicities are available in the literature for all of the
180: stars in Table 1. The [Fe/H] metallicities listed in column 9 come from three
181: different sources referred to in column 10, namely
182: Fulbright (2000), Johnson (2002), and McWilliam \etal\ (1995b), all of whom
183: employed high-resolution spectra. In the case of the latter two
184: of these references the quoted value of [Fe/H] is based on analysis of
185: \ion{Fe}{1} lines. To give a sense of the uncertainties in these metallicities,
186: the values of [Fe/H] from Cowan \etal\ (2005) are listed in column 11 for
187: comparison with the other literature values.
188:
189: The spectra are shown in a stacked presentation in Figures 2 and 3, aligned
190: according to a photospheric wavelength scale.
191: The zero-point for each spectrum (marked by broken lines)
192: is offset from that of the spectrum below it
193: by an amount equal to $2 \times 10^{-13}$ \ecs. In addition,
194: some of the spectra with the lowest flux levels have been further scaled by an
195: arbitrary factor. The expected positions of the
196: interstellar \ion{Mg}{2} {\it h} and {\it k} absorption
197: features are shown by vertical lines. In many cases these are shifted
198: well away from the stellar chromospheric emission profiles as a result of
199: large stellar radial velocities. Thus, the typical central
200: reversal of chromospheric emission line profiles becomes apparent in
201: these objects. However, for some stars (many of the stars in Figure 2 and
202: HD 115444 in Figure 3), the ISM absorption is superimposed
203: on the chromospheric lines, and in giants such as HD 122563 and HD 165195
204: significantly alters the apparent line asymmetry and affects the
205: observed line fluxes.
206:
207: In most cases the stars exhibit asymmetric {\it h} and {\it k} profiles with
208: the sense of the asymmetry being consistent between the two lines. Two notable
209: exceptions are HD 122956 and HD 6833. In the former case the profiles are
210: nearly symmetric and the degree of asymmetry is mild. In the case of HD 6833
211: the {\it k} line is pronouncedly asymmetric ($V/R < 1$) while the {\it h} line
212: shows a modest asymmetry in the opposite sense. By contrast, {\it IUE} spectra
213: obtained in 1984 showed similar {\it h} and {\it k} profiles (see Figure 2
214: of Dupree \etal\ 1990b). Evidently the asymmetry of this star is time
215: variable. In the
216: case of the star $\alpha$ Ori a self-absorbed \ion{Mn}{1} resonance line
217: absorbs the short wavelength emission of the \ion{Mg}{2} {\it k} line
218: (see Figure 5 of Lobel \& Dupree 2000) leading to an apparent asymmetry.
219: However, this effect is not likely to be the cause of the red asymmetry in
220: the \hst\ spectrum of the HD 6833 {\it k} line because this line is
221: narrower than for $\alpha$ Ori so that the \ion{Mn}{1} line lies outside the
222: emission core. The k line is formed at higher atmospheric levels than
223: the h line which could support different velocity fields in a dynamic atmosphere.
224:
225: \section{Emission Line Fluxes}
226:
227: The chromospheric fluxes in the \ion{Mg}{2} emission cores were measured
228: from the spectra by summing directly over the emission profiles between the
229: flux minima on either side without correcting for any
230: photospheric contribution to the line. While this procedure is straightforward,
231: it will overestimate the amount of chromospheric emission by including
232: some photospheric contribution near the line profile minima.
233: The photospheric \ion{Mg}{2} lines are extremely deep, however, practically
234: reaching the zero level near line center, and as such are expected to
235: contribute a relatively small amount to the derived fluxes.
236:
237: To investigate this point, R. Kurucz
238: kindly computed synthetic spectra for two of the hottest stars in our sample,
239: HD 6755 and HD 175305, for which the greatest photospheric fluxes
240: would be expected within the wavelength ranges encompassed by the emission
241: cores. Pure radiative models were constructed using the latest version
242: of Kurucz (2005) model atmospheres. Such models
243: contain an isothermal atmosphere in hydrostatic equilibrium extended
244: until the density falls to
245: negligible values. The assumption is that no additional heating is
246: present to produce a rise in atmospheric temperature. A
247: chromosphere is absent and the synthesized spectrum
248: represents only the photospheric contribution. The Mg and Ca
249: abundances
250: for these models were enhanced by
251: 0.3 dex over the Fe values based on the fine analysis by Fulbright (2000).
252: The atmospheric parameters (Carney \etal\ 2003) adopted for these
253: models were $T_{\rm eff} = 5080$K, [Fe/H] = --1.7, $\log g = 2.7$
254: (corresponding to $M_V=1.5$), and
255: $v_{\rm tot} \sin i = 3.5$ \kms\ for HD 6755, with
256: $T_{\rm eff}=5050$ K, [Fe/H]= --1.4, $\log g =2.8$
257: ($M_V=1.8$), and $v_{\rm rot}\sin i = 1.5$ \kms\ for HD~175305.
258: Although we have adopted a small rotational velocity for HD~6755
259: this star is a spectroscopic binary (Carney \etal\ 2003). However,
260: with a period of 1641 days
261: it shouldn't be rotating very fast, although there is the uncertainty of
262: whether the spectrum contains some light from the companion.
263:
264: The Kurucz photospheric models are overplotted on the observed \ion{Mg}{2}
265: spectra in Figures 4 and 5. In both cases, over the wavelength range
266: bounded by the flux minima on either side of the emission profile,
267: the net flux in the photospheric \ion{Mg}{2} lines is relatively small
268: by comparison with the integrated chromospheric emission.
269: Subtracting the contribution of
270: the synthesized photospheric flux model to the integrated \ion{Mg}{2}
271: emission amounts to a correction of
272: less than 12 percent. For the remainder of the stars, we fit
273: a local continuum, in the form of a second-order polynomial,
274: to the minima on each side of
275: the emission (the k$_1$ and h$_1$ locations). The difference between
276: the total flux measured from zero and the continuum-removed flux can amount to
277: 30\%. One star, HD 122563, is an exception where it is a factor
278: of 2.1 smaller due to the high continuum level (see Figure 2), and this
279: continuum subtracted flux is used here (Table 3).
280: However, this star, and four others have radial velocities
281: comparable to that of the local interstellar medium. As a result,
282: the observed emission is
283: reduced by interstellar \ion{Mg}{2} absorption. A previous study (Dupree
284: \etal\ 1990b) of the
285: effects of interstellar absorption on high latitude metal deficient stars with low
286: reddening, concluded that the stellar
287: flux could be increased by $\sim$25\% if interstellar absorption
288: were absent. We have marked these
289: stars in Table 3, and the fluxes are displayed in the figures as lower limits.
290:
291:
292: The {\it k} and {\it h} emission fluxes measured from the \hst\ spectra
293: are listed in columns 2 and 3 of Table 3.
294: The total observed flux $F_{obs}$ for the two lines combined is listed
295: in column 4.
296: These fluxes at Earth were converted to fluxes at the surface of each star
297: through use of the Barnes \etal\ (1978) relation to calculate the
298: stellar angular diameter $\phi$ (in units of 10$^{-3}$
299: arcsec) from the stellar $(V-R)_0$ color. These diameters are listed in
300: column 5 of Table 3. A stellar surface flux in units of \ecs\ was then
301: calculated from the relation
302: \begin{equation}
303: F_{\rm Mg\ II}=
304: F_{obs}(d/R_{\star})^2 = F_{obs}\times 1.702 \times 10^{17}/\phi^2,
305: \end{equation}
306: where $R_{\star}$ is the stellar radius,
307: and corrected for reddening according to the $E(B-V)$ values from Table 1 and
308: $A_{\lambda 2800} = 6.1 E(B-V)$ (Seaton 1979; Cardelli \etal\ 1989).
309: The values of $F_{\rm Mg\ II}$ are listed in column 6 of Table 3, while column 7
310: gives their ratio to the solar \ion{Mg}{2} emission flux.
311:
312:
313: The \ion{Mg}{2} fluxes of the metal-poor giants are much less than that of the
314: Sun as a star, but the relevant comparison for our purposes is to
315: other red giant stars of higher metallicity.
316: The observed \ion{Mg}{2} surface fluxes measured from the {\it HST} spectra
317: are plotted versus $(V-R)_0$ in Figure 6. This figure also includes data
318: from Dupree \etal\ (1990b) for metal-poor field giants observed
319: with {\it IUE} (Table 4), updated using values of $E(B-V)$ and [Fe/H] from
320: Pilachowski \etal\ (1996), McWilliam \etal\ (1995a,b), Fulbright \etal\ (2000),
321: Beers \etal\ (2000), and Burris \etal\ (2000). In addition, the fluxes
322: compiled by Dupree \etal\ for Population I supergiants, Hyades giants,
323: and M67 giant stars (Tables 5, 6 and 7) are also shown.
324: Dupree \etal\ (1990b) found that the surface fluxes of
325: metal-poor giants in their smaller sample were
326: comparable to those of near-solar abundance giants.
327: A similar effect can be seen in Figure 6 with most of the metal-poor
328: giants falling close to the Population~I sequence except for stars
329: near $(V-R)_0 \sim 0.8$. Four metal-poor
330: stars in this region have lower limits on their surface
331: fluxes. While the correction for interstellar \ion{Mg}{2}
332: absorption will increase the stellar flux, it does not appear likely
333: that the flux will increase by
334: 0.5 dex, or a factor of 3 or more to bring them up to the Population I
335: stars. Interestingly, the M67 giants, which are
336: about 0.2 mag redder in $(V-R)_0$ show a similar trend.
337:
338: Despite the similarity between the \ion{Mg}{2} surface fluxes of metal-poor
339: and Population~I giants,
340: these fluxes do show a correlation with [Fe/H] as
341: shown in Figure 7. In fact they appear to show less scatter in this figure
342: at a given [Fe/H] than in Figure 6 at a fixed $(V-R)_0$, suggesting that
343: the trend seen with [Fe/H] may be more than an artifact of the relation between
344: red giant color and metallicity.\footnote{In fact, a color-metallicity relation
345: coupled with the trend seen in Figure 6, would cause a more metal-poor (bluer)
346: star to have a higher surface flux than a metal-richer giant, the opposite of
347: what is seen in Figure 7.} Based upon this evidence, the giants which fall
348: below the mean sequence in Figure 6 do so as a result of this metallicity
349: effect. The most discrepant metal-poor stars in Figure 6 with
350: $(V-R)_0 < 0.87$ and \ion{Mg}{2} surface
351: fluxes less than $10^5$ \ecs\ are HD 6268, 115444, 122563,
352: 126587, and 186478. They all have metallicities of ${\rm [Fe/H]} < -2.5$.
353: Some of the scatter in Figures 6 and 7 might be due to variability. High
354: quality {\it IUE} spectra of cool stars have shown a difference up to
355: a
356: factor of 2 in fluxes of the \ion{Mg}{2} lines,
357: measured at different times when the expected error of measurement is
358: $\leq \pm 20$\% (Brocius \etal\ 1985; Dupree \etal\ 1987).
359:
360: \section{Line Profiles}
361:
362: \subsection{Asymmetries}
363:
364: The asymmetries of the \ion{Mg}{2} {\it k}-line profiles from the {\it HST}
365: spectra are shown in Figure~8 as a function of position in the color-magnitude
366: diagram using the values in Table 1. Stars are plotted with symbols
367: designating the $V/R$ parameter, which
368: refers to the ratio between the maximum intensities in the violet
369: (short-wavelength) and red (long-wavelength) peaks
370: of the {\it k}-line emission
371: profile. The 8 stars are plotted whose radial velocities place them away
372: from the local interstellar \ion{Mg}{2} absorption, the presence of which
373: could affect the asymmetry. The most luminous stars show outflow signatures,
374: in that the violet emission is less than the red emission. The long
375: broken line at $M_V = -1.7$ marks the lower
376: extent of the region where asymmetries of $V/R < 1$ and K$_3$
377: absorption reversal blue-shifts dominate in the \ion{Ca}{2}
378: K$_2$ emission-line profile as determined from the metal-poor field giant
379: spectroscopy of Smith \etal\ (1992) and Dupree \& Smith (1995).
380: In the globular cluster NGC 2808, which has a metallicity of [Fe/H] = --1.14
381: (Carretta \etal\ 2004), Cacciari \etal\ (2004) find a similar limit to
382: the $V/R < 1$ asymmetry of the Ca II K-line. The \ion{Mg}{2} emission is
383: formed higher in the extended stellar atmosphere, and is consequently
384: more sensitive to accelerating outflows than the emission lines formed in the
385: lower chromosphere. Population I giants show the same signature in
386: \ion{Mg}{2} at similar luminosities (Stencel \& Mullan 1980; B\"ohm-Vitense
387: 1981) indicated by the short-dashed line in Figure~8. Anthony-Twarog
388: \& Twarog (1994) evaluated the reddening and derived absolute
389: magnitudes for these stars from {\it uvby} photometry, and
390: their values are used in Figure~9. The
391: position of our target stars with respect to the Population I objects
392: remains similar to that in Figure~8.
393:
394: \subsection{Wilson-Bappu Emission Line Widths}
395:
396: Widths of the \ion{Mg}{2} {\it k} lines were measured following the precepts
397: of Cassatella \etal\ (2001). The observed width is taken to be the full
398: width of a fitted Gaussian profile (deleting the central absorption
399: reversal from the fit), measured at the half maximum value of the
400: observed line peak. Correction was made for the
401: instrumental broadening by assuming that the measured width is
402: a quadratic sum of the intrinsic stellar width $W_0$
403: and a resolution element. We omitted three stars
404: where the emission was very weak and the spectrum noisy
405: (HD 115444) or the interstellar absorption
406: seriously compromised the wing of the line (HD 6268 and HD 186478).
407: The line widths are plotted as a function of $M_V$ in Figure~10 using
408: the Bond (1980) data, and in Figure~11 for absolute
409: magnitudes from Anthony-Twarog \& Twarog (1994).
410: This is the \ion{Mg}{2} equivalent of the Wilson-Bappu (1957) relation
411: for metal-poor giants.
412: Also shown in these figures as a broken line is the mean Wilson-Bappu relation
413: for Population I giants as derived by Cassatella \etal\ (2001) for a large
414: sample of stars for which high-resolution {\it IUE} spectra and
415: {\it Hipparcos} parallaxes are available. The Cassatella \etal\ relation
416: (their equation 3) is
417: \begin{equation}
418: M_V =(34.56 \pm 0.29) -
419: (16.75 \pm 0.14) \log (W_0\phantom{i}{\rm km}\phantom{i}{\rm s}^{-1}).
420: \end{equation}
421: The dispersions in line width at
422: a given $M_V$ among the Cassatella \etal\ (2001) sample are shown by the
423: shaded regions in Figures~10 and 11, which depict the 50$-$67\% extent in the values of
424: $\log W_0$ among stars within bins of one-magnitude width.
425: The metal-poor red giants have {\it k} emission profiles that are
426: narrower than the bulk of the Cassatella \etal\ sample (and in some
427: cases, narrower than all the stars) and certainly more narrow than the mean relation for the sample.
428:
429: A similar situation was found for the
430: \ion{Ca}{2} K$_2$ lines of metal-poor red giants by Dupree \& Smith (1995).
431: By contrast, from their spectroscopy of red giants in NGC 2808, Cacciari
432: \etal\ (2004) found no evidence for a metallicity dependence of the
433: Wilson-Bappu effect in \ion{Ca}{2}~K$_2$. The stars in this {\it HST} sample
434: all have metallicities
435: much lower than NGC 2808 ([Fe/H]=$-$1.14, Carretta \etal\ 2004), so
436: that our sample would be more likely to reveal a metallicity dependence to
437: the Wilson-Bappu effect. The difference between the observed line
438: width $\log W_0$ and that given by
439: equation (2) is plotted against [Fe/H] in Figure 12 for the metal-poor giants
440: in the {\it HST} sample. This figure does reveal evidence of a correlation,
441: with the deviation from the Cassatella \etal\ relation being greater for
442: lower metallicities. These figures therefore provide evidence for a
443: metallicity dependence to the Wilson-Bappu effect for the \ion{Mg}{2}
444: {\it k} line. Since the line widths of the metal-poor giants are consistently
445: less than predicted by the Cassatella \etal\ (2001) relation for their absolute
446: magnitude, values of $M_V$ would be underestimated if based on the
447: observed \ion{Mg}{2}~{\it k} width.
448:
449: \section{Discussion}
450:
451: Results from the \hst\ spectra reported in this paper extend and confirm
452: those found previously (Dupree \etal\ 1990b) in showing
453: that the \ion{Mg}{2} chromospheric
454: emission fluxes of metal-poor Population II red giants are
455: comparable to or within a factor of $\sim 4$ of Population I
456: giants and supergiants
457: of similar $(V-R)_0$ color. This behavior, in spite of [Fe/H] ratios
458: that are factors of 10 to 1000
459: less than the Population I stars and that our sample is older,
460: suggests a different chromospheric structure and heating mechanisms from
461: Population I stars must be present in these halo red giants.
462: The \hst\ data do provide the first
463: evidence that the \ion{Mg}{2} emission fluxes correlate with [Fe/H],
464: at least for [Fe/H]~$< -2.0$ (Figure 7). At higher metallicities this trend
465: may flatten out although the number of stars in our sample is really too small
466: to be definitive on this point. This metallicity effect may provide an
467: explanation for the range in $\log F_{\rm Mg\ II}$ seen among stars with
468: $(V-R)_0 \sim 0.7$ to 1.0 in Figure 6. One limitation of the present \hst\
469: sample is that the color range is fairly restricted, so we cannot conclude very
470: much about metal-poor giants outside this range, except that the four such
471: stars in Figure 6 do superimpose on the Population I sequence.
472:
473: Although the most metal-poor stars in our sample have ${\rm [Mg/H]}< -2$
474: (such stars have ${\rm [Mg/Fe]} \leq 0.4$; Hanson \etal\ 1998), they
475: have \ion{Mg}{2} emission fluxes no less than about 0.6-0.7 dex that of
476: Population I giants and supergiants. Peterson \& Schrijver (1997)
477: reported similarly that for subdwarfs the \ion{Mg}{2} emission flux did not
478: scale strongly with metallicity. They suggested that subdwarfs may have
479: chromospheres similar to that of the quietest regions of the Sun. The picture
480: that emerges for metal-poor Population II stars is that their chromospheres
481: may evolve analogously to those of quiet Population I stars. The similarity in
482: $F_{\rm Mg\ II}$ between the M67 and the least active metal-poor giants in
483: Figure 6 may be consistent with this suggestion.
484:
485: The metal-poor giants in Figure 6 are both older and much more metal-deficient
486: than the evolved Population I stars to which they are compared. Magnetic
487: processes, direct or indirect, play a role in chromospheric heating of a star
488: like the Sun (Judge \etal\ 2003), but the contribution of heating
489: through dissipation of acoustic waves remains controversial (Cuntz \etal\ 2007;
490: Fossom \& Carlsson 2006;
491: Wedemeyer-B\"ohm \etal\ 2007). Since the stellar dynamo within a main sequence
492: star does decline with time it may be that in Population II stars the trends
493: in chromospheric \ion{Mg}{2} emission are being driven largely by an acoustic
494: heating of the chromosphere, as argued by Peterson \& Schrijver (1997). Variable
495: profile asymmetries of H\gal\ in metal deficient giants led Smith \& Dupree (1988)
496: to identify pulsation as a likely source of heating and momentum deposition.
497: A consideration here could involve the energy balance. If the energy
498: input were similar in both metal-rich and metal-poor atmospheres, the fact
499: that radiative cooling is decreased in the metal-poor chromospheres
500: could lead to a warmer atmosphere and increase the \ion{Mg}{2} emission,
501: thus compensating for the decreased Mg abundance.
502:
503:
504: Cuntz \etal\ (1994) have explained the insensitivity of the \ion{Mg}{2}
505: emission-line fluxes of cool giant stars to metallicity in terms of acoustic
506: shock-wave heated chromospheric models. These models account for a
507: basal, i.e., non-magnetic, component (e.g., Schrijver
508: 1987; Rutten \etal\ 1991) in the fluxes of the \ion{Mg}{2} and \ion{Ca}{2}
509: emission lines of late-type stars (Buchholz \etal\ 1998). Cuntz \etal\
510: computed the run of physical conditions with height within red giant
511: chromospheres of various metallicity into which the same acoustic flux is input
512: in the form of propagating shock waves. As the metallicity
513: of the chromosphere is reduced the bulk of the \ion{Mg}{2} {\it k} line
514: emission tends to be formed deeper at higher mass column densities.
515: Cuntz \etal\ (1994) computed the physical height in their models
516: at which the optical depth across the {\it k} line profile (computed
517: in plane-parallel geometry) was equal to unity
518: for metallicities ranging from $Z_{\odot}$ to $Z_{\odot}/100$.
519: Across this resulting range in heights the
520: temperature and velocity structure of their shock models did not vary greatly
521: with either metallicity or altitude. Thus even though the physical line
522: formation depth does vary with metallicity, the line was always formed under
523: conditions of similar temperature and
524: velocity structure. As a result, the emission flux was found not to vary
525: greatly with metallicity.
526:
527:
528: Despite the success of these acoustic models, magnetic heating may still play
529: a role in maintaining the chromospheres of Population II stars, at least
530: during some phases of their evolution. The evidence of the subdwarf HD 103095
531: ([Fe/H] = --1.2, Balachandran \& Carney 1996) suggests that
532: magnetic activity cycles may be present on the main sequence (Baliunas
533: \etal\ 1995), although with a color of $B-V = 0.75$ this
534: star is much cooler than the main sequence progenitors of the red giants in
535: Table 1. Main sequence turnoff stars with a metallicity of [Fe/H] = $-$2.3,
536: [$\alpha$/Fe] = 0.3, and ages of 12$-$14 Gyr have colors of
537: $(B-V)_0 \sim 0.325$-0.35 and effective temperatures of
538: $\log T_{\rm eff} {\rm (K)} \sim 3.83$ to 3.82 (see Figures 11 and 12 of
539: Bergbusch \& VandenBerg 2001), i.e., $T_{\rm eff} = 6760$-6610 K. These
540: are similar in temperature to low-mass Population I F2-F5 V stars (Bessell 1979).
541: Simon \& Landsman (1991) found that \ion{C}{2} 1335 \AA\ emission among
542: Population I F dwarfs earlier than spectral type F5 is uncorrelated with
543: rotation, supporting the idea that the chromospheres of A and early-F dwarfs
544: are heated by the shock dissipation of acoustic waves.
545: Thus the main sequence progenitors of the metal-poor giants in Table 1
546: are expected to have had acoustically-heated chromospheres.
547: When they evolve off the main sequence towards cooler effective temperatures
548: such stars may transition to chromospheres that have a dynamo contribution
549: to the heating, as suggested by Simon \& Drake (1989) for Population I giants.
550: Clump giants in the Hyades clearly give evidence for magnetic activity
551: (Baliunas \etal\ 1983). By the time that they evolve beyond the
552: luminosity of the horizontal branch, however, it may be that old metal-poor
553: giants have returned to chromospheres that are mainly heated by
554: acoustic processes.
555:
556: Among Population I giants evidence exists for both acoustic and magnetic
557: heating of the chromosphere. Observations of Population I G and K giants in
558: similar evolutionary states do show evidence for
559: correlations between chromospheric emission line
560: fluxes, X-ray fluxes, and rotation (Strassmeier \etal\ 1994; Gondoin 2005).
561: The evolution of X-ray emission with advancing evolution away from the main
562: sequence depends on stellar mass; for low-mass stars ($M < 1.5$ $M_{\odot}$)
563: the X-ray activity decreases with evolution away from the main sequence,
564: while for intermediate-mass giants
565: ($1.5 \phantom{i} M_{\odot} < M < 3.0 \phantom{i} M_{\odot}$)
566: the activity at first increases up to a spectral type of around K1, beyond
567: which it drops (Pizzolato \etal\ 2000; Gondoin 1999, 2005).
568: During the red giant phase of evolution of Population I stars up to
569: colors as red as $B-V \sim 1.2$, a dynamo may contribute to chromospheric
570: activity, but with advanced evolution to the higher luminosities and cooler
571: temperatures of the later K giants the radiative energy losses in the
572: \ion{Mg}{2} and other emission
573: lines can approach the ``basal'' levels observed among red giants by Schrijver
574: (1987) and Judge \& Stencel (1991), which they attribute to acoustic shock
575: heating. The evolution of \ion{Ca}{2} H and K emission fluxes among the red
576: giants of the solar-age open cluster M67 seems consistent with this scenario
577: (Dupree \etal\ 1999).
578: The chromospheric evolution of Population II giants may be analogous to those
579: of Population I, with the modification that the greater ages of the progenitors
580: of the former stars has reduced the magnetic-heating contribution, so that
581: the acoustic component is relatively more important.
582:
583: It will be important to use the next generation of HST instruments to
584: probe the onset of outflows to fainter magnitudes and access a
585: more homogeneous sample of stars represented by globular cluster
586: members.
587:
588:
589:
590: \acknowledgements
591: We are grateful to Chris Sneden who called our attention to these
592: spectra from his program and made them available for analysis. The SIMBAD
593: reference catalogue produced by the Centre de Donn\'ees astronomiques
594: de Strasbourg was extremely useful. The
595: MAST Archive at STScI provided the HST spectra.
596: This work was supported in part by NASA.
597:
598: {\it Facility:} \facility{HST (GHRS, STIS)}, \facility{IUE}
599:
600:
601:
602:
603: \begin{thebibliography}{}
604:
605:
606: \bibitem[]{}Anthony-Twarog, B. J., \& Twarog, B. A. 1994, \aj, 107,
607: 1577 [AT94]
608:
609: \bibitem[]{587}Balachandran, S. C., \& Carney, B. W. 1996, \aj, 111, 946
610:
611: \bibitem[]{589}Baliunas, S. L., Hartmann, L., \& Dupree, A. K. 1983, \apj,
612: 271, 672
613:
614: \bibitem[]{592}Baliunas, S. L., \etal\ 1995, \apj, 438, 269
615:
616: \bibitem[]{594}Barnes, T. G., Evans, D. S., \& Moffett, T. J. 1978, \mnras, 183,
617: 285
618:
619: \bibitem[]{597}Bates, B., Kemp, S. N., \& Montgomery, A. S. 1993, \aap, 97, 937
620:
621: \bibitem[]{599}Beers, T. C., Chiba, M., Yoshii, Y., Platais, I., Hanson, R. B.,
622: Fuchs, B., \& Rossi, S. 2000, AJ, 119, 2866
623:
624: \bibitem[]{602}Bergbusch, P. A., \& VandenBerg, D. A. 2001, \apj, 556, 322
625:
626: \bibitem[]{604}Bessell, M. S. 1979, \pasp, 91, 589
627:
628: \bibitem[]{606}B\"{o}hm-Vitense, E. 1981, \apj, 244, 504
629:
630: \bibitem[]{608}B\"{o}hm-Vitense, E. 1992, \aj, 103, 608
631:
632: \bibitem[]{610}Bond, H. E. 1980, \apjs, 44, 517 [B80]
633:
634: \bibitem[]{612}Brosius, J. W., Mullan, D. J., \& Stencel, R. E. 1985,
635: \apj, 288, 310
636:
637: \bibitem[]{615}Buchholz, B., Ulmschneider, P., \& Cuntz, M. 1998, \apj, 494, 700
638:
639: \bibitem[]{617}Burris, D. L., Pilachowski, C. A., Armandroff, T. E.,
640: Sneden, C., Cowan, J. J., \& Roe, H. 2000, \apj, 544, 302
641:
642: \bibitem[]{620}Cacciari, C., \& Freeman, K. C. 1983, \apj, 268, 185
643:
644: \bibitem[]{622}Cacciari, C., Bragaglia, A., Rossetti, E., Fusi Pecci, F.,
645: Mulas, G., Carretta, E., Gratton, R. G., Momany, Y., \& Pasquini, L.
646: 2004, \aap, 413, 343
647:
648: \bibitem[]{626}Cardelli, J. A., Clayton, G. C., \& Mathis, J. S. 1989,
649: \apj, 345, 245
650:
651: \bibitem[]{629}Cardini, D. 2005, \aap, 430, 303
652:
653: \bibitem[]{631}Carney, B. W. 1980, \aj, 85, 38
654:
655: \bibitem[]{633}Carney, B. W., Latham, D. W., Stefanik, R. P., Laird,
656: J. B., \& Morse, J. A. 2003, AJ, 125, 293
657:
658: \bibitem[]{636}Carretta, E., Bragaglia, A., \& Cacciari, C. 2004, \apj, 610, L25
659:
660: \bibitem[]{638}Cassatella, A., Altamore, A., Badiali, M., \& Cardini, D.
661: 2001, \aap, 374, 1085
662:
663: \bibitem[]{641}Cohen, J. G. 1976, \apj, 203, L127
664:
665: \bibitem[]{}Cousins, A. W. J. 1976, Mem. R. A. S., 81, 25
666:
667: \bibitem[]{643}Cowan, J. J., Sneden, C., Beers, T. C., Lawler, J. E., Simmerer,
668: J., Truran, J. W., Primas, F., Collier, J., \& Burles, S. 2005, \apj, 627, 238
669:
670: \bibitem[]{646}Cuntz, M., Rammacher, W., \& Ulmschneider, P. 1994, \apj, 432, 690
671:
672: \bibitem[]{648}Cuntz, M., Rammacher, W., \& Musielak, Z. E. 2007, \apj, 657, L57
673:
674: \bibitem[]{650}Dravins, D., Linde, P., Ayres, T. R., Linsky, J. L.,
675: Monsignori-Fossi, B., Simon, T., \& Wallinder, F. 1993, \apj, 403, 412
676:
677: \bibitem[]{653}Dupree, A. K., \& Smith, G. H. 1995, \aj, 110, 405
678:
679: \bibitem[]{655}Dupree, A. K., Hartmann, L., \& Avrett, E. H. 1984, \apj,
680: 281, L37
681:
682: \bibitem[]{658}Dupree, A. K., Baliunas, S. L., Guinan, E. F., Hartmann,
683: L., Nassiopoulos, G. E., \& Sonneborn, G. 1987, \apj, 317, L85
684:
685: \bibitem[]{661}Dupree, A. K., Harper, G. M., Hartmann, L., Jordan, C., Rodgers,
686: A. W., \& Smith, G. H. 1990a, \apj, 361, L9
687:
688: \bibitem[]{664}Dupree, A. K., Hartmann, L., \& Smith, G. H. 1990b, \apj,
689: 353, 623
690:
691: \bibitem[]{667}Dupree, A. K., Sasselov, D. D., \& Lester, J. B. 1992, \apj, 387, L85
692:
693: \bibitem[]{669}Dupree, A. K., Hartmann, L., Smith, G. H., Rodgers, A. W.,
694: Roberts, W. H., \& Zucker, D. B. 1994, \apj, 421, 542
695:
696: \bibitem[]{672}Dupree, A. K., Whitney, B. A., \& Pasquini, L. 1999, \apj, 520,
697: 751
698:
699: \bibitem[]{675}Fossum, A., \& Carlsson, M. 2006, \apj, 646, 579
700:
701: \bibitem[]{677}Fulbright, J. P., 2000, \aj, 120, 1841
702:
703: \bibitem[]{679}Gondoin, P. 1999, \aap, 352, 217
704:
705: \bibitem[]{681}Gondoin, P. 2005, \aap, 444, 531
706:
707: \bibitem[]{}Gonzalez, G., \& Wallerstein, G. 1998, \aj, 116, 765
708:
709: \bibitem[]{683}Gratton, R. G., Pilachowski, C. A., \& Sneden, C. 1984, \aap, 132,
710: 11
711:
712: \bibitem[]{686}Hanson, R. B., Sneden, C., Kraft, R. P., \& Fulbright, J.
713: 1998, AJ, 116, 1286
714:
715: \bibitem[]{}Hartmann, L., Dupree, A. K., \& Raymond, J. C. 1982, \apj,
716: 252, 214
717:
718: \bibitem[]{}Hartmann, L., Jordan, C., Brown, A., \& Dupree,
719: A. K. 1985, \apj, 296, 576
720:
721: \bibitem[]{}Janes, K. A., \& Heasley, J. N. 1988, \aj, 95, 762
722:
723: \bibitem[]{}Janes, K. A., \& Smith, G. H. 1984, AJ, 89, 487
724:
725:
726: \bibitem[]{}Johnson, H. L., \& Sandage, A. R. 1955, \apj, 121, 616
727:
728: \bibitem[]{}Johnson, H. L., Iriarte, B., Mitchell, R. I., \&
729: Wisniewskj, W. Z. 1966, Comm. Lunar Planet. Lab., 4, 99
730:
731:
732: \bibitem[]{689}Johnson, J. A. 2002, ApJS, 139, 219
733:
734: \bibitem[]{691}Judge, P. G., \& Stencel, R. E. 1991, \apj, 371, 357
735:
736: \bibitem[]{}Judge, P. G., Carlsson, M., \& Stein, R. F. 1983, \apj, 597, 1158
737:
738: \bibitem[]{693}Kemp, S. N., \& Bates, B. 1995, \aaps, 112, 513
739:
740: \bibitem[]{695} Kurucz, R. L. 2005, Mem. Soc. Astron. Ital. Supp., 8, 14
741:
742: \bibitem[]{697}Lobel, A., \& Dupree, A. K. 2000, \apj, 545, 454
743:
744: \bibitem[]{699}Lyons, M. A., Kemp, S. N., Bates, B., \& Shaw, C. R. 1996, \mnras,
745: 280, 835
746:
747: \bibitem[]{702}Mallia, E. A., \& Pagel, B. E. J. 1978, \mnras, 184, P55
748:
749: \bibitem[]{704}Mauas, P. J. D., Cacciari, C., \& Pasquini, L. 2006, \aap, 454,
750: 609
751:
752: \bibitem[]{}McWilliam, A., Preston, G. W., Sneden, C., \& Shectman,
753: S. 1995a, AJ, 109, 2736
754:
755: \bibitem[]{707}McWilliam, A., Preston, G. W., Sneden, C., \& Searle, L. 1995b, \aj,
756: 109, 2757
757: \bibitem[]{}Murray, C. A., \& Clements, E. D. 1968, Royal Obs. Bull.,
758: No. 139, 309
759:
760: \bibitem[]{711}Peterson, R. C. 1981, \apj, 248, L31
761:
762: \bibitem[]{713}Peterson, R. C., \& Schrijver, C. J. 1997, \apj, 480, L47
763:
764: \bibitem[]{715}Pilachowski, C., Sneden, C., \& Kraft, R., 1996, \aj, 111, 1689
765:
766: \bibitem[]{717}Pizzolato, N., Maggio, A., \& Sciortino, S. 2000, \aap, 361, 614
767:
768: \bibitem[]{}Ridgway, S. J. Joyce, R. R., White, N. M., \& Wing,
769: R. F. 1980, \apj, 235,126
770:
771: \bibitem[]{719}Rutten, R. G. M., Schrijver, C. J., Lemmens, A. F. P., \& Zwaan, C.
772: 1991, \aap, 252, 203
773:
774: \bibitem[]{724}Sandage, A. 1970, \apj, 162, 841
775:
776: \bibitem[]{}Sanders, W. L. 1977, \aaps, 27, 89
777:
778: \bibitem[]{722}Schrijver, C. J. 1987, \aap, 172, 111
779:
780: \bibitem[]{726}Seaton, M. J. 1979, \mnras, 187, 73p
781:
782: \bibitem[]{728}Shetrone, M. D. 1994, \pasp, 106, 161
783:
784: \bibitem[]{730}Shetrone, M. D., \& Keane, M. J. 2000, \aj, 119, 840
785:
786: \bibitem[]{732}Simon, T., \& Drake, S. A. 1989, \apj, 346, 303
787:
788: \bibitem[]{734}Simon, T., \& Landsman, W. 1991, \apj, 380, 200
789:
790: \bibitem[]{}Simon, T., Linsky, J. L., \& Stencel, R. E. 1982, \apj,
791: 257, 225
792:
793: \bibitem[]{736}Smith, G. H., \& Churchill, C. W. 1998, \mnras, 297, 388
794:
795: \bibitem[]{738}Smith, G. H., \& Dupree, A. K. 1988, \aj, 95, 1547
796:
797: \bibitem[]{740}Smith, G. H., \& Dupree, A. K. 1998, \aj, 116, 931
798:
799: \bibitem[]{}Smith, G. H., \& Janes, K. A. 1988 in {\it A Decade of UV
800: Astronomy with IUE}, ed. E. J. Rolfe (ESA SP-281, Vol. 2), 193
801:
802: \bibitem[]{742}Smith, G. H., Dupree, A. K., \& Churchill, C. W. 1992, \aj,
803: 104, 2005
804:
805: \bibitem[]{745}Smith, G. H., Dupree, A. K., \& Strader, J. 2004, \pasp, 116, 819
806:
807: \bibitem[]{747}Soderblom, D. R. 1985, \aj, 90, 2103
808:
809: \bibitem[]{749}Soderblom, D. R., Duncan, D. K., \& Johnson, D. R. H. 1991, \apj,
810: 375, 722
811:
812: \bibitem[]{}Steiman-Cameron, T. Y., Johnson, H. R., \& Honeycutt,
813: R. F. 1985, \apj, 291, L51
814:
815: \bibitem[]{752}Stencel, R., E., \& Mullan, D. J. 1980, \apj, 240, 718
816:
817: \bibitem[]{}Stencel, R., E., Mullan, D. J., Linsky, J. L., Basri,
818: G. S., \& Worden, S. P. 1980, \apjs, 44, 383
819:
820: \bibitem[]{754}Stone, R. P. S. 1983, \pasp, 95, 27
821:
822: \bibitem[]{756}Strassmeier, K. G., Handler, G., Paunzen, E., \& Rauth, M. 1994,
823: \aap, 281, 855
824:
825: \bibitem[]{}Taylor, B. J. 2007, \aj, 133, 370
826:
827:
828: \bibitem[]{759}Wedemeyer-B\"ohm, S., Steiner, O., Bruls, J., \&
829: Rammacher, W. 2007 in {\it The Physics of Chromospheric Plasmas}, ASP
830: Conf. Ser. 368, eds. P. Heinzel, I. Dorotovi\v c, \& R. J. Rutten, 93
831: (astro-ph/0612627)
832:
833: \bibitem[]{763}Wilson, O. C., \& Bappu, M. K. V. 1957, \apj, 125, 661
834:
835: \end{thebibliography}{}
836:
837: \clearpage
838:
839: %FIGURE 1
840: \begin{figure}
841: %\includegraphics[angle=0,scale=0.8]{f1.eps}
842: \plotone{f1.eps}
843: \caption{Color-magnitude diagram showing the location of the stars
844: discussed in this paper. Solid lines mark the
845: location of the red giant branch, asymptotic giant branch, and subgiant
846: branch in the globular cluster M92 according to Sandage (1970).}
847: \end{figure}
848:
849: \clearpage
850:
851: %FIGURE 2
852: \begin{figure}
853: %\includegraphics[angle=0,scale=0.8]{f2.eps}
854: \epsscale{.9}
855: \plotone{f2.eps}
856: \epsscale{1}
857: \caption{\small Chromospheric emission profiles of the
858: \ion{Mg}{2} {\it h} and {\it k} lines for stars listed in Table 1.
859: The spectra are arranged from top to bottom in order of decreasing
860: absolute visual magnitude. HD 22170 is the brightest star in the
861: sample at M$_V$ = $-$1.9 and HD 6268 has M$_V$ = $-$1.2.
862: The spectra are aligned on a photospheric wavelength scale.
863: Vertical lines mark the positions of interstellar medium
864: absorption. The flux zero-points are offset by $2 \times 10^{-13}$ between
865: adjacent spectra, and the spectra of two stars are shown magnified by a
866: factor of 2 or 4. The broken line marks the zero level for each spectrum.}
867: \end{figure}
868:
869: \clearpage
870:
871: %FIGURE 3
872: \begin{figure}
873: %\includegraphics[angle=0,scale=0.8]{fig3_dupree.eps}
874: \plotone{f3.eps}
875: \caption{Chromospheric emission profiles for the remaining stars
876: from Table 1, with details the same as Fig. 2. Here HD 115444 has
877: M$_V$ = $-$1.1
878: and HD 175305 and HD 6755 are the two faintest stars in our sample
879: at M$_V$ = $+$1.2 and M$_V$ = $+$1.5 respectively. }
880: \end{figure}
881:
882: \clearpage
883:
884: %FIGURE 4
885: \begin{figure}
886: \includegraphics[angle=90,scale=0.6]{f4.eps}
887: %\plotone{f4.eps}
888: \caption{Kurucz model synthesis of a photospheric spectrum for HD 6755
889: overplotted on the observed {\it HST STIS} spectrum. The agreement between
890: the two spectra is quite good outside of the chromospheric emission profiles.
891: The interstellar \ion{Mg}{2} lines are marked as IS.}
892: \end{figure}
893:
894: \clearpage
895:
896: %FIGURE 5
897: \begin{figure}
898: \includegraphics[angle=90,scale=0.6]{f5.eps}
899: %\plotone{f5.eps}
900: \caption{Kurucz model synthesis of a photospheric spectrum for HD 175305
901: overplotted on the {\it HST STIS} spectrum. IS marks the interstellar
902: \ion{Mg}{2} absorption lines.}
903: \end{figure}
904:
905: \clearpage
906:
907: %FIGURE 6
908: \begin{figure}
909: \includegraphics[angle=90,scale=0.8]{f6.eps}
910: %\plotone{f6.eps}
911: \caption{Surface \ion{Mg}{2} fluxes from the {\it HST} sample of this paper
912: versus $(V-R)_0$ color. Also shown are data for metal
913: deficient giants reported previously, and Population I field and cluster
914: giants, from Dupree \etal\ (1990b). The length of the arrows for lower
915: limits to the flux represents a 30\% increase in the flux.}
916: \end{figure}
917:
918: \clearpage
919:
920: %FIGURE 7
921: \begin{figure}
922: %\includegraphics[angle=0,scale=0.8]{f7.eps}
923: \plotone{f7.eps}
924: \caption{Surface \ion{Mg}{2} fluxes of the metal deficient giant stars in
925: this sample from {\it HST} and those studied previously with {\it IUE}
926: (Dupree \etal\ 1990b) as a function of metallicity.
927: The error bar shown corresponds to a typical expected error of
928: $\pm 30\%$. Several of the IUE targets had flux errors of $\pm 50\%$ and
929: those are plotted over the symbol. Stars whose profiles may be affected by
930: interstellar \ion{Mg}{2} absorption are shown as lower limits, and the
931: length of the arrow
932: represents
933: a 30\% increase in the flux. }
934: \end{figure}
935:
936: \clearpage
937:
938: %FIGURE 8
939: \begin{figure}
940: %\includegraphics[angle=0,scale=0.9]{f8.eps}
941: \plotone{f8.eps}
942: \caption{Asymmetry of the \ion{Mg}{2} {\it k}-line emission
943: derived from the {\it HST} spectra
944: as a function of position in the color-magnitude diagram. The
945: short-dashed line marks the boundary of \ion{Mg}{2} asymmetries in Population
946: I stars (Stencel \& Mullan 1980; B\"ohm-Vitense 1981); signatures of
947: outflow occur towards higher (V$-$R) values.
948: The long-dashed line marks the lower limit
949: of the region where outflow asymmetries ($V/R < 1$)
950: and K$_3$-line blueshifts were found in the profile of the
951: \ion{Ca}{2} K$_2$ line of metal-poor red giants by
952: Smith \etal\ (1992), Dupree \& Smith (1995), and Cacciari \etal\
953: (2004).
954: Values of $M_V$ and reddening are taken from Bond (1980).}
955: \end{figure}
956:
957: \clearpage
958:
959: %FIGURE 9
960: \begin{figure}
961: \plotone{f9.eps}
962: \caption{Same as Fig. 8, but using the values of $M_V$ and
963: reddening from
964: Anthony-Twarog \& Twarog (1994) where $E(V-R)=1.1E(b-y)$
965: except for HD 6833 where the Bond (1980) values are assumed. The
966: placement of the stars with respect to the asymmetries found in
967: Population I stars is similar to that using the Bond (1980)
968: values.}
969: \end{figure}
970: \clearpage
971:
972: %FIGURE 10
973: \begin{figure}
974: %\includegraphics[angle=0,scale=0.8]{f9.eps}
975: \plotone{f10.eps}
976: \caption{Wilson-Bappu width $W_0$ of the \ion{Mg}{2} {\it k} line
977: for the metal-poor giants observed by {\it HST} (filled
978: circles). The shaded region indicates the
979: middle 50\% to 67\% of the Cassatella \etal\ sample
980: in $\log W_0$ for the stars in one-magnitude $M_V$-bins between
981: $M_V$=$\pm$3. The horizontal lines mark the total extent of the
982: measurements of line width, again
983: for one-magnitude $M_V$-bins. The broken line represents the relation obtained by
984: Cassatella \etal\ (2001) for a large sample of stars observed
985: with {\it IUE} and {\it Hipparcos}, including additional brighter and
986: fainter stars not shown in this figure. The metal-deficient giants in our
987: sample, here using the values of $M_V$ from Bond (1980)
988: systematically exhibit smaller values of $\log W_0$. }
989: \end{figure}
990:
991: \clearpage
992: %FIGURE 11
993: \begin{figure}
994: \plotone{f11.eps}
995: \caption{Same as Fig. 10 but here using the $M_V$ values
996: obtained by Anthony-Twarog \& Twarog (1994) except for
997: HD 6833 where the Bond (1980) values are used. The line widths
998: are systematically smaller than found for Population I field
999: stars.}
1000:
1001: \end{figure}
1002:
1003: \clearpage
1004:
1005: %FIGURE 12
1006: \begin{figure}
1007:
1008: \plotone{f12.eps}
1009: \caption{Difference between the observed \ion{Mg}{2} 2795 \AA\ line
1010: width and that predicted by the Cassatella \etal\ (2001) relation
1011: as a function of [Fe/H] for the metal-poor stars in the {\it HST}
1012: sample
1013: and selecting the Bond (1980) values (Fig. 10).
1014: The trend here suggests that the deviation is metal dependent.}
1015: \end{figure}
1016:
1017: \clearpage
1018:
1019: %TABLES
1020: \input{tab1.tex}
1021:
1022: \clearpage
1023:
1024: \input{tab2.tex}
1025:
1026: \clearpage
1027:
1028: \input{tab3.tex}
1029: \clearpage
1030:
1031:
1032: \input{tab4.tex}
1033:
1034: \clearpage
1035: \input{tab5.tex}
1036:
1037: \clearpage
1038: \input{tab6.tex}
1039:
1040:
1041: \clearpage
1042: \input{tab7.tex}
1043:
1044: \end{document}
1045:
1046:
1047:
1048:
1049: