1: \documentclass[preprint2]{aastex}

2: 

3: \begin{document}

4: 

5: \newcommand{\psra}{\mbox{PSR J0737$-$3039A}}

6: \newcommand{\psrb}{\mbox{PSR J0737$-$3039B}}

7: 

8: \title{The Mean Pulse Profile of \psra}

9: \author{R.\ N.\ Manchester\altaffilmark{1}, M.\ Kramer\altaffilmark{2}, 

10: A. Possenti\altaffilmark{3}, A. G. Lyne\altaffilmark{2},   \\

11: M.\ Burgay\altaffilmark{3}, I. H. Stairs\altaffilmark{4}, 

12: A.\ W.\ Hotan\altaffilmark{5}, M.\ A.\ McLaughlin\altaffilmark{2},   \\

13: D.\ R.\ Lorimer\altaffilmark{2}, G.\ B.\ Hobbs\altaffilmark{1}, 

14: J.\ M. Sarkissian\altaffilmark{1}, N.\ D'Amico\altaffilmark{6},  \\

15: F. Camilo\altaffilmark{7}, B. C. Joshi\altaffilmark{8} \& 

16: P. C. C. Freire\altaffilmark{9}}

17: 

18: \altaffiltext{1}{Australia Telescope National Facility, CSIRO,

19: P.O. Box 76, Epping NSW 1710, Australia}

20: \altaffiltext{2}{Jodrell Bank Observatory, University of Manchester, Macclesfield, 

21: Cheshire, SK11 9DL, UK}

22: \altaffiltext{3}{INAF - Osservatorio Astronomico di Cagliari, Loc. Poggio dei

23: Pini, Strada 54, 09012 Capoterra, Italy}

24: \altaffiltext{4}{Dept. of Physics and Astronomy, University of British

25: Columbia, 6224 Agricultural Road, Vancouver BC V6T 1Z1, Canada}

26: \altaffiltext{5}{Swinburne Centre for Astrophysics \& Supercomputing, Hawthorn, 

27: VIC 3122, Australia}

28: \altaffiltext{6}{Universit\`a degli Studi di

29: Cagliari, Dipartimento di Fisica, 09042 Monserrato, Italy}

30: \altaffiltext{7}{Columbia Astrophysics Laboratory, Columbia University, 

31: 550 West 120th Street,  NY 10027}

32: \altaffiltext{8}{National Centre for Radio Astrophysics, P.O. Bag 3, Ganeshkhind, 

33: Pune 411007, India}

34: \altaffiltext{9}{NAIC, Arecibo Observatory, HC03 Box 53995, PR 00612}

35: 

36: \begin{abstract}

37: General relativity predicts that the spin axes of the pulsars in the

38: double-pulsar system (PSR J0737$-$3039A/B) will precess rapidly, in

39: general leading to a change in the observed pulse profiles. We have

40: observed this system over a one-year interval using the Parkes 64-m

41: radio telescope at three frequencies: 680, 1390 and 3030 MHz. These

42: data, combined with the short survey observation made two years

43: earlier, show no evidence for significant changes in the pulse profile

44: of \psra, the 22-ms pulsar. The limit on variations of the profile

45: 10\% width is about $0\fdg5$ per year. These results imply an angle

46: $\delta$ between the pulsar spin axis and the orbit normal of

47: $\lesssim 60 \degr$, consistent with recent evolutionary studies of

48: the system. Although a wide range of system parameters remain

49: consistent with the data, the model proposed by \citet{jr04} can be

50: ruled out. A non-zero ellipticity for the radiation beam gives

51: slightly but not significantly improved fits to the data, so that a

52: circular beam describes the data equally well within the

53: uncertainties.

54: \end{abstract}

55: 

56: \keywords{pulsars: general --- pulsars: individual (J0737--3039A)}

57: 

58: \section{Introduction} \label{sec:intro}

59: The double-pulsar system PSR J0737$-$3039A/B \citep{bdp+03,lbk+04}

60: provides a wonderful laboratory for investigations of relativistic

61: gravity \citep{klb+04} as well as the physics of pulsar magnetospheres

62: \citep[e.g.,][]{mkl+04}. The system consists of a 22-ms pulsar, \psra,

63: in a 2.4-h binary orbit with \psrb, a younger pulsar with a spin

64: period of 2.7 s. The system is mildly eccentric ($e \sim 0.088$) and

65: is viewed nearly edge-on (orbit inclination $i \sim 88\degr$). With

66: mean orbital speeds $v \sim 0.001 c$, the system is highly

67: relativistic, allowing the detection of four ``post-Keplerian''

68: parameters in just six months of observation \citep{lbk+04}.  The

69: post-Keplerian parameters, together with the mass ratio, uniquely

70: measurable in this double-pulsar system, give accurate values for the

71: masses of the two stars as well as stringent tests of general

72: relativity \citep{klb+04}.

73: 

74: Within the framework of general relativity, the spin vectors of the

75: two pulsars are expected to exhibit geodetic precession about the

76: total angular momentum of the system \citep{dr74,bo75}. Since the

77: total angular momentum is dominated by the orbital motion, the pulsar

78: spin vectors effectively precess about the orbit normal. The

79: precession rate $\Omega$ depends on the Keplerian parameters and the

80: masses of both pulsars, with predicted precessional periods of $\sim

81: 75$ yr and $\sim 71$ yr for \psra~and \psrb~ respectively

82: \citep{lbk+04}. These are about a factor of four shorter than the value

83: for PSR B1913+16, the Hulse-Taylor binary pulsar (Weisberg, Romani \&

84: Taylor 1989)\nocite{wrt89}. In general, precession will result in a

85: change of the width and shape of the observed profile as the angle of

86: the line of sight with respect to the pulsar spin axis ($\zeta$)

87: changes during the precessional period. Long-term evolution of the

88: mean pulse profile has been observed in PSR B1913+16

89: \citep{wrt89,kra98} and PSR B1534+12 \citep{sta04} and interpreted as

90: evidence for geodetic precession. This gives the three-dimensional

91: geometry (within a $180\degr$ ambiguity) of the system and, in the

92: case of PSR B1913+16, a prediction that the pulsar will no longer be

93: visible after the year 2025. \citet{wt02} have used the precessional

94: motion to produce a two-dimensional image of the part of the PSR

95: B1913+16 emission beam traversed so far, suggesting that the beam is

96: elongated in the latitudinal direction.

97: 

98: If the magnetic-pole model \citep{rc69a} is assumed, $\zeta$ and the

99: inclination angle of the magnetic axis with respect to the spin axis

100: ($\alpha$) can be deduced from a fit of the model to the observed

101: variations of polarization position-angle, at least for wide profiles

102: \citep[e.g.,][]{lm88}. \citet{drb+04} used such an analysis to

103: conclude that \psra~is a nearly aligned system with $\alpha \sim

104: 4\degr$, but the fit to the observed position angle variations is

105: poor. The observer angle $\zeta$ is unconstrained by this

106: fit. \citet{jr04} determined possible geometries of the

107: J0737$-$3039A/B system based on a model for the orbital modulation of

108: the \psrb~pulse intensity. Two solutions were obtained with

109: ($\alpha,\delta$) of ($1\fdg6 \pm 1\fdg3, 167\degr \pm 10\degr$) and

110: ($14\degr \pm 2\degr, 90\degr \pm 10\degr$) respectively. The first of

111: these is consistent with the \citet{drb+04} solution. Both solutions

112: predict a rapid evolution of the observed profile width of \psra, with

113: expected changes of $\sim 42\degr$ and $\sim 96\degr$ per year for the

114: two models. Here we report on observations of the mean pulse profile

115: of \psra~over a three-year interval, 2001 August to 2004 August.

116: 

117: \section{Observations and Analysis} \label{sec:obs}

118: We have used the Parkes 64-m radio telescope to observe the PSR

119: J0737$-$3039A/B system since 2003 May 1 (MJD 52670), shortly after its

120: confirmation, to 2004 August 8 (MJD 53225). Observations were made in

121: three frequency bands centered at 680, 1390 and 3030 MHz, with the

122: 680-MHz (50cm) and 3030-MHz (10cm) observations commencing in 2003,

123: December. Before 2003 October, the 1390 MHz observations were made

124: using the center beam of the multibeam receiver, which has an

125: equivalent system noise of approximately 29 Jy. In addition, we have

126: the original survey pointing, a 4-min observation made at 1390 MHz on

127: 2001 August 22 \citep{bdp+03}. After 2003 October, the H-OH receiver

128: with a system noise of about 42 Jy was used. Both the multibeam and

129: H-OH systems were dual polarization and used a filterbank system

130: consisting of $2 \times 512 \times 0.5$ MHz channels. After detection,

131: the two polarizations for each channel were summed, sampled at either

132: 80 $\mu$s or 125 $\mu$s intervals and recorded to tape for subsequent

133: processing. The other two bands were observed using the

134: dual-polarization coaxial 10cm/50cm receiver, with system noises of

135: $\sim 48$ Jy and $\sim 64$ Jy at 10cm and 50cm respectively. A

136: filterbank system with $2 \times 256 \times 0.25$ MHz channels was

137: used with the 50cm receiver and at 10cm the filterbank system had $2

138: \times 192 \times 3$ MHz channels.

139: 

140: Over the 15-month interval, we obtained 59 observations at 1390 MHz,

141: 17 observations at 680 MHz and 27 at 3030 MHz, all with durations

142: ranging from 10 min to 5 hr. Data from each observation

143: were folded at the apparent topocentric period of \psra~to form mean

144: pulse profiles with 256 phase bins, which were then summed in

145: frequency and time to form a single profile using the PSRCHIVE data

146: analysis system (Hotan, van Straten \& Manchester

147: 2004)\nocite{hvm04}. All observations for each frequency band were

148: then summed to form grand-average profiles containing 116, 27 and 34

149: hrs of data, respectively. These profiles are shown in

150: Figure~\ref{fg:3frq}.

151: 

152: \begin{figure}[ht]

153: \includegraphics[scale=0.35,angle=270]{f1.eps}

154: \caption{Mean pulse profiles for \psra~at three frequencies, where

155:   $360\degr$ of longitude corresponds to the pulse period. The 680-MHz

156:   and 3030-MHz profiles were aligned to give maximum cross-correlation

157:   with the 1390-MHz profile. }\label{fg:3frq}

158: \end{figure} 

159: 

160: The 1390-MHz observations over the 15 months from 2003 May were

161: divided into eight chronologically-ordered groups and average profiles

162: were formed for each of these eight groups. Integration times ranged

163: between 5 hrs (MJD 53068) and 20 hrs (MJD 53000). Examination of these

164: average profiles showed little or no evidence for evolution of the

165: mean pulse profile over the 15-month interval. The eight profiles were

166: each aligned with the grand-average profile using a cross-correlation

167: analysis and baselines subtracted, taking the longitude range

168: $310\degr$ to $50\degr$ (Figure~\ref{fg:3frq}) as the baseline

169: region. Each profile was scaled to have the same area as the

170: grand-average profile, which was then subtracted from it to form the

171: difference profiles shown in Figure~\ref{fg:diff}. Also shown at the

172: bottom of the figure is the difference profile corresponding to a

173: stretched version of the grand-average profile, demonstrating the

174: signature of a change in profile width. The adopted stretch factor

175: (1.004) corresponds to an increase of about $1\degr$ in profile width,

176: where the width is defined to be the separation of the steep outer

177: edges of the profile (at roughly $60\degr$~and $300\degr$~in

178: Figures~\ref{fg:3frq} and \ref{fg:diff}) at 10\% of the amplitude of

179: the second (stronger) component.

180: 

181: Significant differences are seen, especially between MJDs of 53000 and

182: 53103. In general, these differences do not have the signature of an

183: increase or decrease in profile width, but rather appear to result

184: from changes in the relative amplitude of the various pulse

185: components, especially at the leading and trailing components of the

186: profile. Polarization observations \citep[e.g.,][]{drb+04} show that

187: these components have high linear polarization. It is likely that the

188: observed changes are an instrumental effect resulting from differences

189: in receiver gain between the two polarization channels combined with

190: parallactic angle variations. 

191: 

192: These difference profiles confirm the lack of significant secular

193: profile evolution. There is no evidence for a systematic change in the

194: overall profile width to a level of much less than $1\degr$ and there

195: is no evidence for any intrinsic changes in the width or shape of the

196: leading and trailing pulse components considered separately. We have

197: also reprocessed the original 4-min survey pointing to give a profile

198: with signal-to-noise ratio of about 25. Again, this shows no evidence

199: for any significant change in profile shape or width.

200: 

201: \begin{figure}[ht]

202: \includegraphics[scale=0.45]{f2.eps}

203: \caption{Difference profiles for \psra~at 1390~MHz relative to the

204: grand-average pulse profile shown at the top of the figure for eight

205: epochs, together with the difference profile corresponding to the

206: grand-average profile stretched by 0.4\%. The vertical scale of the

207: difference profiles is a factor of ten larger than that of the

208: grand-average profile.}\label{fg:diff}

209: \end{figure} 

210: 

211: In order to quantify the profile stability, widths at 10\% of the

212: pulse peak were measured for each of the eight 1390-MHz profiles and

213: the 2001 observation using a cross-correlation technique.  The leading

214: ($50\degr$ -- $130\degr$ in Figures~\ref{fg:3frq} and \ref{fg:diff})

215: and trailing ($240\degr$ -- $310\degr$) pulse components were

216: separately cross-correlated with a standard profile derived from the

217: grand-average profile containing just the relevant component,

218: determining the phase of each component relative to a reference

219: phase. The difference of these phases was then added to the 10\% width

220: of the grand-average profile to give the 10\% width at each

221: epoch. Errors from the cross-correlation analysis were increased in

222: quadrature by 10 $\mu$s to allow for systematic effects. The derived

223: 10\% widths are plotted in Figure~\ref{fg:width} along with the fitted

224: trend line. Clearly, there is no significant trend, with the fitted

225: slope being $-0\fdg2 \pm 0\fdg3$ per year. The 10\% profile width at

226: MJD 53000 (2003 December 27) from the fit is $238\fdg45 \pm

227: 0\fdg12$. Quoted errors are $\pm 2\sigma$ for both the width and

228: slope.

229: 

230: \begin{figure}[ht]

231: \includegraphics[scale=0.3,angle=270]{f3.eps}

232: \caption{Profile widths at 10\% of the peak amplitude for \psra~at

233: 1390 MHz as a function of MJD relative to 53000. Error bars are $\pm 1

234: \sigma$. The dashed line is a formal weighted least-squares fit of a linear

235: trend to these data.}\label{fg:width}

236: \end{figure} 

237: 

238: \section{Discussion and Conclusions} \label{sec:discn}

239: Figure~\ref{fg:3frq} confirms the overall bilateral symmetry of the

240: \psra~pulse profile, consistent with its origin from magnetic field

241: lines associated with a single magnetic pole

242: \citep{bdp+03,drb+04}. There are significant differences in the

243: spectral indices of the various pulse components. The most striking of

244: these is the relatively flat spectrum of the component at the extreme

245: leading edge of the profile which is relatively much stronger at 3030

246: MHz. The 3030-MHz profile also shows a matching flat-spectrum

247: component at the extreme trailing edge of the profile, reinforcing the

248: idea that both components of the observed profile originate

249: on field lines associated with a single magnetic pole. We assume this

250: in the following discussion.

251: 

252: The main result of this paper is the extraordinary stability of the

253: \psra~pulse profile over more than three years, a significant fraction

254: of the expected geodetic precessional period of 75 years. Provided the

255: angle between the spin axis of the pulsar and the orbital angular

256: momentum ($\delta$) is non-zero, we would expect to see a variation in

257: profile width as a function of time. In particular, the system

258: geometries proposed by \citet{jr04} predict changes of many tens of

259: degrees per year in the observed profile width. Clearly, these are not

260: observed, ruling out this model in its present form.

261: 

262: One possible explanation for the observed lack of variation in

263: profile width is that the spin axis of \psra~is aligned with the

264: orbital angular momentum, i.e., $\delta \sim 0$. Since we view the

265: orbit nearly edge-on, even for an orthogonal magnetic axis ($\alpha

266: \sim 90\degr$), this would require an effective beam radius $\rho >

267: 90\degr$, which seems unlikely. Furthermore, given that \psrb~probably

268: suffered a significant natal kick, an aligned rotation axis for \psra~

269: is possible but unlikely (Willems, Kalogera \& Henninger

270: 2004)\nocite{wkh04b}.

271: 

272: We have investigated the limits which can be placed on the geometry of

273: the system by modeling the observed profile width as a function of

274: time \citep[cf.][]{kra98}. An inclination angle $i = 88\degr$

275: is assumed\footnote{Since only $\sin i$ is determined at present,

276: equivalent solutions are possible with $i=92\degr$ and $\alpha

277: \rightarrow 180\degr-\alpha$, etc.}.  For a circular emission beam,

278: four parameters are sufficient to model the expected profile changes:

279: $\alpha$, $\delta$, $\rho$ and an epoch $T_0$ describing the

280: precession phase. We fitted a circular beam model with $\rho<90\degr$

281: to the data and derive a $\chi^2$-sphere plot for $\alpha$ and

282: $\delta$ as shown in Figure~\ref{fg:prec}. This figure shows that

283: solutions with misalignment angle $\delta \lesssim 60\degr$ are

284: preferred, consistent with the conclusions of \citet{wkh04b} based on

285: the observed velocity of the system.

286: 

287: The formal best-fit solution (reduced $\chi^2 = 2.88$) is located at

288: $\alpha\sim 19\degr$ and $\delta \sim 14\degr$ (and at a corresponding

289: mirror-solution of $\alpha\sim 161\degr$ and $\delta \sim

290: 166\degr$). The predicted time variation for the observed profile width

291: for this configuration is shown in Figure~\ref{fg:prec}(a); pulses

292: would be detected over the whole precessional period with an

293: approximately sinusodial variation in the profile width. However, the

294: $\chi^2$-sphere is rather flat, and a statistically satisfactory

295: solution is possible over a wide range of

296: angles. Figure~\ref{fg:prec}(b) shows a solution with small $\alpha$

297: and $\delta$ which is consistent with the preferred solution of

298: \citet{drb+04} based on the observed position angle variations. Both

299: of these solutions have a beam radius of very close to $90\degr$

300: corresponding to a fan beam. In case (b) however, the pulse is visible

301: for a limited period only, roughly 1984 to 2023. Figure~\ref{fg:prec}(c)

302: shows a solution with intermediate $\alpha$ and $\delta$ for which the

303: beam radius is smaller, about $65\degr$. In all solutions, the

304: observed pulse is emitted from field lines associated with a single

305: magnetic pole as indicated by the observed profile symmetry.

306: 

307: \begin{figure}[ht]

308: \includegraphics[scale=0.35,angle=270]{f4.eps}

309: \caption{Results of fits to the observed profile widths of a model for

310:   geodetic precession of the pulsar spin axis, inclined by an angle

311:   $\delta$ to the orbit normal, with a circular beam of radius $\rho$

312:   inclined at an angle $\alpha$ to the pulsar spin axis. The upper

313:   part shows contours of $\chi^2$ at 1-$\sigma$ intervals, relative to

314:   the minimum value, with the darker regions having lower

315:   $\chi^2$. The lower cutoff to the allowed regions represents the

316:   locus of points where the beam radius $\rho = 90\degr$. These

317:   contours are mirrored for $90\degr < \delta < 180\degr$ with

318:   ($\alpha$, $\delta$) corresponding to ($180\degr - \alpha$, $180\degr

319:   - \delta$). In the lower part of the figure, the variation of

320:   profile width is shown as a function of time for three cases: (a) a

321:   point near the global minimum for $\chi^2$, (b) a small $\alpha$,

322:   small $\delta$ case, and (c) a solution with intermediate $\alpha$

323:   and $\delta$. The angle $\zeta$ is the minimum value of the

324:   inclination of the line of sight to the pulsar spin axis which

325:   occurs at precessional phase zero (or $180\degr$) when the pulsar

326:   spin axis is in the plane defined by the orbit normal and the line

327:   of sight.}\label{fg:prec}

328: \end{figure} 

329: 

330: Unless $\delta\sim0$, the essentially constant profile width observed

331: so far implies that the system is at or close to precessional phase

332: zero (or $180\degr$). At these times is ${\rm d}W/{\rm d}t \sim

333: 0$.\footnote{There is a special case with $\cos\rho\sin\delta\cos\Phi

334: = \cos\alpha$, where $\Phi$ is the precessional phase, which results in

335: ${\rm d}W/{\rm d}t = 0$, but this requires an improbable fine-tuning

336: of the system parameters.}  This is possible -- a similar situation

337: evidently applies for PSR B1913+16 \citep{kra98} -- but statistically

338: unlikely. A possible way of relaxing this constraint is to allow a

339: non-circular beam.  \citet{big90b} and \citet{ks98} have shown that,

340: under certain assumptions, the polar emission beam is compressed in

341: the latitudinal direction for $\alpha>0\degr$. On the other hand,

342: polarization observations of young and millisecond pulsars

343: \citep[e.g.,][]{nv83,mh04} suggest beams effectively elongated in the

344: latitude direction. First, we took the model of \citet{ks98} in which

345: the beam compression is a function of $\alpha$ only and recomputed the

346: $\chi^2$-sphere. The results were only marginally different to those

347: shown in Figure~\ref{fg:prec}. We then took the beam ellipticity as a

348: free parameter, first doing a grid search in ellipticity at each point

349: on the ($\alpha$,$\delta$) plane and then searching for a global

350: $\chi^2$ minimum in ($\rho_m$, $T_0$, $\epsilon$) space, where

351: $\rho_m$ is the beam major axis, $T_0$ is the time of precessional

352: phase zero and $\epsilon$ is the beam ellipticity.  Latitudinally

353: compressed beams with axial ratio $\rho_{\rm lat}/\rho_{\rm long}

354: \lesssim 0.5$ are restricted to a fairly small range of $\alpha$ and

355: $\delta$ around $45\degr$. On the other hand, fan-like beams with

356: large axial ratio allow a much larger range of $\alpha$ and $\delta$

357: with values $\lesssim 30\degr$ preferred. However, in none of these

358: cases was the $\chi^2$ value significantly better than those found in

359: the circular-beam case.

360: 

361: In conclusion, we find that the pulse-width variations (or lack of

362: them) observed so far allow a wide range of configurations for the PSR

363: J0737$-$3039A/B system. Models with non-circular beams give somewhat

364: better fits to the data but are statistically indistinguishable from

365: fits with a circular beam. Misalignment angles $\delta \lesssim

366: 60\degr$ are generally preferred, consistent with the conclusions of

367: \citet{wkh04b}, and the configurations discussed by \citet{jr04}

368: can be ruled out. Solutions in which the pulse never disappears from

369: view are possible, as are solutions where it does, in the most extreme

370: cases about 15 years from now. Clearly, a longer time baseline, even

371: one more year, will help to constrain the models, as will a more

372: realistic interpretation of the observed position-angle variations.

373: 

374: \acknowledgments

375: 

376: The Parkes radio telescope is part of the Australia Telescope which is

377: funded by the Commonwealth Government for operation as a National

378: Facility managed by CSIRO. IHS holds an NSERC UFA and is supported by

379: a Discovery Grant. DRL is a University Research Fellow funded by the

380: Royal Society.  FC acknowledges support from NSF grant

381: AST-02-05853. NDA, AP and MB received support from the Italian

382: Ministry of University and Research (MIUR) under the national program

383: {\it Cofin 2003}.

384: 

385: %\bibliographystyle{apj} 

386: %\bibliography{journals,modrefs,psrrefs,crossrefs}

387: 

388: \begin{thebibliography}{20}

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

390: 

391: \bibitem[{Barker \& O'Connell(1975)}]{bo75}

392: Barker, B.~M. \& O'Connell, R.~F. 1975, ApJ, 199, L25

393: 

394: \bibitem[{Biggs(1990)}]{big90b}

395: Biggs, J.~D. 1990, MNRAS, 245, 514

396: 

397: \bibitem[{{Burgay} {et~al.}(2003){Burgay}, {D'Amico}, {Possenti}, {Manchester},

398:   {Lyne}, {Joshi}, {McLaughlin}, {Kramer}, {Sarkissian}, {Camilo}, {Kalogera},

399:   {Kim}, \& {Lorimer}}]{bdp+03}

400: {Burgay}, M., {D'Amico}, N., {Possenti}, A., {Manchester}, R.~N., {Lyne},

401:   A.~G., {Joshi}, B.~C., {McLaughlin}, M.~A., {Kramer}, M., {Sarkissian},

402:   J.~M., {Camilo}, F., {Kalogera}, V., {Kim}, C., \& {Lorimer}, D.~R. 2003,

403:   Nature, 426, 531

404: 

405: \bibitem[{Damour \& Ruffini(1974)}]{dr74}

406: Damour, T. \& Ruffini, R. 1974, C. R. Acad. Sc. Paris, Serie A, 279, 971

407: 

408: \bibitem[{{Demorest} {et~al.}(2004){Demorest}, {Ramachandran}, {Backer},

409:   {Ransom}, {Kaspi}, {Arons}, \& {Spitkovsky}}]{drb+04}

410: {Demorest}, P., {Ramachandran}, R., {Backer}, D.~C., {Ransom}, S.~M., {Kaspi},

411:   V., {Arons}, J., \& {Spitkovsky}, A. 2004, ApJ, 615, L137

412: 

413: \bibitem[{{Hotan} {et~al.}(2004){Hotan}, {van Straten}, \&

414:   {Manchester}}]{hvm04}

415: {Hotan}, A.~W., {van Straten}, W., \& {Manchester}, R.~N. 2004, PASA, 21, 302

416: 

417: \bibitem[{{Jenet} \& {Ransom}(2004)}]{jr04}

418: {Jenet}, F.~A. \& {Ransom}, S.~M. 2004, Nature, 428, 919

419: 

420: \bibitem[{{Kapoor} \& {Shukre}(1998)}]{ks98}

421: {Kapoor}, R.~C. \& {Shukre}, C.~S. 1998, ApJ, 501, 228

422: 

423: \bibitem[{{Kramer}(1998)}]{kra98}

424: {Kramer}, M. 1998, ApJ, 509, 856

425: 

426: \bibitem[{Kramer {et~al.}(2004)Kramer, {Lyne}, Burgay, Possenti, Manchester,

427:   Camilo, McLaughlin, Lorimer, D'Amico, Joshi, Reynolds, \& Freire}]{klb+04}

428: Kramer, M., {Lyne}, A.~G., Burgay, M., Possenti, A., Manchester, R.~N., Camilo,

429:   F., McLaughlin, M.~A., Lorimer, D.~R., D'Amico, N., Joshi, B.~C., Reynolds,

430:   J., \& Freire, P. C.~C. 2004, in {Binary Pulsars}, ed. F.~Rasio \& I.~H.

431:   Stairs (San Francisco: Astronomical Society of the Pacific), astro-ph/0405179

432: 

433: \bibitem[{Lyne {et~al.}(2004)Lyne, {Burgay}, Kramer, Possenti, Manchester,

434:   Camilo, {McLaughlin}, Lorimer, Joshi, Reynolds, \& Freire}]{lbk+04}

435: Lyne, A.~G., {Burgay}, M., Kramer, M., Possenti, A., Manchester, R.~N., Camilo,

436:   F., {McLaughlin}, M., Lorimer, D.~R., Joshi, B.~C., Reynolds, J.~E., \&

437:   Freire, P. C.~C. 2004, Science, 303, 1153

438: 

439: \bibitem[{Lyne \& Manchester(1988)}]{lm88}

440: Lyne, A.~G. \& Manchester, R.~N. 1988, MNRAS, 234, 477

441: 

442: \bibitem[{{Manchester} \& {Han}(2004)}]{mh04}

443: {Manchester}, R.~N. \& {Han}, J.~L. 2004, ApJ, 609, 354

444: 

445: \bibitem[{{McLaughlin} {et~al.}(2004){McLaughlin}, {Kramer}, {Lyne}, {Lorimer},

446:   {Stairs}, {Possenti}, {Manchester}, {Freire}, {Joshi}, {Burgay}, {Camilo}, \&

447:   {D'Amico}}]{mkl+04}

448: {McLaughlin}, M.~A., {Kramer}, M., {Lyne}, A.~G., {Lorimer}, D.~R., {Stairs},

449:   I.~H., {Possenti}, A., {Manchester}, R.~N., {Freire}, P.~C.~C., {Joshi},

450:   B.~C., {Burgay}, M., {Camilo}, F., \& {D'Amico}, N. 2004, ApJ, 613, L57

451: 

452: \bibitem[{Narayan \& Vivekanand(1983)}]{nv83}

453: Narayan, R. \& Vivekanand, M. 1983, A\&A, 122, 45

454: 

455: \bibitem[{Radhakrishnan \& Cooke(1969)}]{rc69a}

456: Radhakrishnan, V. \& Cooke, D.~J. 1969, Astrophys. Lett., 3, 225

457: 

458: \bibitem[{{Stairs} {et~al.}(2004){Stairs}, {Thorsett}, \&

459:   {Arzoumanian}}]{sta04}

460: {Stairs}, I.~H., {Thorsett}, S.~E., \& {Arzoumanian}, Z. 2004, Phys. Rev.

461:   Lett., 93, 141101

462: 

463: \bibitem[{Weisberg {et~al.}(1989)Weisberg, Romani, \& Taylor}]{wrt89}

464: Weisberg, J.~M., Romani, R.~W., \& Taylor, J.~H. 1989, ApJ, 347, 1030

465: 

466: \bibitem[{{Weisberg} \& {Taylor}(2002)}]{wt02}

467: {Weisberg}, J.~M. \& {Taylor}, J.~H. 2002, ApJ, 576, 942

468: 

469: \bibitem[{Willems {et~al.}(2004)Willems, Kalogera, \& Henninger}]{wkh04b}

470: Willems, B., Kalogera, V., \& Henninger, M. 2004, ApJ, 616, astro-ph/0404423

471: 

472: \end{thebibliography}

473: 

474: \end{document}

475: 

476: