1: \documentclass[12pt,preprint]{aastex}
2: \usepackage{psfig} %instead of using \plotone
3:
4: \shorttitle{A Survey of 56 Mid-latitude EGRET Error Boxes for Radio
5: Pulsars}
6: \shortauthors{Crawford et al.}
7:
8: \begin{document}
9:
10: \title{A Survey of 56 Mid-latitude EGRET Error Boxes for Radio
11: Pulsars}
12:
13: \author{Fronefield Crawford\altaffilmark{1,2},
14: Mallory S. E. Roberts\altaffilmark{3,4},
15: Jason W. T. Hessels\altaffilmark{3},
16: Scott M. Ransom\altaffilmark{3,5},
17: Margaret Livingstone\altaffilmark{3},
18: Cindy R. Tam\altaffilmark{3},
19: Victoria M. Kaspi\altaffilmark{3}}
20:
21: \altaffiltext{1}{Department of Physics and Astronomy, Franklin \&
22: Marshall College, Lancaster, PA 17604, USA; email: fcrawfor@fandm.edu}
23:
24: \altaffiltext{2}{Department of Physics, Haverford College, Haverford,
25: PA 19041, USA}
26:
27: \altaffiltext{3}{Department of Physics, McGill University, Montreal,
28: QC H3A 2T8, Canada}
29:
30: \altaffiltext{4}{Eureka Scientific, Inc., 2452 Delmer Street, Suite
31: 100, Oakland, CA 94602, USA}
32:
33: \altaffiltext{5}{National Radio Astronomy Observatory, 520 Edgemont
34: Rd., Charlottesville, VA 22903, USA}
35:
36: \begin{abstract}
37: We have conducted a radio pulsar survey of 56 unidentified
38: $\gamma$-ray sources from the 3rd $EGRET$ catalog which are at
39: intermediate Galactic latitudes ($5^{\circ} < |b| < 73^{\circ}$). For
40: each source, four interleaved 35-minute pointings were made with the
41: 13-beam, 1400-MHz multibeam receiver on the Parkes 64-m radio
42: telescope. This covered the 95\% error box of each source at a
43: limiting sensitivity of $\sim 0.2$~mJy to pulsed radio emission for
44: periods $P\ga 10$~ms and dispersion measures $\la 50$ pc cm$^{-3}$.
45: Roughly half of the unidentified $\gamma$-ray sources at $|b| >
46: 5^{\circ}$ with no proposed active galactic nucleus counterpart were
47: covered in this survey. We detected nine isolated pulsars and four
48: recycled binary pulsars, with three from each class being new. Timing
49: observations suggest that only one of the pulsars has a spin-down
50: luminosity which is even marginally consistent with the inferred
51: luminosity of its coincident $EGRET$ source. Our results suggest that
52: population models, which include the Gould belt as a component,
53: overestimate the number of isolated pulsars among the mid-latitude
54: Galactic $\gamma$-ray sources and that it is unlikely that Gould belt
55: pulsars make up the majority of these sources. However, the
56: possibility of steep pulsar radio spectra and the confusion of
57: terrestrial radio interference with long-period pulsars ($P \ga
58: 200$~ms) having very low dispersion measures ($\la 10$ pc cm$^{-3}$,
59: expected for sources at a distance of less than about 1~kpc) prevent
60: us from strongly ruling out this hypothesis. Our results also do not
61: support the hypothesis that millisecond pulsars make up the majority
62: of these sources. Non-pulsar source classes should therefore be
63: further investigated as possible counterparts to the unidentified
64: $EGRET$ sources at intermediate Galactic latitudes.
65: \end{abstract}
66:
67: \keywords{pulsars: general, searches --- gamma rays: observations}
68:
69: \section{Introduction}
70:
71: Determining the nature of Galactic $\gamma$-ray sources with energies
72: above 100 MeV is one of the outstanding problems in high-energy
73: astrophysics. The $EGRET$ telescope on the Compton Gamma-Ray
74: Observatory, which was active from 1991 to 1999, identified about half
75: a dozen of the brightest $\gamma$-ray sources in the Galactic plane as
76: young pulsars \citep{tbb+99}. It also demonstrated that most of the
77: sources at low Galactic latitudes ($|b| \la 5^{\circ}$) are associated
78: with star forming regions, and hence may be pulsars, pulsar wind
79: nebulae, supernova remnants, winds from massive stars, or high-mass
80: X-ray binaries \citep{kc96,yr97,rbt99}. In addition, molecular clouds
81: can either be sources of $\gamma$-rays or enhance the production of
82: $\gamma$-rays by particles produced by the source classes mentioned
83: above \citep{aha01}. Various targeted multi-wavelength campaigns to
84: identify low-latitude sources have discovered a number of likely
85: counterparts \citep{rrk01,hcg+01, rhr+02,brrk02,hgc+04}. The recent
86: Parkes Multibeam Survey has also discovered several new pulsars
87: coincident with $EGRET$ $\gamma$-ray sources; these pulsars have spin
88: characteristics that are similar to those of the known $\gamma$-ray
89: pulsars \citep{dkm+01,kbm+03}.
90:
91: While there are many candidate counterparts to $EGRET$ sources at low
92: latitudes, there are few firm identifications owing to the large
93: positional uncertainties of the sources (typically $\sim 1^{\circ}$
94: across). In general, a timing signature, such as a pulse detection, is
95: necessary to be certain of a source identity. Since young pulsars
96: tend to be noisy rotators, extrapolating a pulse ephemeris reliably
97: back to the era of the $EGRET$ observation is generally not possible.
98: With the improved resolution and sensitivity of the upcoming $AGILE$
99: and $GLAST$ missions, the low-latitude $EGRET$ sources should be more
100: easily identified.
101:
102: There are estimated to be between 50 and 100 sources detected by
103: $EGRET$ at mid-Galactic latitudes which are associated with our
104: Galaxy. As a class, these sources tend to be fainter and have steeper
105: spectra than those at low latitudes \citep{hbb+99}. Their positional
106: uncertainty is therefore on average even greater ($\sim 1.5^{\circ}$
107: across) than it is for the low-latitude sources. These mid-latitude
108: sources have a spatial distribution which is similar to the Gould belt
109: of local regions of recent star formation plus a Galactic Halo
110: component \citep{gre00,gre01}. The Gould belt provides a natural
111: birth place for many nearby ($\la 0.5$~kpc), middle-aged pulsars
112: similar to Geminga \citep{hh92}. Both the outer gap \citep{yr95} and
113: polar cap \citep{hz01} models of pulsar emission suggest that many of
114: these pulsars should be detectable in $\gamma$-rays but that the
115: majority should have their radio beams missing Earth. However, if
116: predictions from recent models are realistic, then between 25\% and
117: 50\% of $\gamma$-ray pulsars might still be visible to us as radio
118: pulsars \citep{gvh04,czlj04}.
119:
120: The mid-latitude $EGRET$ source distribution is also similar to the
121: distribution of recycled pulsars in the Galactic field
122: \citep{rom01}. The fastest millisecond pulsars (MSPs) can have
123: spin-down luminosities ($\dot{E} \propto \dot{P}/P^{3}$) and
124: magnetospheric potentials similar to those of young pulsars. There has
125: been one possible detection of $\gamma$-ray pulsations from an MSP
126: \citep{khv+00} and some preliminary modeling of that emission
127: \citep{hum05}. If a significant fraction of the mid-latitude sources
128: are MSPs at typical Galactic distances, many should be detectable as
129: radio pulsars \citep{sgh05}. Since MSPs tend to be in binary systems,
130: $GLAST$ will not be sensitive to them in blind searches (owing to
131: computational reasons associated with the very long integration times
132: and the large number of trials required to search the parameter
133: space).
134:
135: Here we describe a radio pulsar survey of 56 unidentified sources from
136: the 3rd $EGRET$ catalog (3EG) \citep{hbb+99} which are at intermediate
137: Galactic latitudes ($5^{\circ} < |b| < 73^{\circ}$). The survey used
138: the 1400-MHz, 13-beam multibeam receiver \citep{swb+96} on the 64-m
139: radio telescope in Parkes, Australia to search for pulsed
140: emission. This receiver has been used very successfully to find
141: pulsars in a number of recent radio pulsar surveys \citep{mlc+01,
142: ebs+01, kbm+03, mfl+06, bjd+06}. Discovery of radio pulsar
143: counterparts to these $EGRET$ sources would not only provide
144: interesting systems for individual study and establish the
145: identifications of the target sources (e.g., Roberts et
146: al. 2002\nocite{rhr+02}), but it would also help resolve outstanding
147: questions about the pulsar emission mechanism and the physical origin
148: of pulsar radiation at different wavelengths (see, e.g., Harding et
149: al. 2004 and references therein\nocite{hgg+04}).
150:
151: \section{Survey Parameters and Data Processing}
152:
153: We used four criteria in the selection of target $EGRET$ sources for
154: our survey. First, a source was included only if it was not in the
155: range of the Parkes Multibeam Survey \citep{mlc+01}, which covered
156: Galactic latitudes $|b| < 5^{\circ}$. Since our targeted survey had a
157: comparable sensitivity to the Parkes Multibeam Survey, there was no
158: reason to repeat that coverage. Second, a source had to have no
159: strong candidate for an active galactic nucleus (AGN) as determined by
160: the study of \citet{mhr01}. Third, a source had to have been easily
161: observable by the Parkes telescope, corresponding to a declination
162: range $\delta < +20^{\circ}$. Finally, the positional uncertainty from
163: the 3EG catalog had to be sufficiently small that a single
164: four-pointing tessellation pattern with the multibeam receiver would
165: cover virtually the entire 95\% confidence region of the source. Using
166: these criteria, we selected 56 unidentified $EGRET$ $\gamma$-ray
167: sources to survey. Figure \ref{fig-1} shows the sky locations of the
168: 56 target sources and the locations of known pulsars. Table
169: \ref{tbl-1} lists the 56 $EGRET$ sources with their nominal 3EG
170: positions. These positions were used as the target centers in the
171: first pointing of each pointing cluster. Since the beams of the
172: multibeam receiver are spaced two beamwidths apart, four pointings are
173: required for full coverage of a region on the sky (e.g., Manchester et
174: al.~2001\nocite{mlc+01}). This is illustrated in Figure \ref{fig-2}.
175:
176: We recorded a total of 3016 beams in the survey between June 2002 and
177: July 2003.\footnote{Nine telescope pointings were repeated in the
178: survey, and one pointing was missed. All other pointings were unique
179: (see Table \ref{tbl-1}).} For each telescope pointing, we used a
180: 35-minute observation sampled at 0.125 ms with 1-bit per sample. 96
181: contiguous frequency channels of 3 MHz each were recorded during each
182: observation, providing a total observing bandwidth of 288 MHz centered
183: at 1374 MHz. The observing setup was similar to the one described in
184: detail by \citet{mlc+01} for the Parkes Multibeam Survey, except that
185: twice the sample rate was used here in order to increase sensitivity
186: to MSPs. Each resulting beam contained $\sim 200$~MB of raw data,
187: corresponding to a total of $\sim 600$~GB of raw survey data to be
188: processed for pulsar signals.
189:
190: The raw data from the survey were originally processed at McGill
191: University using the Borg computer cluster and the PRESTO suite of
192: pulsar analysis tools \citep{r01,
193: rem02}\footnote{http://www.cv.nrao.edu/$\sim$sransom/presto} with
194: acceleration searches. In the search, we dedispersed each data set at
195: 150 trial dispersion measures (DMs) ranging from 0 to 542 pc
196: cm$^{-3}$, which easily encompassed the expected maximum DM for
197: Galactic pulsars in the directions observed \citep[][see
198: Table~1]{cl02}. The values of the DM trials were chosen such that the
199: spacing did not add to the dispersive smearing already caused by the
200: finite frequency channels. Since radio frequency interference (RFI)
201: can mask pulsar signals, we searched for RFI in particular spectral
202: channels and time bins for each observation, and a mask was created to
203: exclude these data from the subsequent reduction and analysis.
204: Typically about 10-20\% of the data were rejected in this process.
205:
206: For each trial DM, we summed the frequency channels with appropriate
207: delays to create a time series. The time series was then Fourier
208: transformed using a Fast Fourier Transform (FFT), and a red noise
209: component of the power spectrum (i.e., low-frequency noise in the
210: data) was removed. This was done by dividing the spectral powers by
211: the local median of the power spectrum, increasing the number of bins
212: used in the average logarithmically with frequency. We masked known
213: interference signals in the power spectrum, corresponding to less than
214: 0.05\% of the spectrum, and used harmonic summing with up to 8
215: harmonics to enhance sensitivity to highly non-sinusoidal signals. In
216: the acceleration search, we were sensitive to signals in which the
217: fundamental drifted linearly by up to 100 Fourier bins during the
218: course of the observation, providing sensitivity to pulsars in tight
219: binaries; the maximum detectable acceleration was $a_{\rm max} = 6.8
220: P$ m s$^{-2}$, where $P$ is the pulsar spin period in
221: milliseconds. This is about 40\% of the maximum acceleration searched
222: in the Parkes Multibeam Survey processing, which used a segmented
223: linear acceleration search \citep{fsk+04,l05}. We estimate that our
224: acceleration search would have been sensitive to all but one of the
225: known pulsars in double neutron star binary systems (the one exception
226: being PSR J0737$-$3039A). We performed folding searches around
227: candidate periods and period derivatives and examined the results by
228: eye. The characteristic signal of interest was a dispersed, wideband,
229: extremely regular series of pulsations.
230:
231: Averaged over the survey, the sensitivity to pulsars in an RFI free
232: environment was $\sim 0.2$ mJy for most periods and DMs (see Figure
233: \ref{fig-3}). The sensitivity calculation is outlined in \citet{c00}
234: and \citet{mlc+01} and was determined for a blind FFT search. RFI
235: tends to introduce sporadic, highly variable red noise in the power
236: spectra, especially at low dispersion measures (DM $\la 10$ pc
237: cm$^{-3}$). Therefore, sensitivity to slow pulsars ($P \ga 200$~ms)
238: with low DMs is reduced in a way which is difficult to quantify. In
239: addition, the DM peaks of long-period pulsars are broader than those
240: of MSPs and hence are more difficult to distinguish from zero DM when
241: the DM is very low. During this first processing run, we discovered
242: six new pulsars and redetected all previously known pulsars that were
243: within the full-width half-maximum area of the survey beams (see Table
244: \ref{tbl-2}).
245:
246: We conducted a second processing pass at Haverford College using the
247: pulsar search packages SEEK and SIGPROC (e.g., Lorimer et al
248: 2000\nocite{lkm+00}).\footnote{http://sigproc.sourceforge.net} The
249: re-processing of the data with a different analysis package aimed to
250: see whether there were pulsars that were missed during the first
251: processing pass. Of particular interest were long-period pulsars ($P
252: \ga 20$ ms), since fewer than expected were found in the first
253: processing run. We therefore decimated the data prior to processing
254: to reduce their size and thus significantly decrease the processing
255: time while still maintaining sensitivity to longer-period pulsars.
256: The data were decimated by a factor of four in frequency and a factor
257: of 16 in time, resulting in effective frequency channels of 12~MHz
258: sampled every 2.0~ms. This reduced the size of each data set by a
259: factor of 64. We were in practice sensitive to pulsars with periods
260: greater than about 20 ms in the re-processing of the data.
261:
262: These data were dedispersed at 450 trial DMs between 0 and 700 pc
263: cm$^{-3}$. The large number of DM trials ensured that no weak
264: candidates with fast periods ($P \sim 20$-30 ms) were missed between
265: DM steps. Each resulting time series was Fourier transformed, excised
266: of RFI, and searched for candidate signals. We then dedispersed and
267: folded the raw data at DMs and periods around the candidate values. We
268: redetected all of the pulsars that had been detected in the first
269: processing run (except for PSR J1614$-$2230, which has a period of
270: $\sim 3$~ms), but no additional pulsars were found. We also searched
271: the data for dispersed single pulses. Dispersed radio bursts have
272: recently been observed from a newly discovered class of transient
273: radio sources; these sources are believed to be associated with
274: rotating neutron stars \citep{mll+06}. Our single pulse search
275: revealed no new candidates, but several known pulsars were redetected
276: in this way. We also constructed an archive of the raw data from the
277: survey on DVD \citep{ccd+04}. A complete index of the survey and
278: instructions for requesting raw data from the archive is accessible
279: via the world wide web.\footnote{http://cs.haverford.edu/pulsar}
280:
281: \section{Results}
282:
283: We detected a total of 13 pulsars in the survey, six of which were
284: new. Timing observations quickly established that three of the six new
285: pulsars are isolated and three are in binary systems.
286: Table~\ref{tbl-2} lists all 13 pulsars detected in the survey.
287:
288: The three new isolated pulsars, PSRs J1632$-$1032, J1725$-$0732, and
289: J1800$-$0125, were timed at Parkes in 2003 and 2004 with some
290: supplemental observations taken with the Green Bank Telescope
291: (GBT). We conducted timing observations at roughly monthly intervals
292: at several central observing frequencies (mostly 1374 MHz, but also
293: 680, 820, 1400, 1518, and 2934 MHz, depending on the receivers
294: available at different times) and produced times-of-arrival from the
295: observations. The observing setup was similar to the one used for
296: timing pulsars discovered in the Parkes Multibeam Survey
297: \citep{mlc+01}. These data were fit to a model which included spin
298: parameters, sky position, and DM using the TEMPO software
299: package.\footnote{http://www.atnf.csiro.au/research/pulsar/tempo} We
300: used supplemental GBT observations taken in the middle of 2004 along
301: with the original Parkes survey observations to obtain phase-connected
302: timing solutions which spanned more than a year. Table~\ref{tbl-3}
303: gives the full timing solutions for these three new isolated pulsars
304: (including 1400-MHz flux densities), and Figure~\ref{fig-4} shows
305: their 20-cm pulse profiles.
306:
307: The three new binary pulsars, PSRs J1614$-$2315, J1614$-$2230, and
308: J1744$-$3922,\footnote{One of the new binary pulsars, PSR
309: J1744$-$3922, was independently discovered in the re-processing of the
310: Parkes Multibeam Survey data \citep{fsk+04}.} were regularly timed
311: with Parkes and the GBT over a similar period of time
312: \citep{hrr+05}. These pulsars will be discussed in detail by
313: \citet{rrh+06}. We also detected a fourth binary pulsar, PSR
314: J0407+1607, in the survey. This pulsar was previously discovered in an
315: Arecibo drift scan survey by \citet{lxk+05}.
316:
317: If the pulsar distances estimated from the DMs using the NE2001
318: Galactic electron density model \citep{cl02} are approximately correct
319: (to within about a factor of two), then none of the pulsars detected
320: has a spin-down luminosity which is large enough to clearly account
321: for the $\gamma$-ray luminosity of its coincident $EGRET$ source. Only
322: the MSP PSR J1614$-$2230 has a spin-down luminosity of a similar
323: magnitude to the estimated $\gamma$-ray luminosities of our sources,
324: which, given the DM distances and $EGRET$ fluxes, are in the $10^{34}$
325: to $10^{35}$ erg s$^{-1}$ range. Even PSR J1614$-$2230 would have to
326: be highly efficient to be the counterpart to its coincident
327: $\gamma$-ray source (this will be discussed in more detail by Ransom
328: et al. 2006\nocite{rrh+06}). Therefore, none of the pulsars is a
329: strong candidate for an $EGRET$ association based on its spin-down
330: luminosity. All of the DM-estimated distances to the detected pulsars
331: ($d \ga 1.3$~kpc; see Table \ref{tbl-2}) are too large to be part of a
332: Gould Belt population, which is expected to have a distance $\la
333: 0.5$~kpc. In fact, one of the new pulsars, PSR J1632$-$1013, has a DM
334: which is larger than the maximum expected DM along its line of sight.
335: Although only about half of the surveyed $EGRET$ sources were within
336: $30^{\circ}$ of the Galactic center, only PSR J1821+1715 and the
337: long-period binary PSR J0407+1607 were detected outside this region.
338:
339: \section{Discussion}
340:
341: The majority of identified $EGRET$ sources at high Galactic latitudes
342: are of the blazar sub-class of AGN. As stated above, we selected
343: against these sources based on the work of \citet{mhr01}. However,
344: more recent radio and optical work by Sowards-Emmerd and collaborators
345: \citep{srm03,srm+04} on the complete sample of 3EG sources north of
346: $-40^{\circ}$ declination has significantly expanded the number of
347: potential AGN identifications. 33 sources remaining with no potential
348: AGN counterparts (corresponding to roughly half of all such
349: unidentified sources at Galactic latitudes $|b| > 5^{\circ}$) were
350: included in our search. We included about one quarter of the sources
351: with only weak AGN candidates by their criterion as well. Six of our
352: sources were identified in their work as having firm AGN associations
353: (see Table \ref{tbl-1}). Therefore, for discussion purposes, we
354: assume that 50\% of all unidentified Galactic sources with $|b| >
355: 5^{\circ}$ were covered in our survey.
356:
357: One well-discussed model suggests that the mid-latitude $EGRET$
358: sources are primarily nearby, middle-aged pulsars born in the Gould
359: belt. This has been motivated by an apparently statistically
360: significant spatial correlation between the unidentified $\gamma$-ray
361: sources and the Gould belt \citep{gmb+00,gre01}. \citet{gvh04} have
362: modeled the pulsar population using estimated pulsar birth rates in
363: the Gould Belt in addition to simulating the Galactic population as a
364: whole, and their simulations suggest that $\sim 15$ pulsars ought to
365: be detectable by $EGRET$ at mid-latitudes, roughly half of which are
366: radio loud (assuming a particular luminosity law and beaming model for
367: the radio emission which is consistent with the total known population
368: of isolated radio pulsars). However, since their simulation accounts
369: for only $\sim 1/4$ of the total unidentified $\gamma$-ray population,
370: the hypothesis that all of the sources are pulsars would suggest that
371: $\sim 15$ radio loud pulsars ought to have been detectable in our
372: sample of $EGRET$ sources. A similar study by \citet{czlj04}, based
373: on the outer gap emission model, finds $\sim 4$ radio loud pulsars
374: from the Gould belt and another 4 from the remainder of the Galaxy at
375: $|b| > 5^{\circ}$. The total number of pulsars at mid-latitudes from
376: this simulation accounts for $\sim 1/2$ the total unidentified
377: population, indicating that our survey should have detected $\sim 8$
378: associated radio pulsars. Both of these simulations were done using
379: estimates of the limiting sensitivities of a variety of previous radio
380: surveys which were mostly performed at $\sim 400$ MHz and do not
381: include the various multibeam surveys at mid- and high-latitudes. Our
382: survey covered $\sim 50$\% of the potential $EGRET$ pulsars at $|b| >
383: 5^{\circ}$, and yet no plausible radio candidates were discovered.
384: The absence of detections in our survey is significant given the
385: discrepancy between our results and the $\sim 8$ and $\sim 15$
386: detectable radio pulsars predicted in the two models under the
387: assumption of a single source class consisting of pulsars. For a
388: source distance of 0.5 kpc, our 1400-MHz luminosity limit was about
389: 0.05~mJy~kpc$^{2}$; the radio luminosity, $L_{1400}$, is defined as
390: $L_{1400} = S_{1400} d^{2}$, where $S_{1400}$ is the 1400-MHz flux
391: density and $d$ is the pulsar distance. This luminosity limit is
392: lower than the 1400-MHz luminosity of all but two pulsars for which
393: this quantity has been measured and published
394: \citep{mht+05}.\footnote{http://www.atnf.csiro.au/research/pulsar/psrcat}
395: The surveys used for the studies mentioned above were typically $\sim
396: 4$ times less sensitive than our survey (assuming an average spectral
397: index of $-2$ for pulsars, as was assumed by Cheng et
398: al. 2004\nocite{czlj04}). Our results suggest that the simulations
399: significantly overestimate the radio-loud $\gamma$-ray pulsar
400: population at mid-latitudes and do not support the hypothesis that
401: middle-aged, nearby pulsars make up the majority of the unidentified
402: sources.
403:
404: There are several important caveats to this conclusion. The first is
405: that the average radio spectral index of middle-aged, $\gamma$-ray
406: emitting pulsars is unknown. If, for whatever reason, these sources
407: preferentially have very steep radio spectra, we might not be
408: sensitive to them at the relatively high observing frequency of this
409: survey. The second caveat is the difficulty in distinguishing a peak
410: at a small but nonzero DM in the data at this frequency. A clear
411: indication of a dispersed signal is one of the important ways of
412: distinguishing a celestial signal from local RFI. Since Gould belt
413: pulsars are expected to be very close to Earth ($d \la 0.5$ kpc), the
414: expected DM is less than about 10 pc cm$^{-3}$ along many lines of
415: sight. This often cannot be clearly differentiated from zero DM with
416: the high observing frequency of the multibeam system. This is
417: especially true of long-period pulsars. In fact, we detected a large
418: number of promising candidates with pulsar-like characteristics which
419: peaked at a DM of zero. Although we attempted (and failed) to confirm
420: some of the most pulsar-like of these candidates at 680 MHz, we still
421: cannot definitely rule out that some of these candidates may be
422: astronomical sources. Observations of these sources at lower
423: frequencies (300-400 MHz) with modern, wide-bandwidth systems (50-64
424: MHz) may be able to resolve these low-DM and spectral index
425: issues. However, a recent 327-MHz search of 19 mid-latitude $EGRET$
426: error boxes visible from the Arecibo telescope found no new pulsar
427: counterparts \citep{cml05}, lending support to the conclusion that
428: pulsars are not powering the majority of these $\gamma$-ray sources.
429:
430: Although this survey detected more pulsars in binary systems per
431: square degree (0.032 deg$^{-2}$) outside of globular clusters than any
432: previous survey, PSR J1614$-$2230 was the only MSP we detected which
433: is even a marginal counterpart candidate. Recent modeling of
434: high-energy spectra of MSPs \citep{hum05} suggests that most MSPs
435: visible to $EGRET$ would be active radio pulsars with significant
436: radio luminosity. Therefore, the number of observable radio MSPs
437: detectable by our survey should only depend on the relative radio and
438: $\gamma$-ray beaming fractions. At large DMs (DM $\ga 100$ pc
439: cm$^{-3}$), our sensitivity to MSPs is severely compromised owing to
440: dispersive smearing. However, Table \ref{tbl-1} indicates that less
441: than half of our $EGRET$ targets have a maximum expected DM greater
442: than 100 pc cm$^{-3}$, and, of these, only the most distant pulsars
443: near the edge of the Galactic electron layer would actually have such
444: large DMs. Dispersive smearing is therefore likely not the reason why
445: a majority of MSPs would have been missed in our survey. For a
446: distance of $\sim 3$~kpc, most of the $\gamma$-ray sources would have
447: luminosities of $\sim 10^{35}$ ergs s$^{-1}$, and so we deem it
448: unlikely that MSPs could be powering $EGRET$ sources at distances much
449: further than this. At 3~kpc, our 1400-MHz luminosity limit for a 2-ms
450: pulsar with a DM of 50 pc ${\rm cm}^{-3}$ is $\sim 5$~mJy
451: kpc$^2$. While the dependence of radio luminosity on spin-down
452: luminosity is not well known for MSPs, this level of sensitivity would
453: have allowed us to detect the majority of known MSPs. Therefore, our
454: results do not support the hypothesis that recycled pulsars having
455: radio luminosities similar to those of the known population make up
456: the majority of the unidentified $EGRET$ source population. On the
457: other hand, the detection of a total of four binary systems in this
458: survey indicates that deeper surveys for binary pulsars, especially
459: within $30^{\circ}$ of the Galactic center, appear warranted.
460:
461: The detection of only 3 new isolated pulsars was somewhat surprising,
462: especially since we discovered an equal number of new binary pulsars
463: and detected 6 previously known isolated pulsars within the survey
464: area (Table \ref{tbl-2}). Since our survey was $\sim 3$ to 4 times
465: more sensitive than previous surveys (assuming a typical spectral
466: index), we might have expected to discover a dozen or so new isolated
467: pulsars. As noted above, most of the previous surveys at high
468: latitudes were conducted at lower observing frequencies, and therefore
469: such a simple estimate is subject to uncertainties in the spectral
470: index and the influence of RFI. However, the strong detections of all
471: previously known pulsars argues that these uncertainties may not be
472: very significant.
473:
474: We therefore estimate the total number of pulsars we could expect to
475: detect at our observing frequency by comparing our results with those
476: of the Swinburne mid-latitude surveys \citep{ebs+01,jac04}. These
477: surveys covered Galactic longitudes $-100^{\circ} < l < 50^{\circ}$
478: using the Parkes multibeam receiver and an identical observing setup
479: to ours, but with only 1/8 the integration time. The first of these
480: surveys covered Galactic latitudes $5^{\circ} < |b| < 15^{\circ}$ and
481: detected 170 pulsars, including 12 binaries. By simply scaling by the
482: area covered in this survey, the integration time, and assuming a
483: ${\rm d}\log N / {\rm d}\log S$ distribution of $-1$ for Galactic
484: plane pulsars at 20~cm \citep{bam+03}, we would expect to have
485: detected a total of $\sim 24$ pulsars instead of 13. However, we
486: should have detected only 2-3 binary pulsars, while we detected 4.
487: The second Swinburne survey, covering $15^{\circ} < |b| < 30^{\circ}$,
488: detected only 62 pulsars, 11 of which were binaries
489: \citep{jac04}. This, along with the fact that 11 of our 13 detections
490: were within $\sim 30^{\circ}$ of the Galactic center, suggests a
491: strong spatial dependence to the pulsar population out of the plane,
492: which is hardly surprising. We therefore calculated the number of
493: isolated pulsars we would have expected to detect within the error
494: boxes overlapping the coverage of the Swinburne surveys given the
495: total area covered by our survey within each Swinburne survey and
496: within $|l| < 30^{\circ}$. Scaling from the surveys and assuming a
497: ${\rm d}\log N / {\rm d}\log S$ distribution of $-1$, we should have
498: detected $\sim 7$ isolated pulsars but only $\sim 1$ binary pulsar,
499: when we actually detected 8 and 3, respectively, in this region. In
500: the $EGRET$ boxes within the Swinburne latitudes but outside their
501: longitude range (presuming no further longitudinal dependence for
502: $|l|> 30^{\circ}$), we would have expected $\sim 1$ isolated and $\sim
503: 0$ binary pulsars, while we detected 1 of each. At higher latitudes,
504: if the detection rate remained the same for $|b|> 30^{\circ}$ as for
505: the second Swinburne survey ($15^{\circ} < |b| < 30^{\circ}$), we
506: would have expected to detect $\sim 2$ pulsars. No pulsars were
507: detected in our survey at high latitudes. We therefore conclude that
508: our results are consistent with an extrapolation from the Swinburne
509: observations only if we take into account a strong latitudinal
510: dependence of the isolated pulsar distribution, as expected for a disk
511: based population, and the apparent concentration of binary pulsars
512: within $\sim 30^{\circ}$ of the Galactic center. This supports the
513: trend in the spatial distribution of MSPs suggested by \citet{bjd+06}
514: obtained by combining data from the Parkes High-Latitude pulsar survey
515: and the two Swinburne surveys. This suggests that we have not yet
516: reached the lower luminosity limit of either the isolated or binary
517: pulsar populations at mid-Galactic latitudes toward the Galactic
518: center, since we found approximately what would be expected from a
519: simple ${\rm d}\log N / {\rm d}\log S$ extrapolation. However, we may
520: be reaching the luminosity limit toward the anti-center.
521:
522: \section{Conclusions}
523:
524: There are now 20 pulsars that are known to lie within 1.5 times the
525: radius of the 95\% confidence contours of $EGRET$ sources at $|b| >
526: 5^{\circ}$. Of these, only the Crab pulsar and PSR B1055$-$52 have
527: confirmed associations with the coincident $\gamma$-ray emission. Of
528: the remaining 18 pulsars, including the 13 detected in our survey and
529: the recently discovered PSR J2243+1518 \citep{cml05}, none is
530: energetic enough for a clear association. Other than PSR J1614$-$2230,
531: which is at best a marginal candidate, no pulsars from any survey have
532: been found which can be associated with unidentified $EGRET$ error
533: boxes at mid-Galactic latitudes. Non-pulsar source classes should
534: therefore be investigated further. \citet{gkr05} discuss the viability
535: of low-mass microquasars as $EGRET$ sources. Recently, there has been
536: the suggestion that much of the $\gamma$-ray emission at mid-latitudes
537: is due to gas not being included in the models used for calculating
538: the $\gamma$-ray background maps \citep{gct05}. In this case, many of
539: the cataloged sources may not be truly point-like. Regardless, as
540: suggested by spectral and variability studies of the population (e.g.,
541: Grenier 2003\nocite{gre03}), the likelihood of pulsars being able to
542: account for a majority of the cataloged unidentified $EGRET$ sources
543: at intermediate Galactic latitudes seems remote.
544:
545: \acknowledgements
546:
547: This work was supported by the Canada Foundation for Innovation, the
548: Haverford College Faculty Research Fund, the Haverford Faculty Support
549: Fund, the Keck Northeast Astronomy Consortium, and the NRAO Foreign
550: Telescope Travel Grants program. The Parkes radio telescope is part
551: of the Australia Telescope, which is funded by the Commonwealth of
552: Australia as a National Facility operated by CSIRO. VMK is a Canada
553: Research Chair, and JWTH is an NSERC PGS-D Fellow. VMK received
554: support from an NSERC Discovery Grant and Steacie Fellowship
555: Supplement, and by the FQRNT and CIAR. We have made use of the $ROSAT$
556: data archive of the Max-Planck-Institut f\"{u}r extraterrestrische
557: Physik (MPE) at Garching, Germany. We thank the referee, Fernando
558: Camilo, for helpful suggestions for the revised manuscript, and Dunc
559: Lorimer for providing key components of the software used in the
560: re-analysis (SEEK and SIGPROC). We also thank Andrew Cantino, Allison
561: Curtis, Saurav Dhital, Steve Gilhool, Megan Roscioli, Gabe Roxby, Ryan
562: Sajac, Reid Sherman, and Aude Wilhelm for contributions to the data
563: processing and analysis.
564:
565: \begin{thebibliography}{}
566:
567: \bibitem[Aharonian(2001)]{aha01} Aharonian, F.~A.\ 2001, Space Science
568: Reviews, 99, 187
569:
570: \bibitem[Bhattacharya et al.(2003)]{bam+03} Bhattacharya, D.,
571: Aky{\"u}z, A., Miyagi, T., Samimi, J., \& Zych, A.\ 2003, \aap, 404,
572: 163
573:
574: \bibitem[Braje et al.(2002)]{brrk02} Braje, T.~M., Romani, R.~W.,
575: Roberts, M.~S.~E., \& Kawai, N.\ 2002, \apjl, 565, L91
576:
577: \bibitem[Burgay et al.(2006)]{bjd+06} Burgay, M., et al.\ 2006,
578: \mnras, 368, 283
579:
580: \bibitem[Cantino et al.(2004)]{ccd+04} Cantino, A., Crawford, F.,
581: Dhital, S., Dougherty, J., \& Sherman, R. 2004, Eleventh SIAM
582: Conference on Parallel Processing for Scientific Computing, San
583: Francisco, CA
584:
585: \bibitem[Champion, McLaughlin, \& Lorimer(2005)]{cml05} Champion,
586: D.~J., McLaughlin, M.~A., \& Lorimer, D.~R.\ 2005, \mnras, 364, 1011
587:
588: \bibitem[Cheng et al.(2004)]{czlj04} Cheng, K.~S., Zhang, L., Leung,
589: P., \& Jiang, Z.~J.\ 2004, \apj, 608, 418
590:
591: \bibitem[Condon et al.(1998)]{ccg+98} Condon, J.~J., Cotton, W.~D.,
592: Greisen, E.~W., Yin, Q.~F., Perley, R.~A., Taylor, G.~B., \&
593: Broderick, J.~J.\ 1998, \aj, 115, 1693
594:
595: \bibitem[Cordes \& Lazio(2002)]{cl02} Cordes, J. M. \& Lazio,
596: T. J. W. 2002, astro-ph/0207156
597:
598: \bibitem[Crawford(2000)]{c00} Crawford, F. 2000, Ph.D. thesis,
599: Massachusetts Institute of Technology
600:
601: \bibitem[D'Amico et al.(2001)]{dkm+01} D'Amico, N., et al.\ 2001,
602: \apjl, 552, L45
603:
604: \bibitem[Edwards et al.(2001)]{ebs+01} Edwards, R.~T., Bailes, M., van
605: Straten, W., \& Britton, M.~C.\ 2001, \mnras, 326, 358
606:
607: \bibitem[Faulkner et al.(2004)]{fsk+04} Faulkner, A.~J., et al.\ 2004,
608: \mnras, 355, 147
609:
610: \bibitem[Gehrels et al.(2000)]{gmb+00} Gehrels, N., Macomb, D.~J.,
611: Bertsch, D.~L., Thompson, D.~J., \& Hartman, R.~C.\ 2000, \nat, 404,
612: 363
613:
614: \bibitem[Gonthier, Van Guilder, \& Harding(2004)]{gvh04} Gonthier,
615: P.~L., Van Guilder, R., \& Harding, A.~K.\ 2004, \apj, 604, 775
616:
617: \bibitem[Grenier(2000)]{gre00} Grenier, I.~A.\ 2000, American
618: Institute of Physics Conference Series, 515, 261
619:
620: \bibitem[Grenier(2001)]{gre01} Grenier, I.~A.\ 2001, ASSL Vol.~267:
621: The Nature of Unidentified Galactic High-Energy Gamma-Ray Sources, 51
622:
623: \bibitem[Grenier(2003)]{gre03} Grenier, I.~A.\ 2003, Texas in
624: Tuscany.~XXI Symposium on Relativistic Astrophysics, 397
625:
626: \bibitem[Grenier, Casandjian, \& Terrier(2005)]{gct05} Grenier, I.~A.,
627: Casandjian, J.-M., \& Terrier, R.\ 2005, Science, 307, 1292
628:
629: \bibitem[Grenier, Kaufman Bernad{\'o}, \& Romero(2005)]{gkr05}
630: Grenier, I.~A., Kaufman Bernad{\'o}, M.~M., \& Romero, G.~E.\ 2005,
631: \apss, 297, 109
632:
633: \bibitem[Halpern \& Holt(1992)]{hh92} Halpern, J.~P., \& Holt, S.~S.\
634: 1992, \nat, 357, 222
635:
636: \bibitem[Halpern et al.(2001)]{hcg+01} Halpern, J.~P., Camilo, F.,
637: Gotthelf, E.~V., Helfand, D.~J., Kramer, M., Lyne, A.~G., Leighly,
638: K.~M., \& Eracleous, M.\ 2001, \apjl, 552, L125
639:
640: \bibitem[Halpern et al.(2004)]{hgc+04} Halpern, J.~P., Gotthelf,
641: E.~V., Camilo, F., Helfand, D.~J., \& Ransom, S.~M.\ 2004, \apj, 612,
642: 398
643:
644: \bibitem[Harding \& Zhang(2001)]{hz01} Harding, A.~K., \& Zhang, B.\
645: 2001, \apjl, 548, L37
646:
647: \bibitem[Harding et al.(2004)]{hgg+04} Harding, A. K., Gonthier,
648: P. L., Grenier, I. A., \& Perrot, C. A. 2004, Advances in Space
649: Research, 33, 571
650:
651: \bibitem[Harding, Usov, \& Muslimov(2005)]{hum05} Harding, A.~K.,
652: Usov, V.~V., \& Muslimov, A.~G.\ 2005, \apj, 622, 531
653:
654: \bibitem[Hartman et al.(1999)]{hbb+99} Hartman, R.~C.~et al.\ 1999,
655: \apjs, 123, 79
656:
657: \bibitem[Haslam et al.(1982)]{hss+82} Haslam, C.~G.~T., Stoffel, H.,
658: Salter, C.~J., \& Wilson, W.~E.\ 1982, \aaps, 47, 1
659:
660: \bibitem[Hessels et al.(2005)]{hrr+05} Hessels, J., Ransom,
661: S., Roberts, M., Kaspi, V., Livingstone, M., Tam, C., \& Crawford, F.\
662: 2005, ASP Conf.~Ser.~328: Binary Radio Pulsars, 328, 395
663:
664: \bibitem[Jacoby(2004)]{jac04} Jacoby, B. A. 2004, Ph.D. thesis,
665: California Institute of Technology
666:
667: \bibitem[Kaaret \& Cottam(1996)]{kc96} Kaaret, P., \& Cottam, J.\
668: 1996, \apjl, 462, L35
669:
670: \bibitem[Kramer et al.(2003)]{kbm+03} Kramer, M.~et al.\ 2003, \mnras,
671: 342, 1299
672:
673: \bibitem[Kuiper et al.(2000)]{khv+00} Kuiper, L., Hermsen, W.,
674: Verbunt, F., Thompson, D.~J., Stairs, I.~H., Lyne, A.~G., Strickman,
675: M.~S., \& Cusumano, G.\ 2000, \aap, 359, 615
676:
677: \bibitem[Lorimer et al.(2000)]{lkm+00} Lorimer, D.~R., Kramer, M.,
678: M{\" u}ller, P., Wex, N., Jessner, A., Lange, C., \& Wielebinski, R.
679: \ 2000, \aap, 358, 169
680:
681: \bibitem[Lorimer et al.(2005)]{lxk+05} Lorimer, D.~R., et al.\ 2005,
682: \mnras, 359, 1524
683:
684: \bibitem[Lyne(2005)]{l05} Lyne, A.~G.\ 2005, ASP Conf.~Ser.~328:
685: Binary Radio Pulsars, 328, 37
686:
687: \bibitem[Manchester et al.(2001)]{mlc+01} Manchester, R.~N.~et al.\
688: 2001, \mnras, 328, 17
689:
690: \bibitem[Manchester et al.(2005)]{mht+05} Manchester, R.~N., Hobbs,
691: G.~B., Teoh, A., \& Hobbs, M.\ 2005, \aj, 129, 1993
692:
693: \bibitem[Manchester et al.(2006)]{mfl+06} Manchester, R.~N., Fan, G.,
694: Lyne, A. G., Kaspi, V. M., \& Crawford, F. 2006, \apj, accepted
695:
696: \bibitem[Mattox, Hartman, \& Reimer(2001)]{mhr01} Mattox, J.~R.,
697: Hartman, R.~C., \& Reimer, O.\ 2001, \apjs, 135, 155
698:
699: \bibitem[McLaughlin et al.(2006)]{mll+06} McLaughlin, M.~A.,
700: et al.\ 2006, \nat, 439, 817
701:
702: \bibitem[Morris et al.(2002)]{mhl+02} Morris, D.~J.~et al.\ 2002,
703: \mnras, 335, 275
704:
705: \bibitem[Ransom(2001)]{r01} Ransom, S.~M. 2001, Ph.D. thesis, Harvard
706: Univ.
707:
708: \bibitem[Ransom, Eikenberry, \& Middleditch(2002)]{rem02} Ransom,
709: S.~M., Eikenberry, S.~S., \& Middleditch, J.\ 2002, \aj, 124, 1788
710:
711: \bibitem[Ransom et al.(2006)]{rrh+06} Ransom, S., Roberts, M.,
712: Hessels, J., Livingstone, M., Crawford, F., Tam, C., \& Kaspi,
713: V. 2006, \apj, in preparation
714:
715: \bibitem[Ransom et al.(2003)]{rlh+03} Ransom, S. M., Livingstone,
716: M. A., Hessels, J. W. T., Roberts, M. S. E., Tam, C., Kaspi, V. M., \&
717: Crawford, F. 2003, American Astronomical Society, HEAD Meeting, 35,
718: \#20.10
719:
720: \bibitem[Roberts, Romani, \& Kawai(2001)]{rrk01} Roberts, M.~S.~E.,
721: Romani, R.~W., \& Kawai, N.\ 2001, \apjs, 133, 451
722:
723: \bibitem[Roberts et al.(2002)]{rhr+02} Roberts, M.~S.~E., Hessels,
724: J.~W.~T., Ransom, S.~M., Kaspi, V.~M., Freire, P.~C.~C., Crawford, F.,
725: \& Lorimer, D.~R.\ 2002, \apjl, 577, L19
726:
727: \bibitem[Roberts et al.(2003)]{rrh+03} Roberts, M. S. E., Ransom,
728: S. M., Hessels, J. W. T., Livingstone, M. A., Tam, C., Kaspi, V. M.,
729: \& Crawford, F. 2003, IAU Symposium 218: Young Neutron Stars and Their
730: Environments, Sydney, Australia, 156
731:
732: \bibitem[Romani(2001)]{rom01} Romani, R.~W.\ 2001, ASSL Vol.~267: The
733: Nature of Unidentified Galactic High-Energy Gamma-Ray Sources, 153
734:
735: \bibitem[Romero, Benaglia, \& Torres(1999)]{rbt99} Romero, G.~E.,
736: Benaglia, P., \& Torres, D.~F.\ 1999, \aap, 348, 868
737:
738: \bibitem[Sowards-Emmerd, Romani, \& Michelson(2003)]{srm03}
739: Sowards-Emmerd, D., Romani, R.~W., \& Michelson, P.~F.\ 2003, \apj,
740: 590, 109
741:
742: \bibitem[Sowards-Emmerd et al.(2004)]{srm+04} Sowards-Emmerd, D.,
743: Romani, R.~W., Michelson, P.~F., \& Ulvestad, J.~S.\ 2004, \apj, 609,
744: 564
745:
746: \bibitem[Staveley-Smith et al.(1996)]{swb+96} Staveley-Smith, L.~et
747: al.\ 1996, Publications of the Astronomical Society of Australia, 13,
748: 243
749:
750: \bibitem[Story, Gonthier, \& Harding(2005)]{sgh05} Story, S.~A.,
751: Gonthier, P.~L., \& Harding, A.~K.\ 2005, American Astronomical
752: Society Meeting Abstracts, 207, \#183.10
753:
754: \bibitem[Thompson et al.(1999)]{tbb+99} Thompson, D.~J., et al.\ 1999,
755: \apj, 516, 297
756:
757: \bibitem[Yadigaroglu \& Romani(1995)]{yr95} Yadigaroglu, I.-A., \&
758: Romani, R.~W.\ 1995, \apj, 449, 211
759:
760: \bibitem[Yadigaroglu \& Romani(1997)]{yr97} Yadigaroglu, I.-A., \&
761: Romani, R.~W.\ 1997, \apj, 476, 347
762:
763: \end{thebibliography}
764:
765: \clearpage
766:
767: \begin{figure}
768: \centerline{\psfig{figure=f1.ps,angle=90,width=7in}}
769: \caption{Aitoff plot in Galactic coordinates of the locations of the
770: 56 unidentified $EGRET$ $\gamma$-ray error boxes surveyed (open
771: circles) and the known pulsars listed in the public pulsar catalog
772: (solid dots) \citep{mht+05}. The dashed lines correspond to Galactic
773: latitudes $\pm 5^{\circ}$, the latitude limits of the Parkes Multibeam
774: Survey \citep{mlc+01}, which had a comparable sensitivity to the
775: survey described here. The centers of the surveyed $EGRET$ targets
776: lie outside this region.}
777: \label{fig-1}
778: \end{figure}
779:
780: \clearpage
781:
782: \begin{figure}
783: \centerline{\psfig{figure=f2.ps,width=6in}}
784: \caption{Target $EGRET$ source 3EG J1627$-$2419, showing the
785: $\gamma$-ray error box (contour lines), the multibeam survey coverage
786: in our search for radio pulsations (circles), X-ray emission from the
787: $ROSAT$ All-Sky Survey (pixelated squares), and 1.4 GHz emission from
788: the NRAO VLA Sky Survey (grayscale) \citep{ccg+98}. The radio and
789: X-ray images were obtained from NASA's {\it SkyView} facility
790: (http://skyview.gsfc.nasa.gov). The contours represent 68\%, 95\%, and
791: 99\% uncertainties in the $\gamma$-ray source position, and the
792: circles indicate the Parkes half-power beam size. Four tiled multibeam
793: pointings are shown (labeled a,b,c,d) with 13 beams each.}
794: \label{fig-2}
795: \end{figure}
796:
797: \clearpage
798:
799: \begin{figure}
800: \centerline{\psfig{figure=f3.ps,width=6in}}
801: \caption{Minimum detectable 1400-MHz flux density (in the absence of
802: RFI) as a function of pulsar period for our survey of $EGRET$
803: targets. A range of DMs was assumed in the calculation, with the
804: sensitivity curve for each DM labeled (in units of pc cm$^{-3}$). An
805: intrinsic duty cycle of 5\% for the pulsed emission was assumed in the
806: sensitivity calculation as was a sky temperature of 5~K at 1400 MHz;
807: this is the maximum sky temperature for any of our sources
808: \citep{hss+82}. In the calculation, we used the gain of the center
809: beam of the multibeam receiver, which is the most sensitive of the 13
810: beams. Averaging over the gains of the 13 beams of the receiver
811: slightly increases the baseline limit to $\sim 0.2$~mJy. Assuming a
812: duty cycle smaller than 5\% lowers it. The inclusion of higher-order
813: harmonics in the search is the cause of the sudden jumps in the
814: sensitivity curves at small periods. The details of the observing
815: system parameters and the sensitivity calculation, which is for a
816: blind FFT search, are outlined in \citet{c00} and \citet{mlc+01}. For
817: the second processing run using the resampled data, the baseline limit
818: of $\sim 0.2$~mJy remains, but the sensitivity to pulsars with periods
819: below about 20 ms is sharply degraded for all DMs (see Section
820: 2). Note that a significant red noise component in the FFT from RFI
821: begins to degrade the sensitivity for periods $\ga 200$~ms and is not
822: included in the model of the sensitivity.}
823: \label{fig-3}
824: \end{figure}
825:
826: \clearpage
827:
828: \begin{figure}
829: \centerline{\psfig{figure=f4.ps,width=6in,angle=270}}
830: \caption{Integrated 20-cm profiles for PSRs J1632$-$1013,
831: J1725$-$0732, and J1800$-$0125, the three isolated pulsars discovered
832: in the survey. Each profile is the sum of many timing observations and
833: has a total of 256 bins. One full period is shown in each case. Timing
834: parameters for these pulsars, including flux densities and pulse
835: widths, are presented in Table \ref{tbl-3}.}
836: \label{fig-4}
837: \end{figure}
838:
839:
840: \clearpage
841:
842: \input{tab1} %TABLE 1: EGRET sources info
843: \input{tab2} %TABLE 2: All detected pulsars
844: \input{tab3} %TABLE 3: Timing solutions for three isolated psrs
845:
846:
847: \end{document}
848: