astro-ph0201355/ms.tex
1: %\documentclass[12pt,preprint]{aastex}
2: \documentclass{article}
3: \usepackage{emulateapj}
4: 
5: 
6: \slugcomment{To appear in ApJL}
7: 
8: %\shorttitle{Convective Wavelength Shifts}
9: %\shortauthors{Allende Prieto et al.}
10: 
11: \begin{document}
12: \title{Convective Wavelength Shifts in the Spectra of Late-Type Stars}
13: 
14: \author{Carlos Allende Prieto, David L. Lambert, Robert G. Tull, 
15: and Phillip J. MacQueen}
16: \affil{McDonald Observatory and Department of Astronomy \\ University of Texas
17:   \\ RLM 15.308, Austin, TX 78712-1083, \\ USA}
18:   
19: %\email{callende,dll,rgt,pjm@astro.as.utexas.edu}
20: 
21: 
22: \begin{abstract}
23: 
24: 
25: We present  ultra-high resolution  spectra for a set of
26: nearby F-G-K stars on, or close to, the main sequence. The wavelength shifts
27:  of stellar lines relative to their laboratory wavelengths 
28:  are measured for more than a thousand Fe {\sc i} lines per star, 
29:  finding a clear correlation
30:  with line depth. The observed patterns are  interpreted as
31:  convective blue-shifts that become more prominent for weaker lines,
32:  which are formed in deeper atmospheric layers. A morphological 
33:  sequence with spectral type
34:  or effective temperature is apparent.
35:  Two K giant stars have also been studied. The velocity span between 
36: weak and strong lines  for these stars is larger  than for the 
37:  dwarfs and subgiants of similar spectral types.
38: % which may simply reflect larger convective velocities due to lower densities. 
39:  Our results show that  convective 
40:   wavelength shifts may seriously compromise the accuracy of 
41:   absolute spectroscopic radial velocities, but  that an empirical 
42:   correction may be applied to measured velocities. 
43:   
44: % {\it Subject headings:   convection 
45: % --- line: formation 
46: % --- stars: atmospheres
47: % --- techniques: radial velocities 
48: % --- techniques: spectroscopic  
49: %}
50: 
51: \end{abstract}
52: 
53: 
54: \section{Introduction}
55: 
56: Convection in late-type stars penetrates into the photosphere, 
57: producing inhomogeneities. Temperature variations 
58: of up to a few hundred Kelvin 
59: are apparent in optical images of the Sun, 
60: and a velocity field of  several kilometers per second 
61: is observed in spatially resolved 
62:  solar spectra. The 
63: warmer upflows appear as bright granules surrounded by narrower, cooler
64: downflows. While present technology cannot  resolve stellar granulation,
65: there is evidence  of its presence in
66: all late-type stars. The  
67: temperature and velocity inhomogeneities produce absorption
68: line profiles which are shifted and asymmetric, effects 
69: that are readily noticed in  ultra-high dispersion stellar spectra.
70: 
71: An extensive literature exists on solar line asymmetries and shifts
72: (see, e.g., Neckel \& Labs 1990; Asplund et al. 2000; Pierce \& Lopresto 2000).
73: Modern stellar studies were pioneered by Dravins (1974, 1985) 
74: and Gray (1980, 1981). Dravins (1999) provides a short review of recent work.
75: Line asymmetries and line shifts are consequences of the same
76: phenomenon and, therefore, give largely equivalent information. When the
77: number of lines measured is limited, but the available features are mostly
78: unblended, line asymmetries carry  more information. On the contrary, if
79: the spectrum is heavily crowded, line shifts, which can be measured
80: for many lines, are likely to provide 
81:  more information.
82: A limited wavelength coverage, largely dictated by the detector size and
83: the scarcity of high-accuracy laboratory wavelengths, steered stellar 
84: studies toward the analysis of line asymmetries.
85: 
86: Line bisectors -- a measure of the line asymmetry -- vary smoothly
87:  across the HR 
88: diagram (Gray \& Toner 1986; Dravins 1987; Gray \& Nagel 1989; 
89: Dravins \& Nordlund 1990). Other parameters, such as
90: rotation, chemical composition, magnetic fields, or binarity, are likely to
91: play a role, but the limited data available have prevented an in-depth study.
92: A recent comprehensive work on the Fe {\sc i}
93:  spectrum  (Nave et al.
94: 1994) has provided accurate laboratory wavelengths for 9501 lines. 
95: These data and 
96:  improvements in astronomical instrumentation make it feasible and practical 
97:  to turn to line shifts
98: as a complement to line bisectors. 
99: 
100: Gravitational and convective wavelength shifts systematically affect absolute 
101: determinations of radial velocities.
102: % Correlations between radial velocity and
103: %spectral type have been reported repeatedly in the literature 
104: %(see Madsen, Dravins, 
105: %\& Lindegren 2002).
106:  Other effects can also introduce systematic radial velocity 
107: errors, but they are expected to be  less important for most stars 
108: (Lindegren, Dravins, \&
109: Madsen 1999). Gravitational redshifts are proportional
110: to the mass-to-radius ratio. For spectral lines formed in the photospheres of
111: dwarf stars, the gravitational shift is the same for all lines, 
112: in the range between 0.4 and 2 km s$^{-1}$. 
113: When accurate parallaxes and photometry
114: are available, comparison with evolutionary models allows us to determine dwarf
115: masses and radii within 8 \% and 6 \% , respectively 
116: (Allende Prieto \& Lambert 1999).
117: It is therefore possible to estimate the gravitational shift
118: within 10 \%, yielding worst-case uncertainties of $\sim 0.2$ km s$^{-1}$.
119: Convective line shifts may vary with spectral type. 
120: The velocity difference between weak and strong lines is about
121:  $\sim 0.6$ km s$^{-1}$ for the Sun (Allende Prieto \& Garc\'{\i}a L\'opez
122: 1998) and probably more than 1 km s$^{-1}$ for the F5 subgiant  $\alpha$ CMi
123: (Procyon; Allende Prieto et al. 2002), but  no systematic study across
124: the HR diagram has been published. Obviously, accurate spectroscopic studies of
125: absolute radial velocities 
126: cannot afford to  neglect these shifts. 
127: 
128: In this paper, we analyze optical 
129: spectra of several late-type stars  
130: obtained with the High Resolution Spectrograph (HRS) coupled to the Hobby-Eberly
131: Telescope (HET). The high quality and large spectral coverage of the spectra
132: allow us to measure line shifts for a large number of 
133: Fe {\sc i} lines. Section \ref{obs} describes the observations and the data. In 
134: Section \ref{doit} we  describe the analysis,
135:  and in \S \ref{end} we discuss the results.
136: 
137: 
138: \section{Observations}
139: \label{obs}
140: 
141: Our observations were obtained with the 9.2m HET at 
142: McDonald Observatory and the HRS (Tull 1998).
143: The HRS saw first light on  March 17, 2001, starting science observations
144: shortly afterwards. This fiber-coupled echelle spectrograph uses an R-4
145: echelle, cross-dispersing gratings, and a mosaic of two 2k$\times$4k CCDs.
146: %The HRS was used with a 3.2$\times$0.25 arcseconds slit that was 
147: %fed by a 2 arcseconds fiber.
148:  In the selected mode, HRS provides a 
149: FWHM resolving power 
150: $R \simeq 120,000$ and complete spectral coverage between 
151: 4094 and 7890 \AA, with the exception
152: of the  gap (5977--5999 \AA)  between the CCDs.
153: The observations were performed in {\it queue} mode in May and 
154: June 2001 (see Table 1). An  observation consisted of two 
155: exposures that were extracted and combined. 
156: The final signal-to-noise (S/N) per pixel was $\ge 400$ at most wavelengths.
157: 
158: The spectra were processed with the tasks in the IRAF {\it echelle} package. 
159: We removed the bias signal, flatfielded and extracted 
160: the spectra,  subtracting the low background  in the wide 
161: interorder regions. 
162: %The spectra were wavelength calibrated 
163: %with about 
164: %2000 Th-Ar lines, and  continuum normalized dividing each spectral order by  
165: %a cubic spline interpolation with ten segments. 
166: In addition to the program stars
167: observed with the HET, we have included in this study the spectra
168: of the Sun (Kurucz et al. 1984) and $\alpha$ CMi (Allende Prieto et al. 2002).
169: 
170: \section{Analysis}
171: \label{doit}
172: 
173: \subsection{Stellar parameters}
174: 
175: The effective temperature ($T_{\rm eff}$) has been estimated 
176: for all the program stars  
177:  by one or more authors using the Infrared Flux Method
178: (Blackwell \& Lynas-Gray 1994; Alonso et al. 1996, 1999). 
179: %In the case of
180: %Alonso et al., we have also used their (V-K)-[Fe/H] calibration, whenever 
181: %(V-K) was available . 
182: The $T_{\rm eff}$ of $\alpha$ Boo (Arcturus) 
183: has  been independently derived by 
184: di Benedetto (1998) and Griffin \& Lynas-Gray (1999). 
185: We have adopted the
186: values in Table 1. All our stars  were
187: analyzed by Allende Prieto \& Lambert (1999), and we accepted the 
188: surface gravities derived there from the (\bv) colors and the 
189: {\it Hipparcos} trigonometric parallaxes. 
190: The iron abundances in Table 1
191: are a straight average of previous spectroscopic analyses
192: compiled by Cayrel de Strobel et al. (2001).
193: A recent discussion of the preferred stellar parameters of $\alpha$ CMi 
194: can be found in Allende Prieto et al. (2002).
195: Fe\,{\sc i} lines at 4602.00 and 4602.94 \AA\ with accurate
196: $gf$-values and damping parameters were analyzed with
197: synthetic spectra computed from
198: flux-constant LTE model  atmospheres to estimate the projected
199: rotational velocities in Table 1.
200: 
201: 
202: \subsection{Measuring the line centers}
203: 
204: A variety of procedures  have been used previously to
205: determine the central wavelength of a spectral line. The most popular methods
206: involve some type of least-squares polynomial fitting or polynomial
207: interpolation (see, e.g., Neckel \& Labs 1990; Allende Prieto \& Garc\'{\i}a
208: L\'opez 1998).  
209: As the stellar spectral lines are asymmetric, 
210:  the details of the measuring procedure 
211:  affect the derived central wavelength.
212: Polynomial least-squares fits are well-suited for error estimation, 
213: however, there is some arbitrariness in the 
214: selection of the ideal order
215: and wavelength interval to use in the fits. For the slowly rotating stars 
216: analyzed here, we have found that a reasonable choice, the same for the 
217: whole sample, provides stable results regardless of the exact 
218: selection for the
219: order and wavelength interval.  Our 
220: particular choice  is to fit 35 m\AA\ around the line center with
221: a third order polynomial. Preceding the fitting by a cubic-spline interpolation
222: with a step of 5 m\AA\ proved useful for the HRS spectra, but degraded
223: slightly the quality of the results for the superior solar atlas.
224: 
225: To estimate the error in a 
226: derived central wavelength, we determine the horizontal
227: scatter between the polynomial fit F($\lambda$) and the 
228: observed spectrum $f(\lambda)$. For each observed wavelength $\lambda_i$ 
229: considered in the fit, we  find numerically the roots of the polynomial 
230:  $F(\lambda) - f(\lambda_i)$ using Laguerre's 
231:  method, select the appropriate root, 
232:  and then derive the rms scatter between the observed and fit wavelengths. 
233:  Assuming the errors are Gaussian, 
234:  this scatter is divided by the square root of the 
235:  number of points entering the polynomial fit to
236:  derive the uncertainty $\sigma_{\rm s}$. As an example, application 
237:  to the solar flux spectrum provides a nearly normal distribution 
238:  of uncertainties  that peaks at 0.00025 \AA\ and shows  
239:   bumps on the long-value tail, signaling blends. 
240:   Nave et al. (1994) classify their wavelengths into four 
241: categories,
242: depending on their uncertainty ($\sigma_{\rm l}$), which ranges 
243: from 0.4 m\AA\ to more than 
244: 10 m\AA. Most lines have errors below 5 m\AA\ and about 
245: half of the 
246: lines have errors below 1 m\AA\footnote{This comparison indicates that an
247: extension and improvement of the laboratory wavelengths for neutral iron
248:  and other species is desirable. A spectrograph used
249:  to obtain stellar spectra may be well suited for this purpose 
250:  (see Allende Prieto 2001)}.
251:  
252: \begin{figure*}
253: \begin{center}
254: \includegraphics[width=4cm,angle=90]{f1a.eps}
255: \includegraphics[width=4cm,angle=90]{f1b.eps}
256: \includegraphics[width=4cm,angle=90]{f1c.eps}
257: \includegraphics[width=4cm,angle=90]{f1d.eps}
258: \includegraphics[width=4cm,angle=90]{f1e.eps}
259: \includegraphics[width=4cm,angle=90]{f1f.eps}
260: \includegraphics[width=4cm,angle=90]{f1g.eps}
261: \includegraphics[width=4cm,angle=90]{f1h.eps}
262: \figcaption{
263: The left-hand panels show the relative velocity shifts of Fe {\sc i}
264:  lines with respect
265: to their rest wavelengths. The velocity scale for each star 
266: is shifted to have a null velocity
267: for the median shift of the strongest lines considered. The dots show the
268: shifts for the individual lines. The asterisks
269:  with error bars are mean values
270: after applying a median filter, and
271: the solid line is a polynomial fit to them. The dashed curve shows the median
272: shifts for the different bins. The right-hand panels
273: show histograms of the scatter of the velocity shifts around the polynomial
274: fit. The expected scatter (see text) is shown by the symmetric solid  curve. 
275: The mean velocity shifts determined from the analysis of the $\alpha$ Boo atlas
276: of Hinkle et al. (2000) are displayed as filled circles in the left-hand panel
277: for this star.
278: \label{shifts}}
279: \end{center}
280: \end{figure*} 
281: 
282: \tablenum{1}
283: \tablecolumns{7}
284: \tablewidth{0pc}
285: \begin{deluxetable}{cccccccc}
286: \tiny
287: \tablecaption{Observed sample \label{table1}}
288: \tablehead{
289: \colhead{Star} & \colhead{Obs. date} &  \colhead{Sp. Type} 
290: & \colhead{$T_{\rm eff}$\tablenotemark{a}} & \colhead{$\log g$} 
291: & \colhead{[Fe/H]}  &  \colhead{$v \sin i$}  & \colhead{$S$(v)} \\
292:  \colhead{}  & \colhead{} &   \colhead{}   &  \colhead{(K)} 
293:  &   \colhead{(dex)}  & \colhead{(dex)} & \colhead{(km s$^{-1}$)}  & \colhead{(km s$^{-1}$)}\\ }
294: \startdata
295: $\alpha$ CMi    & McDonald atlas & F5 IV & 6530 &    3.96 (0.02) &  $-$0.05 (0.03) & 2.8 (0.1)  	& 0.88 \\
296: Sun		& FTS atlas & G2 V  &	5770	&	4.44 (0.01) &   0.00	   & 1.9 (0.1) 		& 0.54 \\
297: 70 Vir		& June 5  & G4 V    &	5455	&	4.04 (0.11)  &  $-$0.11\tablenotemark{b}        & 3.2 (0.1) 		& 0.72 \\
298: $\mu$ Her A	& May 28  & G5 IV   &	5523	&	3.87 (0.04)  &	0.23 (0.07) & 4.1 (0.2) 	& 0.59 \\
299: $\beta$ Aql	& June 11 & G8 IV   &	5040	&	3.56 (0.07)  &  $-$0.02 (0.10) & 3.1 (0.3) 	& 0.57 \\
300: $\eta$ Cep	& May 29  & K0 IV   &	4944	&	3.40 (0.18)  &  $-$0.12 (0.14) & 3.1 (0.1)	& 0.51 \\
301: \tableline
302: $\alpha$ Boo 	& June 9  & K1.5 III &	4290	&	2.00 (0.56)  &  $-$0.51 (0.03) & 3.0 (0.5)	& 0.93 \\
303: $\xi$ Dra	& May 29  & K2 III   &  4430    & 	2.31 (0.32)  &  $-$0.13 (0.04) & 5.1 (0.5)	& 1.04 \\
304: \enddata
305: \tablenotetext{a}{Uncertainties in $T_{\rm eff}$ are 100 K or smaller}
306: \tablenotetext{b}{A single measurement is available}
307: \end{deluxetable}
308:  
309:  
310: The uncertainties in the measured  wavelength shifts, 
311: $\sigma = \sqrt{\sigma_{\rm s}^2 + \sigma_{\rm l}^2}$, are then converted to
312: velocity, and their  distribution $n(\sigma)$ is determined. 
313: Then, we  use $n(\sigma)$ 
314: to predict the distribution of 
315: errors ($\delta$) in the observed velocity shifts  as
316: 
317: \begin{equation}
318: N(\delta) = \frac{1}{C} \int_{0}^{\infty}  \frac{n(\sigma)}{\sigma} 
319: \exp \left(-\frac{\delta^2}{2\sigma^2}\right) d\sigma,
320: \label{ndelta}
321: \end{equation}
322: 
323: \noindent where $C$ is chosen such  that 
324: $\int_{-\infty}^{\infty} N(\delta) d\delta = 1$. The derived $N(\delta)$ 
325: can be compared to the observations 
326: in order to constrain the intrinsic scatter. 
327: In all cases, the derived and measured histograms are similar.
328: %indicating a lack of intrinsic scatter, as might arise, for example,
329: %from the effect of strong magnetic fields.
330: 
331: \subsection{Searching for correlations}
332: 
333: %We measured velocity  shifts for Fe I lines in each spectrum.
334: The number of iron lines identified and measured for each star varies from
335: 1340 to 1569\footnote{With the exception of $\alpha$ CMi, for which 
336: only 588 lines were measured}.
337: Solar analyses  have shown clear 
338: correlations of the line shifts with line strength  (Allende Prieto \& 
339: Garc\'{\i}a L\'opez 1998)
340: and wavelength (Hamilton \& Lester 1999). We explore here the variation
341: of the shifts with line strength, as this is the dominant effect. The line
342: equivalent width and the central line depth have been the most
343: widely used  line strength indicators. As pointed out by 
344: Pierce \& Lopresto (2000), 
345: the use of the equivalent width avoids saturation  
346:  for the strongest lines.
347: On the other hand, equivalent width  measurements are distorted 
348: by blends in the
349: line wings which may hardly affect line depths.
350: Blends are a  particularly serious problem
351: when dealing with cool stars, as some of those in our sample. Therefore, we 
352: have adopted the line depth.
353: 
354: 
355: The left-hand panels in Fig. \ref{shifts} show the line 
356: velocity shifts  for 
357: the stars in our sample as a function of the residual flux
358: at the center of the line $f_{\rm c}$ ($= 1 -$ line depth). In these panels, 
359: the shifts of the 
360: individual lines are marked with dots. Grouping the lines in bins of
361: 0.1 in depth, and 
362: after applying a median filter, we derive the average values shown as
363: asterisks with error bars. The solid
364:  lines are 3$^{\rm th}$-order polynomial fits to
365: guide the eye, and the 
366: broken lines trace the original median values. 
367: We have not attempted to disentangle
368: the stellar radial velocities from the gravitational shifts and the 
369:  atmospheric motions.
370: The absolute values on the vertical
371: axis have been shifted to have a null median 
372: velocity shift for the strongest lines; a condition that 
373: the polynomial fits were as well forced to satisfy.
374: The right-hand panels show the scatter of the line shifts
375: about the average points. The scatter expected from Eq. (1), mainly the result
376: of the uncertainties in the laboratory wavelengths, the finite width of the 
377: lines, and the presence of photometric noise, is also displayed
378:  (thick symmetric curve). The $\sigma(\delta)$  
379:  given in the right-hand panels 
380:  %and in Table 1
381: corresponds to a Gaussian  fit by least-squares 
382: to the observed scatter. The velocity span of the polynomial fits, 
383: $S$(v), is  listed in Table 1.
384: 
385: 
386: 
387:  A  spectroscopic 
388: atlas of $\alpha$ Boo at a superior resolving power and S/N  was 
389: published by Hinkle et al. (2000). The corresponding left-hand panel in 
390:  Figure \ref{shifts}  compares
391: the  average shifts from our measurements from the
392: atlas (filled circles) with those from 
393: the HRS spectrum (asterisks with error bars). The agreement is quite
394: satisfactory.  The $\sigma$ of the scatter in our spectrum and in the atlas
395: agree within 10 m s$^{-1}$.
396: %Not only do  $\alpha$ Boo and $\xi$ Dra have a lower gravity, 
397: %but they are also cooler than the dwarfs and subgiants in our sample. 
398: 
399: 
400: \section{Discussion}
401: \label{end}
402: 
403: The pattern of the  velocity shifts is similar for all the stars,
404:  and in qualitative 
405: agreement with solar results. The net 
406: convective blue shifts of the lines strengthen toward deeper photospheric
407: layers, and therefore affect the weak lines more than the strong ones. 
408: Despite
409: our limited sample, there is an indication of a smooth dependence of
410: the average velocity shifts with spectral type. 
411: This effect  is clearer in our analysis of line shifts than 
412: in previous studies using bisectors. As argued in the introduction, this is
413: likely the result of line shifts measurements 
414: being less affected by blending than line bisectors in late-G and K type stars.
415: 
416: Our preliminary results indicate that the trend of  
417: line shifts as a function of line strength can
418: be determined as a function of spectral type and gravity. 
419: At least for late-type dwarfs, assuming the strongest lines in the 
420: spectrum are free from convective
421: shifts (as is the case for the solar photosphere), 
422: it is possible to  
423: correct for convective wavelength shifts to within $\simeq 0.2$ km s$^{-1}$. 
424: 
425: 
426: The measured range of wavelength shifts for neutral iron lines is larger in 
427: $\alpha$ CMi than in cooler stars of classes IV and V.
428:  The range of wavelength shifts
429: is also larger,  reaching up to $\sim$ 1 km s$^{-1}$,  for the K giants 
430: than for the G-K dwarfs and subgiants in our
431: sample, which is not 
432: surprising. Although their lower
433: effective temperatures provide less flux to transport, their lower
434: atmospheric pressure implies larger convective velocities. 
435:  Fig. 1 suggests that, once we account for the observational errors,
436:   there is little room for intrinsic line-to-line scatter 
437: in the  F-G-K stars sampled here.
438: %   in the spectra of G-K dwarfs and subgiants, but not so much
439: %  for K giants and F-type stars.
440: Convection is supposed to cease in main sequence
441: stars with spectral types earlier than about F2, but other velocity fields
442:  must
443: be present in those atmospheres, judging from the asymmetry of the spectral 
444: lines (e.g. Gray \& Nagel 1989). Determining accurate
445:  absolute radial velocities 
446: demands an understanding of the wavelength shifts in the 
447: spectra of these stars too.
448: 
449: Solar observations  have shown differences
450: in the center-to-limb variation of the granulation along the central 
451: meridian and the equator 
452: (e.g. Beckers \& Taylor 1980; 
453: Rodr\'{\i}guez Hidalgo, Collados, \& V\'azquez 1992). 
454: Observations of a large number of stars may
455: then show peculiarities at a given spectral type 
456: depending on the orientation of the rotational axis.  Stars with
457: enhanced  magnetic fields are also expected to be
458: peculiar in terms of observed wavelength shifts, as the enhanced magnetic
459: fields may hinder the convective motions. Furthermore, line shifts
460: may vary with time for stars exhibiting an analog of the solar
461: 11 year cycle. 
462: 
463: Observations of line shifts for a significant number of
464: stars with high quality are required for a deeper understanding of 
465: granulation in stellar atmospheres, its relationship with other stellar
466: phenomena,  and the role of surface convection 
467: in the structure and evolution of stars. Our first results show that
468: measurements
469: of convective 
470: line shifts are a must in order to derive accurate absolute radial 
471: velocities.
472: 
473: We thank the HET staff for their outstanding job making possible science
474: observations with HRS since day one. 
475: %Resident astronomers 
476: %Jeff Mader, Brian Roman, and Matt Shetrone; and telescope operators Frank
477: %Deglman, Teddy
478: %George, and Vickie Riley acquired the HRS spectra used here. 
479: The Hobby-Eberly Telescope is operated by McDonald Observatory on behalf
480: of The University of Texas at Austin, the Pennsylvania State University,
481: Stanford University, Ludwig-Maximilians-Universit\"at M\"unchen, and
482: Georg-August-Universit\"at G\"ottingen.
483: NSO/Kitt Peak FTS data used here were produced
484: by NSF/NOAO.
485: This research was supported in part by the NSF (grant AST-0086321).
486: 
487: 
488: 
489: \begin{thebibliography}{}
490: 
491: \bibitem[]{} Allende Prieto, C. 2001, 
492: The Spectrum of the Th-Ar Hollow-Cathode Lamp
493: Used with the 2dcoud\'e Spectrograph, 
494: {\tt http://hebe.as.utexas.edu/2dcoude/thar/} (astro-ph/0111172)
495: 
496: \bibitem[]{} Allende Prieto, C., Asplund, M., Garc\'{\i}a L\'opez, R. J., \& Lambert, D. L. 2002, \apj, in press (astro-ph/011055)
497: 
498: \bibitem[Allende Prieto \& Garcia Lopez(1998)]{1998A&AS..129...41A} Allende Prieto, C., \& Garc\'{\i}a L\'opez, R.~J.\ 1998, \aaps, 129, 41
499: 
500: \bibitem[Allende Prieto \& Lambert(1999)]{1999A&A...352..555A} Allende Prieto, C., \& Lambert, D.~L.\ 1999, \aap, 352, 555
501: 
502: \bibitem[Alonso, Arribas, \& Martinez-Roger(1996)]{1996A&A...313..873A} Alonso, A., Arribas, S., \& Mart\'{\i}nez-Roger, C.\ 1996, \aap, 313, 873
503: 
504: \bibitem[Alonso, Arribas, \& Mart{\' i}nez-Roger(1999)]{1999A&AS..140..261A} Alonso, A., Arribas, S., \& Mart\'{\i}nez-Roger, C.\ 1999, \aaps, 140, 261
505: 
506: \bibitem[Asplund et al.(2000)]{2000A&A...359..729A} Asplund, M., Nordlund, {\AA}., Trampedach, R., Allende Prieto, C., \& Stein, R.~F.\ 2000, \aap, 359, 729
507: 
508: \bibitem[Beckers \& Taylor(1980)]{1980SoPh...68...41B} Beckers, J.~M.,~\& Taylor, W.~R.\ 1980, \solphys, 68, 41
509: 
510: \bibitem[Blackwell \& Lynas-Gray(1994)]{1994A&A...282..899B} Blackwell, D.~E.,~\& Lynas-Gray, A.~E.\ 1994, \aap, 282, 899
511: 
512: \bibitem[Cayrel de Strobel, Soubiran, \& Ralite(2001)]{2001A&A...373..159C} Cayrel de Strobel, G., Soubiran, C., \& Ralite, N.\ 2001, \aap, 373, 159
513: 
514: \bibitem[di Benedetto(1998)]{1998A&A...339..858D} di Benedetto, G.~P.\ 1998, \aap, 339, 858
515: 
516: \bibitem[Dravins(1974)]{1974A&A....36..143D} Dravins, D.\ 1974, \aap, 36, 143
517: 
518: \bibitem[Dravins(1985)]{1985srv..proc..311D} Dravins, D.\ 1985, IAU Colloq.~88: Stellar Radial Velocities, 311
519: 
520: \bibitem[Dravins(1987)]{1987A&A...172..211D} Dravins, D.\ 1987, \aap, 172, 211
521: 
522: \bibitem[Dravins(1999)]{1999psrv.conf..268D} Dravins, D.\ 1999, ASP Conf.~Ser.~185: IAU Colloq.~170: Precise Stellar Radial Velocities, 268
523: 
524: \bibitem[Dravins \& Nordlund(1990)]{1990A&A...228..203D} Dravins, D.,~\& Nordlund, \AA.\ 1990, \aap, 228, 203
525: 
526: \bibitem[Gray(1980)]{1980ApJ...235..508G} Gray, D.~F.\ 1980, \apj, 235, 508
527: 
528: \bibitem[Gray(1981)]{1981ApJ...251..583G} Gray, D.~F.\ 1981, \apj, 251, 583
529: 
530: \bibitem[Gray \& Nagel(1989)]{1989ApJ...341..421G} Gray, D.~F.,~\& Nagel, T.\ 1989, \apj, 341, 421
531: 
532: \bibitem[Gray \& Toner(1986)]{1986PASP...98..499G} Gray, D.~F.,~\& Toner, C.~G.\ 1986, \pasp, 98, 499
533: 
534: \bibitem[Griffin \& Lynas-Gray(1999)]{1999AJ....117.2998G} Griffin, R.~E.~M.,~\& Lynas-Gray, A.~E.\ 1999, \aj, 117, 2998
535: 
536: \bibitem[Hamilton \& Lester(1999)]{1999PASP..111.1132H} Hamilton, D.,~\& Lester, J.~B.\ 1999, \pasp, 111, 1132
537: 
538: \bibitem[Hinkle, Wallace, Valenti, \& Harmer(2000)]{2000vnia.book.....H} Hinkle, K., Wallace, L., Valenti, J., \& Harmer, D.\ 2000, 
539: Visible and Near Infrared Atlas of the Arcturus Spectrum 3727-9300 
540: \AA\  (San Francisco: ASP)
541: 
542: \bibitem[Kurucz, Furenlid, \& Brault(1984)]{1984sfat.book.....K} Kurucz, R.~L., Furenlid, I., \& Brault, J.\ 1984, National Solar Observatory Atlas (Sunspot, New Mexico: NSO)
543: 
544: \bibitem[Lindegren, Dravins, \& Madsen(1999)]{1999psrv.conf...73L} Lindegren, L., Dravins, D., \& Madsen, S.\ 1999, ASP Conf.~Ser.~185: IAU Colloq.~170: Precise Stellar Radial Velocities, 73
545: 
546: %\bibitem[]{} Madsen, S., Dravins, D., \&  Lindegren, L. \ 2002, \aap, in press
547: 
548: \bibitem[Nave et al.(1994)]{1994ApJS...94..221N} Nave, G., Johansson, S., Learner, R.~C.~M., Thorne, A.~P., \& Brault, J.~W.\ 1994, \apjs, 94, 221
549: 
550: \bibitem[Neckel \& Labs(1990)]{1990SoPh..126..207N} Neckel, H.,~\& Labs, D.\ 1990, \solphys, 126, 207
551: 
552: \bibitem[Pierce \& Lopresto(2000)]{2000SoPh..196...41P} Pierce, A.~K.,~\& Lopresto, J.~C.\ 2000, \solphys, 196, 41
553: 
554: \bibitem[Rodriguez Hidalgo, Collados, \& Vazquez(1992)]{1992A&A...264..661R} Rodr\'{\i}guez Hidalgo, I., Collados, M., 
555: \& V\'azquez, M.\ 1992, \aap, 264, 661
556: 
557: \bibitem[Tull(1998)]{1998SPIE.3355..387T} Tull, R.~G.\ 1998, \procspie, 3355, 387
558: 
559: \end{thebibliography}
560: 
561: 
562: \end{document}
563: 
564: 
565: