1: %\documentclass[12pt,preprint]{aastex}

2: \documentclass{emulateapj}

3: \usepackage{apjfonts}

4: 

5: % \usepackage{amsmath}

6: 

7: %\usepackage[,longnamesfirst]{natbib}

8: 

9: \newcommand{\Teff}{\mbox{$T_{\rm eff}$}}

10: \newcommand{\Dnu}{\mbox{$\Delta \nu$}}

11: \newcommand{\acena}{\mbox{$\alpha$~Cen~A}}

12: \newcommand{\acenb}{\mbox{$\alpha$~Cen~B}}

13: \newcommand{\acen}{\mbox{$\alpha$~Cen}}

14: \newcommand{\bhyi}{\mbox{$\beta$~Hyi}}

15: \newcommand{\cms}{\mbox{cm\,s$^{-1}$}}

16: \newcommand{\eboo}{\mbox{$\eta$~Boo}}

17: \newcommand{\ms}{\mbox{m\,s$^{-1}$}}

18: \newcommand{\muHz}{\mbox{$\mu$Hz}}

19: \newcommand{\mynote}[1]{{\bf [#1]}}

20: \newcommand{\new}[1]{{\bf #1}}

21: \renewcommand{\new}[1]{{#1}}

22: \newcommand{\myomit}[1]{{\bf Omit: #1}}

23: \renewcommand{\myomit}[1]{\relax}

24: %\let\epsilon\varepsilon

25: \newcommand{\half}{{\textstyle\frac{1}{2}}}

26: \newcommand{\sixth}{{\textstyle\frac{1}{6}}}

27: \newcommand{\tenth}{{\textstyle\frac{1}{10}}}

28: 

29: \slugcomment{To appear in ApJ Letters}

30: 

31: \shorttitle{Near-surface corrections for stellar oscillations}

32: \shortauthors{Kjeldsen et al.}

33: 

34: 

35: \begin{document}

36: 

37: \title{Correcting stellar oscillation frequencies for near-surface effects}

38:  

39: \author{

40: Hans~Kjeldsen\altaffilmark{1},

41: Timothy R. Bedding\altaffilmark{2} and

42: J{\o}rgen Christensen-Dalsgaard\altaffilmark{1}

43: }

44: 

45: \altaffiltext{1}{Danish AsteroSeismology Centre (DASC), Department of

46: Physics and Astronomy, University of Aarhus, DK-8000 Aarhus C, Denmark;

47: hans@phys.au.dk, jcd@phys.au.dk}

48: 

49: \altaffiltext{2}{Institute of Astronomy, School of Physics, University of

50: Sydney, NSW 2006, Australia; bedding@physics.usyd.edu.au}

51: 

52: 

53: 

54: \begin{abstract} 

55: %% Version: \today

56: In helioseismology, there is a well-known offset between observed and

57: computed oscillation frequencies.  This offset is known to arise from

58: improper modeling of the near-surface layers of the Sun, and a similar

59: effect must occur for models of other stars.  Such an effect impedes

60: progress in asteroseismology, which involves comparing observed oscillation

61: frequencies with those calculated from theoretical models.  Here, we use

62: data for the Sun to derive an empirical correction for the near-surface

63: offset, which we then apply three other stars (\acena{}, \acenb{} and

64: \bhyi).  The method appears to give good results, in particular providing

65: an accurate estimate of the mean density of each star.

66: \end{abstract}

67: 

68: \keywords{stars: individual (\bhyi, \acena, \acenb) --- stars:~oscillations

69:   --- Sun:~helioseismology}

70: 

71: \section{Introduction}

72: 

73: Both helio- and asteroseismology involve comparing observed oscillation

74: frequencies with those calculated from theoretical models.  However, for

75: the Sun there is a long-standing systematic offset between observed and

76: computed frequencies that is known to arise from improper modeling of the

77: near-surface layers \citep{ChDDL88,DPV88,ChDDA96,ChD+T97}.  This offset is

78: independent of the angular degree of the mode ($l$) and increases with

79: frequency.  A similar offset must occur for models of other stars, and

80: should be taken into account whenever observations of stellar oscillations

81: are compared with theory.  In this Letter we use data for the Sun to derive

82: an empirical correction for these near-surface effects and show how to

83: apply this to other stars.  The method appears to give good results, in

84: particular providing an accurate estimate of the mean density of each star.

85: 

86: \section{Method}

87: 

88: The p-mode oscillations in solar-type stars for a given angular degree~$l$

89: are approximately equally spaced in frequency, with a separation of $\Dnu$

90: (the so-called large separation; see \citealt{ChD2004} for a review of the

91: theory of solar-like oscillations).  Since the offset from incorrect

92: modelling of the near-surface layers is independent of~$l$, we can derive

93: the correction by considering only the radial modes ($l=0$) and then apply

94: it to all modes. 

95: 

96: Suppose we have a set of observed frequencies for radial modes, $\nu_{\rm

97: obs}(n)$, where $n$ is the radial order.  Suppose also that $\nu_{\rm best}(n)$

98: are the frequencies from the best model, by which we mean the one

99: that best describes the parameters and internal structure of the star, but

100: which still fails to model correctly the surface layers,

101: 

102: For the Sun, the difference between observed and best model frequencies

103: turns out to be well fitted by a power law, \new{which has the convenient

104: property of being free of a frequency scale (see also \citealt{ChD+G80})}:

105: \begin{equation}

106:   \nu_{\rm obs}(n) - \nu_{\rm best}(n) = a \left(\frac{\nu_{\rm

107:   obs}(n)}{\nu_0}\right)^b, \label{eq.fobs-fref}

108: \end{equation}

109: where $\nu_0$ is a suitably chosen reference frequency, and $a$ and $b$ are

110: parameters to be determined.  

111: 

112: Suppose further that we have calculated a reference model that has

113: frequencies $\nu_{\rm ref}(n)$ and is close to the best model.  From homology

114: scaling it then follows that, to a good approximation,

115: \begin{equation}

116:   \nu_{\rm best}(n) = r \nu_{\rm ref}(n),\label{eq.fbest.fref}

117: \end{equation}

118: where the scaling factor $r$ is related to the mean densities

119: $\bar \rho_{\rm best}$ and $\bar \rho_{\rm ref}$ of the best and reference

120: models by

121: \begin{equation}

122:   \bar{\rho}_{\rm best} = r^2 \bar{\rho}_{\rm ref}.  \label{eq.rho}

123: \end{equation}

124: Given a determination of $r$, equation~(\ref{eq.rho}) provides our best

125: estimate of the mean density of the star.

126: 

127: Substituting equation~(\ref{eq.fbest.fref}) into

128: equation~(\ref{eq.fobs-fref}) gives

129: \begin{equation}

130:   \nu_{\rm obs}(n) - r \nu_{\rm ref}(n) = a \left(\frac{\nu_{\rm

131:   obs}(n)}{\nu_0}\right)^b \label{eq.fobs-rfref} 

132: \end{equation}

133: and differentiating with respect to $n$ gives

134: \begin{equation}

135:   \Delta\nu_{\rm obs}(n) - r \Delta\nu_{\rm ref}(n) = a b \left(\frac{\nu_{\rm

136:             obs}(n)}{\nu_0}\right)^{b-1} \frac{\Delta\nu_{\rm

137:       obs}(n)}{\nu_0}.\label{eq.Deltanu}

138: \end{equation}

139: Combining and rearranging these last two equations gives

140: \begin{equation}

141:   r = (b-1)\left(b \frac{\nu_{\rm ref}(n)}{\nu_{\rm

142: 	obs}(n)} - \frac{\Delta\nu_{\rm ref}(n)}{\Delta\nu_{\rm

143: 	obs}(n)}\right)^{-1} \label{eq.r}

144: \end{equation}

145: and

146: \begin{equation}

147:   b = \left(r \frac{\Delta\nu_{\rm ref}(n)}{\Delta\nu_{\rm

148:       obs}(n)} - 1\right) 

149:       \left(r \frac{\nu_{\rm ref}(n)}{\nu_{\rm

150: 	obs}(n)} - 1\right)^{-1}. \label{eq.b}

151: \end{equation}

152: If we know $b$ then we can calculate $r$ using equation~(\ref{eq.r}), or

153: {\em vice versa\/} using equation~(\ref{eq.b}).  We can then obtain~$a$

154: using equation~(\ref{eq.fobs-rfref}).

155: 

156: We now show how to apply the above method to a set of observed and

157: calculated frequencies.  Suppose we have frequencies for $N$ radial modes

158: with orders $n_1, n_2, \ldots, n_N$ (not necessarily consecutive).  We use

159: these to calculate the four terms needed to evaluate equation~(\ref{eq.r})

160: or~(\ref{eq.b}).  For $\nu_{\rm obs}(n)$ and $\nu_{\rm ref}(n)$, we simply use

161: the means of the given sets of frequencies, which we denote by $\langle

162: \nu_{\rm obs}(n) \rangle$ and $\langle \nu_{\rm ref}(n) \rangle$.

163: 

164: To estimate the large separations, we calculate the slope of a linear

165: least-squares fit to the given frequencies (as a function of~$n$):

166: \begin{eqnarray}

167:   \langle \Delta\nu_{\rm obs}(n) \rangle &=& 

168:   \frac{\sum_{i=1}^{N} 

169:     \left(\nu_{\rm obs}(n_i) - \langle \nu_{\rm obs}(n)\rangle\right)

170:     \left(n_i-\langle n\rangle\right)}

171:   {\sum_{i=1}^{N} \left(n_i-\langle n\rangle\right)^2}\\

172: %

173:   \langle \Delta\nu_{\rm ref}(n) \rangle &=& 

174:   \frac{\sum_{i=1}^{N} 

175:     \left(\nu_{\rm ref}(n_i) - \langle \nu_{\rm ref}(n)\rangle\right)

176:     \left(n_i-\langle n\rangle\right)}

177:   {\sum_{i=1}^{N} \left(n_i-\langle n\rangle\right)^2}.

178: \end{eqnarray}

179: We must then assume a value for either $b$ or $r$, and use

180: equation~(\ref{eq.r}) or~(\ref{eq.b}) to estimate the other.  Finally, the

181: value of~$a$ is found from equation~(\ref{eq.fobs-rfref}), as follows:

182: \begin{equation}

183:   a = \frac{\langle \nu_{\rm obs}(n) \rangle - r \langle \nu_{\rm ref}(n) \rangle}

184:       {N^{-1} \sum_{i=1}^{N} \left(\nu_{\rm obs}(n_i)/\nu_0\right)^b}.

185: %%   \left\langle \left(\frac{\nu_{\rm obs}(n)}{\nu_0}\right)^b \right\rangle = 

186: %%      \frac{1}{N} \sum_{i=1}^{N} \left(\frac{\nu_{\rm obs}(n)}{\nu_0}\right)^b.

187: \end{equation}

188: 

189: We now proceed to apply this method to the Sun in order to measure~$b$

190: (\S\ref{sec.sun}), and then adopt this value of~$b$ for other stars

191: (\S\ref{sec.stars}).

192: 

193: \section{Application to the Sun}\label{sec.sun}

194:    

195: For the Sun we took Model~S of \citet{ChDDA96}, as listed in the first row

196: of Table~\ref{tab.models}.  We assumed this to be the ``best'' solar model,

197: in the sense defined above, which means we can set $r=1$ (see

198: equation~\ref{eq.fbest.fref}).  For the observed solar frequencies, we used

199: those measured by \citet{LBB97} with the GOLF instrument on the SOHO

200: spacecraft.

201: 

202: We followed the procedure described above, choosing $\nu_0 = 3100\,\muHz$ and

203: setting $r=1$, and using the data to measure~$b$ and~$a$.  We have

204: chosen to use the nine modes centred at the peak of the oscillation power,

205: from which we obtained the results shown in the first line of

206: Table~\ref{tab.calcs}, and a value of $b=4.90$.  The differences between

207: observed and Model~S frequencies are plotted as the squares in

208: Fig.~\ref{fig.solar}, and the solid curve is the function given by

209: equation~(\ref{eq.fobs-fref}).

210: 

211: \new{ The above fit was made for the strongest 9 radial modes in the Sun.

212: We repeated the analysis for different numbers of modes (all values from 7

213: to 13), and found the derived value of $b$ to range from 4.4 to 5.25.

214: Clearly, the frequency differences do not exactly follow a power law, and

215: so the exponent in the power-law fit depends substantially on the frequency

216: range.  Importantly, the value for $a$ varies by less than 0.1\,\muHz{} in

217: all cases.  }

218: 

219: In addition to Model~S, we also considered models denoted~S$^-$ and~S$^+$

220: from the same evolution sequence as Model~S, but with ages of 2.25 and

221: 7.44\,Gyr, respectively.  The parameters of these models are given in

222: Table~\ref{tab.models}.  We kept $b$ fixed at the value found for the

223: ``best'' model ($4.90$) and used equation~(\ref{eq.r}) to estimate $r$.

224: The results are shown in Table~\ref{tab.calcs}, and also in

225: Fig.~\ref{fig.solar}.  Importantly, the derived density of the Sun ($r^2

226: \bar{\rho}_{\rm ref}$) is correct for both these calculations, despite the

227: very different densities of the models themselves, giving us confidence

228: that the method has been successful.  \new{Even more importantly, the

229: derived density of the Sun is completely insensitive to the choice of $b$.

230: This reflects the fact that the value of~$r$ obtained by fitting to

231: equation~\ref{eq.fobs-rfref} is not sensitive to the exact form of the

232: function on the right-hand side, so long as that function tends to zero

233: with decreasing frequency.}

234: 

235: \section{Application to \acena, \acenb{} and \bhyi}\label{sec.stars}

236: 

237: We have considered three stars for which observations and models have been

238: published.  For \acena{} we took observed frequencies (radial modes only)

239: from four sources: \citet{B+C2002}, \citet{BKB2004}, \citet{BazBK2007} and

240: \citet{FCE2006}.  This gave a set of 33 measured frequencies for 11 orders,

241: all of which were given equal weight in the fitting process.  For \acenb{}

242: the observed frequencies were taken from \citet{KBB2005} and those for

243: \bhyi{} from \citet{BKA2007}.  \new{In each case, we used all the

244: detected $l=0$ frequencies.}

245: 

246: The models that we have used for these stars are listed in

247: Table~\ref{tab.models}.  These include published models of \acena{} and~B

248: by \citet{MPL2000} and \citet{TPM2002}, \new{of \acenb{} by

249: \citet{TBG2008},} and of \bhyi{} by \citet{F+M2003}.  In addition to

250: published models, we have considered several computed with the Aarhus

251: stellar evolution code (ASTEC, \citealt{ChD2008}).  Models~A and~B are

252: models of the \acen{} system, computed with essentially the same physics as

253: Model~S and fitted (T. Teixeira et al.\ 2008, in preparation) to the

254: observations of \citet{BKB2004} and \citet{KBB2005}.  For \bhyi, Model~H

255: matches the parameters reported by \citet{NDB2007} and was computed with

256: similar physics, but neglecting diffusion and settling.  Finally,

257: Models~H$^-$ and~H$^+$ are from the same evolution sequence as Model~H, but

258: with substantially different ages and densities, which bracket those of

259: Model~H.

260: 

261: For each model we used the value of $b$ found for the Sun, and used

262: equation~(\ref{eq.r}) to estimate $r$.  The results are shown in

263: Table~\ref{tab.calcs}, and also in Figs.~\ref{fig.bhyi}--\ref{fig.acenb}.

264: \new{Note that there is considerable scatter in the observed frequencies of

265: \acena{} and~B due to the relatively short span of the observations

266: relative to the mode lifetime.  Taking this into account, } it is once

267: again encouraging to see that the power law \new{(with a single value of

268: $b$)} provides a good fit to the frequency differences, and that for each

269: star there is good agreement between the densities derived from the

270: different models.  

271: 

272: \new{To estimate uncertainties, we have repeated the fits for the same

273: range of $b$ values considered in~\S\ref{sec.sun} (4.4--5.25) and again

274: found that there is no effect on the calculated density, to the precision

275: quoted in Table~\ref{tab.calcs}.  Over this range, the change in~$a$ is

276: less than 1\,\muHz{} for \bhyi{}, less than 0.4\,\muHz{} for \acena{} and

277: less than 0.7\,\muHz{} for \acenb.  The changes in the correction terms

278: plotted in Figs.~\ref{fig.bhyi}--\ref{fig.acenb} are comparable or slightly

279: bigger.  }

280: 

281: For each star, we can identify the model that is closest to being the

282: ``best'' model as the one having $r$ closest to unity (note this is not

283: necessarily the model with the smallest near-surface offset).  These give

284: our best estimate of the stellar density for each star.

285: \myomit{and the corresponding frequency differences are plotted in

286: Fig.~\ref{fig.best}, this time as a function of normalized frequency

287: ($\nu_{\rm obs}/\nu_0$).}

288: 

289: \citet{NDB2007} recently used interferometry to measure the angular

290: diameter of \bhyi{} to be $2.257\pm0.019$\,mas.  They combined this with

291: the parallax (from {\em Hipparcos}) and the mean density (from

292: asteroseismology) to determine the radius and mass of the star.  Using our

293: best estimate of the mean density of \bhyi{} ($0.258 \pm

294: 0.001$\,g\,cm$^{-3}$) and the revised {\em Hipparcos} parallax ($134.070

295: \pm 0.110$\,mas; \citealt{vanLee2007}), we derive slightly updated values,

296: finding a radius of $1.809 \pm 0.015\,R_\sun$ (0.85\%) and a mass of $1.085

297: \pm 0.028\,M_\sun$ (2.6\%).

298: 

299: 

300: \section{Discussion and Conclusions}

301: 

302: The method outlined here for correcting near-surface effects can be applied

303: to model frequencies before they are compared with observations.  As in the

304: case of the Sun, we expect that the correction is independent of degree at

305: a given frequency, for low-degree acoustic modes, and thus the correction

306: determined from radial modes can be applied to all such modes.

307: 

308: There is, however, an important exception.  In evolved stars, mixed

309: modes may be observed that have the character of gravity modes in the deep

310: interior of the star.  Observational evidence for mixed modes has been

311: found in \eboo{} \citep{KBV95,KBB2003,CEB2005} and \bhyi{} \citep{BKA2007}.

312: Owing to the larger amplitude of these modes in the stellar interior and

313: hence their higher inertia, their frequencies are less affected by the

314: surface effects.  Techniques need to be developed to take this into account

315: in the application of the surface correction before the frequencies are

316: analyzed.

317: 

318: \new{The use of a single power law is made plausible by the knowledge that

319: the offsets we are modelling are caused by the properties of the

320: near-surface layers and hence presumably depend only on surface gravity,

321: effective temperature and composition, and not on the details of the

322: internal properties of the star.  This may be particularly true of the

323: exponent which, in the simple analysis by \citet{ChD+G80}, is determined by

324: an effective average polytropic index (i.e., the relation between pressure

325: and density) in the near-surface layers. }

326: 

327: It is important to note that, owing to its strong frequency dependence, the

328: offset also affects the large frequency separation $\Delta \nu$, as is

329: indeed implicit in equation~(\ref{eq.Deltanu}). \new{For example, although

330: Model~S is one of the best available models of the Sun, it has a large

331: separation that is 1\,\muHz{} greater than the observed value.}  Thus,

332: attempting to fit to stellar models based on $\Delta \nu$ will introduce

333: systematic errors, unless the corresponding correction is applied.

334: 

335: We have shown how to identify the model for each star that is closest to

336: the ``best'' model, by requiring that $r$ be as close as possible to unity.

337: This gives us an extremely accurate estimate of the mean stellar density.

338: However, it is important to point out that, while a model with $r$ close to

339: unity gives a good match to the overall structure of the star, it does not

340: necessarily reproduce the structure of the core or give a reliable estimate

341: of the stellar age.  Determination of those properties requires taking into

342: account the frequencies of the non-radial modes (including the small

343: frequency separations).  That is the next stage of model fitting, which can

344: be done after the surface correction has been applied to all modes.

345: 

346:    

347: \acknowledgments

348: 

349: We thank Teresa Teixeira for assistance in determining Models~A and~B, and

350: Mario Monteiro for providing the frequencies of Model~$S_0$ for \bhyi{} in

351: electronic form.  This work was supported financially by the Danish Natural

352: Science Research Council and the Australian Research Council.

353: 

354: %% \input{bibfiles-simple}

355: %% \bibliographystyle{natbib/mynatbib}

356: 

357: \begin{thebibliography}{23}

358: \expandafter\ifx\csname natexlab\endcsname\relax\def\natexlab#1{#1}\fi

359: \expandafter\ifx\csname url\endcsname\relax

360:   \def\url#1{{\tt #1}}\fi

361: 

362: \bibitem[{Bazot} et~al.(2007){Bazot}, {Bouchy}, Kjeldsen, Charpinet, Laymand,

363:   \& {Vauclair}]{BazBK2007}

364: {Bazot}, M., {Bouchy}, F., Kjeldsen, H., Charpinet, S., Laymand, M., \&

365:   {Vauclair}, S., 2007, A\&A, 470, 295.

366: 

367: \bibitem[{Bedding} et~al.(2007){Bedding}, {Kjeldsen}, Arentoft,

368:   et~al.]{BKA2007}

369: {Bedding}, T.~R., {Kjeldsen}, H., Arentoft, T., et~al., 2007, ApJ, 663, 1315.

370: 

371: \bibitem[Bedding et~al.(2004)Bedding, Kjeldsen, Butler, et~al.]{BKB2004}

372: Bedding, T.~R., Kjeldsen, H., Butler, R.~P., et~al., 2004, ApJ, 614, 380.

373: 

374: \bibitem[Bouchy \& {Carrier}(2002)Bouchy, \& {Carrier}]{B+C2002}

375: Bouchy, F., \& {Carrier}, F., 2002, A\&A, 390, 205.

376: 

377: \bibitem[{Carrier} et~al.(2005){Carrier}, {Eggenberger}, \& {Bouchy}]{CEB2005}

378: {Carrier}, F., {Eggenberger}, P., \& {Bouchy}, F., 2005, A\&A, 434, 1085.

379: 

380: \bibitem[Christensen-Dalsgaard(2004)]{ChD2004}

381: Christensen-Dalsgaard, J., 2004, Sol. Phys., 220, 137.

382: 

383: \bibitem[Christensen-Dalsgaard(2008)]{ChD2008}

384: Christensen-Dalsgaard, J., 2008, Ap\&SS.

385: \newblock in press (astro-ph/0710.3114).

386: 

387: \bibitem[Christensen-Dalsgaard et~al.(1996)Christensen-Dalsgaard, D\"appen,

388:   Ajukov, et~al.]{ChDDA96}

389: Christensen-Dalsgaard, J., D\"appen, W., Ajukov, S.~V., et~al., 1996, Sci, 272,

390:   1286.

391: 

392: \bibitem[{Christensen-Dalsgaard} et~al.(1988){Christensen-Dalsgaard},

393:   {D\"appen}, \& {Lebreton}]{ChDDL88}

394: {Christensen-Dalsgaard}, J., {D\"appen}, W., \& {Lebreton}, Y., 1988, Nat, 336,

395:   634.

396: 

397: \bibitem[{Christensen-Dalsgaard} \& {Gough}(1980){Christensen-Dalsgaard}, \&

398:   {Gough}]{ChD+G80}

399: {Christensen-Dalsgaard}, J., \& {Gough}, D.~O., 1980, Nat, 288, 544.

400: 

401: \bibitem[{Christensen-Dalsgaard} \& {Thompson}(1997){Christensen-Dalsgaard}, \&

402:   {Thompson}]{ChD+T97}

403: {Christensen-Dalsgaard}, J., \& {Thompson}, M.~J., 1997, MNRAS, 284, 527.

404: 

405: \bibitem[{Dziembowski} et~al.(1988){Dziembowski}, {Patern\'o}, \&

406:   {Ventura}]{DPV88}

407: {Dziembowski}, W.~A., {Patern\'o}, L., \& {Ventura}, R., 1988, A\&A, 200, 213.

408: 

409: \bibitem[Fernandes \& {Monteiro}(2003)Fernandes, \& {Monteiro}]{F+M2003}

410: Fernandes, J., \& {Monteiro}, M.~J.~P.~F.~G., 2003, A\&A, 399, 243.

411: 

412: \bibitem[{Fletcher} et~al.(2006){Fletcher}, {Chaplin}, {Elsworth}, {Schou}, \&

413:   {Buzasi}]{FCE2006}

414: {Fletcher}, S.~T., {Chaplin}, W.~J., {Elsworth}, Y., {Schou}, J., \& {Buzasi},

415:   D., 2006, MNRAS, 371, 935.

416: 

417: \bibitem[Kjeldsen et~al.(2003)Kjeldsen, Bedding, Baldry, et~al.]{KBB2003}

418: Kjeldsen, H., Bedding, T.~R., Baldry, I.~K., et~al., 2003, AJ, 126, 1483.

419: 

420: \bibitem[Kjeldsen et~al.(2005)Kjeldsen, Bedding, Butler, et~al.]{KBB2005}

421: Kjeldsen, H., Bedding, T.~R., Butler, R.~P., et~al., 2005, ApJ, 635, 1281.

422: 

423: \bibitem[Kjeldsen et~al.(1995)Kjeldsen, Bedding, Viskum, \& Frandsen]{KBV95}

424: Kjeldsen, H., Bedding, T.~R., Viskum, M., \& Frandsen, S., 1995, AJ, 109, 1313.

425: 

426: \bibitem[Lazrek et~al.(1997)Lazrek, {Baudin}, {Bertello}, et~al.]{LBB97}

427: Lazrek, M., {Baudin}, F., {Bertello}, L., et~al., 1997, Sol. Phys., 175, 227.

428: 

429: \bibitem[Morel et~al.(2000)Morel, {Provost}, {Lebreton}, {Th{\'e}venin}, \&

430:   {Berthomieu}]{MPL2000}

431: Morel, P., {Provost}, J., {Lebreton}, Y., {Th{\'e}venin}, F., \& {Berthomieu},

432:   G., 2000, A\&A, 363, 675.

433: 

434: \bibitem[North et~al.(2007)North, {Davis}, {Bedding}, et~al.]{NDB2007}

435: North, J.~R., {Davis}, J., {Bedding}, T.~R., et~al., 2007, MNRAS, 380, L83.

436: 

437: \bibitem[Tang et~al.(2008)Tang, Bi, Gai, \& Xu]{TBG2008}

438: Tang, Y.-K., Bi, S.-L., Gai, N., \& Xu, H.-Y., 2008, Chin. J. A\&A.

439: \newblock in press.

440: 

441: \bibitem[Th{\' e}venin et~al.(2002)Th{\' e}venin, {Provost}, {Morel},

442:   {Berthomieu}, {Bouchy}, \& {Carrier}]{TPM2002}

443: Th{\' e}venin, F., {Provost}, J., {Morel}, P., {Berthomieu}, G., {Bouchy}, F.,

444:   \& {Carrier}, F., 2002, A\&A, 392, L9.

445: 

446: \bibitem[van Leeuwen(2007)]{vanLee2007}

447: van Leeuwen, F.

448: \newblock {\em {Hipparcos, the New Reduction of the Raw Data}}.

449: \newblock Springer: Dordrecht, 2007.

450: 

451: \end{thebibliography}

452: 

453: \clearpage

454: \input{table1}

455: \input{table2}

456: 

457: \clearpage

458: 

459: \begin{figure}

460: \epsscale{0.9}

461: %\plotone{idl/fig/sun04.ps}

462: \plotone{f1.eps}

463: \caption[]{\label{fig.solar} The difference between observed and calculated

464: frequencies for radial modes in the Sun.  The squares are for Model~S, with

465: the solid curve showing a fit to equation~(\ref{eq.fobs-rfref}) with $r=1$,

466: which gives $b=4.90$ (see~\S\ref{sec.sun}).  Also shown are the results of

467: applying the same corrections to Model~S$^-$ (crosses) and Model~S$^+$

468: (pluses).  The dotted curves show the corrections calculated from

469: equation~(\ref{eq.fobs-rfref}).}

470: \end{figure}

471: 

472: \begin{figure}

473: \epsscale{0.9}

474: %\plotone{idl/fig/bhyi02.ps}

475: \plotone{f2.eps}

476: \caption[]{\label{fig.bhyi} The difference between observed and calculated

477: frequencies for radial modes in \bhyi.  The models shown are: Model~H

478: (squares), Model~H$^-$ (triangles), Model~H$^+$ (diamonds) and FM2003

479: (asterisks). The dotted curves show the corrections calculated from

480: equation~(\ref{eq.fobs-rfref}).}

481: \end{figure}

482: 

483: \begin{figure}

484: \epsscale{0.9}

485: %\plotone{idl/fig/acena06.ps}

486: \plotone{f3.eps}

487: \caption[]{\label{fig.acena} Same as Fig.~\ref{fig.bhyi}, but for \acena.

488: The models shown are: Model~A (squares), M2000~A (triangles), Model~S$^+$

489: (pluses) and Th2002~A (asterisks).}

490: \end{figure}

491: 

492: \begin{figure}

493: \epsscale{0.9}

494: %\plotone{idl/fig/acenb01.ps}

495: \plotone{f4.eps}

496: \caption[]{\label{fig.acenb} Same as Fig.~\ref{fig.bhyi}, but for \acenb.

497: The models shown are: Model~B (squares), M2000~B (triangles), Model~S$^-$

498: (crosses), Th2002~B (asterisks) \new{and T2008 (diamonds)}.}

499: \end{figure}

500: 

501: \end{document}

502: