0808.1594/ms.tex
1: \documentclass[12pt,preprint]{aastex}
2: \shorttitle{Accurate VLBI Astrometry of PSR J0437--4715}
3: \shortauthors{Deller, Verbiest, Bailes \& Tingay}
4: 
5: \begin{document}
6: 
7: \newcommand{\pbdot}{\ensuremath{\dot{P}_{\mathrm b}}}
8: \newcommand{\Gdot}{\ensuremath{\dot{G}}}
9: \newcommand{\degrees}{\ensuremath{^{\circ}}}
10: 
11: \title{Extremely high precision VLBI astrometry of PSR J0437--4715 and implications for 
12: theories of gravity}
13: 
14: \author{A.T. Deller\altaffilmark{1,2}, J.P.W. Verbiest\altaffilmark{1,2},  S.J. Tingay\altaffilmark{3} \& M. Bailes\altaffilmark{1}}
15: \altaffiltext{1}{Centre for Astrophysics and Supercomputing, Swinburne University of Technology, Mail H39, P.O. Box 218, Hawthorn, VIC 3122, Australia}
16: 
17: \altaffiltext{2}{co-supervised through the Australia Telescope National Facility, P.O. Box 76, Epping, NSW 1710, Australia}
18: 
19: \altaffiltext{3}{Department of Imaging and Applied Physics, Curtin University of Technology, Bentley, WA, Australia}
20: 
21: \begin{abstract}
22: Using the recently upgraded Long Baseline Array, we have measured the trigonometric parallax
23: of PSR J0437--4715 to better than 1\% precision, the most precise pulsar
24: distance determination made to date.  Comparing this VLBI distance measurement to
25: the kinematic distance obtained from pulsar timing, which is calculated from the pulsar's 
26: proper motion and apparent rate of change of orbital period,
27: gives a precise limit on the unmodeled relative acceleration between the Solar System
28: and PSR J0437--4715, which can be used in a variety of applications. Firstly,
29: it shows that Newton's gravitational constant $G$ is stable with time
30: ($\Gdot/G  = (-5 \pm 26) \times 10^{-13}$\ yr$^{-1}$, 95\% confidence). Secondly,
31: if a stochastic gravitational wave background existed at the 
32: currently quoted limit, this null result would fail
33: $\sim\!50$\% of the time. Thirdly, it excludes Jupiter-mass planets within 226 AU
34: of the Sun in 50\% of the sky (95\% confidence). Finally, the $\sim\!1$\% agreement of the parallax
35: and orbital period derivative distances provides a fundamental confirmation
36: of the parallax distance method upon which all astronomical distances are
37: based.
38: \end{abstract}
39: 
40: \keywords{astrometry --- pulsars: individual(J0437--4715) --- gravitation}
41: 
42: \section{Introduction}
43: PSR J0437--4715 has been observed intensively since its discovery by \citet{johnston93a}.
44: It is the brightest and nearest observed millisecond pulsar, and has also been 
45: studied in the optical \citep{bell93a}, ultraviolet \citep{kargaltsev04a}, and X--ray 
46: \citep{zavlin02a} bands.  The high rotational stability and close
47: proximity of this pulsar--white dwarf binary system make it an excellent
48: probe of General Relativity (GR) and alternative forms of gravitational theories.  The measurement
49: of its Shapiro delay by \citet{van-straten01a}, where the radio waves from the pulsar are delayed
50: as they pass through the gravitational potential of the companion, is one such test which has
51: shown consistency with GR predictions.  The search for the low frequency stochastic gravitational
52: wave background (GWB) using pulsar timing arrays 
53: \citep[e.g.][]{jenet05a} is another test of GR which is 
54: facilitated in part by timing of J0437--4715.
55: 
56: Although variation of Newton's gravitational constant $G$\ with time is forbidden in GR, 
57: this is not required in alternate formalisms of gravity \cite[e.g.][]{brans61a}.
58: \citet{verbiest08a} have measured the rate of change of orbital period \pbdot\ in the 
59: J0437--4715 system
60: to better than 2\% precision and shown that the agreement of the derived kinematic distance with  
61: the parallax distance derived from timing limits the time variation of  
62: $G$\ to less than 3 parts in $10^{11}$.
63:  In this Letter, we present a new Very Long Baseline Interferometry (VLBI) determination
64: of the position, proper motion, and parallax of PSR J0437--4715 and show that this improves the
65: previously published \Gdot\ limit derived from this system by an order of magnitude, approaching the
66: best published limit ($\Gdot/G  = (4\pm9)\times 10^{-13}$\ yr$^{-1}$), which is derived from 
67: Lunar Laser Ranging 
68: \citep[LLR;][]{williams04a}.  Our VLBI observations and results are presented in 
69: \S\ref{sec:vlbi} and \S\ref{sec:results} respectively and the limitations on apparent accelerations
70: (due to \Gdot\ or other possible causes such as unseen massive planets) are 
71: derived in \S\ref{sec:anomacc}.
72: In \S\ref{sec:gwb}, we investigate the impact of the stochastic GWB
73: on our results and derive an independent
74: limit on the GWB amplitude.  We summarize our conclusions in \S\ref{sec:conclusions}.
75: 
76: \section{VLBI observations}
77: \label{sec:vlbi}
78: Details of the observational, correlation, and post--processing strategies used are
79: covered in detail in \citet{deller08a} and only a brief summary is presented here.
80: Four observational epochs (MJD 53868, 54055, 54181, 54416) 
81: spanned 1.5 years, with each epoch lasting 12 hours and utilizing
82: all available antennas from the Australian Long Baseline Array (LBA).  
83: These were the three ATNF antennas (ATCA, Parkes,
84: Mopra), the two University of Tasmania antennas (Hobart and Ceduna) and, when available,
85: one of the NASA Deep Space Network (DSN) antennas at the Tidbinbilla 
86: station\footnote{DSS43 (70m) participated on MJD 54055 and 54181, while
87: DSS34 (34m) was used on MJD 54416}.  Observations were made at 8.4 GHz, and
88: four 16 MHz bands were used, quantized at two bit precision to give a total data rate of 256 Mbps.
89: 
90: Phase referencing was conducted using the International Celestial Reference Frame
91: (ICRF) source J0439--4522, a bright (several hundred mJy flux at 8 GHz) and compact radio 
92: source less than 2\degrees\ from PSR J0437--4715, whose position is known to 
93: 1 mas \citep{ma98a}.  A six minute target/calibrator cycle 
94: was employed, with observing time divided equally between the target 
95: and calibrator.  The well--studied blazar PKS 0537--441 was used for bandpass calibration.
96: 
97: The data were correlated, using matched filtering (gating) on pulse profiles, with the DiFX 
98: software correlator \citep{deller07a}.  Due to the extremely narrow pulse profile of PSR J0437--4715
99: at 8.4 GHz, this gating improves the output signal to noise ratio (SNR) by 
100: a factor of 7 compared to ungated data.  The pulsar ephemeris used to set gates was taken from
101: \citet{hotan06a}.
102: 
103: Due to the relatively high frequency used for these observations, several of the calibration steps 
104: described in \citet{deller08a} proved unnecessary.  Data reduction was performed with and
105: without GPS--based ionospheric correction, and the small ($\sim100\,\mu$as) corrections which
106: resulted from ionospheric correction gave a poorer parallax fit overall, with double the
107: estimated systematic errors.  Given the uncertainty of 
108: GPS--based corrections at these Southern latitudes, the ionospheric corrections were therefore
109: not applied.  Additionally, the effect of scintillation was
110: negligible at these higher frequencies, so the scintillation correction detailed in 
111: \citet{deller08a} was not applied.
112: 
113: Attaining single epoch accuracies of $\sim\!100\,\mu$as, as shown in \S\ref{sec:results},
114: required post-correlation corrections to the visibilities due to the proper motion of the pulsar
115: ($\sim\!200\,\mu$as in each 12 hour observation).  Additionally, the orbital motion of
116: PSR J0437--4715 causes a peak to peak displacement of $\sim\!110\,\mu$as.
117: Compensation for this displacement in the single--epoch positions
118: improved the quality of the proper motion and parallax fit, reducing the estimated systematic
119: error contribution by 3\%.
120: 
121: \section{Astrometric results}
122: \label{sec:results}
123: As discussed in \citet{deller08a}, the optimal weighting for visibility points when imaging 
124: the pulsar to fit for position depends on the ratio 
125: of thermal noise to residual systematic errors.  For strong targets systematic errors dominate the
126: error budget, meaning the use of equally weighted visibilities (as opposed to weighting visibilities by
127: baseline sensitivity) produces superior results.  For PSR J0437--4715 the SNR of a typical
128: single--band, single--epoch detection was $\sim\!20$, well above the threshold of 
129: $\sim\!10$\ identified in \citet{deller08a}, and thus we would expect equally weighted
130: visibilities to produce superior results.  To confirm this, the parallax was determined
131: from two separate datasets, which were produced using sensitivity--weighted 
132: and equally weighted visibilities.  As expected,
133: the equally weighted dataset produced a better fit, with smaller errors on fitted parameters and
134: smaller estimated systematic error contributions.
135: 
136: The data were fitted using an iterative approach
137: designed to estimate and account for systematic errors, described in \citet{deller08a}.  The
138: resulting parameter values are shown in Table~\ref{tab:results} along with the \citet{verbiest08a}
139: timing measurements\footnote{The timing positions of \citet{verbiest08a} have been re--fitted
140: at the VLBI reference epoch of MJD 54100}; 
141: the VLBI fits to pulsar motion
142: and observed positions are shown in Figure~\ref{fig:0437radec}.  
143: All errors are 1$\sigma$\ unless otherwise stated, and include the covariances between fitted 
144: parameters.  Over this short timespan, the covariance between parallax and proper motion is 
145: significant, and constitutes approximately half of the quoted parallax uncertainty.
146: 
147: The average total single--epoch error of 132\,$\mu$as is the best relative astrometry performed 
148: by the LBA to date, and is similar to other recent VLBI astrometric results \citep[e.g][]{loinard07a}.  
149: By way of comparison, the limiting accuracy for EVN observations at 8.4 GHz has been simulated 
150: by \citet{pradel06a} to be 83\,$\mu$as with a 1\degrees\,calibrator--target separation, at a 
151: declination of 50\degrees.  The EVN has similar sensitivity to the LBA, but
152: somewhat better uv coverage and longer baselines.
153: 
154: Our derived distance of $156.3 \pm 1.3$\,pc is the most accurate distance
155: measurement (in both absolute and fractional distance) 
156: for a pulsar to date and approaches the most accurate distance measurements 
157: made of objects outside the solar system \citep[T Tauri, $147.6 \pm 0.6\,{\mathrm{pc}}$:][]{loinard07a}.  
158: Previously, the highest--precision VLBI pulsar distance determinations were 
159: those made by \citet{brisken02a}, who measured the distance of PSR J0953+0755 to 
160: an accuracy of 5 pc, along with 8 other Northern Hemisphere pulsars.
161: The two previous parallax measurements made using a Southern array, of PSR J0835--4510 
162: \citep{dodson03a} and PSR J1456--6843 \citep{bailes90a}, had 1$\sigma$\,distance
163: errors of 19 and 70 pc respectively.  
164: 
165: \subsection{Comparison to timing astrometry}
166: 
167: To compare the VLBI and timing positions, we have re--fitted the timing data
168: of \citet{verbiest08a} to obtain the position of 
169: PSR J0437--4715 at our reference epoch (MJD 54100).
170: Table~\ref{tab:results} shows that the timing and VLBI positions at MJD 54100 differ by
171: over two mas, many times the formal errors shown. However, the formal errors are negligible
172: compared to the 1 mas uncertainty in the VLBI phase reference calibrator position
173: and the potential constant offsets due to phase--referencing
174: errors such as station position errors \citep[known to exist at the cm level for the LBA;][]{deller08a}.  
175: Discrepancies between interferometric and timing positions of even larger magnitudes have
176: been found using the DE200 frame \citep{bartel96a}, 
177: and differences at the mas level still exist for the position
178: PSR J0437--4715 calculated using the newer DE414 solar system ephemeris, as compared to the
179: DE405 ephemeris used by \citet{verbiest08a}. Thus, we conclude that the VLBI and timing 
180: position difference is consistent with the uncertainty in the calibrator position 
181: and the offset between the solar system frame and the ICRF.
182: 
183: The values obtained for the VLBI proper motion differ by $\sim4\sigma$\ in both right
184: ascension and declination from the timing values.  A likely cause for this discrepancy is small 
185: changes in the centroid position of the phase reference source due to intrinsic source variability, 
186: which would be absorbed into the proper motion fit.  The phase reference source, which is
187: modeled by a $\sim1\,$mas FWHM Gaussian with additional delta components within
188: 5 mas of strength 2\% and 0.2\% of peak flux, 
189: shows no gross evidence of variability (width/positions of primary/secondary components
190: constant to $\sim200\,\mu$as, and secondary fluxes are constant to $\sim1\,$mJy), but as it is only 
191: barely resolved (the beamsize is $\sim3\,$mas) variability at the $\sim100\,\mu$as level 
192: would be difficult to detect.  \citet{titov07a} show that some ICRF sources exhibit apparent
193: proper motions of hundreds of $\mu$as yr$^{-1}$ for periods of several years, 
194: and while a detailed VLBI history of this source is 
195: not available, calibrator measurements from ATCA show that the flux density has declined by a 
196: factor of three over a five year period, implying some variability.  
197: %\citet{gontier01a} show a median centroid drift rate of $\sim 50\,\mu$as yr$^{-1}$\ for ICRF sources
198: %on timescales comparable to our observational duration, 
199: %with some sources showing over $500\,\mu$as yr$^{-1}$.
200: The higher proper motion precision obtained with the timing data 
201: (a factor of 5--7 times better than the VLBI results) reflects the fact that the timing data spans a 
202: time baseline 7 times longer than the VLBI dataset.
203: 
204: The parallax value obtained from VLBI is consistent with that derived from timing, and yields a
205: distance which is consistent with the kinematic distance of $157.0 \pm 2.4$\ pc 
206: derived from the orbital period derivative \pbdot.  However, the
207: VLBI parallax measurement is an order of magnitude more precise than the timing measurement,
208: and yields a distance which is a factor of two more precise than the kinematic distance.
209: 
210: \section{Limits on anomalous accelerations}
211: \label{sec:anomacc}
212: The newly measured parallax of $\pi = 6.396 \pm 0.054$\,mas allows an improved 
213: measurement of any anomalous acceleration of either the Solar System or 
214: PSR J0437--4715. Specifically, the apparent acceleration due to time variability of 
215: Newton's gravitational constant $G$ and the mass of an undetected trans-Neptunian 
216: planet near the line of sight to the pulsar can be limited.
217: 
218: \subsection{Constraints on \Gdot}
219: \label{subsec:gdot}
220: As first described by \citet{damour91a}, a precise measurement of a binary pulsar's orbital 
221: period derivative, \pbdot, can be used to constrain a variation of the 
222: gravitational constant, $G$. However, as \citet{bell96a} pointed out, for 
223: PSR J0437$-$4715 a precise distance needs to be known in order to 
224: correct \pbdot\ for the Shklovskii acceleration caused by its proper motion
225: \citep{shklovskii70a}. 
226: This analysis has been performed, based exclusively on timing data, by
227: \citet{verbiest08a}, whose limit was dominated by the uncertainty in their 
228: parallax measurement. Our VLBI parallax value, however, improves this limit by 
229: a factor of nearly 10 down to $\dot{G}/G = (-5 \pm 26)\times 10^{-13}$\ yr$^{-1}$\ at 
230: 95\% certainty. This value compares well to the most stringent limit currently 
231: published: $(4\pm 9) \times 10^{-13}$\ yr$^{-1}$, derived through lunar laser 
232: ranging \citep{williams04a}. Since $\pi$ and \pbdot\ are now both measured to similar
233: precision, both measurements will have to be improved for a further 
234: significant increase in \Gdot\ sensitivity.  Additional VLBI observations and continued timing
235: could see this limit improve upon the existing LLR limit early in the next decade.
236: 
237: \subsection{Acceleration due to massive bodies}
238: 
239: An alternative source of anomalous acceleration is heavy planets in a 
240: wide orbit around the Sun or the pulsar. Building upon the initial analysis of 
241: \citet{zakamska05a}, our parallax measurement can be combined with the timing results 
242: from \citet{verbiest08a} to derive the following result: 
243: $a_{\odot}/c = (3 \pm 16 ) \times 10^{-20}\,{\mathrm s}^{-1}$\ at the $2 \sigma$ level. 
244: This improves the limit published in \citet{verbiest08a} by an order of magnitude and 
245: makes PSR J0437--4715 a more sensitive Solar System accelerometer than PSR J1713+0747, 
246: the most precise pulsar listed by \citet{zakamska05a}. From this, the limit for 50\% of the 
247: sky\footnote{within 60\degrees\,of the line of sight towards and away from PSR J0437--4715} 
248: can be calculated as: 
249: $|a_{\odot, 50\% {\mathrm{sky}}}/c| \leq 3.9\times 10^{-19}\,{\mathrm s}^{-1}$ (95\% certainty).
250: This acceleration limit can be used to exclude massive bodies within
251: a given radius of the Sun; for example, at Kuiper--belt radii (50 AU) it
252: excludes a planet more massive than Uranus over 50\% of the sky, while  
253: Jupiter--mass planets are excluded within 226\,AU over 50\% of the sky.
254: 
255: \section{Impact of the stochastic GWB}
256: \label{sec:gwb}
257: 
258: Using tools recently developed by \citet{hobbs08a}, we have simulated the effect of a GWB with spectral index $-2/3$ (as predicted for GWBs caused by black hole--black hole mergers) and dimensionless amplitude $1.1\times10^{-14}$\ (the best published GWB limit; \citealt{jenet06a}) on the observed value of \pbdot\ from pulsar timing. The simulated GWB causes the kinematic distance to be 
259: inconsistent with the VLBI parallax distance at the 2$\sigma$\, level in 
260: $\sim50\%$ of trials.  Thus, although these observations cannot improve 
261: upon the present GWB limit, they are consistent with it.
262: Simulations with a GWB with amplitude of $1.1\times 10^{-13}$ show inconsistencies between the kinematic and VLBI distances at the $2\sigma$ level in $95\%$\ of trials, providing an
263: independent exclusion of a GWB with an amplitude at or above this value.
264: 
265: It is also interesting to note that the precise limit on $\dot{G}$\ presented in \S\ref{subsec:gdot} would be impossible in a Universe with a strong GWB. In the simulations with GWB amplitude $1.1\times 10^{-13}$, the observed \Gdot\ value is inconsistent with 0 in $99\%$\ of cases, merely due to the GWB-induced corruption of the timing measurements.  Thus, the stochastic GWB must eventually limit the accuracy of measurements of \Gdot\ in the fashion outlined in this paper.
266: 
267: \section{Conclusions}
268: \label{sec:conclusions}
269: We have observed PSR J0437--4715 using the LBA and obtained the most precise pulsar 
270: distance determination to date, with an error $<1.5$\,pc.  Combined with accurate timing data, this
271: has enabled us to confirm that  $|\Gdot/G| < 3.1 \times 10^{-12}$\ yr$^{-1}$, 
272: the most stringent limit not obtained though
273: Solar System tests. Alternatively, assuming an unchanging gravitational force, the 
274: results can be interpreted as excluding any unseen Jupiter--mass planets within 226 AU of the Sun
275: in 50\% of the sky.
276: Finally, the agreement between VLBI and timing results in this single case implies an upper limit to the 
277: stochastic GWB amplitude which is within an order of magnitude of 
278: the best limit derived from observations of ensembles of pulsars.
279: 
280: \acknowledgements
281: 
282: This work has been supported by the Australian Federal Government's Major National Research Facilities program.  ATD is supported via a Swinburne University of Technology Chancellor's Research Scholarship and a CSIRO postgraduate scholarship.  The Long Baseline Array is part of the Australia Telescope which is funded by the Commonwealth of Australia for operation as a National Facility managed by CSIRO.
283: 
284: \clearpage
285: 
286: \begin{thebibliography}{25}
287: \expandafter\ifx\csname natexlab\endcsname\relax\def\natexlab#1{#1}\fi
288: 
289: \bibitem[{{Bailes} {et~al.}(1990){Bailes}, {Manchester}, {Kesteven}, {Norris},
290:   \& {Reynolds}}]{bailes90a}
291: {Bailes}, M., {Manchester}, R.~N., {Kesteven}, M.~J., {Norris}, R.~P., \&
292:   {Reynolds}, J.~E. 1990, \nat, 343, 240
293: 
294: \bibitem[{{Bartel} {et~al.}(1996){Bartel}, {Chandler}, {Ratner}, {Shapiro},
295:   {Pan}, \& {Cappallo}}]{bartel96a}
296: {Bartel}, N., {Chandler}, J.~F., {Ratner}, M.~I., {Shapiro}, I.~L., {Pan}, R.,
297:   \& {Cappallo}, R.~J. 1996, \aj, 112, 1690
298: 
299: \bibitem[{{Bell} \& {Bailes}(1996)}]{bell96a}
300: {Bell}, J.~F., \& {Bailes}, M. 1996, \apjl, 456, L33+
301: 
302: \bibitem[{{Bell} {et~al.}(1993){Bell}, {Bailes}, \& {Bessell}}]{bell93a}
303: {Bell}, J.~F., {Bailes}, M., \& {Bessell}, M.~S. 1993, \nat, 364, 603
304: 
305: \bibitem[{Brans \& Dicke(1961)}]{brans61a}
306: Brans, C., \& Dicke, R.~H. 1961, Phys. Rev., 124, 925
307: 
308: \bibitem[{{Brisken} {et~al.}(2002){Brisken}, {Benson}, {Goss}, \&
309:   {Thorsett}}]{brisken02a}
310: {Brisken}, W.~F., {Benson}, J.~M., {Goss}, W.~M., \& {Thorsett}, S.~E. 2002,
311:   \apj, 571, 906
312: 
313: \bibitem[{{Damour} \& {Taylor}(1991)}]{damour91a}
314: {Damour}, T., \& {Taylor}, J.~H. 1991, \apj, 366, 501
315: 
316: \bibitem[{{Deller} {et~al.}(2007){Deller}, {Tingay}, {Bailes}, \&
317:   {West}}]{deller07a}
318: {Deller}, A.~T., {Tingay}, S.~J., {Bailes}, M., \& {West}, C. 2007, \pasp, 119,
319:   318
320: 
321: \bibitem[{{Deller} {et~al.}(2008){Deller}, {Tingay}, \& {Brisken}}]{deller08a}
322: {Deller}, A.~T., {Tingay}, S.~J., \& {Brisken}, W.~F. 2008, \apj, submitted
323: 
324: \bibitem[{{Dodson} {et~al.}(2003){Dodson}, {Legge}, {Reynolds}, \&
325:   {McCulloch}}]{dodson03a}
326: {Dodson}, R., {Legge}, D., {Reynolds}, J.~E., \& {McCulloch}, P.~M. 2003, \apj,
327:   596, 1137
328: 
329: \bibitem[{{Hobbs} {et~al.}(2008){Hobbs}, {Jenet}, {Lee}, {Verbiest}, {Yardley},
330:   {Manchester}, {Lommen}, {Coles}, {Edwards}, \& {Shettigara}}]{hobbs08a}
331: {Hobbs}, G.~B., {Jenet}, F.~A., {Lee}, K.~J., {Verbiest}, J.~P.~W., {Yardley},
332:   D., {Manchester}, R.~N., {Lommen}, A., {Coles}, W.~A., {Edwards}, R.~T., \&
333:   {Shettigara}, C. 2008, MNRAS, submitted
334: 
335: \bibitem[{{Hotan} {et~al.}(2006){Hotan}, {Bailes}, \& {Ord}}]{hotan06a}
336: {Hotan}, A.~W., {Bailes}, M., \& {Ord}, S.~M. 2006, \mnras, 369, 1502
337: 
338: \bibitem[{{Jenet} {et~al.}(2005){Jenet}, {Hobbs}, {Lee}, \&
339:   {Manchester}}]{jenet05a}
340: {Jenet}, F.~A., {Hobbs}, G.~B., {Lee}, K.~J., \& {Manchester}, R.~N. 2005,
341:   \apjl, 625, L123
342: 
343: \bibitem[{{Jenet} {et~al.}(2006){Jenet}, {Hobbs}, {van Straten}, {Manchester},
344:   {Bailes}, {Verbiest}, {Edwards}, {Hotan}, {Sarkissian}, \& {Ord}}]{jenet06a}
345: {Jenet}, F.~A., {Hobbs}, G.~B., {van Straten}, W., {Manchester}, R.~N.,
346:   {Bailes}, M., {Verbiest}, J.~P.~W., {Edwards}, R.~T., {Hotan}, A.~W.,
347:   {Sarkissian}, J.~M., \& {Ord}, S.~M. 2006, \apj, 653, 1571
348: 
349: \bibitem[{{Johnston} {et~al.}(1993){Johnston}, {Lorimer}, {Harrison}, {Bailes},
350:   {Lyne}, {Bell}, {Kaspi}, {Manchester}, {D'Amico}, \&
351:   {Nicastro}}]{johnston93a}
352: {Johnston}, S., {Lorimer}, D.~R., {Harrison}, P.~A., {Bailes}, M., {Lyne},
353:   A.~G., {Bell}, J.~F., {Kaspi}, V.~M., {Manchester}, R.~N., {D'Amico}, N., \&
354:   {Nicastro}, L. 1993, \nat, 361, 613
355: 
356: \bibitem[{{Kargaltsev} {et~al.}(2004){Kargaltsev}, {Pavlov}, \&
357:   {Romani}}]{kargaltsev04a}
358: {Kargaltsev}, O., {Pavlov}, G.~G., \& {Romani}, R.~W. 2004, \apj, 602, 327
359: 
360: \bibitem[{{Loinard} {et~al.}(2007){Loinard}, {Torres}, {Mioduszewski},
361:   {Rodr{\'{\i}}guez}, {Gonz{\'a}lez-L{\'o}pezlira}, {Lachaume}, {V{\'a}zquez},
362:   \& {Gonz{\'a}lez}}]{loinard07a}
363: {Loinard}, L., {Torres}, R.~M., {Mioduszewski}, A.~J., {Rodr{\'{\i}}guez},
364:   L.~F., {Gonz{\'a}lez-L{\'o}pezlira}, R.~A., {Lachaume}, R., {V{\'a}zquez},
365:   V., \& {Gonz{\'a}lez}, E. 2007, \apj, 671, 546
366: 
367: \bibitem[{{Ma} {et~al.}(1998){Ma}, {Arias}, {Eubanks}, {Fey}, {Gontier},
368:   {Jacobs}, {Sovers}, {Archinal}, \& {Charlot}}]{ma98a}
369: {Ma}, C., {Arias}, E.~F., {Eubanks}, T.~M., {Fey}, A.~L., {Gontier}, A.-M.,
370:   {Jacobs}, C.~S., {Sovers}, O.~J., {Archinal}, B.~A., \& {Charlot}, P. 1998,
371:   \aj, 116, 516
372: 
373: \bibitem[{{Pradel} {et~al.}(2006){Pradel}, {Charlot}, \&
374:   {Lestrade}}]{pradel06a}
375: {Pradel}, N., {Charlot}, P., \& {Lestrade}, J.-F. 2006, \aap, 452, 1099
376: 
377: \bibitem[{{Shklovskii}(1970)}]{shklovskii70a} 
378: {Shklovskii}, I.~S. 1970, Soviet Astronomy, 13, 562 
379: 
380: \bibitem[{{Titov}(2007)}]{titov07a}
381: {Titov}, O.~A. 2007, Astronomy Letters, 33, 481
382: 
383: \bibitem[{{van Straten} {et~al.}(2001){van Straten}, {Bailes}, {Britton},
384:   {Kulkarni}, {Anderson}, {Manchester}, \& {Sarkissian}}]{van-straten01a}
385: {van Straten}, W., {Bailes}, M., {Britton}, M., {Kulkarni}, S.~R., {Anderson},
386:   S.~B., {Manchester}, R.~N., \& {Sarkissian}, J. 2001, \nat, 412, 158
387: 
388: \bibitem[{{Verbiest} {et~al.}(2008){Verbiest}, {Bailes}, {van Straten},
389:   {Hobbs}, {Edwards}, {Manchester}, {Bhat}, {Sarkissian}, {Jacoby}, \&
390:   {Kulkarni}}]{verbiest08a}
391: {Verbiest}, J.~P.~W., {Bailes}, M., {van Straten}, W., {Hobbs}, G.~B.,
392:   {Edwards}, R.~T., {Manchester}, R.~N., {Bhat}, N.~D.~R., {Sarkissian}, J.~M.,
393:   {Jacoby}, B.~A., \& {Kulkarni}, S.~R. 2008, \apj, 679, 675
394: 
395: \bibitem[{{Williams} {et~al.}(2004){Williams}, {Turyshev}, \&
396:   {Boggs}}]{williams04a}
397: {Williams}, J.~G., {Turyshev}, S.~G., \& {Boggs}, D.~H. 2004, Physical Review
398:   Letters, 93, 261101
399: 
400: \bibitem[{{Zakamska} \& {Tremaine}(2005)}]{zakamska05a}
401: {Zakamska}, N.~L., \& {Tremaine}, S. 2005, \aj, 130, 1939
402: 
403: \bibitem[{{Zavlin} {et~al.}(2002){Zavlin}, {Pavlov}, {Sanwal}, {Manchester},
404:   {Tr{\"u}mper}, {Halpern}, \& {Becker}}]{zavlin02a}
405: {Zavlin}, V.~E., {Pavlov}, G.~G., {Sanwal}, D., {Manchester}, R.~N.,
406:   {Tr{\"u}mper}, J., {Halpern}, J.~P., \& {Becker}, W. 2002, \apj, 569, 894
407: 
408: \end{thebibliography}
409: 
410: 
411: \clearpage
412: 
413: \begin{deluxetable}{lrr}
414: \tabletypesize{\tiny}
415: \tablecaption{Fitted VLBI results for PSR J0437--4715 and comparative timing values (positions
416: re--referenced to the VLBI proper motion epoch)}
417: \tablewidth{0pt}
418: \tablehead{
419: \colhead{Parameter} & \colhead{Fitted value and error} & \colhead{\citet{verbiest08a} timing values\phm{\tablenotemark{A}}}}
420: \startdata
421: Right Ascension (J2000)	&  	04$^{\mathrm h}$37$^{\mathrm m}$15.883250$^{\mathrm s}$ $\pm$ 0.000003	& 
422: 				04$^{\mathrm h}$37$^{\mathrm m}$15.883185$^{\mathrm s}$ $\pm$ 0.000006\phm{\tablenotemark{A}} 	\\
423: Declination (J2000) 	&  	$-$47\degrees15'09.031863" $\pm$ 0.000037	& 
424: 				$-$47\degrees15'09.034033" $\pm$ 0.000070\phm{\tablenotemark{A}}	\\
425: $\mu_{\alpha}$	(mas/yr)	&  	121.679 $\pm$ 0.052\phn\phn\phn	&
426: 						121.453 $\pm$ 0.010\phn\phn\phn\phm{\tablenotemark{A}}	\\
427: $\mu_{\delta}$	(mas/yr)	&  	$-$71.820 $\pm$ 0.086\phn\phn\phn	&
428: 						$-$71.457 $\pm$ 0.012\phn\phn\phn\phm{\tablenotemark{A}}	\\
429: Parallax $\pi$ (mas)	 		&  	6.396 $\pm$ 0.054\phn\phn\phn	&
430: 						6.65 $\pm$ 0.51\phn\phn\phn\phn\phm{\tablenotemark{A}}	\\
431: Distance (pc)	&  	156.3 $\pm$ 1.3\phn\phn\phn\phn\phn		&
432: 				157.0 $\pm$ 2.4\tablenotemark{A}\phn\phn\phn\phn\phn\\		
433: Transverse velocity $v_{\mathrm T}$ (km/s)	&  	104.71 $\pm$ 0.95\phn\phn\phn\phn	&
434: 				104.9 $\pm$ 1.6\tablenotemark{A}\phn\phn\phn\phn\phn	\\
435: Reduced chi--squared 	&  	1.0 \phs \phn\phd\phn\phn\phn\phn\phn\phn$\  $ &
436: 					 \phs \phn\phd\phn\phn\phn\phn\phn\phn$\  $\phm{\tablenotemark{A}}\\
437: Average epoch mean fit error (mas)			& 0.059 \phs \phn\phd\phn\phn\phn\phn\phn\phn$\  $ &
438: 					 \phs \phn\phd\phn\phn\phn\phn\phn\phn$\  $\phm{\tablenotemark{A}}\\
439: Average intra--epoch systematic error (mas) 	& 0.068 \phs \phn\phd\phn\phn\phn\phn\phn\phn$\  $ &
440: 					 \phs \phn\phd\phn\phn\phn\phn\phn\phn$\  $\phm{\tablenotemark{A}}\\
441: Average inter--epoch systematic error (mas) 	& 0.103 \phs \phn\phd\phn\phn\phn\phn\phn\phn$\  $  &
442: 					 \phs \phn\phd\phn\phn\phn\phn\phn\phn$\  $\phm{\tablenotemark{A}}\\
443: Reference epoch for proper motion	(MJD)	& 54100.0 \phs \phn\phd\phn\phn\phn\phn\phn\phn$\  $&
444: 			54100.0 \phs \phn\phd\phn\phn\phn\phn\phn\phn$\  $\phm{\tablenotemark{A}}\\
445: \enddata
446: \tablenotetext{A}{Derived from the kinematic distance obtained from \pbdot, 
447: not the less precise parallax values}
448: \label{tab:results}
449: \end{deluxetable}
450: 
451: \clearpage
452: 
453: \begin{figure}
454: \begin{center}
455: \begin{tabular}{cc}
456: \includegraphics[width=0.4\textwidth]{f1a.eps} &
457: \includegraphics[width=0.4\textwidth]{f1b.eps} \\
458: \includegraphics[width=0.4\textwidth]{f1c.eps} &
459: \includegraphics[width=0.4\textwidth]{f1d.eps} \\
460: \end{tabular}
461: \caption{Motion of PSR J0437--4715, with measured positions
462: overlaid on the best fit.  Clockwise from top left: Motion in declination vs right ascension; as before 
463: but with proper motion subtracted; right ascension vs time with proper motion subtracted; and
464: declination vs time with proper motion subtracted.}
465: \label{fig:0437radec}
466: \end{center}
467: \end{figure}
468: 
469: %\bibliographystyle{apj}
470: %\bibliography{deller_thesis}
471: 
472: \end{document}