0805.4324/ms.tex
1: \documentclass{emulateapj}
2: %
3: %\documentclass[12pt,preprint]{aastex}
4: %\documentclass[manuscript]{aastex}
5: %\documentclass[preprint2]{aastex}
6: %\documentclass[preprint2,longabstract]{aastex}
7: 
8: \shorttitle{Stratification of sunspot umbral dots}
9: \shortauthors{Riethm\"uller et al.}
10: 
11: \begin{document}
12: 
13: \title{Stratification of sunspot umbral dots from inversion of Stokes profiles recorded by $Hinode$}
14: 
15: \author{T. L. Riethm\"uller, S. K. Solanki, and A. Lagg}
16: \affil{Max-Planck-Institut f\"ur Sonnensystemforschung, Max-Planck-Str. 2, 37191 Katlenburg-Lindau, Germany;}
17: \email{[riethmueller;solanki;lagg]@mps.mpg.de}
18: 
19: \begin{abstract}
20:    This work aims to constrain the physical nature of umbral dots (UDs) using high-resolution
21:    spectropolarimetry. Full Stokes spectra recorded by the spectropolarimeter on $Hinode$ of 51 UDs
22:    in a sunspot close to the disk center are analyzed. The height dependence of the temperature,
23:    magnetic field vector, and line-of-sight velocity across each UD is obtained from an inversion
24:    of the Stokes vectors of the two Fe~I lines at 630~nm. No difference is found at higher altitudes
25:    ($-3 \le \log(\tau_{500}) \le -2$) between the UDs and the diffuse umbral background. Below that level
26:    the difference rapidly increases, so that at the continuum formation level ($\log(\tau_{500}) = 0$)
27:    we find on average a temperature enhancement of 570~K, a magnetic field weakening of 510~G, and
28:    upflows of 800~m~s$^{-1}$ for peripheral UDs, whereas central UDs display an excess temperature of on
29:    average 550~K, a field weakening of 480~G, and no significant upflows. The results for, in particular,
30:    the peripheral UDs, including cuts of magnetic vector and velocity through them, look remarkably similar
31:    to the output of recent radiation MHD simulations. They strongly suggest that UDs are produced by
32:    convective upwellings.
33: \end{abstract}
34: 
35: \keywords{Sun: photosphere --- Sun: sunspots --- techniques: spectroscopic}
36: 
37: %__________________________________________________________________
38: \section{Introduction}
39: 
40:    The energy transport immediately below the solar surface is mainly determined by convective
41:    processes that are visible as granulation patterns in white-light images of the quiet photosphere.
42:    This convection is suppressed inside sunspot umbrae due to the strong vertical magnetic field, but some
43:    form of magnetoconvection \citep{Weiss2002} is needed to explain the observed umbral brightnesses.
44:    Umbral fine structure such as light bridges or umbral dots, dotlike bright features inside umbrae,
45:    may well be manifestations of magnetoconvection. Different models have been proposed to explain UDs,
46:    e.g., columns of field-free hot gas in between a bundle of thin magnetic flux ropes
47:    \citep{Parker1979,Choudhury1986}, or spatially modulated oscillations in a strong magnetic field
48:    \citep{Weiss1990}. Recent numerical simulations of three-dimensional radiative magnetoconvection
49:    \citep{Schuessler2006} reveal convective plumes that penetrate through the solar surface and look very
50:    much like UDs. Although recent broadband images may have spatially resolved UDs
51:    \citep[][in preparation]{Sobotka2005,Riethmueller2008}, spectropolarimetry is needed to learn more about their physical nature.
52:    Previous spectroscopic observations led to heterogeneous results. \citet{Kneer1973} found that
53:    UDs exhibit upflows of 3~km~s$^{-1}$ and a 50\% weaker magnetic field compared to the nearby umbra,
54:    whereas \citet{Lites1991} and \citet{Tritschler1997} reported
55:    little field weakening. Finally, \citet{SocasNavarro2004} observed a weakening
56:    of 500~G and upflows of a few 100~m~s$^{-1}$. More details can be found in the reviews of umbral fine
57:    structure by \citet{Solanki2003} and \citet{Sobotka2006}. One reason for the
58:    difference in results has been the influence of scattered light and variable seeing, which affect
59:    the different analyzed data sets to varying degrees. It therefore seems worthwhile to invert
60:    Stokes profiles obtained by the spectropolarimeter (SP) on the $Hinode$ spacecraft. The usefulness
61:    of $Hinode$ data for the study of UDs was demonstrated by \citet{Bharti2007}, who found
62:    that large UDs show dark lanes whose existence had been predicted by \citet{Schuessler2006}.
63: 
64: %__________________________________________________________________
65: \section{Observations and data reduction}
66: 
67:    The data employed here were acquired by the spectropolarimeter \citep{Lites2001}
68:    of the Solar Optical Telescope \citep[SOT,][]{Suematsu2008} onboard $Hinode$. They are
69:    composed of full Stokes spectra in the Fe~I line pair around 6302~{\AA} and the nearby continuum
70:    of a sunspot of NOAA AR~10933 recorded from 12:43 to 12:59~UT on 2007 January 5
71:    using the 0.16$^{\prime\prime}$x164$^{\prime\prime}$ slit. At this time the sunspot was located at
72:    a heliocentric angle of 4$^\circ$, i.e. very close to disk center. The observations covered the
73:    spectral range from 6300.89 to 6303.26~{\AA}, with a sampling of 21~{m\AA}~pixel$^{-1}$. The SP
74:    was operated in its normal map mode, i.e. both the sampling along the slit and the slit-scan
75:    sampling were 0.16$^{\prime\prime}$, so that the spatial resolution should be close to the
76:    diffraction limit of $1.22~\lambda/D = 0.32^{\prime\prime}$. The integration time per slit position
77:    was 4.8~s which reduced the noise level to $10^{-3}~I_c$.
78: 
79:    The data were corrected for dark current, flat field, and instrumental polarization with the help of the
80:    SolarSoft package.\footnote{See \url{http://www.lmsal.com/solarsoft}.} A continuum intensity image (put together from the
81:    slit scan) of the chosen umbra is shown in Figure~\ref{FigUmbra}. Due to the large slit length
82:    we are always able to find a sufficiently extensive region of quiet Sun that is used to normalize
83:    intensities.
84: 
85:    \begin{figure}
86:    \centering
87:    \includegraphics[width=\linewidth]{f1.eps}
88:    \caption{Continuum intensity map of the sunspot NOAA~10933 as observed by the $Hinode$ SOT/SP on 2007
89:    January 5. Heliocentric angle is $\theta$~=~4$^\circ$. Intensities are normalized to the intensity
90:    level of the quiet photosphere $I_{ph}$. The white line at (4,4)~Mm marks the cut through an umbral
91:    dot (UD) that is discussed in greater detail.}
92:    \label{FigUmbra}
93:    \end{figure}
94: 
95: %________________________________________________________________
96: \section{Data analysis}
97: 
98:    To obtain atmospheric stratifications of temperature ($T$), magnetic field strength ($B$), and line-of-sight
99:    velocity ($v_{LOS}$) we use the inversion code SPINOR described by \citet{Frutiger2000b}.
100:    This code incorporates the STOPRO routines \citep{Solanki1987}, which compute synthetic Stokes
101:    profiles of one or more lines upon input of their atomic data and one or more model atmospheres.
102:    Local thermodynamic equilibrium conditions are assumed and the Unno-Rachkovsky radiative transfer
103:    equations are solved. The inversions use an optical depth scale as the appropriate coordinate
104:    for radiative transfer problems. For reasons of comparability we use the optical depth at 500~nm
105:    ($\tau_{500}$). Starting with an initial guess model, the synthetic profiles were iteratively
106:    fitted to observed data using response functions (RFs) and the merit function $\chi^2$
107:    \citep{RuizCobo1992,Frutiger2000a} is minimized. With the help of the  RFs we find that the
108:    Fe~I line pair at 6302~{\AA} is mainly formed within the $\log(\tau_{500})$ interval [$-3,0$], which
109:    corresponds to a height range of about 400~km under hydrostatic equilibrium conditions in the umbra.
110:    The free parameters are defined at the four nodes $-3$, $-2$, $-1$, and $0$ of the $\log(\tau_{500})$ grid.
111:    The atmospheric stratification is then interpolated using splines onto a 10 times finer $\log(\tau_{500})$
112:    grid.
113: 
114:    The first step of our analysis is the wavelength calibration required to determine line-of-sight
115:    (LOS) velocities. For every slit position we average the Stokes $I$ profiles of all locations
116:    along the slit whose total polarization $P = \int(Q^2 + U^2 + V^2)^{1/2}d\lambda$ is negligible, since
117:    those locations are assumed to represent the quiet Sun. This mean $I$ profile is used to fit Voigt profiles
118:    to the two Fe~I lines from which the line center wavelengths are determined. The convective blueshift
119:    of 140~m~s$^{-1}$ \citep[see][]{MartinezPillet1997,Dravins1981} is then removed.
120: 
121:    The next step is to find an appropriate model atmosphere. Since we are interested in the atmospheric
122:    stratification of temperature, magnetic field strength, and LOS velocity within a UD, these
123:    three atmospheric parameters are assumed to be height dependent, whereas field inclination and azimuth angle,
124:    micro-turbulence, and macro-turbulence are assumed to be height independent. We experimented intensively
125:    with adding a second model component to represent the stray light, but the inversion results did not
126:    improve significantly, confirming the almost negligible stray light in the SP. Therefore, in the
127:    interests of a robust inversion, we forbore from adding a stray light component, thus reducing the number
128:    of free parameters.
129: 
130:    Lastly, we have to find initial guesses for all free parameters. We use an initial temperature stratification
131:    according to the umbral core model L of \citet{Maltby1986} and assume a vertical magnetic
132:    field of 2000~G and zero LOS velocity at all heights. Initial guesses for micro-turbulence and macro-turbulence
133:    are 0.1 and 2~km~s$^{-1}$, respectively. Other initial guesses gave very similar results, except for
134:    a limited number of outliers. For these, repeating the inversion with an initial guess close to the final
135:    result of one of the neighboring pixels returned values consistent with those obtained for the other pixels.
136: 
137: 
138: %________________________________________________________________
139: \section{Inversion results}
140: 
141:    We analyzed a total of 51 UDs, which were identified by applying the multilevel tracking (MLT) algorithm
142:    \citep[][in preparation]{Bovelet2001,Riethmueller2008}. For each UD
143:    the location of its core was identified, a cut was made through it, reaching to the neighboring diffuse
144:    background (DB), and the profiles from all the pixels along this cut were inverted. We first discuss the
145:    results for the UD marked in Figure~\ref{FigUmbra}, chosen because of its brightness, which leads
146:    to particularly small error bars. A comparison of the measured profiles with the best-fit profiles resulted
147:    from the inversion can be seen in Figure~\ref{FigProfileUdBright} for the UD and in Figure~\ref{FigProfileUdDark}
148:    for the DB selected as the location of lowest continuum intensity in a 1.4~$\times$~1.4~Mm$^2$ environment
149:    of the UD center. Due to the low signal in the dark background the measured DB profiles are much noisier
150:    than the UD center's profiles, but in general, the Stokes spectra can be fitted remarkably well.
151: 
152:    \begin{figure}
153:    \centering
154:    \includegraphics[width=\linewidth]{f2.eps}
155:    \caption{Stokes $I$, $V$, $Q$ and $U$ profiles from the center of the UD marked in Figure~\ref{FigUmbra}.
156:    Red lines are the measured, blue lines the best-fit profiles, i.e. the inversion result. The bottom parts
157:    of each panel show the difference between the two on an expanded scale.}
158:    \label{FigProfileUdBright}
159:    \end{figure}
160: 
161:    \begin{figure}
162:    \centering
163:    \includegraphics[width=\linewidth]{f3.eps}
164:    \caption{The same as Figure~\ref{FigProfileUdBright}, but for Stokes $I$, $V$, $Q$ and $U$ profiles of
165:    the diffuse background near the UD.}
166:    \label{FigProfileUdDark}
167:    \end{figure}
168: 
169:    The stratification of the retrieved atmospheric parameters $T$, $v_{LOS}$, and $B$ in the center of the UD
170:    and in the DB are plotted in Figure~\ref{FigSingleAtm}. In the upper photosphere ($-3 \le \log(\tau_{500})
171:    \le -2$) the error bars overlap; i.e. we find little significant difference between UD and DB. In the
172:    deeper photosphere, however, the inversions return strongly different stratifications. Thus, the UD
173:    temperature is higher than the DB temperature, consistent with the intensity enhancement of the UD in
174:    the continuum map. The LOS velocity (which is identical to the vertical velocity due to the small
175:    heliocentric angle) exhibits strong upflows in the UD center, whereas the DB is nearly at rest.
176:    The magnetic field strength is roughly 2~kG for the heights $-3 \le \log(\tau_{500}) \le -1$. Below
177:    $\log(\tau_{500}) = -1$ the field strength of the UD decreases strongly with depth, whereas the field
178:    strength of the DB increases moderately.
179: 
180:    \begin{figure}
181:    \centering
182:    \includegraphics[width=\linewidth]{f4.eps}
183:    \caption{Atmospheric stratification obtained from the Stokes profiles at the location of the UD's center
184:    (red lines) and from the Stokes profiles of the diffuse background near the UD (blue lines).
185:    The formal errors of the inversion at the used optical depth nodes are indicated by bars. Negative
186:    LOS velocity values indicate upflows.}
187:    \label{FigSingleAtm}
188:    \end{figure}
189: 
190:    The vertical cuts of magnetic field strength and LOS velocity through 13 pixels lying along the white
191:    line in Figure~\ref{FigUmbra} are shown in Figure~\ref{FigVerticalCut}. Jumps from one pixel to the next
192:    were smoothed through interpolation. There is clear evidence for a localized decrease in UD field
193:    strength in the low photosphere, co-located with an upflow that extends higher into the atmosphere and
194:    a weak downflow on at least one side. The magnetic fields are 4$^\circ$ more inclined in the UD than
195:    they are in the DB around the UD. Figure~\ref{FigVerticalCut} looks remarkably like Fig.~2 of
196:    \citet{Schuessler2006}, in spite of the fact that Figure~\ref{FigVerticalCut}
197:    is plotted on an optical depth scale in the vertical direction and is thus distorted by an unknown
198:    amount relative to a corresponding figure on a geometrical scale.
199: 
200:    \begin{figure}
201:    \centering
202:    \includegraphics[width=\linewidth]{f5a.eps}\\
203:    \includegraphics[width=\linewidth]{f5b.eps}
204:    \caption{Vertical cut through the UD marked in Figure~\ref{FigUmbra} in the direction indicated by the
205:    white line. Colors of the top panel indicate magnetic field strength. The bottom panel shows LOS velocity.
206:    Negative velocities are upflows.}
207:    \label{FigVerticalCut}
208:    \end{figure}
209: 
210:    Next we discuss all 51 analyzed UDs. In the literature we often find a separation into two UD regimes.
211:    For example, \citet{Grossmann1986} differentiate between peripheral
212:    UDs (PUDs) and central UDs (CUDs), i.e. between UDs that are born close to the umbra-penumbra boundary
213:    and UDs that are born deep in the umbra. We follow this distinction and plot the obtained stratifications
214:    of the 30 PUDs (distance to umbra-penumbra boundary less than 2000~km) in the top panels of
215:    Figure~\ref{FigMultipleAtms}, while the remaining 21 CUDs are represented in the bottom panels of
216:    Figure~\ref{FigMultipleAtms}. The results largely mirror those obtained for the UD discussed above. In the
217:    upper atmosphere UDs center and DB do not differ in their mean values of $T$, $v_{LOS}$, and $B$.
218:    On average, the CUDs are about 150~K cooler than the PUDs in the upper atmosphere, just as the DB
219:    around the CUDs is cooler than the DB around the PUDs. At $\log(\tau_{500}) = 0$ we find that PUDs are
220:    570~K hotter than the local DB and CUDs are 550~K hotter than the DB in their vicinity. The magnetic
221:    field strength at $\log(\tau_{500}) = 0$ is weakened by about 510~G for PUDs and 480~G for CUDs, whereas
222:    only PUDs exhibit significant upflows of about 800 m~s$^{-1}$. The mean LOS velocity shows no difference between
223:    CUD centers and DB. In order to make sure that an upflow is not being missed due to the lower S/N ratio
224:    of the CUD Stokes profiles, we have also averaged the Stokes profiles of all the CUDs. An inversion
225:    of there averaged Stokes profiles gave a result that agrees with the averaged stratifications ($green~line$)
226:    in the bottom panels of Figure~\ref{FigMultipleAtms} within the error bars. This suggests that any upflow
227:    velocity in CUDs is mostly restricted to layers below the surface or is too concentrated or too weak to be
228:    detected by the inversions. Finally, we find that the magnetic field of the PUDs is on average
229:    4$^\circ$ more horizontal than for their DB. We see no inclination difference for CUDs.
230: 
231:    \begin{figure}
232:    \centering
233:    \includegraphics[width=\linewidth]{f6.eps}
234:    \caption{Atmospheric stratifications of peripheral umbral dots (top 3 panels) and central umbral dots
235:    (bottom 3 panels). The red lines show the stratification at the location of the UD's center and the
236:    blue lines correspond to the nearby diffuse background. The green line is the weighted average of all
237:    red lines and the yellow line is the weighted average of all blue lines, where we used the reciprocal
238:    error bars as weighting factors.}
239:    \label{FigMultipleAtms}
240:    \end{figure}
241: 
242: %______________________________________________________________
243: 
244: \section{Discussion}
245: 
246:    We identified 30 peripheral and 21 central umbral dots in $Hinode$ spectropolarimetric data of a sunspot
247:    within 4$^\circ$ of disk center. With the help of Stokes profile inversions of the Fe~I lines at
248:    630~nm we determined the stratifications of temperature, magnetic field strength, and LOS velocity.
249:    The present work differs from that of \citet{SocasNavarro2004} in the superior quality
250:    of the employed data with twice the spatial resolution and practically no scattered light. This allows
251:    a detailed determination of the atmospheric stratification. The higher spatial resolution of the
252:    $Hinode$ SP data also allows us to, for the first time, reconstruct both the horizontal and the vertical
253:    structure of UDs. We also extended the analysis to a more numerous statistical ensemble of 51 UDs.
254: 
255:    Vertical cuts through UDs provide a remarkable confirmation of the results of MHD simulations of
256:    \citet{Schuessler2006}: both show that UDs differ from their surroundings
257:    mainly in the lowest visible layers, where the temperature is enhanced and the magnetic field is
258:    weakened. We found a temperature enhancement of 550~K and a magnetic field reduction of about 500~G
259:    (at optical depth unity). In addition, PUDs display upflow velocities of 800~m~s$^{-1}$ on average,
260:    again in good agreement with the simulations. There are also some differences between our results
261:    and those of \citet{Schuessler2006}. Thus, according to our inversions the
262:    magnetic field strength of the DB is somewhat depth dependent. This was not the case for the MHD
263:    simulations due to the used periodic boundary conditions. Furthermore, although some of the UDs
264:    display a weak downflow bounding the strong central upflow (see Figure~\ref{FigVerticalCut}),
265:    these are neither as narrow nor as strong as the downflows at the ends of dark lanes as reported
266:    by \citet{Schuessler2006}, probably due to the limited spatial resolution of our data. We may also
267:    be missing some of the narrow downflows by considering only single cuts across individual UDs.
268: 
269:    \citet{SocasNavarro2004} reported 10$^\circ$ more inclined magnetic fields in PUDs.
270:    This result is qualitatively confirmed by our work; we find an inclination increase of 4$^\circ$ for PUDs
271:    but no increase for CUDs, which can be assumed as a further hint that the main part of the CUD structure
272:    is below the surface. These results can be interpreted in terms of the strong DB fields expanding with
273:    height and closing over the UD, as proposed by \citet{SocasNavarro2004}.
274: 
275: %______________________________________________________________
276: \begin{thebibliography}{}
277: 
278:    \bibitem[Bharti et al.(2007)]{Bharti2007} Bharti, L., Joshi, C. \& Jaaffrey, S. N. A. 2007,
279:       ApJ, 669, L57
280: 
281:    \bibitem[Bovelet \& Wiehr(2001)]{Bovelet2001} Bovelet, B., \& Wiehr, E. 2001,
282:       Solar Phys., 201, 13  
283: 
284:    \bibitem[Choudhury(1986)]{Choudhury1986} Choudhury, A. R. 1986,
285:       ApJ, 302, 809
286: 
287:    \bibitem[Dravins et al.(1981)]{Dravins1981} Dravins, D., Lindegren, L., \& Nordlund, {\AA}. 1981,
288:       A\&A, 96, 345
289: 
290:    \bibitem[Frutiger(2000)]{Frutiger2000a} Frutiger, C. 2000,
291:       Inversion of Zeeman Split Stokes Profiles: Application to solar and stellar surface structures,
292:       Ph.D. Thesis, Institute of Astronomy, ETH Z\"urich, No.~13896
293: 
294:    \bibitem[Frutiger et al.(2000)]{Frutiger2000b} Frutiger, C., Solanki, S. K., Fligge, M., \& Bruls, J. H. M. J. 2000,
295:       A\&A, 358, 1109
296: 
297:    \bibitem[Grossmann-Doerth et al.(1986)]{Grossmann1986} Grossmann-Doerth, U., Schmidt, W., \& Schr\"oter, E. H. 1986,
298:       A\&A, 156, 347
299: 
300:    \bibitem[Kneer(1973)]{Kneer1973} Kneer, F. 1973,
301:       Solar Phys., 28, 361
302: 
303:    \bibitem[Lites et al.(1991)]{Lites1991} Lites, B. W., Bida, T. A., Johannesson, A., \& Scharmer, G. B. 1991,
304:       ApJ, 373, 683
305: 
306:    \bibitem[Lites et al.(2001)]{Lites2001} Lites, B. W., Elmore, D. F., \& Streander, K. V. 2001,
307:       in ASP Conf. Ser. 236, Advanced Solar Polarimetry, ed.\ M. Sigwarth (San Francisco: ASP), 33
308: 
309:    \bibitem[Maltby et al.(1986)]{Maltby1986} Maltby, P., Avrett, E. H., Carlsson, M., Kjeldseth-Moe, O., Kurucz, R. L., et al. 1986,
310:       ApJ, 306, 284
311: 
312:    \bibitem[Mart\'inez Pillet et al.(1997)]{MartinezPillet1997} Mart\'inez Pillet, V., Lites, B. W. \& Skumanich, A. 1997,
313:       ApJ, 474, 810
314: 
315:    \bibitem[Parker(1979)]{Parker1979} Parker, E. N. 1979,
316:       ApJ, 234, 333
317: 
318:    \bibitem[T. L. Riethm\"uller et al.(2008)]{Riethmueller2008} Riethm\"uller, T. L., Solanki, S. K., Zakharov, V. \& Gandorfer, A. 2008,
319:       A\&A, 492, 233
320: 
321:    \bibitem[Ruiz Cobo \& del Toro Iniesta(1992)]{RuizCobo1992} Ruiz Cobo, B., \& del Toro Iniesta, J. C. 1992,
322:       ApJ, 398, 375
323: 
324:    \bibitem[Sch\"ussler \& V\"ogler(2006)]{Schuessler2006} Sch\"ussler, M., \& V\"ogler, M. 2006,
325:       ApJ, 641, L73
326: 
327:    \bibitem[Sobotka(2006)]{Sobotka2006} Sobotka, M. 2006,
328:       Dissertation for Doctor Scientiarum, Acad. Sci. Czech Republic
329: 
330:    \bibitem[Sobotka \& Hanslmeier(2005)]{Sobotka2005} Sobotka, M., \& Hanslmeier, A. 2005,
331:       A\&A, 442, 323
332: 
333:    \bibitem[Socas-Navarro et al.(2004)]{SocasNavarro2004} Socas-Navarro, H., Mart\'inez Pillet, V., Sobotka, M., \& V\'azquez, M. 2004
334:       ApJ, 614, 448
335: 
336:    \bibitem[Solanki(1987)]{Solanki1987} Solanki, S. K. 1987,
337:       The Photospheric Layers of Solar Magnetic Fluxtubes,
338:       Ph.D. Thesis, Institute of Astronomy, ETH Z\"urich, No.~8309
339: 
340:    \bibitem[Solanki(2003)]{Solanki2003} Solanki, S. K. 2003,
341:       A\&A Rev., 11, 153
342: 
343:    \bibitem[Suematsu et al.(2008)]{Suematsu2008} Suematsu, Y., Tsuneta, S., Ichimoto, K., Shimizu, T., Otsubo, M., et al. 2008,
344:       Solar Phys., in press
345: 
346:    \bibitem[Tritschler \& Schmidt(1997)]{Tritschler1997} Tritschler, A., \& Schmidt, W. 1997,
347:       A\&A, 321, 643
348: 
349:    \bibitem[Weiss(2002)]{Weiss2002} Weiss, N. O. 2002,
350:       Astron. Nachr., 323, 371
351: 
352:    \bibitem[Weiss et al.(1990)]{Weiss1990} Weiss, N. O., Brownjohn, D. P., Hurlburt, N. E., \& Proctor, M. R. E. 1990,
353:       Mon. Not. R. astr. Soc., 245, 434
354: 
355: \end{thebibliography}
356: 
357: \end{document}
358: 
359: