1: %\documentclass[12pt,preprint]{aastex}
2: %\documentclass{aastex}
3: \documentclass{emulateapj}
4: %\usepackage{psfig}
5:
6: \slugcomment{Accepted by the Astrophysical Journal}
7:
8: \shorttitle{Radio Searches of Four AXPs}
9: \shortauthors{Crawford, Hessels, \& Kaspi}
10:
11: \newcommand{\transaxp}{XTE~J1810$-$197}
12:
13: \begin{document}
14:
15: \title{Deep Searches for Radio Pulsations and Bursts
16: from Four Southern Anomalous X-ray Pulsars}
17:
18: \author{Fronefield Crawford\altaffilmark{1}, Jason
19: W. T. Hessels\altaffilmark{2,3}, \& Victoria M. Kaspi\altaffilmark{2}}
20:
21: \altaffiltext{1}{Department of Physics and Astronomy, Franklin and
22: Marshall College, P.O. Box 3003, Lancaster, PA 17604; email:
23: fcrawfor@fandm.edu}
24:
25: \altaffiltext{2}{Department of Physics, McGill University, Montreal,
26: QC H3A 2T8, Canada}
27:
28: \altaffiltext{3}{Astronomical Institute ``Anton Pannekoek'',
29: University of Amsterdam, Kruislaan 403, 1098 SJ Amsterdam, The
30: Netherlands}
31:
32: \begin{abstract}
33: We have searched for persistent radio pulsations, bright single
34: pulses, and bursts from four Southern anomalous X-ray pulsars
35: (AXPs). Deep observations were conducted at 1.4\,GHz in 1999 July and
36: August with the Parkes 64-m telescope. For all of the target AXPs,
37: the upper limits on integrated pulsed emission are $\sim0.02$\,mJy,
38: while the limits on the flux of single pulses are $\sim1$\,Jy or
39: better. The corresponding radio luminosity limits are significantly
40: below the observed luminosities of the majority of the observed radio
41: pulsars, and they are significantly below the radio luminosity of
42: \transaxp, the only AXP known thus far to emit radio pulsations, at
43: the epoch when radio pulses were first detected from this source. Our
44: null results support the hypothesis that pulsed radio emission from
45: AXPs is only present in conjunction with X-ray outburst activity.
46: However, we cannot rule out the possibility that pulsed radio emission
47: is present, but that it is too weak to be detected in our
48: observations, or that unfavorable viewing geometries prevent this
49: emission from being seen by terrestrial observers. Given the possible
50: association between the radio emission and transient X-ray behavior of
51: \transaxp, continued radio searches of these and other AXPs at
52: different epochs are warranted, particularly after periods of X-ray
53: burst activity.
54: \end{abstract}
55:
56: \keywords{stars: neutron --- pulsars: general}
57:
58: \section{Introduction}
59:
60: Anomalous X-ray Pulsars (AXPs), thought to be magnetars
61: \citep[see][for a review]{wt06,kas06}, are a rare class of slowly
62: rotating ($P_{\rm spin} = 5-12$\,s) neutron stars whose X-ray emission
63: may be powered by the decay of their huge inferred magnetic fields
64: ($B_{\rm surf} = 10^{14} - 10^{15}$\,G). Though they are relatively
65: bright X-ray sources ($10^{34} - 10^{36}$\,erg~s$^{-1}$), previous
66: attempts to detect radio pulsations from these objects have been
67: unsuccessful \citep[e.g.][]{ims02,cpkm02,bri+06}. Recently, however,
68: \citet{crh+06} reported the discovery of pulsed radio emission from
69: the transient AXP \transaxp\ \citep{ims+04} in 2006 March, subsequent
70: to the discovery of a radio point source in VLA imaging of the area
71: \citep{hgb+05}. These discoveries have provided an exciting new
72: spectral window on magnetar physics and another link between the AXPs
73: and conventional, rotation-powered radio pulsars \citep[especially the
74: small population of radio pulsars with high inferred magnetic fields,
75: e.g.][]{msk+03}. Such a connection was previously suggested by the
76: similarity seen in the timing behavior of these two classes of
77: objects, namely the presence of timing noise and glitches
78: \citep[e.g.][]{kcs99,klc00}.
79:
80: Interestingly, radio emission from \transaxp, which has shown
81: tremendous X-ray brightening\footnote{Note that other AXPs have also
82: shown flux variation \citep[e.g. 1E~2259+586,][]{wkt+04}, and the most
83: recently established AXP, CXO~J164710.2$-$455216 \citep{mcc+06}, had
84: an outburst on 2006 September 21 \citep{kbc+06}.}, has only been
85: observed subsequent to its one known X-ray outburst, which occurred
86: sometime between 2002 November and 2003 January \citep{ims+04}. This
87: indicates that its radio emission is not a stable phenomenon, but
88: rather is somehow associated with its X-ray activity. Archival radio
89: pulsation searches prior to this X-ray outburst, in which the AXP was
90: not detected in radio, show that the pulsed radio flux of \transaxp\
91: increased by more than an order of magnitude from its quiescent level
92: \citep{crh+06}. Current radio monitoring \citep{ccr+07} reveals that
93: there remains an intrinsic flux variability of about a factor of two
94: on day timescales, and that there has been a gradual decrease in
95: brightness and a broadening of the pulse profile since its discovery
96: as a pulsed radio source. Such large-scale intrinsic flux variability
97: has come as a surprise and is not seen in normal radio pulsars
98: \citep{ssh+00}. \transaxp\ is also peculiar in that it has a very
99: flat radio spectrum and was at the time of its discovery the brightest
100: radio pulsar in the sky above 20\,GHz.
101:
102: If the recent emergence of radio emission from \transaxp\ is indeed
103: directly related to its X-ray behavior, then that may explain why the
104: ``persistent'' AXPs\footnote{See the McGill University magnetar
105: catalog at
106: http://www.physics.mcgill.ca/$\sim$pulsar/magnetar/main.html.} are (at
107: the very least) radio dim. On the other hand, the small beaming
108: fraction of long-period radio pulsars, presumably owing to their large
109: light cylinders and correspondingly small polar caps
110: \citep[e.g.][]{ran90,ymj99}, and the narrow radio beam of \transaxp\
111: itself \citep{crh+06}, suggest that AXPs, if they are generically radio
112: emitters, should have narrow emission beams and therefore would be
113: unlikely to be detected. In the latter case, that the only confirmed
114: transient AXP is also the only known radio-emitting AXP could be a
115: geometrical coincidence, although it would still be unclear why it was
116: not seen before its X-ray outburst. Only additional radio
117: observations of the AXPs can distinguish between these possibilities
118: and determine the phenomenological link between X-ray and radio
119: activity. Furthermore, given the observed radio flux variability of
120: \transaxp\ on both day and year timescales, {\it repeated}
121: observations of the AXPs are certainly warranted.
122:
123: Here we report a sensitive search for radio emission, both periodic
124: and bursting, from three southern AXPs and one southern AXP candidate
125: (potentially also a transient AXP) using the Parkes radio telescope.
126: We detect no emission of either kind and provide the most stringent
127: upper limits to date on such emission from these objects. Our
128: results, together with those recently reported by \citet{bri+06} for
129: these same objects, establish that any periodic or bursting radio
130: emission from these sources is either non-existent, very weak, or very
131: sporadic at the epoch of these observations, as compared with
132: \transaxp, most of the known rotating radio transients
133: \citep[RRATs,][]{mll+06}, and the bulk of the known radio pulsar
134: population. These null results lay the foundation for future radio
135: observations of these AXPs in the event that one exhibits a
136: significant X-ray outburst like that seen in \transaxp, as this could
137: in principle establish whether AXP radio emission is associated in
138: some way with X-ray outbursts.
139:
140: \section{Observations}
141:
142: Three established AXPs (1E 1048.1$-$5937, 1E 1841$-$045, and 1RXS
143: J170849.0$-$400910) and one AXP candidate (AX J1845$-$0258\footnote{AX
144: J1845$-$0258 is a transient AXP candidate only, see \S4 for further
145: discussion.}) were observed with the Parkes telescope in 1999 July and
146: August (Table~\ref{tbl-1}) as part of a larger search campaign to find
147: young, energetic radio pulsars \citep{cpkm02}. To our knowledge, none
148: of these sources was in a state of X-ray outburst at the time of our
149: observations, although such outbursts are known to be short-lived and
150: some may have been missed, despite the X-ray monitoring of these
151: objects \citep[e.g.][]{kcs99,kgc+01}.
152: Each AXP was observed continuously with a single
153: 16800-s (4.7-h) integration using the center beam of the multibeam
154: receiver \citep{swb+96, mlc+01}. This receiver has been successfully
155: used in a number of targeted pulsar searches
156: \citep[e.g.][]{cmg+02,crg+06} and large-scale pulsar surveys
157: \citep{mlc+01,ebvb01,bjd+06}. The central observing frequency was
158: 1374\,MHz with 288\,MHz of bandwidth, consisting of 96 3-MHz
159: channels. Each channel was 1-bit sampled every 0.25\,ms and recorded
160: onto DLT magnetic tape. At the start of each observing session, a
161: bright test pulsar (PSR J1359$-$6038) was observed to verify that the
162: observing system was working properly. Details of the observing setup
163: are similar to those of the Parkes Multibeam Survey \citep{mlc+01} and
164: the radio search of these same four AXPs by \citet{bri+06}. Although
165: the observations of \citet{bri+06} were conducted with the same
166: observing setup as ours, our observations were longer, analyzed
167: somewhat differently, and, most importantly, were taken at a different
168: epoch. The latter point is particularly relevant given the transient
169: nature and day-to-day variability seen in \transaxp.
170:
171: \section{Data Analysis and Results}
172:
173: Each AXP observation was searched for periodicities, single pulses,
174: and bursts, and was folded at the known neutron star spin period using
175: the X-ray timing ephemeris when available. Standard Fourier-based
176: periodicity searches were reported by \citet{cpkm02}; no AXP
177: counterparts or periodicities indicative of another previously
178: undiscovered pulsar in the same field were found in those searches.
179: Here we describe the single pulse/burst and direct folding analyses,
180: which were conducted with the {\tt PRESTO}\footnote{Available at
181: http://www.cv.nrao.edu/$\sim$sransom/presto.} suite of pulsar analysis
182: tools \citep{ran01, rem02}.
183:
184: The first step in both of these analyses was to excise narrow-band and
185: transient radio frequency interference (RFI) from the data by
186: identifying bad frequency channels or time intervals and then masking
187: these. Given the slow rotation rates of the AXPs, and the pernicious
188: effect of low frequency, ``red'' noise in radio data, this step was
189: particularly important for ensuring that a relatively weak signal was
190: not being hidden by RFI. Typically 15-25\% of the frequency channels
191: and $\sim 5$\% of the integration time were masked in this process.
192:
193: \subsection{Folding Search}
194:
195: In the folding search, the entire 288\,MHz of observing bandwidth was
196: subdivided into 32 9-MHz subbands, each of which was transformed to
197: the Solar System barycenter and dedispersed at a dispersion measure
198: (DM) of 500\,pc cm$^{-3}$. Each subband was folded at the precisely
199: predicted spin period of the AXP at the observation epoch using a
200: contemporaneous ephemeris from X-ray observations (see
201: Table~\ref{tbl-1} for references; no ephemeris was available for
202: AX~J1845$-$0258 so the discovery period was used). Trial DMs were
203: created by shifting the subbands with respect to each other in pulse
204: phase. DMs were searched in spacings of $36-72$\,pc cm$^{-3}$ from zero
205: up to a maximum DM of $4000-8000$\,pc cm$^{-3}$, depending on the spin
206: period. We also folded the data allowing for a search in period which
207: ranged $\pm 5-10$\,ms from the predicted pulse period. In the case of
208: AX J1845$-$0258, we searched a period range between its discovery
209: period and a period 212 ms greater than this. The upper limit of this
210: range was determined by using twice the period derivative that results
211: from assuming that AX J1845$-$0258 has a magnetic field equal to that
212: of SGR J1806$-$20, the magnetar with the largest inferred field.
213: Separate folds were also made for each target from nine overlapping
214: shorter segments of the full observation. These segments started at
215: intervals of 10\% of the data length and consisted of 20\% of the
216: total integration time. This was done to account for the presence of
217: significant scintillation (unlikely given the high predicted DMs of
218: these sources), strong transient RFI near the AXP spin period, or
219: large-scale intrinsic pulse strength variability on a timescale
220: shorter than the observation length.
221:
222: We found no convincing periodic signals from any of the AXPs in any of
223: the folds we made. Apart from the statistical significance of the
224: measured $\chi^{2}$ values in the folded profiles, the criteria for
225: differentiating signals potentially related to the AXPs from RFI were
226: that the AXP signal should be broadband, have a steady pulse phase,
227: and be dispersed. None of the observed signals met these criteria,
228: and we therefore associate them with RFI. We note that the frequency
229: decimation of the data into subbands, which required an assumed DM,
230: did not introduce a significant amount of dispersive smearing compared
231: to the long spin periods of these neutron stars. Furthermore,
232: dispersive smearing within the frequency channels and multipath
233: scattering effects are not likely to be significant factors in our
234: non-detections: the estimated smearing and scattering times (obtained
235: from the NE2001 Galactic electron model of \citet{cl02} assuming DM
236: values up to $\sim 2000$ pc cm$^{-3}$) are less than 1.3\% of the
237: pulse period in each case. This is significantly less than the 2.7\%
238: duty cycle assumed in our sensitivity calculations, which is the duty
239: cycle of \transaxp's integrated radio profile in its discovery paper
240: \citep[see Figure 1 of][]{crh+06}.
241:
242: We used the modified radiometer equation, presented by \citet{dtws85}
243: and more recently by \citet{lfl+06}, to estimate the 1400-MHz upper
244: limits on the pulsed radio emission from these AXPs. This equation is
245: appropriate to use for the folding searches we conducted \citep[see,
246: e.g., Figure 2 of][which shows the fidelity of using this equation for
247: folding-search sensitivity estimates]{lfl+06}. The sky temperature at
248: 1374\,MHz was estimated for each target AXP using the model of
249: \citet{hssw82} and assuming a radio spectral index of $-2.6$. As
250: stated above, a duty cycle of 2.7\% was also assumed. The sensitivity
251: limit can easily be computed for different assumed duty cycles in the
252: modified radiometer equation. High-pass filtering is also present in
253: the observing system; this filter has a time constant of $\sim 0.9$\,s
254: \citep{mlc+01}. However, if the radio pulses from our target sources
255: are as narrow as those seen from \transaxp\ ($\la 0.3$\,s for all four
256: of our targets, assuming a 2.7\% duty cycle), many harmonics of the
257: fundamental will not suffer any filtering. We therefore do not
258: include any attenuation factor from this filtering in our sensitivity
259: estimates, in contrast to \citet{bri+06}. If our assumption is
260: correct, then \citet{bri+06} have underestimated the true sensitivity
261: of their observations by a factor of $2-2.5$. Table \ref{tbl-1} lists
262: our calculated upper limits for pulsed radio emission at 1400\,MHz. It
263: should be noted that these limits do not include loss of data from RFI
264: masking effects and do not account for any red noise contribution from
265: RFI, which is time-variable and thus difficult to quantify precisely.
266: Roughly speaking, however, we estimate that the RFI environment may
267: worsen these limits by a factor of two. We also used the estimated
268: distances to the AXPs to calculate the corresponding 1400-MHz
269: luminosity limits (Table \ref{tbl-1}).
270:
271: \subsection{Single Pulse Search}
272:
273: To search for dispersed single pulses and bursts, the raw data were
274: dedispersed at a set of trial DMs ranging from zero to $4000-8000$\,pc
275: cm$^{-3}$, depending on the spin period. The DM spacing varied as a
276: function of DM (being wider at higher DMs) and was chosen so as not to
277: increase the dispersive smearing already present in the frequency
278: channels. Using the native 0.25-ms time resolution of the data, each
279: dedispersed time series was searched for candidate single pulses
280: having a signal-to-noise ratio (S/N) greater than 6.5. This S/N
281: threshold was chosen to avoid confusion with the RFI background. To
282: maintain sensitivity to pulses broader than 0.25\,ms,
283: matched-filtering with a boxcar function of varying widths (ranging
284: from 1 to 30 samples) was also done.\footnote{The boxcar widths used
285: were 1, 2, 3, 4, 6, 9, 14, 20, and 30 samples.} To enhance sensitivity
286: to even longer-duration single pulses and bursts, the dedispersed time
287: series were then downsampled by combining the original samples into
288: contiguous blocks of 2, 4, 8, 16, and 32 samples. Each set of
289: downsampled time series was then searched using the same set of boxcar
290: filters, providing sensitivity to signals of duration $\la 240$\,ms
291: (this is equal to the maximum boxcar length, 30 bins, times the
292: maximum downsampled sample time, 8\,ms). This range encompasses and
293: significantly exceeds the $\la 10$\,ms single-pulse widths seen by
294: \cite{crh+06} for \transaxp. The sensitivity to pulses in the
295: downsampled data increases as the square root of the number of
296: combined samples (and further increases with the boxcar filtering),
297: but for each downsampling, a slightly higher S/N threshold was used to
298: eliminate spurious candidates from RFI. Specifically, we used S/N
299: thresholds of 7.0, 8.0, 9.0, 11.0, and 12.0 for the downsampling
300: factors described above. To differentiate dispersed single pulses from
301: RFI, results from the different trial DMs were stacked on top of each
302: other in order to identify signals that peaked in S/N at a non-zero DM
303: \citep[e.g.,][]{lk04,mll+06}. No candidate signals were detected from
304: any of the target AXPs in any of these search trials. Throughout the
305: search, we maintained excellent sensitivity to single pulses for all
306: reasonable DMs.
307:
308: We have determined single-pulse flux density and luminosity detection
309: limits over the range of pulse time scales we considered ($0.25 -
310: 240$\,ms), using the modified radiometer equation (e.g., Lorimer et
311: al. 2006\nocite{lfl+06}) and the S/N thresholds described above.
312: These limits are presented in Table \ref{tbl-1}. As in the case of
313: the folding search, the deleterious effects of red noise and data loss
314: from RFI masking were not included in the sensitivity limits.
315:
316: \section{Discussion and Conclusions}
317:
318: We place a flux density limit of $\sim 0.02$\,mJy on integrated pulsed
319: radio emission from our four AXP targets. This is about ten times
320: lower than the upper limits placed on \transaxp's pulsed emission in
321: archival Parkes Multibeam Survey observations prior to the outburst of
322: that source \citep[see Table~1 in][]{crh+06}. Our greater sensitivity
323: compared with the Parkes Multibeam Survey data is the result of $8
324: \times$ longer integration time and higher gain because our sources
325: were specifically placed at the center of the receiver beam. However,
326: the corresponding 1400-MHz luminosity limits from our observations are
327: only $2-3 \times$ lower than those established for \transaxp\ prior to
328: outburst because that source is estimated to be $2-3 \times$ closer
329: than our target AXPs. The public pulsar catalog
330: \citep{mhth05}\footnote{Available at
331: http://www.atnf.csiro.au/research/pulsar/psrcat.} shows that of the
332: known pulsars with measured 1400-MHz radio luminosities, $\sim 5$\%
333: have values below 1\,mJy kpc$^{2}$, which is comparable to our
334: limits. Based on the work of \citet{lml+98}, \citet{gsgv01} estimate
335: that $\ga 60$\% of potentially observable pulsars have 1400-MHz
336: luminosities less than 1\,mJy kpc$^{2}$. It is therefore conceivable
337: that weak radio pulses could be emitted by our targets but that they
338: are below our detection threshold.
339:
340: However, the average 1400-MHz flux density reported by \citet{crh+06}
341: for \transaxp\ soon after it was first detected was $\sim
342: 8$\,mJy. This corresponds to a 1400-MHz luminosity of $\sim 80$\,mJy
343: kpc$^{2}$, which is almost two orders of magnitude larger than the
344: luminosity limits of our searches. Thus, comparably strong radio
345: emission from our target AXPs would have been easily detectable if it
346: were beamed in our direction. Long-period pulsars typically have
347: smaller duty cycles and narrower emission beams than faster pulsars
348: \citep{ymj99}, thereby making line-of-sight detection more unlikely.
349: It is possible that all of our target AXPs could be strong radio
350: emitters but that none is detectable by us owing to unfavorable
351: viewing geometries. We note however, that the emission from
352: \transaxp\ is different than for normal rotation-powered radio
353: pulsars. \citet{crh+06} argue that it is possible that the radio
354: emission from \transaxp\ is generated on closed field lines (as
355: opposed to conventional radio pulsar emission, which streams along
356: open magnetic field lines above the magnetic poles) and beamed into a
357: much broader range of angles. If radio emission from magnetars is
358: emitted in a variety of directions, then detectability is less
359: contingent on beaming.
360:
361: One of the sources in our sample, AX~J1845$-$0258, is a candidate
362: transient AXP having had an X-ray outburst in 1993 \citep{gv98,tkk+98}
363: but not having been seen to emit pulsed X-ray emission since then. In
364: this sense, it may be like \transaxp, but in a quiescent state in
365: 1999, when the observations reported here were taken \cite[see][for a
366: more detailed summary of this source's X-ray behavior]{tkgg06}. If
367: AX~J1845$-$0258 is indeed an AXP and undergoes another flare in the
368: future, it will be important to compare any subsequently detected
369: radio emission with that seen from \transaxp\ subsequent to its X-ray
370: outburst. In that event, the upper limits presented here during
371: AX~J1845$-$0258's quiescent phase will be a critical piece of
372: information. The observation of an outburst on 2006 September 21
373: \citep{kbc+06} from the most recently established AXP,
374: CXO~J164710.2$-$455216 \citep{mcc+06,wkg06}, suggests that this
375: outburst was similar to the one seen from \transaxp. No pulsed radio
376: emission was detected a week after the burst, to a limit of 0.04\,mJy
377: \citep[][further observations are planned]{brip06}.
378:
379: For a comparison of our single-pulse upper limits with the observed
380: single pulses reported for \transaxp\ by \citet{crh+06}, we compared
381: the corresponding single-pulse luminosity limits for our sources and
382: for \transaxp\ (Table~\ref{tbl-1}). Our luminosity limits in the most
383: conservative case range from 22 to 69 Jy kpc$^{2}$, below the
384: luminosity of $\ga 100$\, Jy kpc$^{2}$ derived from the $\ga 10$\,Jy
385: pulses for \transaxp\ reported by \citet{crh+06}. Single pulses were
386: also detected from almost every rotation of \transaxp. This suggests
387: that we would have expected to detect a large number of pulses during
388: the course of our observations if comparable radio emission were being
389: emitted by our targets.
390:
391: Our non-detection of single pulses further weakens the hypothesis that
392: some of the sporadically emitting radio sources known as RRATs
393: \citep{mll+06} are magnetars. This link has also been weakened by two
394: other recent results. First, the discovery of X-ray emission
395: associated with the most ``magnetar-like'' of the RRATs, J1819$-$1458,
396: shows that this emission is more typical of middle-aged
397: rotation-powered pulsars than that seen from the magnetars
398: \citep{rbg+06}. Secondly, \citet{wsrw06} argue that the nearby
399: rotation-powered pulsar PSR~B0656+14 shows emission that would cause
400: it to be identified as an RRAT if its distance were greater. We note,
401: however, that single radio pulse/burst searches towards the AXPs are
402: also motivated for other reasons. \transaxp\ has shown strong
403: individual pulses, with an intensity distribution that follows a power
404: law. This presents the possibility of detecting individual bright
405: pulses even if the majority of the radio emission is below the
406: detection threshold. If magnetars can emit giant pulses, this is also
407: potentially interesting, as such pulses may be generated in a
408: different part of the magnetosphere than ``normal'' pulses
409: \citep[][and references therein]{kbm+06}. If beaming plays an
410: important role in the detectability of radio emission from magnetars,
411: then giant pulses could provide another window in which to detect
412: radio emission. Furthermore, AXPs are known to emit rare X-ray
413: bursts. One of the AXPs in our sample, 1E~1048.1$-$5937, was the
414: first shown to have such behavior \citep{gkw02}. Such X-ray bursts
415: may be accompanied by radio bursts \citep{l02}.
416:
417: In conclusion, our non-detections support the hypothesis that pulsed
418: radio emission from AXPs is connected to X-ray outburst activity, as
419: may be the case for \transaxp. However, unfavorable viewing
420: geometries and very weak radio emission cannot be dismissed as reasons
421: for our non-detections. Given that the transient radio emission
422: observed from \transaxp\ appears to be linked with its transient X-ray
423: behavior, continued radio monitoring of all the AXPs is warranted,
424: particularly after periods of X-ray burst activity.
425:
426: \acknowledgements
427:
428: We thank Scott Ransom for creating the {\tt PRESTO} suite of software
429: tools, which were used in this analysis, and we acknowledge use of the
430: McGill University magnetar catalog. J.W.T.H. holds an NSERC
431: Postdoctoral Fellowship. V.M.K. is the Lorne Trottier Canada Research
432: Chair and acknowledges support from NSERC, CIAR, and FQRNT. The
433: Parkes radio telescope is part of the Australia Telescope, which is
434: funded by the Commonwealth of Australia for operation as a National
435: Facility managed by CSIRO.
436:
437: %\begin{thebibliography}{}
438: \bibliographystyle{apj}
439: \bibliography{journals1,modrefs,psrrefs,chk07_ref}
440:
441: \clearpage
442:
443: \input{tab1}
444:
445: \end{document}
446:
447:
448:
449:
450:
451: