1: \documentclass[12pt,preprint]{aastex}
2:
3:
4: \usepackage{psfig,times,emulateapj5}
5: \usepackage{natbib}
6: \usepackage[below]{placeins}
7:
8:
9:
10:
11: \newcommand{\rxj}{RX~J0720.4-3125}
12: \newcommand{\rxjb}{RX~J1856.5-3754}
13: \newcommand{\xspec}{{\em XSPEC}}
14: \newcommand{\rosat}{{\em ROSAT}}
15: \newcommand{\chandra}{{\em Chandra}}
16: \newcommand\asca{{\em ASCA}}
17: \newcommand\xmm{{\em XMM-Newton}}
18: \newcommand\sax{{\em BeppoSAX}}
19: \newcommand\rxte{{\em RXTE}}
20: \newcommand\msun{{$M_{\odot}$}}
21: \newcommand{\kms}{{km\,s$^{-1}$}}
22: \newcommand{\nh}{{$N_{\rm H}$}}
23:
24:
25: \begin{document}
26:
27: \title{The continued spectral evolution of the neutron star {RX~J0720.4-3125}}
28: \author{Jacco Vink, Cor P. de Vries, Mariano M\'{e}ndez}
29:
30: \affil{SRON National Institute for Space Research,
31: Sorbonnelaan 2, NL-3584 CA, Utrecht, The Netherlands}
32: \email{j.vink@sron.nl}
33: \and
34: \author{Frank Verbunt}
35: \affil{Astronomical Institute, Utrecht University,
36: PO Box 80000, NL-3508 TA Utrecht, The Netherlands}
37:
38:
39:
40: \begin{abstract}
41: We observed the isolated neutron star \rxj\ with
42: \chandra's Low Energy Transmission Grating Spectrometer,
43: following the \xmm\ discovery
44: of long term spectral evolution of this source.
45: The new observation shows that the spectrum of \rxj\
46: has continued to change in the course of 5 months.
47: It has remained hard, similar
48: to the last \xmm\ observation, but the strong depression observed with \xmm\
49: at long wavelengths has disappeared. Contrary to the \xmm\ observations,
50: the new \chandra\ observation shows that the flux increase at short
51: wavelengths and the decrease at long wavelengths do not necessarily occur
52: simultaneously.
53: \end{abstract}
54:
55: \keywords{
56: stars: individual ({\rxj}) --
57: stars: neutron --
58: X-rays: stars
59: }
60:
61:
62: \maketitle
63: \section{Introduction}
64: \rxj\ \citep{haberl97},
65: is one of the best studied isolated
66: neutron stars whose X-ray emission is
67: dominated by radiation from the hot neutron star surface.
68: Their X-ray spectra are best described by blackbody emission with
69: $kT_\infty \sim 70$~eV, but the emitting areas seem
70: to be significantly smaller than the canonical neutron star surface.
71: This suggests that the X-ray
72: emission is coming from a small, hot, fraction
73: of the neutron star \citep[][]{motch03,kaplan03b},
74: or may not be pure blackbody
75: radiation, but instead the result of
76: emission from a reflective condensed matter surface
77: \citep[][]{lenzen78,turolla04}.
78: Recently it was observed that the isolated neutron stars
79: RX J1308.6+2127 and RX J1605.3+3249
80: exhibit broad absorption features that may be caused by
81: proton cyclotron absorption \citep{haberl03a,vankerkwijk04}.
82:
83: The spectral behavior of \rxj\ is even more surprising.
84: Early observations with \chandra\ \citep{kaplan03b}
85: and \xmm\ \citep{paerels01} are consistent with a blackbody-like spectrum,
86: but subsequent observations by \xmm\ indicate a slow evolution
87: of the spectrum between 10-38~\AA,
88: resulting in deviations from a Planckian spectrum.
89: The deviations consist of a flux decrease at wavelengths longer
90: than $\sim 23$\AA,
91: and a flux increase at shorter wavelengths
92: \citep{devries04}.
93: The nature of the emerging spectrum is not clear (modified blackbody,
94: cyclotron absorption in combination with a hotter blackbody?),
95: nor is it clear what causes
96: the emission properties to change. In \citet{devries04}
97: we suggested that \rxj\ is precessing,
98: which causes us to observe the hot region of the star under a
99: continuously different angle.
100: This requires that the surface emission is anisotropic,
101: and may alter as result of traversing a magnetized atmosphere
102: \citep{ho04}.
103:
104:
105: Whatever the underlying mechanism,
106: it is clear that the spectral evolution of \rxj\
107: potentially provides clues to the nature of the X-ray surface emission
108: from isolated neutron stars, whereas the fact that the spectrum evolves
109: may have implications for the structure of neutron stars, e.g.
110: if the evolution is a consequence of precession.
111: In order to further monitor its spectral evolution,
112: and to investigate the spectrum for $\lambda > 38$~\AA\ with high
113: spectral resolution,
114: we were granted Director's Discretionary Time
115: (DDT) on \chandra\ for observing
116: \rxj\ with the Low Energy Transmission Grating
117: Spectrometer (LETGS) in combination with the HRC-S (High Resolution
118: Camera) microchannel plate detector.
119: In this letter we report on the analysis of this observation.
120: We show that the spectrum has continued to evolve: it has further hardened,
121: but the attenuation at long wavelengths has disappeared.
122:
123: \begin{figure*}
124: \centerline{
125: \psfig{figure=f1.eps,width=0.77\textwidth}}
126: \figcaption{
127: \chandra\ LETGS spectrum of \rxj\ as observed on February 27, 2004.
128: The colored lines show various models folded through the detector
129: response.
130: The green line and the error weighted residuals are for the best fit model
131: listed in Table~\ref{bbody} column 3.
132: The blue line is the best fit model to the spectrum of February 2, 2000
133: (Table~\ref{bbody} column 2). The red line shows the best
134: fit model to the \xmm-RGS spectrum of October 27, 2003
135: \citep{devries04}; it has been cut off at the wavelength limit of the
136: RGS instruments (38~\AA).
137: \label{ddt}}
138: \end{figure*}
139:
140: \section{Observation and data analysis}
141:
142: \chandra\ observed \rxj\ for 35~ks
143: on February 27, 2004 (Obs. ID 5305) as part of its DDT program.
144: For our analysis we used the cleaned event list available from the
145: \chandra\ X-ray Center. Spectral extraction and
146: ancillary response files were made using the
147: standard \chandra\ reduction package CIAO v.3.0.2.
148: For comparison we also analyzed
149: the longest
150: archival LETGS observation, made on February 2, 2000
151: \citep[Obs. ID 745, see also][]{kaplan03b}, and the six
152: archival \xmm\ Reflective Grating Spectrometer (RGS) data sets discussed
153: in \citet{devries04}.
154:
155: The LETGS DDT spectrum is shown in Fig.~\ref{ddt}.
156: The spectrum is clearly harder than the LETGS spectrum of February 2000.
157: This confirms the discovery of spectral evolution of \rxj\ \citep{devries04}.
158: However, it is clear that the source has continued to evolve since the
159: last \xmm\ observation of 27 October 2003, as the
160: attenuation of the emission above 23~\AA, which increased with time and
161: was very prominent in the last RGS spectrum,
162: has almost disappeared.
163: In fact, unlike the last RGS spectrum both LETGS spectra
164: can be fitted reasonably well with pure blackbody models, but with different
165: temperatures. The new spectrum requires a hotter blackbody temperature
166: of $kT_\infty \sim 100$~eV than that of February 2000, which
167: is best fitted with a temperature of $\sim 85$~eV. The precise values for
168: temperature depend on the modeling assumptions (Table~\ref{bbody}),
169: e.g. whether we allow
170: the interstellar absorption to vary from one observation to another
171: (columns \nh\ free) or whether we fit the two spectra simultaneously,
172: forcing the absorption parameters for the two models to be equal.
173:
174: \begin{table*}[b]
175: \begin{center}
176: \caption{Best fit blackbody models.\label{bbody}}
177:
178: \begin{tabular}{lllll}
179: \tableline\tableline\noalign{\smallskip}
180: {} & \multicolumn{2}{c}{2/2/2000} & \multicolumn{2}{c}{27/2/2004}\\
181: Parameter & \nh\ free &
182: \multicolumn{2}{c}{jointly fitted \nh} &
183: \nh\ free\\
184: \tableline\noalign{\smallskip}
185:
186: $kT_{\infty}$\ (eV)& $82.9\pm1.5$ &$86.4\pm2.4$ &$100.6\pm1.1$& $103.4\pm1.3$\\
187: $R$ (km) & $5.36\pm0.12$ &$4.63\pm0.07$&$3.43\pm0.04$& $3.13\pm0.05$\\
188: \nh\ ($10^{20}\,{\rm cm}^{-2}$)
189: & $1.28\pm0.12$ &\multicolumn{2}{c}{$0.90\pm0.07$}
190: & $0.63\pm0.09$\\
191: C-statistic/\#bins& 727.5/720 & 735.8/719 & 780.6/719
192: &771.3/719\\
193:
194: \noalign{\smallskip}\tableline
195: \end{tabular}
196: \tablecomments{
197: The models were fitted using the spectral analysis program
198: \xspec\, \citep{xspec}.
199: Errors are 68\% confidence limits.
200: The statistic used to optimize the fit
201: was the C-statistic \citep{cash79}.
202: We assume a distance of 300~pc \citep{kaplan03b}.
203: }
204: \end{center}
205: \end{table*}
206:
207:
208: We note that the spectral fitting can be further improved for both LETGS
209: spectra by an additional soft {\em emission} component,
210: which contributes to $\lambda > 50$~\AA.
211: However, we
212: do not want to overemphasize this. First, it may be the result of
213: calibration uncertainties regarding the instrument sensitivity.
214: Secondly,
215: assuming that the additional component
216: is real and caused by thermal emission,
217: a very cool blackbody is required ($kT_\infty \sim 23$~eV),
218: with an emitting area too large for a neutron star,
219: corresponding to a radius of $\sim 400$~km
220: \citep[following][we assume here, and throughout the
221: rest of the text a distance of 300~pc]{kaplan03b}.
222:
223:
224: \vbox{
225: \centerline{
226: \psfig{figure=f2.eps,width=0.85\columnwidth}}
227: \figcaption{The flux evolution of \rxj\ for the wavelength ranges
228: 10-13 \AA, 23-38 \AA\ and the combined range, determined from
229: \xmm-RGS (open squares) and \chandra-LETGS spectra (triangles)
230: The fluxes are normalized to the fluxes of
231: the first RGS and LETGS observations respectively, which are thus
232: per definition 1.
233: \label{flux}}
234: %\smallskip
235: }
236:
237:
238: The observed increase of the emission above 23~\AA\ with respect to the
239: last RGS observation is very interesting,
240: but also somewhat unfortunate,
241: as the broad spectral range of the LETGS would have allowed to put
242: better constraints on the spectral shape of the low energy attenuation.
243: For instance,
244: \citet{devries04}
245: showed that the deviations from a blackbody spectrum apparent from
246: the \xmm-RGS spectra,
247: could be either modeled by a broad Gaussian absorption component, or
248: by a blackbody modified by a multiplicative power law.
249: \citet{devries04}
250: favored the characterization by a power law times a blackbody,
251: as it provided good fits to the data with fewer parameters:
252: Apart from the blackbody parameters, only the power-law slope had to be
253: determined.
254: Although there is no clear physical
255: rational for such a model, the spectral evolution could be characterized
256: by just one parameter: the power-law slope. Changing the power-law
257: slope results in a simultaneous hardening and supression of long wavelength
258: emission, which
259: describes well the spectral evolution up to October, 2003,
260: see Fig.~\ref{flux}.
261: \footnote{In order to reduce the effects of
262: x[systematic uncertainties in the sensitivity of the two instruments,
263: we have normalized the fluxes to those of the first observation
264: with each instrument. The normalized fluxes can be converted to
265: absolute fluxes by multiplication with
266: $(4.3\pm0.1)\times10^{-12}$~erg\,s$^{-1}$cm$^{-2}$ (23-38~\AA)
267: and $(1.7\pm0.2)\times10^{-12}$~erg\,s$^{-1}$cm$^{-2}$\ (10-23~\AA).}
268:
269:
270: The alternative spectral model discussed in \citet{devries04},
271: a blackbody with
272: Gaussian absorption, e.g. caused by proton cyclotron absorption, has
273: three additional parameters: central energy, $E_0$, width, $\sigma_E$,
274: and normalization.
275: Such a model has been considered by \citet{vankerkwijk04} for RX J1605.3+3249,
276: and for \rxj\ by \citet{haberl03b},
277: who found a central energy of 271~eV and width $\sigma_E = 64$~eV.
278: Reanalyzing the RGS spectra for this study,
279: and simultaneously fitting all spectra with a blackbody model with
280: Gaussian absorption with one central energy and width gives
281: $E_0 = 306\pm 8$~eV (40.5~\AA) and $\sigma_E = 137\pm8$~eV.
282: The central energy is outside the RGS spectral range
283: and the width is so large that, as far as the RGS
284: spectra are concerned, the Gaussian absorption component
285: is indistinguishable from an exponential absorption feature. However,
286: the RGS spectra are inconsistent with the best-fit Gaussian parameters
287: found by \citet{haberl03b}, mainly because the RGS spectra require
288: a broader absorption feature than a Gaussian absorption with
289: $\sigma_E = 64$~eV.
290:
291:
292: Whatever the best model to describe the deviations from a blackbody
293: spectrum, the new LETGS spectrum breaks the trend that
294: a spectral hardening is accompanied by a decrease of long wavelength
295: emission.
296: The flux in the spectral range
297: from 10-90 \AA\ has slightly increased since February 2000 by 8\%,
298: but the best-fit parameters of the blackbody model imply
299: that the total blackbody flux has hardly changed.
300: Taken at face value this would imply that the emission area must have
301: decreased (Table~\ref{bbody}).
302:
303:
304: It is of interest that apart from a long term change in the spectrum,
305: \xmm\ CCD spectroscopy indicates that the spectral shape is also a function
306: of pulse phase \citep{cropper01,haberl03b},
307: and is in itself also subject to evolution \citep{devries04}.
308: Unfortunately, the RGS timing resolution ($\sim$4~s)
309: is not sufficient to
310: check the shape of the long wavelength attenuation as a function of pulse
311: phase.
312: On the other hand,
313: the timing resolution of \chandra's HRC-S detector is $\sim4$~ms.
314: We therefore folded the events of both the February 2000 and 2004
315: observation with the 8.3911~s period
316: \citep[see][for the latest timing results]{cropper04}, and added a timing
317: offset in order to align the folded light curves
318: so that the maximum emission occurs at phase 0.25 and the
319: minimum at 0.75. We then extracted spectra for the phase bins 0.125 - 0.375
320: and 0.625-0.875 (Fig.~\ref{phases}).
321:
322:
323:
324: Apart from an obvious variation in brightness between on and off peak,
325: the overall shapes of the spectra do not change much.
326: In order to test this we statistically compared the two spectra
327: for each observation by direct comparison of the countrates per spectral bin.
328: We assumed that the spectra are the same in both phase bins, except for
329: a normalization factor.
330: The best fitting normalization factors, i.e. the ratio
331: in flux between on and off peak emission, are
332: $1.19\pm0.06$\ (Feb. 2000) and $1.22\pm0.05$ (Feb 2004),
333: which is consistent with previous published pulse profiles
334: \citep{cropper01,devries04}.
335: Comparing the on peak spectra with the off peak spectra multiplied by
336: the normalization factor we obtained
337: respectively $\chi^2/dof = 23.38/15$\ and $\chi^2/dof = 17.77/15$,
338: which corresponds to probabilities of 8\% and 27\% that on and off peak
339: spectra are similar.
340: Our limits for variations are still consistent with the results
341: of \citet{cropper01},
342: who found variable absorption of the order of
343: $\Delta$\nh$ = 4\times10^{19}$~cm$^{-2}$.
344: Note that
345: although the LETGS has a higher spectral resolution,
346: the \xmm\ CCD instruments have a much higher effective area.
347: Phase resolved spectroscopy with the LETGS does, however,
348: indicate that whatever the phase dependent spectral changes, they must be
349: rather subtle and are probably caused by broad features.
350:
351:
352: \begin{figure*}
353: \centerline{
354: \psfig{figure=f3a.eps,angle=-90,width=0.5\textwidth}
355: \psfig{figure=f3b.eps,angle=-90,width=0.5\textwidth}}
356: \caption{
357: Phase resolved LETGS spectra obtained from
358: the February 2000 (left) and February 2004 observations (right).
359: The spectral extraction was done for phases 0.125 - 0.375 (open squares)
360: and 0.625-0.875 (filled triangles). The residuals take into account the
361: normalization differences.
362: \label{phases}}
363: \end{figure*}
364:
365:
366:
367: \section{Discussion}
368: Following the discovery of long term spectral changes in the X-ray emission
369: of \rxj\ by \xmm, the new \chandra\ DDT observation shows that the
370: X-ray spectrum of \rxj\ has continued to evolve.
371: In particular
372: the spectrum has remained hard from October 2003 to February 2004, but,
373: somewhat unexpected,
374: the attenuation of emission at wavelengths longer than 23~\AA\
375: has disappeared.
376:
377: A simple minded explanation would be that part of \rxj\ surface must have
378: heated up in the course of the last 4 years, initially accompanied by
379: absorption, but which disappeared between October 2003 and February 2004.
380: The absorption mechanism could be proton cyclotron absorption, as
381: was invoked to explain the broad absorption features in the spectra of
382: other isolated neutron stars.
383:
384: However, such an explanation is not completely
385: consistent with the evolution of the spectrum of \rxj. First of all,
386: if the surface is heated up by a starquake \citep{larson02}, or by
387: increased accretion from the interstellar medium,
388: we would expect that an additional hot spot at the surface would appear,
389: or that the existing hot spot would become hotter. As a result the total
390: blackbody flux would increase. This is, however, in
391: disagreement with our best-fit blackbody parameters and the flux
392: decrease between 45-55~\AA\ (Fig.~\ref{ddt}).
393: Note that heating as a result of accretion from the interstellar medium
394: is also unlikely given the the high proper motion
395: of the neutron star \citep[$97\pm12$~mas/yr,][]{motch03}.
396:
397: Our suggestion that the spectral evolution of \rxj\ is due
398: to precession \citep{devries04},
399: can at least qualitatively explain the results.
400: It assumes that the emission remains unchanged, but that we see the
401: hot spot under different viewing angles.
402: Free precession has been observed in some pulsars \citep[e.g.][]{stairs}.
403: Recently \citet{wasserman03} has calculated the expected precession period
404: caused by an oblique magnetic field and magnetic stresses in the neutron
405: crust. His model requires a Type II superconducting interior.
406: The expected precession period scales as $P_p\propto P_0B^{-1}$, with $P_0$ the
407: rotation period.
408: For \rxj, which has an inferred magnetic field of
409: $B=3\times 10^{13}$~G \citep{cropper04},
410: the expected precession period is $P_p\sim4$~yr, which is consistent
411: with the timescale of the spectral evolution.
412: Moreover, \xmm\ CCD spectra indicate that a modulation of the spectrum
413: with pulse phase is present, and the fact that pulsation
414: can be observed at all requires anisotropic surface emission \citep{cropper01}.
415: As the phase modulation of the spectrum also implies a spectral change
416: with viewing angle, it is not unreasonable to assume that precession gives
417: rise to a similar modulation of the spectrum with precession phase.
418:
419: Theoretically a strong effect of the viewing angle on the observed spectrum
420: from a highly magnetized atmosphere is to be expected.
421: The reason is that protons and electrons in the neutron star atmosphere
422: are constrained to move along the magnetic field lines.
423: This results in
424: radiation that is strongly polarized and angle dependent.
425: This is further enhanced for magnetic fields $B \gtrsim 10^{13}$~G,
426: for which vacuum polarization becomes important. This is
427: a quantum electrodynamics effect that changes
428: the polarization mode as the photon traverses a density gradient
429: \citep{oezel01,ho04}.
430:
431: The new \chandra\ observation is not able to confirm that the
432: attenuation has an approximate Gaussian shape, as may be expected for
433: proton cyclotron absorption. However, as the spectrum of \rxj\
434: continues to evolve, future \chandra-LETGS observations
435: may be able to constrain the spectral model further.
436: Long term monitoring is necessary to test the precession idea,
437: since it implies a cyclic spectral evolution.
438: \citet{kaplan03b} suggested that \rxj\ is an off-beam radio pulsar.
439: If it is indeed precessing, there is a possibility that at some point
440: during its cycle the radio beam will be directed toward the earth.
441:
442: Finally, we point out that other isolated neutron stars may show similar
443: behavior, indicative of precession. Further observations of those sources
444: is therefore important for increasing our understanding of
445: both the hot surface of neutron stars and for probing
446: their internal structure.
447:
448:
449: \acknowledgements
450: We thank Dr. Harvey Tananbaum for allowing
451: \rxj\ to be observed as part of the
452: DDT program, and for his continued interest
453: in the results. We thank Rob van der Meer for his assistance with the LETGS
454: data analysis.
455:
456:
457:
458: \begin{thebibliography}{19}
459: \expandafter\ifx\csname natexlab\endcsname\relax\def\natexlab#1{#1}\fi
460:
461: \bibitem[{{Arnaud}(1996)}]{xspec}
462: {Arnaud}, K.~A. 1996, in ASP Conf. Ser. 101: Astronomical Data Analysis
463: Software and Systems V, Vol.~5, 17
464:
465: \bibitem[{{Cash}(1979)}]{cash79}
466: {Cash}, W. 1979, \apj, 228, 939
467:
468: \bibitem[{Cropper {et~al.}(2004)Cropper, Haberl, Zane, \& Zavlin}]{cropper04}
469: Cropper, M., Haberl, F., Zane, S., \& Zavlin, V. 2004
470:
471: \bibitem[{{Cropper} {et~al.}(2001){Cropper}, {Zane}, {Ramsay}, {Haberl}, \&
472: {Motch}}]{cropper01}
473: {Cropper}, M., {Zane}, S., {Ramsay}, G., {Haberl}, F., \& {Motch}, C. 2001,
474: \aap, 365, L302
475:
476: \bibitem[{{de Vries} {et~al.}(2004){de Vries}, {Vink}, {M{\' e}ndez}, \&
477: {Verbunt}}]{devries04}
478: {de Vries}, C.~P., {Vink}, J., {M{\' e}ndez}, M., \& {Verbunt}, F. 2004, \aap,
479: 415, L31
480:
481: \bibitem[{{Haberl} {et~al.}(1997){Haberl}, {Motch}, {Buckley}, {Zickgraf}, \&
482: {Pietsch}}]{haberl97}
483: {Haberl}, F., {Motch}, C., {Buckley}, D.~A.~H., {Zickgraf}, F.-J., \&
484: {Pietsch}, W. 1997, \aap, 326, 662
485:
486: \bibitem[{{Haberl} {et~al.}(2003){Haberl}, {Schwope}, {Hambaryan}, {Hasinger},
487: \& {Motch}}]{haberl03a}
488: {Haberl}, F., {Schwope}, A.~D., {Hambaryan}, V., {Hasinger}, G., \& {Motch}, C.
489: 2003, \aap, 403, L19
490:
491: \bibitem[{Haberl {et~al.}(2003)Haberl, Zavlin, Truemper, \&
492: Burwitz}]{haberl03b}
493: Haberl, F., Zavlin, V.~E., Truemper, J., \& Burwitz, V. 2003
494:
495: \bibitem[{Ho \& Lai(2004)}]{ho04}
496: Ho, W. C.~G. \& Lai, D. 2004, \apj \ (in press)
497:
498: \bibitem[{{Kaplan} {et~al.}(2003){Kaplan}, {van Kerkwijk}, {Marshall},
499: {Jacoby}, {Kulkarni}, \& {Frail}}]{kaplan03b}
500: {Kaplan}, D.~L., {van Kerkwijk}, M.~H., {Marshall}, H.~L., {Jacoby}, B.~A.,
501: {Kulkarni}, S.~R., \& {Frail}, D.~A. 2003, \apj, 590, 1008
502:
503: \bibitem[{{Larson} \& {Link}(2002)}]{larson02}
504: {Larson}, M.~B. \& {Link}, B. 2002, \mnras, 333, 613
505:
506: \bibitem[{{Lenzen} \& {Truemper}(1978)}]{lenzen78}
507: {Lenzen}, R. \& {Truemper}, J. 1978, \nat, 271, 216
508:
509: \bibitem[{{Motch} {et~al.}(2003){Motch}, {Zavlin}, \& {Haberl}}]{motch03}
510: {Motch}, C., {Zavlin}, V.~E., \& {Haberl}, F. 2003, \aap, 408, 323
511:
512: \bibitem[{{{\" O}zel}(2001)}]{oezel01}
513: {{\" O}zel}, F. 2001, \apj, 563, 276
514:
515: \bibitem[{{Paerels} {et~al.}(2001)}]{paerels01}
516: {Paerels}, F. {et~al.} 2001, \aap, 365, L298
517:
518: \bibitem[{{Stairs} {et~al.}(2000){Stairs}, {Lyne}, \& {Shemar}}]{stairs}
519: {Stairs}, I.~H., {Lyne}, A.~G., \& {Shemar}, S.~L. 2000, \nat, 406, 484
520:
521: \bibitem[{{Turolla} {et~al.}(2004){Turolla}, {Zane}, \& {Drake}}]{turolla04}
522: {Turolla}, R., {Zane}, S., \& {Drake}, J.~J. 2004, \apj, 603, 265
523:
524: \bibitem[{van Kerkwijk {et~al.}(2004)van Kerkwijk, Kaplan, Durant, Kulkarni, \&
525: Paerels}]{vankerkwijk04}
526: van Kerkwijk, M.~H., Kaplan, D.~L., Durant, M., Kulkarni, S.~R., \& Paerels, F.
527: 2004
528:
529: \bibitem[{{Wasserman}(2003)}]{wasserman03}
530: {Wasserman}, I. 2003, \mnras, 341, 1020
531:
532: \end{thebibliography}
533:
534:
535:
536:
537:
538:
539: \end{document}
540: