1: %\documentclass[12pt,preprint]{aastex}

2: \documentclass[manuscript]{aastex}

3: \usepackage{emulateapj5}

4: %\documentclass[preprint2]{aastex}

5: 

6: \shorttitle{DA\,495 -- an aging PWN}

7: \shortauthors{Kothes et al.}

8: 

9: \begin{document}

10: 

11: \title{DA\,495 -- an aging pulsar wind nebula}

12: 

13: \author{R. Kothes\altaffilmark{1,2}, T.L. Landecker\altaffilmark{1},

14: W. Reich\altaffilmark{3}, S. Safi-Harb\altaffilmark{4}, 

15: Z. Arzoumanian\altaffilmark{5}}

16: \altaffiltext{1}{National Research Council of Canada,

17:               Herzberg Institute of Astrophysics,

18:               Dominion Radio Astrophysical Observatory,

19:               P.O. Box 248, Penticton, British Columbia, V2A 6J9, Canada}

20: 

21: \altaffiltext{2}{Department of Physics and Astronomy, University of Calgary,

22:              2500 University Drive N.W., Calgary, AB, Canada}

23: 

24: \altaffiltext{3}{Max-Planck-Institut f\"ur Radioastronomie, Auf dem H\"ugel

25:              69, 53121 Bonn, Germany}

26: 

27: \altaffiltext{4}{Canada Research Chair, Department of Physics and Astronomy, 

28: University of Manitoba, Winnipeg, MB, R3T 2N2, Canada}

29: 

30: \altaffiltext{5}{CRESST and X-ray Astrophysics Laboratory, NASA-GSFC, 

31: Greenbelt MD 20771, U.S.A.}

32: 

33: \email{roland.kothes@nrc-cnrc.gc.ca, tom.landecker@nrc-cnrc.gc.ca,

34: wreich@mpifr-bonn.mpg.de, safiharb@cc.umanitoba.ca, 

35: zaven@milkyway.gsfc.nasa.gov}

36: 

37: \begin{abstract}

38: We present a radio continuum study of the 

39: pulsar wind nebula (PWN) DA\,495 (G65.7+1.2), including images of total 

40: intensity and linear

41: polarization from 408 to 10550\,MHz based on the Canadian Galactic

42: Plane Survey and observations with the Effelsberg 100-m Radio

43: Telescope. Removal of flux density contributions from a superimposed

44: \ion{H}{2} region and from compact extragalactic sources reveals a

45: break in the spectrum of DA\,495 at 1.3\,GHz, with a spectral index

46: ${\alpha}={-0.45~\pm~0.20}$ below the break and

47: ${\alpha}={-0.87~\pm~0.10}$ above it (${S}_\nu \propto{\nu^{\alpha}}$).  The

48: spectral break is more than three times lower in frequency than the

49: lowest break detected in any other PWN. The break in the

50: spectrum is likely the result of synchrotron cooling, and DA\,495, at

51: an age of $\sim$20,000 yr, may have evolved from an object similar to

52: the Vela~X nebula, with a similarly energetic pulsar. We find a 

53: magnetic field of $\sim$1.3~mG inside the nebula. After correcting for the 

54: resulting high internal rotation measure, the magnetic field structure is 

55: quite

56: simple, resembling the inner part of a dipole field projected onto the plane 

57: of the sky,

58: although a toroidal component is likely also present. The dipole field

59: axis, which should be parallel to the spin axis of the putative pulsar,

60: lies at an angle of ${\sim}50\degr$ east of the North Celestial Pole

61: and is pointing away from us towards the south-west. The upper limit 

62: for the radio surface brightness of any shell-type supernova remnant

63: emission around DA\,495 is 

64: $\Sigma_{1\,GHz} \sim 5.4 \times 10^{-23}$~Watt~m$^{-2}$~Hz$^{-1}$~sr$^{-1}$

65: (assuming a radio spectral index of $\alpha = -0.5$),

66: lower than the faintest shell-type remnant known to date.

67: \end{abstract}

68: 

69: \keywords{ISM: individual (G65.7+1.2), magnetic fields, polarization, 

70: supernova remnants}

71: 

72: \section{Introduction}

73: 

74: The Crab Nebula, the remnant of the supernova explosion of 1054, is an

75: astrophysical laboratory of immense value where we can study the

76: development of a young pulsar and the synchrotron nebula generated by

77: the particles that it injects. The Crab Nebula has become the

78: archetype of a class of objects earlier known as filled-center,

79: plerionic, or Crab-like supernova remnants (SNRs), but now called

80: pulsar wind nebulae (PWNe) in recognition of their energy source. PWNe

81: are a minority in SNR catalogs, less than 10\% in number, but if we

82: include the composite SNRs, synchrotron nebulae centrally placed

83: within the more usual SNR shell, their ranks swell to 15\% of the

84: Galactic SNR catalog (Green 2004).

85: 

86: PWNe are characterized by a flat radio spectral index, $\alpha$, in

87: the range ${0} > {\alpha} > {-0.3}$ (where the flux density $S$ at

88: frequency $\nu$ varies as ${S} \propto {{\nu}^{\alpha}}$) and a

89: central concentration of emission gradually declining to the outer

90: edge. Their appearance reflects the central energy source and their

91: radio spectrum reflects the energy spectrum of the injected

92: particles. The more common shell SNRs exhibit steep outer edges and

93: steep radio spectra (generally ${\alpha} < {-0.3}$).

94: 

95: We have historical records of fewer than ten supernova events

96: (Stephenson \& Green 2002).  By chance two of those have left

97: PWNe, the Crab Nebula and 3C 58, with no detectable trace of

98: accompanying shells, and one a composite remnant, G11.2$-$0.3. Thus we

99: have three examples in which we can closely follow the evolution of

100: young PWNe, but the relatively small number of known

101: PWNe has provided little observational basis for our

102: understanding of the evolution of these objects through later

103: stages. This paper presents and analyzes new observations of DA\,495

104: (G65.7+1.2), which seems likely to be a PWN of advanced age.

105: 

106: DA\,495, while an outlier among PWNe, is without doubt a member

107: of that class. Its radio appearance, first seen in the high-resolution

108: images of \citet{land83}, has the signature of a PWN, a central

109: concentration of emission smoothly declining with increasing radius,

110: with no evidence of a surrounding shell (although there is a

111: depression in intensity near its center). The only indication that it might

112: not be a PWN is its measured radio spectral index, ${\alpha}

113: \approx {-0.5}$ \citep{koth06b}, much steeper than is usual for this

114: class of objects, and more closely resembling that of a shell-type

115: SNR. Nevertheless, shell remnants have steep outer edges where kinetic

116: energy is being converted to radio emission near the shock front,

117: while in DA\,495 the emissivity distribution is centrally concentrated

118: \citep[Fig.~3]{land83} indicating that the energy source is

119: central. Nevertheless, on the basis of the central depression in

120: intensity, \citet{velu89} argue that DA\,495 might be a shell remnant

121: with an exceptionally thick shell formed by a reverse shock

122: interacting with an unusual amount of slowly moving ejecta.

123: 

124: Recent analysis of X-ray data for DA\,495, obtained from the ROSAT and

125: ASCA archives, dramatically confirms the PWN interpretation of

126: this object by revealing a compact central object surrounded by an

127: extended non-thermal X-ray region \citep{arzo04}. Analysis of Chandra

128: data \citep{arzo08} shows a non-thermal X-ray nebula of extent

129: $\sim$30\arcsec\ surrounding the central compact object. The compact

130: source is presumably the pulsar (required to power the

131: synchrotron nebula), and we shall refer to it as such (even though

132: pulsed emission has not been detected). In this paper we present new

133: radio images of DA\,495 which include linear polarization

134: information. Analysis of high-resolution radio images allows us to

135: remove superimposed sources. After this the integrated spectrum of the

136: source exhibits a spectral break characteristic of a synchrotron

137: nebula, and the observed parameters allow an estimate of the strength

138: of the magnetic field within the nebula. Polarization images reveal

139: the structure of the field within this PWN.

140: 

141: \section{Observations and Data Analysis}

142: 

143: The DRAO observations at 408 and 1420 MHz are part of the Canadian

144: Galactic Plane Survey \citep[CGPS,][]{tayl03}. Data from single-antenna

145: telescopes were incorporated with the Synthesis

146: Telescope data to give complete coverage of all structures down to the

147: resolution limit \citep[details of the process can be found in][]{tayl03}. 

148: The one exception is the polarization data at 1420~MHz,

149: for which no single-antenna data were available. However, the

150: telescope is sensitive to all structures from the resolution limit

151: (${\sim}1'$) up to $45'$. The absence of single-antenna data is not a

152: concern for observations of DA\,495, whose full extent is $\le 30'$.

153: The CGPS polarization data were obtained in four bands (1407.0, 1413.0, 1427.6,

154: and 1434.5~MHz) each of width 7.5~MHz, allowing derivation of 

155: rotation measures (RMs) between those frequencies.

156: 

157: The Effelsberg observations were carried out well before this

158: investigation began, and were retrieved from archived

159: data. Observations at 4850 and 10550 MHz were made at different times,

160: but in each case in excellent weather.  The receiver systems,

161: installed at the secondary focus of the telescope, received both hands

162: of circular polarization, with two feeds at 4850 MHz and four feeds at

163: 10550 MHz. Standard data-reduction software based on the NOD2 package

164: \citep{hasl82} was employed throughout.  Data reduction relied on

165: ``software beam switching'' to reject atmospheric emission above the

166: telescope \citep{mors86}. Here every feed is equipped with its

167: own amplifier to allow parallel recording of data. For each pair of

168: feeds a difference map is calculated in which disturbing atmospheric

169: effects are largely subtracted.  These dual-beam maps are restored to

170: the equivalent single-beam map with the algorithm described by

171: \citet{emer79}. To reduce scanning effects, multiple maps -- coverages

172: -- of the field were observed at different parallactic angles. These

173: scanning effects arise from instrumental offsets and drifts that

174: produce different baselevels for each individual scan, a typical

175: result from raster scanning. Individual coverages were combined using

176: the ``plait'' algorithm \citep{emer88}, by destriping the maps in the

177: Fourier domain. We also used 2695 MHz data from the Effelsberg

178: Galactic Plane Survey \citep{reic90}, whose angular resolution is

179: 4\farcm3.

180: 

181: Characteristics of the new observations of DA\,495, made with the DRAO Synthesis

182: Telescope and the Effelsberg 100-m radio telescope, are presented in

183: Table~\ref{obspara}. 

184: 

185: \section{Results}

186: 

187: \subsection{DA\,495 in total intensity}

188: 

189: Figure~\ref{tp} presents total intensity images of DA\,495 at 408, 1420,

190: 4850, and 10550~MHz as well as data at 2695~MHz from the Effelsberg

191: 11\,cm survey \citep{reic90} and a 60$\mu$m image from the IRAS Galaxy

192: Atlas \citep[IGA,][]{cao97}, which is part of the CGPS database. The 1420~MHz

193: image is completely consistent with the earlier DRAO image published

194: by \citet{land83}, although the new image has lower noise by a factor

195: of $\sim3$. All the radio images show a diffuse source of full extent

196: about $25\arcmin$. The high-resolution images at 1420, 4850, and

197: 10550~MHz show the central depression reported by \citet{land83}. Its

198: position is RA(2000) = $19^h 52^m 10.4^s$, DEC(2000) = $+29\degr

199: 26\farcm2$, and it is now apparent that the pulsar (1WGA~J1952.2+2925)

200: is not at the center of the depression (as postulated by

201: \citet{land83}), but lies almost $2\arcmin$ east of it at RA(2000) =

202: $19^h 52^m 17.0^s$, DEC(2000) = $+29\degr 25\arcmin 53\arcsec$

203: \citep{arzo08}. Clearly the explanation of the hole in the nebula

204: cannot be quite as simple as a decline in emission from an aging

205: pulsar. At larger radii, the emission drops off smoothly towards the

206: outer edge without any sign of an outer shell at any frequency.

207: 

208: At 408~MHz DA\,495 is quite circular in appearance. However, an

209: extension of DA\,495 to the south-west is evident at 1420~MHz and

210: becomes increasingly prominent at higher frequencies. The IGA data

211: shown in Figure~\ref{tp} reveal an infrared source coincident with

212: this part of the radio source. The simplest interpretation is that an

213: \ion{H}{2} region is superimposed on the PWN, and in the analysis

214: we assume this to be the case.

215: 

216: \subsection{The integrated spectrum of DA\,495}

217: 

218: We derived the integrated emission of DA\,495 from each of the radio

219: images in Figure~\ref{tp}, and identified a number of flux densities

220: from the literature. All flux densities which we believe to be

221: reliable are listed in Table~\ref{fluxes}. Flux density measurements

222: at low radio frequencies are not easy, especially for sources of low

223: surface-brightness; those that are available for DA\,495 need to be

224: discussed individually.

225: 

226: The area around DA\,495 is covered by the 34.5 MHz survey of

227: \citet{dwar90}, which has an angular resolution of $26\arcmin \times

228: 44\arcmin$ at this declination. An apparently unresolved source of

229: flux density 35 Jy is detected, offset from the position of DA\,495 by

230: 12\farcm5 in right ascension (to the west) and $5'$ in declination (to

231: the north). The offset may arise from compact sources included in the broad

232: beam (see below). The nearby SNR G65.1+0.6 \citep{land90} can be

233: identified in the 34.5 MHz image. Its expected flux density at 34.5

234: MHz, extrapolated from higher frequency measurements, is of the order

235: of 50 Jy, but its surface brightness is lower than that of DA\,495

236: because of its larger size ($1.5\degr \times 0.8\degr$). We consider

237: the 34.5 MHz data to be generally reliable in this area, and adopt an

238: upper limit of 35 Jy for possible emission from DA\,495 at 34.5

239: MHz. This is $\sim$7 times the rms sensitivity of 5 Jy quoted for the

240: survey.

241: 

242: The best low-frequency data come from the new VLA Low-Frequency

243: Sky Survey \citep[VLSS,][]{cohe07} at 74~MHz. We obtained an image of the area

244: surrounding DA\,495 from the 

245: VLSS data base\footnote{web address: http://lwa.nrl.navy.mil/VLSS/}.  The angular resolution is

246: $80\arcsec$, by far the best of any of our data points below 327~MHz,

247: and fully adequate to isolate the small-diameter sources in the

248: field. The survey is sensitive to structures up to $\sim$1$^{\circ}$

249: in size, larger than DA\,495. The survey noise is 100~mJy/beam, but

250: in the area around DA\,495 the noise is higher, about 150~mJy/beam,

251: because of the proximity to Cygnus~A. DA\,495 is not visible in the

252: 74~MHz image, not even after smoothing to a resolution of

253: $3\arcmin$. To obtain an upper limit for the flux density at this

254: frequency, we convolved our 1420~MHz image to this resolution, and

255: added a scaled version to the 74~MHz image.  The scaled image was just

256: detectable when its integrated flux density was 14~Jy, and we adopted

257: this value as the upper limit of the 74~MHz flux density of DA\,495.

258: We tested this upper limit by adding our extrapolated 1420~MHz image

259: to areas around the position of DA\,495 as well. This convinced us

260: that we were not artificially lowering the upper limit by adding the 1420~MHz 

261: image to 74~MHz emission from DA\,495 that is just slightly below the 

262: detection threshhold.

263: 

264: Flux densities for DA\,495 at 83 and 111 MHz appear in the list

265: published by \citet{kova94}. These observations were made with fan

266: beams of extent a few degrees in the N--S direction. They should perhaps

267: also be regarded as upper limits because of the possibility of

268: confusion in the wide fan beams. We note that \citet{kova94}, while

269: listing flux densities for DA\,495, give only an

270: upper limit for the flux density of G65.1+0.6.

271: 

272: Taking the measured flux densities in Table~\ref{fluxes} at face value

273: we derive the integrated spectrum shown in the left panel of

274: Figure~\ref{intspec}, fitted reasonably well over the entire frequency

275: range by a spectral index of ${\alpha}={-0.59~\pm~0.1}$. However, the

276: compact sources in the field around DA\,495 and the superimposed

277: \ion{H}{2} region contaminate this spectrum. The compact sources are

278: probably extragalactic and should mostly have spectral indices

279: steeper than ${\alpha}={-0.59}$ \citep[e.g.][]{cond84}; they will 

280: consequently contribute

281: strongly to the low-frequency flux densities. The \ion{H}{2} region

282: will significantly contaminate the high-frequency flux densities

283: because its spectral index is flatter than that of the non-thermal

284: emission.  We now proceed to derive an integrated spectrum for DA\,495

285: corrected for these extraneous contributions.

286: 

287: We have obtained the flux densities of the three most prominent

288: compact sources in the field shown in Figure~\ref{tp} at frequencies

289: between 74~MHz and 10550~MHz, using our new data and values from

290: the literature. Flux densities of these sources are listed in

291: Table~\ref{sourcefd} and their positions and fitted spectra in

292: Table~\ref{sourceps}. The 74~MHz flux density for Source 1 is

293: taken from the VLSS point source catalog; the flux densities for

294: Sources 2 and 3 were derived by integration from the 74~MHz image.

295: Flux densities for DA\,495 after correction for compact sources are

296: shown in the fourth column of Table~\ref{fluxes}.

297: 

298: We expect the \ion{H}{2} region to affect flux densities at high

299: frequencies, because its spectral index should be

300: ${\alpha}\approx{-0.1}$ for optically thin thermal emission.  The

301: contribution from the \ion{H}{2} region was easily measured using our

302: images at 1420 MHz and 10550 MHz. After convolving to an angular

303: resolution of 2\farcm45, we plotted brightness temperature at 1420 MHz

304: against that at 10550 MHz (using images from which compact sources had

305: been subtracted) selecting only data from the north-east section of

306: the PWN which is unlikely to be contaminated by \ion{H}{2} region

307: emission. The comparison implied a spectral index of $-$0.85 for this

308: part of the PWN. Assuming that this is the spectral index of the

309: non-thermal emission between 1420 and 10550~MHz, and that the

310: \ion{H}{2} region has a spectral index of $-$0.1, we made a

311: point-by-point separation of emission into thermal and non-thermal

312: contributions. The result is shown in Figure~\ref{nthth}. The

313: non-thermal emission has strong circular symmetry around the central

314: depression, and closely resembles our 408~MHz image, where we expect

315: the contribution of the thermal component to be negligible. The

316: thermal component is confined to the south-west, and coincides closely

317: with the infrared source, as expected. Because of the similarity of the 

318: non-thermal map in Fig.~\ref{nthth} and the 408~MHz image in 

319: Fig.~\ref{tp}, the coincidence of the location of the thermal source 

320: in Fig.~\ref{nthth} and the infrared source in Fig.~\ref{tp}, and the 

321: successful separation, we can be confident that our assumption of a

322: constant spectral index for both the thermal and the non-thermal emission 

323: components is appropriate.

324: 

325: The contribution of the \ion{H}{2} region

326: to the flux density of DA\,495 is $440\pm 100$~mJy at 1420

327: MHz. We used this value with a spectral index of $-$0.1 to calculate

328: the contribution of emission from the \ion{H}{2} region at frequencies

329: of 1420 MHz and higher, and subtracted the calculated value to obtain

330: an integrated flux density for the emission from the PWN alone.  At

331: lower frequencies we did not take the emission from the \ion{H}{2}

332: region into account, since the contribution of the thermal emission

333: should become negligible and we cannot determine the frequency at

334: which the thermal emission becomes optically thick.  Flux densities

335: for DA\,495 after correction for the contribution of the \ion{H}{2}

336: region are shown in the fifth column of Table~\ref{fluxes}.

337: 

338: The right panel of Figure~\ref{intspec} shows the integrated spectrum

339: of DA\,495 after correction for both compact sources and the superimposed

340: \ion{H}{2} region. The corrected data strongly suggest a break in the

341: spectrum of DA\,495 at about 1 GHz.  Fitting spectra to frequency points

342: below 1 GHz yields a spectral index of

343: ${\alpha}={-0.45~\pm~0.10}$. Fitting to flux densities above 1 GHz

344: yields ${\alpha}={-0.87~\pm~0.10}$. We consider the spectrum above

345: 1~GHz to be well determined with a two-parameter fit to 6 well defined

346: flux density measurements. Below the break we have seven measured flux

347: densities two of which are upper limits. There are possible errors in the

348: spectral index below the break because of the poor angular resolution

349: in that frequency range. The quoted error ($\pm~0.10$) is the formal 

350: $1 \sigma$-error

351: in the fit, but there are probably larger systematic errors which we have

352: difficulty estimating; somewhat arbitrarily, we adopt a probable error

353: of 0.20.

354: 

355: These reservations notwithstanding, we believe that this analysis has

356: demonstrated convincingly that DA\,495 has a spectral break at

357: $1.3^{+0.3}_{-0.2}$~GHz of ${\Delta\alpha} = 0.42 \pm 0.22$. The spectral 

358: break may be

359: larger than 0.4, since the low-frequency flux densities are probably

360: upper limits, not definite measurements. Further, we have subtracted

361: the contributions of only three compact sources. Some contribution

362: from such sources will remain in our data, so that the plotted values

363: are again to some extent upper limits. The location of the 

364: break frequency, however, should be well defined. Its determination 

365: is dominated by the higher frequency fluxes of the low frequency part 

366: of the spectrum. These values are well constrained, while some of the low

367: frequency fluxes are upper limits. These affect only the slope of

368: the spectrum below the break.

369: 

370: \subsection{Polarization images and rotation measure}

371: 

372: Figure~\ref{pi} shows our polarization images for the area around

373: DA\,495, together with data from the Effelsberg 2695 MHz survey

374: \citep{dunc99}. At all frequencies the polarized emission is confined

375: to a roughly circular region about $15\arcmin$ in diameter. No

376: polarized emission is seen corresponding to the south-west extension,

377: supporting the identification of this as thermal emission. No

378: depolarization effects can be associated with this thermal emission

379: either, and the \ion{H}{2} region which generates it is likely more 

380: distant than DA\,495. 

381: 

382: Integrated over the entire source, the fractional polarization at the

383: higher frequencies is about 25\% (Table~\ref{intpol}). Even at

384: 1420~MHz the integrated polarization has dropped only to 12\%, still

385: half the value at the higher frequencies. At 1420~MHz a RM variation

386: of only 36~rad/m$^2$ is required to rotate the polarization angle by

387: $90\degr$, and superposition of two signals whose polarization angles

388: differ by $90\degr$ produces a net signal with no apparent polarization. 

389: (At our other frequencies RM variations of about

390: 130~rad/m$^2$ at 2695~MHz, 400~rad/m$^2$ at 4850~MHz, and

391: 1900~rad/m$^2$ at 10550~MHz are required to generate a Faraday

392: rotation of $90\degr$.) This implies that there cannot be any rapid

393: changes in rotation measure on scales smaller than the 1420~MHz beam,

394: because otherwise the 1420~MHz emission would be beam depolarized.

395: 

396: At the two highest frequencies the polarized intensity has a

397: remarkable bipolar distribution, seen particularly well in the

398: 4850~MHz image (Fig.~\ref{pi}).  Inspection shows that the bipolar

399: structure is centered on the pulsar, immediately suggesting that the

400: polarization structure is determined by the geometry of the magnetic

401: field, itself tied to the pulsar. We see two prominent lobes of

402: polarized emission, north-east and south-west of the pulsar. At the

403: three high frequencies, from 10550 to 2695~MHz, the location of these

404: lobes seems to rotate systematically clockwise, in the sense opposite

405: to the rotation of polarization angle that arises from the positive RM

406: (see below). Symmetrically on either side of the central ridge at 4850

407: and 10550~MHz there are small regions of low polarized intensity.

408: The total intensity remains high at these locations, and we

409: interpret the low fractional polarization as an indication that

410: polarized emission generated at deeper layers within the source is

411: superimposed on emission generated in closer layers with a

412: significantly different polarization angle; vector averaging reduces

413: the net polarization. This is largely an emission effect rather than a

414: Faraday rotation effect because it is seen at high radio frequencies

415: where Faraday rotation is low.

416: 

417: With linear polarization observations at four frequencies we should be

418: able to investigate the structure of RM across the source.  To this

419: end, all observations were convolved to the lowest resolution,

420: $4\farcm3$, that of the 2695~MHz measurement. However, it proved

421: impossible to calculate a RM map: at many points the calculation

422: failed to fit a slope to the four measured polarization angles within

423: acceptable error limits. The probable cause is ``Faraday thickness''

424: at the lower frequencies, an effect analogous to optical thickness.

425: Depth depolarization - also known as differential Faraday rotation

426: \citep{burn66,soko98} -

427: causes the emission from the

428: deeper levels of the source to be totally depolarized at low

429: frequencies, so that the polarized emission we observe comes only from

430: the nearer layers of the source.  Faraday thickness seems to affect

431: the polarization images at 2695~MHz and 1420~MHz. 

432: 

433: Confining our attention to 4850 and 10550~MHz, we calculated a map of

434: RM at a resolution of 4\farcm3, displayed in Figure~\ref{rmpib}.  RM

435: varies from 200 to 250~rad/m$^2$ on the two major emission lobes,

436: north-east and south-west of the pulsar, to as much as 400~rad/m$^2$

437: in the structures to the east and west.  To identify the Faraday thin

438: and Faraday thick areas we calculated the expected polarization angles

439: at 2695~MHz by extrapolating from the observed angles at 4850 and

440: 10550~MHz and compared the results with values actually observed at

441: the lower frequency. The polarization angle in the south-western lobe

442: is as predicted, while the angle in the north-eastern lobe is about

443: $50^{\circ}$ from the predicted value. Observed polarization angle in

444: two patches lying approximately $4'$ east and west of the pulsar

445: (these patches are prominent in the 10550~MHz image of

446: Figure~\ref{pi}) is about $90^{\circ}$ from the predicted

447: value. Faraday thickness is clearly significant in this object.

448: We also calculated a RM map between 4850 and 10550~MHz at the

449: higher resolution of $2\farcm45$ to examine RM fluctuations on smaller

450: scales. The RM maps at the lower and higher resolution look virtually

451: identical, confirming our impression that there are no strong RM

452: fluctuations on small scales. We also calculated intrinsic

453: polarization angles at the highest resolution and derived a map of the

454: magnetic field (projected to the plane of the sky). This is also shown

455: in Fig.~\ref{rmpib}.

456: 

457: Even though DA\,495 is Faraday thick at frequencies of 2695~MHz and

458: below, we calculated a RM map from the four bands of the CGPS 1420~MHz

459: polarization data (Fig.~\ref{rmpib}). Surprisingly, these values of

460: RM, which probe only the nearer layers of the source, are of the same

461: order ($\sim 200$~rad/m$^2$) as the values derived between 4850 and

462: 10550~MHz where the entire line of sight through the PWN is

463: contributing. If RM was positive right through the source the values

464: at high frequencies should be higher than the values near

465: 1420~MHz. This indicates that there must be regions of negative RM in

466: the deeper layers whose effects are canceled by positive rotation

467: nearer the front surface. A possible field configuration that

468: could produce this is a dipole field with the dipole axis pointing

469: away from us. Almost the entire front part of the nebula would contain

470: magnetic field lines that point towards us, explaining why we observe

471: only positive rotation measure at 1420~Mhz.  However, deeper inside

472: the nebula we would find magnetic field lines pointing away from us,

473: and the Faraday rotation produced by that part of the field would

474: partially cancel rotation produced in the near part. The exact RM

475: configuration seen would depend on how deep we can look inside the

476: nebula at 1420~MHz or other frequencies.

477: 

478: \section{Discussion}

479: 

480: \subsection{Is DA\,495 a PWN or a Shell SNR?}

481: 

482: In the introduction we reviewed the arguments for the classification

483: of DA\,495 as a PWN. In brief, the emissivity is centrally

484: concentrated and drops smoothly to zero at the outer edge without any

485: sign of a steep shell, indicating a central energy source. This

486: conclusion is corroborated by X-ray observations \citep{arzo04,arzo08} 

487: that show a central compact object surrounded by a

488: non-thermal nebula. \citet{velu89} suggest that DA\,495 could be a 

489: shell remnant with an unusually thick shell. While this is a conceivable 

490: interpretation, there is no evidence of a steep outer edge to the shell

491: and the central compact source discovered in X-ray is not located

492: inside the radio hole, but about 2\,$\arcmin$ away from it suggesting

493: that the radio hole is not in the centre of the nebula.

494: 

495: This paper provides further evidence. The structure in polarized

496: intensity is bipolar with mostly radial B-vectors at the outer edge,

497: while for a shell-type remnant we would expect a shell structure in

498: polarized emission with tangential B-vectors. A young shell-type SNR

499: would have radial B-vectors, but integrated polarization would be very

500: low and a young remnant would have bright X-ray emission coming from

501: the expanding shell. The break in the integrated spectrum of DA\,495

502: also provides strong evidence that the object is a PWN. Every

503: known example of such a nebula exhibits such a break, while

504: virtually all shell remnants have spectra characterized by a single

505: power law. While spectral breaks have been claimed for some shell

506: remnants, these claims are gradually melting away in the face of new

507: data with improved sensitivity to extended emission.  The only SNR

508: where such a break has been convincingly established is S\,147

509: (G180.0$-$1.7) which has a break in its integrated spectrum at

510: $\sim$1.5~GHz \citep{furs86,xiao08}. S\,147 is a very old SNR, probably well

511: into the isothermal stage of its evolution; it bears no resemblance to

512: DA\,495.

513: 

514: The only tenable conclusion is that DA\,495 is a pulsar wind nebula, and we

515: will proceed on that basis. 

516: 

517: \subsection{Distance, Size, and Emission Structure}

518: 

519: A distance to DA\,495 was measured by \citet{koth04} on the basis of

520: absorption by \ion{H}{1} of the polarized emission from the PWN.

521: The distance obtained, $d \approx  1.5$\,kpc, was derived kinematically using

522: a flat rotation model for the Milky Way with a Galactocentric

523: radius of R$_\sun = 8.5$\,kpc for the Sun and its velocity

524: v$_\sun = 220$\,km/s around the Galactic center. The latest measurements 

525: of the Sun's

526: galactocentric distance give R$_\sun \approx 7.6$\,kpc

527: which is used in the new distance determination method of

528: \citet[and reference therein]{fost05}.  This method uses a model for the

529: spatial density and velocity field traced by the distribution of

530: \ion{H}{1} in the disk of the Galaxy to derive a distance--velocity

531: relation that can be used to calculate a distance to any object with a

532: known systemic velocity.  The \ion{H}{1} absorption measurements indicate

533: for DA\,495 a systemic velocity of about $+12$\,km/s. Two distances, about

534: 1 and 5\,kpc, correspond to this velocity, one closer than the tangent

535: point and the other beyond it.  Since \citet{koth04} found no

536: absorption of the PWN at the tangent point and DA\,495 exhibits low

537: foreground absorption in the X-ray observations \citep{arzo04,arzo08}

538: we adopt the closer

539: distance as the more likely result.  We proceed to use a distance

540: of $1.0\pm 0.4$\,kpc to determine physical characteristics of the PWN.

541: The error contains the uncertainty in the fit and a velocity dispersion of

542: 4\,km\,s$^{-1}$.

543: 

544: To determine a reliable value for the radius we fitted a

545: three-dimensional model of emissivity to the total-intensity

546: observations at 1420~MHz. Our method is similar to that used by

547: \citet{land83}, a method developed by \citet{hill67}. While

548: \citet{land83} centered their calculation on the depression near

549: the middle of the nebula, we centered our new calculation on the

550: position of the pulsar, as measured by \citet{arzo08}. In the center 

551: there is an emissivity depression

552: with a radius of about $2\arcmin$, and so a physical radius of

553: 0.6\,pc.  Outside, the emissivity is nearly constant to a

554: radius of $7\farcm5$, where it drops abruptly, almost to zero. This

555: translates to an outer radius of 2.2\,pc. \citet{land83} found that

556: the emissivity increased from the outer edge towards the

557: centre. This difference can be explained by the different location

558: assumed for the centre and the contamination by the \ion{H}{2}

559: region (removed from our calculation but included in that of

560: \citet{land83}).

561: 

562: \citet{land83} assumed that

563: the prominent depression marked the center of the 

564: PWN, but we

565: know now that the pulsar lies at one edge of that depression

566: \citep{arzo04,arzo08}, not at its center. One might suppose that the pulsar

567: has moved since it was formed, but we see from Figs.\,\ref{pi} and

568: \ref{rmpib} that the structure in polarized intensity, and so the

569: magnetic field structure, is centered on the pulsar. Since the radio

570: appearance of the nebula is dominated by particles injected early in

571: the life of the pulsar (see Section 4.3), it appears that the pulsar

572: has {\it{not}} moved significantly, and we must find another

573: explanation.  

574: 

575: In order to achieve higher angular resolution at

576: 1420~MHz, we merged the 1420 MHz map from Fig.\,\ref{tp} with data

577: from the NRAO VLA Sky Survey \citep[NVSS;][]{cond98}. The NVSS has a

578: resolution of about $45\arcsec$.  The resulting map is shown in

579: Fig.\,\ref{tppi}.

580: Inspection of Fig.\,\ref{tppi} shows evidence of a

581: second less obvious depression in total intensity, at

582: RA(2000)\,=\,19$^h$ 52$^m$ 33$^s$, DEC(2000)\,=\,29$^{\circ}$

583: 26$'$. This eastern depression is further from the pulsar than the

584: more obvious, and deeper, hole to the west, and the nebula appears to extend to

585: slightly larger radius beyond the eastern hole. Nevertheless, we

586: regard the eastern depression as a significant feature.  In order to

587: produce these two emissivity depressions, there must be a region around the

588: pulsar which is relatively free of magnetic field and/or particles.

589: This brings to mind the Crab Nebula's ``synchrotron

590: bays''. \citet{fese92} ascribe them to an equatorial torus of material

591: ejected by the progenitor star before it exploded. They suggest that

592: the magnetic field of the nebula is wrapped around this torus

593: preventing relativistic particles from entering it. A saddle-shaped

594: depression in total power emission that connects the two holes in

595: DA\,495 (see Fig.~\ref{tppi}) supports a torus as an explanation for

596: these features. In the Crab Nebula the synchrotron bays are located at

597: the edge of the emission region, but DA\,495 is probably much older

598: and the nebula has had time to flow around the torus. The distance

599: between the two holes is about $5\arcmin$ translating to 1.45~pc at a

600: distance of 1~kpc. This indicates an average expansion velocity of

601: about 70~$\frac{10000\,yr}{t}$~km/s, where $t$ is the age of DA\,495.

602: For an age of 10000~yr this expansion velocity would be about 10~\% of

603: that of the Crab Nebula's synchrotron bays, which are expanding at

604: about 900 and 600 km/s for the western and eastern bay respectively

605: \citep{fese92}. We cannot completely rule out that movement

606: of the pulsar is responsible for the off-center location of the

607: pulsar between the two holes, but a difference in expansion velocity

608: between the two sides, similar to that observed in the Crab Nebula,

609: could explain this asymmetry.

610: 

611: \subsection{The Nature of the Spectral Break}

612: 

613: To investigate the nature of the break in the radio spectrum of

614: DA\,495 we first estimate the nebula's age by comparing

615: its energetics with the properties of other PWNe of known age.

616: We assume for the moment that the pulsar has a symmetrical pulsar wind

617: and that equipartition between particle

618: and magnetic field energies has been established. We follow the analysis

619: made by \citet{koth06a} of the PWN G106.65+2.96, although, unlike that case,

620: we have no information on the characteristic age of the DA\,495 pulsar,

621: which has not been detected. Nevertheless, we can get an estimate of

622: the current energy loss rate of the pulsar from \citet{arzo08}:

623: $\dot{E}_{\rm DA\,495} \approx 10^{35}$~erg/s. 

624: We now confine

625: our attention to pulsars in historical SNRs for which we know the

626: exact age, the Crab Nebula (SN 1054), 3C~58 (SN 1181), and G11.2$-$0.3

627: (SN 386) \citep{gree04}. We also add three pulsar wind nebulae for which

628: \citet{koth06a} have calculated the real ages from the location of

629: the cooling breaks in their

630: radio spectra. We extrapolate their evolution to

631: $\dot{E}_{\rm DA\,495}$ to determine possible properties of the pulsar

632: in DA\,495.  

633: 

634: The evolution of the energy loss rate, $\dot{E}$, of a

635: pulsar is defined by \citep[e.g.][]{paci73}:  

636: \begin{equation} 

637: \dot{E}(t) = \frac{\dot{E_0}}{(1+\frac{t}{\tau_0})^\beta}, 

638: \end{equation} 

639: where

640: $\beta$ is related to the nature of the braking torque {($\beta =

641: \frac{n + 1}{n-1}$, and $n$ is the braking index of the pulsar)}, 

642: $t$ is the real

643: age of the pulsar, $\dot{E_0}$ is its energy loss rate

644: at the time it was born, and $\tau_0$ is its intrinsic characteristic age

645: ($\tau_0 =

646: \frac{P_0}{2\dot{P}_0}$, where $P_0$ is the period of the pulsar at

647: the time it was born). Using the information about the current characteristic 

648: age $\tau$ and

649: $\dot{E}$ for the six pulsars taken from the ATNF Pulsar

650: Catalogue \citep{manc05}\footnote{web address:

651: http://www.atnf.csiro.au/research/pulsar/psrcat/} and their real age

652: we can derive their initial parameters, $\tau_0$ and $\dot{E}_0$. For

653: the Crab Pulsar and the pulsar in Vela~X the braking indices are

654: known, yielding a $\beta$ of 2.3 \citep{lyne88} and 6.0

655: \citep{lyne96}, respectively. For the other four pulsars we assume a

656: pure dipole field, implying a $\beta$ of 2. For Vela~X the age depends

657: on the assumed value of $\beta$, since we calculate it from the

658: spectral break frequency. The resulting intrinsic pulsar properties

659: are listed in Table~\ref{pulsars}. To indicate how much our

660: calculations depend on this factor we used different values of $\beta$

661: as indicated in Table~\ref{pulsars}.  Using equation 1 and taking the

662: energy loss rate of the pulsar in DA\,495 as $\dot{E}_{\rm DA\,495}

663: \approx 10^{35}$~erg/s, we can calculate the current age, $t_{\rm

664: DA\,495}$, of the nebula and the current characteristic age,

665: $\tau_{\rm DA\,495}$, of its pulsar using the initial values for each

666: of the six pulsars.  The results are given in

667: Table~\ref{DA495_pulsar}. For the Crab and the Vela~X pulsars we

668: extrapolated the energy loss rate with two different $\beta$ values as

669: indicated in Table~\ref{DA495_pulsar}. Whatever basis we adopt for the

670: calculation, the age of DA\,495 exceeds 20,000 yr, which makes this

671: nebula an old object. This derived age is valid regardless of the

672: mechanism responsible for the break in the radio spectrum. 

673: All known PWNe detected at radio wavelengths contain energetic pulsars, so 

674: the low energy loss rate of the DA\,495 pulsar is in all likelihood the 

675: result of

676: age, not of intrinsically low energy.

677: 

678: We will now proceed by examining the different possibilites for a spectral

679: break in the synchrotron spectrum of a pulsar wind nebula.

680: Virtually all PWNe exhibit breaks in their spectra, at

681: frequencies ranging from X-rays to a few GHz.  Table~\ref{break_freq}

682: lists the break frequencies that have been identified to date. The

683: spectral break in DA\,495 occurs at 1.3\,GHz, significantly lower than

684: for any other pulsar wind nebula. A spectral break in a PWN can 

685: arise in two

686: ways: it may reflect a break in the injected electron spectrum, and so

687: be a mostly pulsar-dependent effect, or it may 

688: arise from synchrotron

689: cooling, in which case the break frequency depends on the magnetic

690: field strength and the age of the relativistic electrons in the

691: nebula. Of course,

692: both mechanisms should be at work in all PWNe. The suspected nature of

693: each spectral break is indicated in Table~\ref{break_freq}.

694: 

695: We first consider the possibility that the break in the radio

696: spectrum results from a break in the injected electron spectrum.

697: Such breaks in the spectra of PWNe are known (e.g. in the Crab

698: Nebula and 3C~58), but they are imperfectly understood.  Since the

699: electron population in the nebula is dominated by early injection, the

700: break must have been created early in the life of the pulsar.

701: The energy at which the break occurs depends on conditions in the

702: area in which electrons are accelerated. Its location in the electron

703: spectrum remains relatively fixed throughout the life of the nebula. In 

704: contrast, a break due to

705: cooling, starts at high energy and moves with time towards lower

706: energies \citep[e.g.][]{weil80} ultimately perhaps passing by the

707: break produced by the injected spectrum. If the magnetic field decays

708: with time, through expansion or other causes, both kinds of break will

709: migrate towards lower frequencies in the radio spectrum. Their 

710: relative positions will not be affected.

711: 

712: A likely location for an injected break is at the energy where the

713: electron acceleration timescale equals the synchrotron loss time,

714: which is in essence a kind of cooling break. The difference between

715: this situation and the normal synchrotron cooling break is that the

716: number of high energy electrons is never high at any stage. Thus, the

717: synchrotron spectrum should steepen by much more than the 0.5 typical

718: of normal synchrotron cooling, which is actually the case for the

719: intrinsic breaks identified in other PWNe. For DA\,495 we found that

720: the spectrum steepens by about 0.4.

721: 

722: There are two other arguments that make an injected break an

723: unlikely interpretation for the break in the radio spectrum of

724: DA\,495.  First, the X-ray synchrotron spectrum has a spectral index

725: of $-0.63 \pm 0.27$

726: \citep[][calculated from the photon index $\Gamma$:

727: $\alpha = - (\Gamma - 1)$]{arzo08}, flatter than the synchrotron radio spectrum above the

728: break ($-$0.87).  This is impossible if the radio break arises from an

729: injected break. The only mechanism that can affect the synchrotron

730: spectrum beyond an intrinsic break is cooling or some other energy

731: loss effect, which can only steepen the spectrum but cannot flatten

732: it. A new round of particle acceleration late in the life of the PWN,

733: an unlikely event, would affect low-energy and high-energy electrons

734: alike. Second, there is no injected break frequency known that is even

735: remotely as low as the 1.3~GHz we have found in DA\,495.

736: 

737: We now consider the possibility that the break at 1.3~GHz is caused by 

738: synchrotron cooling. We will assume that the magnetic field 

739: throughout DA\,495 is relatively constant or at least that there is

740: a dominating magnetic field creating the cooling break. There are several 

741: lines of

742: evidence supporting this assumption. First, our calculation of the

743: synchrotron emissivity from the total-intensity images shows that

744: emissivity is nearly constant in the majority of the nebula (Section

745: 4.2). Second, the break in the total spectrum is very sharp and there

746: is no significant variation of spectral index across the source

747: (DA\,495 has the same appearance at all radio frequencies, see also

748: section 3.2: the separation into thermal and non-thermal components).  

749: Whatever

750: the cause of the spectral break, it is the magnetic field that

751: translates a break in the electron energy spectrum to a break in the

752: emission spectrum.  If the magnetic field varied significantly, the break

753: would be smeared across a wide range of frequency; the break frequency

754: has a very strong dependence on field strength (see Equation (2)

755: below).

756: 

757: \citet{chev00} found that the cooling frequency, $\nu_c$, corresponding 

758: to $\gamma$ (the Lorentz factor of the electron) at which 

759: the electrons are able to radiate their energy over the age of a 

760: PWN, is defined by

761: \begin{equation}

762:    \nu_c [{\rm GHz}] = 1.187\cdot B^{-3} [{\rm G}] \cdot t^{-2}

763: [{\rm yr}],

764: \end{equation}

765: where $B$ is the magnetic field strength inside the synchrotron nebula. Using

766: the break frequency of $\nu_c

767: \approx 1.3$~GHz and the various interpolated ages, $t_{\rm DA\,495}$, we can

768: calculate $B_{\rm{req}}$, the required magnetic field inside the

769: nebula (Table~\ref{DA495_pulsar}). Integrating equation (1) over the age

770: of the nebula gives the total energy inside it and, assuming

771: equipartition and a circular structure, we can calculate the magnetic

772: field, $B_{\rm{max}}$. We denote this field value $B_{\rm{max}}$

773: because we have not accounted for any energy loss in the course of the

774: evolution of the PWN. Comparing the results based on the six pulsars

775: listed in Table~\ref{DA495_pulsar}, we see that it is only in 

776: the calculation based on the Crab pulsar and

777: the pulsar in Vela~X that the maximum magnetic 

778: field exceeds the magnetic field required for the spectral break.

779: If our assumption is valid, then the pulsar in DA\,495 

780: must have 

781: been very energetic with a high energy loss rate at the time of formation,

782: like the Crab pulsar or the pulsar in Vela~X. Given the age of the

783: Vela supernova remnant and the fact that $B_{\rm{max}}$ is 

784: significantly higher than

785: $B_{\rm{req}}$ for any $\beta$ it is more reasonable to assume that

786: the pulsar in DA\,495 is an older version of the Vela pulsar. If we shrank

787: the Vela~X nebula with its current energy content to the size of

788: DA\,495, its cooling break at 100~GHz would decrease to about 2.5~GHz

789: due to the resulting increase in the magnetic field strength.

790: A few thousand years later the break frequency would

791: reach 1.3~GHz.

792: 

793: Even though we cannot completely rule out the possibility that the

794: break in the radio spectrum of DA\,495 at 1.3~GHz is caused by a break

795: in the injected electron spectrum, all the available evidence

796: indicates that synchrotron cooling is the most likely explanation, and

797: that DA\,495 is probably an aging Vela~X nebula. In this case DA\,495

798: is about 20,000~yr old and contains a strong magnetic field of

799: about 1.3~mG. This can easily account for the high rotation

800: measure we have observed.

801: 

802: \subsection{The magnetic field structure inside DA\,495}

803: 

804: The structure of the PWN in total intensity is very regular,

805: especially once the confusing \ion{H}{2} region is removed (see

806: Figure~\ref{nthth}). However, in polarized intensity the appearance is

807: very different, and at the high frequencies DA\,495 is very

808: structured.  Percentage polarization, even at the high frequencies,

809: varies over a wide range across the source. The challenge faced in

810: this section is to deduce something about the structure of the

811: magnetic field inside DA\,495 and to explain as many of the observed

812: characteristics as possible.  While we cannot uniquely determine

813: the field configuration we show that one relatively simple and

814: plausible model, a dipole field with a superimposed toroidal

815: component, is consistent with the observations.

816: 

817: \subsubsection{The foreground Faraday rotation}

818: 

819: In Figure~\ref{pulsarrm} we have plotted the foreground RM observed to

820: all pulsars within $15\degr$ of DA\,495 as a function of distance 

821: from us in kpc.  The pulsar data were taken from the ATNF Pulsar

822: Catalog \citep{manc05}\footnote{web address:

823: http://www.atnf.csiro.au/research/pulsar/psrcat/}.  All pulsars within

824: about 5~kpc have negative RM values, except for one for which

825: ${\rm{RM}}={+7{\rm{~rad/m}}^2}$; we will disregard this exception

826: because it lies at a Galactic latitude of $+14\degr$. The

827: negative RM values are consistent with the known local magnetic field

828: orientation, away from the Sun. Pulsars more distant than 5~kpc have

829: positive RMs, as expected from the well-known field reversal between

830: the local and Sagittarius spiral arms 

831: \citep[e.g.][]{sima79,lyne89}.  Averaged over the nearby pulsars the

832: line-of-sight component of the magnetic field is around $-$2.5~$\mu$G.

833: Using the models of \citet{tayl93} and \citet{cord02} for the

834: distribution of free electrons in the Galaxy, and the distance to

835: DA\,495 of 1.0~kpc we determine the average foreground dispersion

836: measure to be ${\rm DM} \approx 15$~cm$^{-3}$~pc.  With an average

837: foreground magnetic field of $-$2.5~$\mu$G parallel to the line of

838: sight this produces a foreground RM of $-$30~rad/m$^2$. The positive

839: RM values determined from the emission of DA\,495 must arise from

840: Faraday rotation within the PWN itself or occur in a discrete

841: foreground magneto-ionic object which does not itself produce enough

842: emission to be detected in our observations. Since we find that almost

843: all of the polarized emission from DA\,495 is Faraday thick at

844: 2695~MHz, most of the positive RM must originate in DA\,495, because

845: Faraday thickness can only be produced if the rotation is generated in

846: the same volume elements as the polarized emission itself. A

847: foreground object can merely rotate or beam depolarize the entire

848: emission coming from behind it.

849: 

850: \subsubsection{Magnetic field configurations inside DA\,495}

851: 

852: The magnetic field distribution seen in Figure~\ref{rmpib} is the

853: projection onto the plane of the sky of the three-dimensional field

854: within DA\,495, as deduced from our observations at 4850 and 10550~MHz

855: on the assumption that the source is Faraday thin at these

856: frequencies. The first impression is of a highly regular 

857: elongated magnetic field configuration with field lines following a

858: central bar and spreading out radially to the north and south. This

859: strongly resembles the inner part of a dipole field projected onto the

860: plane of the sky. The field lines must close, presumably at large

861: radius.  However, in Figure~\ref{rmpib} we are seeing only the

862: component perpendicular to the line of sight.  The relatively low

863: percentage polarization even at very high frequencies of about 25\,\%

864: - intrinsically it should be about 70\,\% - cannot be explained by

865: depolarization caused by Faraday rotation, because at high frequencies

866: this effect is negligible: a RM of 400 rad/m$^2$, the highest we have

867: detected, rotates the polarization angle by a mere $20^{\circ}$ at

868: 10550~MHz. Hence, the depolarization must be caused by the

869: superposition of polarized emission generated at different points

870: along the line of sight with intrinsically different polarization

871: angles, averaging to a low net polarized fraction.  In a pure dipole

872: field the component perpendicular to the line of sight, which

873: determines the observed polarization angle, does not change

874: significantly along different lines of sight through the nebula.

875: There must be an additional magnetic field present that also creates

876: synchrotron emission but has an intrinsic magnetic field configuration

877: perpendicular to, or at least very different from, the dipole field. A

878: toroidal field component wrapped around the axis of the dipole would

879: match that description. Toroidal field structures have been observed

880: in Vela~X \citep{dods03} and in the PWN in SNR G106.3$+$2.7

881: \citep{koth06a}, although most other pulsar wind nebulae 

882: show only an elongated

883: radial field, indicating the presence of a dipole.  The patches

884: of polarized emission to the east and west of the pulsar have a

885: magnetic field structure perpendicular to that expected for a pure

886: dipole field and could be a hint of a toroidal magnetic field. The two

887: prominent ``holes'' in polarized intensity would then mark areas where

888: the polarized emission generated by the dipole component and that

889: generated by the toroidal component are equally strong but have

890: orthogonal polarization angles so that the net emission is intense but

891: unpolarized. Even though both magnetic field configurations may be

892: present it is clear that the dipole field is the dominant one for

893: emission, as seen in the magnetic field projected to the plane of the

894: sky, and for Faraday rotation. If most of the RM was produced by the

895: toroidal field one half of the nebula would show a negative RM and the

896: other a positive one.

897: 

898: One possible magnetic field configuration that could explain the

899: radio polarization observations of DA\,495 is shown in

900: Fig.~\ref{magsketch}.  This is no more than a naive sketch, but it

901: fits the available observations remarkably well.  A similar

902: combination of dipole and toroidal field components can explain the

903: polarization appearance of other PWNe; the polarized emission

904: structure, the magnetic field vectors projected to the plane of the

905: sky, the fractional polarization, and the Faraday rotation structure

906: all depend on the viewing angle and which one of the two components is

907: dominant \citep{koth06a}.

908: 

909: \subsubsection{The apparent rotation of the polarization structure}

910: 

911: One curious feature of the radio emission from DA\,495 is the

912: seemingly systematic rotation with frequency of almost the entire

913: polarized emission structure. At the three higher frequencies we can

914: see the two prominent lobes rotating from north-east and south-west at

915: 10550~MHz to west and east at 2695~MHz. Near the center the two

916: polarization holes (or, equivalently, the emission ridge) also rotate

917: clockwise. At 2695~MHz we can no longer see this rotation near the

918: center, probably because of the larger beam. This effect must be the

919: product of depth depolarization in a highly ordered magnetic field

920: structure.

921: 

922: A possible explanation is the following.  At the

923: intrinsic (zero wavelength) location of the polarization holes the

924: magnetic field lines of the toroidal and dipole fields are

925: perpendicular, and the polarization averages out. Close to the center

926: of the PWN the dipole field lines are almost all parallel, while the

927: toroidal field lines change with azimuthal angle within the torus. If

928: the Faraday rotation of emission generated in the toroidal field

929: differs from that generated in the dipole field then, at lower

930: frequencies, this averaging no longer occurs at the ``intrinsic''

931: location but at a location where the dipole and toroidal field lines

932: are at some angle other than 90$^{\circ}$. The location of the holes

933: then moves systematically with frequency, in the sense opposite to

934: Faraday rotation, as observed.

935: 

936: If the apparent rotation of the polarized structure is a Faraday

937: rotation effect, then it should show a $\lambda^2$ dependence, and we

938: can test this. We define $\psi$ as the position angle of a line

939: through the centers of the lobes ($\psi$ increasing from north to

940: east).  For 10550~MHz ${\psi} = {45\degr}$ and for 4850~MHz ${\psi}

941: = {15\degr}$.  Drawing a second line through the polarization holes,

942: and assuming it is perpendicular to the line through the lobes, we

943: find ${\psi}={38\degr}$ at 10550~MHz and ${\psi}={10\degr}$ at

944: 4850~MHz.  Averaging these estimates, we derive a ``structural

945: rotation'' of about $-170$~rad/m$^2$ between 10550 and 4850~MHz and an

946: intrinsic $\psi$ of $50\degr$. As a test, we extrapolated to

947: 2695~MHz, and predict ${\psi}={-70\degr}$, which is very close to the

948: $-90\degr$ we observe.  A second test comes from comparison with our

949: RM map (Figure 5).  Faraday rotation seems to be dominated by the

950: dipole field, and the RM should then show a maximum or minimum at the

951: center of the lobes because in those directions the angle between the

952: line of sight and the field lines is smallest. A line from the RM peak

953: in the north-eastern lobe to the pulsar has an angle of $55\degr$ with

954: north, comparable with the intrinsic $\psi$ of $50\degr$ calculated

955: above.

956: 

957: Since the spin axis of the pulsar is the only axis of symmetry

958: in the rotating pulsar frame, we can assume that it is parallel to the

959: axis of the nebula's dipole field, which then would reflect the

960: rotation-averaged pulsar field. 

961: We conclude that the spin axis of the pulsar, the main axis

962: of the toroidal field, and the axis of the dipole magnetic field

963: component all have an angle $\psi\approx {50\degr}$ projected onto

964: the plane of the sky. Further, the viewing angle, $\theta$, between the 

965: line of

966: sight and the equatorial plane of the dipole (see Fig.~\ref{dipole}), must 

967: be larger than

968: $0\degr$. At ${\theta}={0\degr}$ the toroidal field component would

969: be constant projected onto the plane of the sky and we would not

970: observe the structural rotation effect.

971: 

972: \subsubsection{Orientation of the dipole field}

973: 

974: In this section we develop a simple model of the magnetic field

975: configuration in DA\,495. We test the model by computing RM along 

976: a few selected lines of sight through the nebula, and achieve plausible

977: agreement with measurements. We have not attempted detailed modeling

978: of the field.

979: 

980: At a distance of 1.0~kpc the diameter of DA\,495 is 4.4~pc; we assume

981: this is the maximum path length through the nebula. We deduced an

982: internal field of about 1.3~mG from our comparisons with the Vela~X

983: nebula (Section 4.1 and Table~\ref{DA495_pulsar}) and we assume that

984: the field is constant at this value throughout the nebula. We know the

985: projected field distribution (Figure~\ref{rmpib}) and we know that the

986: dipole field should be the dominating factor for the internal Faraday

987: rotation. For these calculations we assume a dipole field as displayed

988: in Figure~\ref{dipole}. 

989: We confine our attention to the main axis of the dipole, because

990: any toroidal component will be perpendicular to the line of sight

991: and will not contribute to RM (and so we ignore the complication of the

992: toroidal component). Magnetic field strength and synchrotron 

993: emissivity are assumed

994: constant throughout the nebula (the evidence for these assumptions 

995: is presented in Sections 4.2 and 4.3), and the intrinsic polarization angle

996: is set to be constant at all points along every line of sight through the

997: nebula. We calculate the electron density $n_e$ within the nebula needed

998: to generate the observed RM after correction for the foreground.

999: We make this calculation for various viewing angles, $\theta$. In our 

1000: calculations we are looking from the

1001: right hand side of Fig.~\ref{dipole} into the nebula. We consider we 

1002: have succeeded if we can find an electron density within

1003: the nebula, assumed constant, and a viewing angle which together provide 

1004: RM values reasonably consistent with the observed RM along different

1005: lines of sight. The foreground RM of approximately $-$30 rad/m$^2$

1006: is taken as a measure of the acceptable uncertainty.

1007: 

1008: Consider first the appearance at ${\theta}={0^{\circ}}$.  In the

1009: northern half the emission generated in the deeper levels first passes

1010: through a magnetic field that is directed away from us on the far side

1011: of the axis and then through a region where it is pointing towards us

1012: on the near side.  Since the emission is generated uniformly along the

1013: line of sight, the bulk of the emission passes through the magnetic

1014: field that is pointing towards us and the resulting rotation measure

1015: is positive. However, in the bottom half the structure is reversed,

1016: and this should lead to a negative RM. We would expect to see a

1017: symmetrical RM map, positive to the north and negative to the south

1018: with peaks of RM of equal absolute value. Now consider the appearance

1019: at an angle ${\theta} > {0^{\circ}}$ as portrayed in

1020: Figure~\ref{dipole}. For lines of sight above the center of the

1021: nebula, positive RM dominates, as above. For lines of sight passing

1022: below the center the emission from the far side of the nebula emerges

1023: with a strong positive RM because the path is long; there is some path

1024: through the near side that contributes negative rotation, and there is

1025: some reduction of net RM, but the path through the near side is

1026: relatively short, so that the final RM value remains strongly

1027: positive, indeed more positive than in the upper lobe. Our observations

1028: do show positive RM throughout, but with a higher RM value on the

1029: north-east lobe. Therefore to represent the observed situation, the

1030: field distribution in Figure~\ref{dipole} must be rotated by

1031: 180$^{\circ}$ around the line of sight, so that the dipole is directed

1032: downwards and away from us.

1033: 

1034: In order to test whether this simple model can predict the observed

1035: properties of DA\,495, we calculated the RM values that would be

1036: observed between 10550 and 4850~MHz for three different paths through

1037: the nebula, through the center and 1.2~pc above and below the

1038: center. At the center we observe a rotation measure of about

1039: 170~rad/m$^2$, which amounts to 200~rad/m$^2$ after correction for the

1040: foreground RM. At the RM peak on the north-eastern lobe, 1.2~pc above

1041: the center, we observe about 310~rad/m$^2$ and at the same distance on

1042: the south-western lobe we find about 250~rad/m$^2$. We calculated the

1043: electron density, n$_e$, required to produce these RM values for the

1044: three lines of sight through the nebula for varying viewing angles

1045: $\theta$.  The results are displayed in Fig.~\ref{rmcalc}.  We find

1046: that the model, although approximate, can reproduce the observed RM

1047: values with reasonable values for the parameters.  The small

1048: rectangles identify the region of satisfactory agreement. 

1049: Within the limitations of our assumptions, the electron

1050: density in the nebula lies between 0.3 and 0.4 cm$^{-3}$ and the

1051: viewing angle $\theta$ is between 33$^{\circ}$ and 40$^{\circ}$.

1052: 

1053: In our model we assumed constant magnetic field throughout the nebula.

1054: Even though the total emissivity and hence the total magnetic field

1055: strength inside the nebula must be relatively constant, we know that

1056: the magnetic field strength in a dipole field decreases with distance

1057: from the centre. We use another simple model, going to the other

1058: extreme, to assess the effect of this assumption: we collapse all the

1059: rotation measure along the line of sight onto the central axis of the

1060: dipole. All the emission generated beyond the axis is rotated and then

1061: added to the emission from the near side. Due to the symmetry of the

1062: dipole, the collapsed RMs at the two positions above and below the

1063: centre used in our calculations would be the same. However, by

1064: changing the viewing angle $\theta$ we would change the pathlength

1065: through the nebula in front of and beyond the central axis.  For

1066: constant emissivity we derive an angle $\theta$ of $10\degr$, in

1067: contrast to the ${\sim}36\degr$ derived above. Under both assumptions,

1068: and also under conditions of varying emissivity within the nebula, the

1069: dipole must be directed away from us towards the south.

1070: 

1071: Finally, in Fig.~\ref{oridip} we illustrate the three dimensional

1072: orientation of the dipole field inside DA\,495 as determined in this

1073: section. On the left hand side we display the magnetic field projected

1074: to the plane of the sky. The magnetic field axis has an angle of about

1075: $50\degr$ east of north and the magnetic field is pointing

1076: south-west. On the right hand side the dipole field orientation along

1077: the line of sight is shown, indicating that the dipole is pointing

1078: away from us towards the south-west.

1079: 

1080: \subsection{Limits on a surrounding shell-type SNR}

1081: 

1082: Is DA\,495 a pure pulsar wind nebula, like the Crab

1083: Nebula, or does it have a surrounding SNR shell? With our new data we

1084: were able to search for such a shell. None was found, and we can place

1085: sensitive upper limits on any such emission.  The noise in the area

1086: around DA\,495 in our 1420~MHz image, which is the most sensitive we

1087: have for such a search, is about 40~mK rms. A $3\sigma$ signal would

1088: be about 1.0~mJy/beam, and we would need this peak flux density for a

1089: positive detection.  If the inner edge of this shell is at the outer

1090: edge of the PWN and its shell thickness is 10\%, typical for an

1091: adiabatically expanding SNR shell, the upper limit for the total flux

1092: density is about 90~mJy at 1420~MHz. An older remnant would have

1093: higher compression and a thinner shell, resulting in an even lower

1094: flux density limit. The 1~GHz flux density limit is about 100~mJy, using a

1095: spectral index of $\alpha = -0.5$, which is a typical value for a 

1096: shell-type SNR.  This translates to a maximum radio

1097: surface brightness of $\Sigma_{\rm 1~GHz}^{\rm max} = 5.4 \times

1098: 10^{-23}$~Watt~m$^{-2}$~Hz$^{-1}$~sr$^{-1}$, slightly

1099: lower than the lowest radio surface-brightness known, that of the SNR

1100: G156.2$+$5.7 which has ${\Sigma_{\rm 1~GHz}}={6 \times

1101: 10^{-23}}$~Watt~m$^{-2}$~Hz$^{-1}$~sr$^{-1}$ \citep{reic92}.

1102: 

1103: The 60$\mu$m image of the DA\,495 field (Fig.~\ref{tp}) shows infrared

1104: emission to the north-east of the SNR. 

1105: We found \ion{H}{1} in the CGPS data, which is likely related to this 

1106: emission, because it exhibits a similar shape. This \ion{H}{1}, however,

1107: is found at velocities near the tangent point, far removed from the systemic 

1108: velocity of DA\,495.

1109: %\ion{H}{1} which is likely related to this emission, because it exhibits a 

1110: %similar shape, can be found in the CGPS data at velocities near the tangent 

1111: %point, far removed from the systemic velocity of DA\,495. 

1112: We do not believe that any of this

1113: infrared emission is related to the SNR.

1114: 

1115: \section{Conclusions}

1116: 

1117: We have imaged the radio emission from the pulsar wind nebula 

1118: DA\,495 in total intensity

1119: and polarized emission across a range of frequencies from 408 to

1120: 10550~MHz.  After correcting for the flux contribution of extraneous

1121: superimposed sources we have demonstrated that there is a break in the

1122: spectrum at about 1.3\,GHz which we can convincingly attribute to

1123: synchrotron cooling. The spectral break at such a low frequency, the

1124: lowest known for any PWN, can only be produced if the pulsar

1125: originally was very energetic, with a high energy loss rate.  We

1126: consider that DA\,495 is an aging pulsar wind nebula which has a pulsar that 

1127: in its 

1128: earlier stages resembled the Crab pulsar or the pulsar in Vela~X. 

1129: The magnetic field inside the nebula is

1130: remarkably well-organized and has the topology of a dipole field

1131: centered on the pulsar; the field probably also has a toroidal

1132: component, wrapped around the dipole axis. The rotation measure is 

1133: high, as would be expected in a

1134: nebula where the magnetic field has the high value of

1135: $\sim$1.3~mG, and the nebula becomes Faraday thick at a very high

1136: frequency.  We have set a very low upper limit for the surface

1137: brightness for any shell of SNR emission surrounding DA\,495. The

1138: absence of a shell implies that the SN explosion occurred in a

1139: low-density environment, and this also explains how the magnetic field

1140: structure has remained so regular to an age of $\sim$20,000 yr.

1141: 

1142: \section{Acknowledgments}

1143: The Dominion Radio Astrophysical Observatory is a National Facility

1144: operated by the National Research Council. The Canadian Galactic Plane

1145: Survey is a Canadian project with international partners, and is

1146: supported by the Natural Sciences and Engineering Research Council

1147: (NSERC).  This research is based on observations with the 100-m

1148: telescope of the MPIfR at Effelsberg. SSH acknowledges support by 

1149: the Natural Sciences and Engineering Research Council and the 

1150: Canada Research Chairs program. ZA was supported by

1151: NASA grant NRA-99-01-LTSA-070. The VLSS is being carried out by 

1152: the (USA) National Radio Astronomy Observatory (NRAO) and the Naval Research 

1153: Lab. The NRAO is operated by Associated Universities, Inc. and is a 

1154: Facility of the (USA) National Science Foundation.

1155: 

1156: \appendix

1157: 

1158: \begin{thebibliography}{}

1159: 

1160: %\bibitem[Aller \& Reynolds(1985)]{alle85} Aller H.D., Reynolds S.P., 

1161: %   1985, \apj, 293, L73

1162: 

1163: \bibitem[Arzoumanian et al.(2004)]{arzo04} Arzoumanian Z., Safi-Harb

1164: S., Landecker T.L., Kothes R., 2004, \apjl, 610, L101

1165: 	

1166: \bibitem[Arzoumanian et al.(2008)]{arzo08} Arzoumanian Z., Safi-Harb

1167: S., Landecker T.L., Kothes R., Camilo, F., 2008, \apj, accepted,

1168: ArXiv Astrophysics e-prints arXiv:0806.3766

1169: 	

1170: %\bibitem[Biggs \& Lyne(1996)]{bigg96} Biggs J.D., Lyne A.G., 1996,

1171: %   \mnras, 282, 691

1172: 

1173: \bibitem[Blondin et al.(2001)]{blon01} Blondin J.M., Chevalier R.A.,

1174:    Frierson D.M., 2001, \apj, 563, 806

1175: 

1176: \bibitem[Bock \& Gaensler(2005)]{bock05} Bock D. C.-J., Gaensler, B.M.,

1177: 2005, \apj, 626, 343

1178: 

1179: \bibitem[Burn(1966)]{burn66} Burn, B.J., 1966, \mnras, 133, 67

1180: 

1181: \bibitem[Cao et al.(1997)]{cao97} Cao, Y., Terebey, S., Prince, T.A., 

1182:    Beichman, C.A., 1997, \apjs, 111, 387

1183: 

1184: \bibitem[Chevalier(2000)]{chev00} Chevalier R.A., 2000, \apj, 539, L45

1185: 

1186: \bibitem[Cohen et al.(2007)]{cohe07} Cohen A.S., Lane W.M., Cotton W.D., 

1187:    Kassim N.E., Lazio T.J.W., Perley R.A., Condon J.J., Erickson W.C.,

1188:    2007, \aj, 134, 1245

1189: 

1190: \bibitem[Condon et al.(1998)]{cond98} Condon J.J., Cotton W.D.,

1191: Greisen E.W., Yin Q.F., Perley R.A., Taylor G.B., Broderick J.J., 1998,

1192: \aj, 115 1693

1193: 

1194: \bibitem[Condon(1984)]{cond84} Condon J.J., 1984, \apj, 287, 461

1195: 

1196: \bibitem[Cordes \& Lazio(2002)]{cord02}

1197: Cordes, J.~M. \& Lazio, T.~J.~W., 2002, ArXiv Astrophysics e-prints,

1198: astro-ph/0207156

1199: 

1200: \bibitem[Dickel \& DeNoyer(1975)]{dick75} Dickel J.R., DeNoyer L.K.,

1201: 1975, \aj, 80, 437

1202: 

1203: \bibitem[Dickel et al.(1971)]{dick71} Dickel J.R., Webber J.C., Yang

1204: K.S, Staff, 1971, \aj, 76, 294

1205: 

1206: \bibitem[Dodson et al.(2003)]{dods03} Dodson R., Lewis D., McConnell D.,

1207:     Deshpande A. A., 2003, \mnras, 343, 116

1208: 

1209: \bibitem[Douglas et al.(1996)]{doug96} Douglas J.N., Bash F.N.,

1210: Bozyan F.A., Torrence G.W., Wolfe C., 1996, \aj, 111, 1945

1211: 

1212: \bibitem[Duncan et al.(1999)]{dunc99} Duncan, A.R., Reich P., Reich W.

1213: F\"urst E., 1999, \aap, 350, 447

1214: 

1215: \bibitem[Dwarakanath \& Udaya Shankar(1990)]{dwar90} Dwarakanath

1216: K.S., Udaya Shankar N., 1990, J. Astrophys. Astr., 11, 323

1217: 

1218: \bibitem[Emerson \& Gr\"ave (1988)]{emer88} Emerson D.T., Gr\"ave R., 

1219:    1988, \aap 190, 353

1220: 

1221: \bibitem[Emerson et al.(1979)]{emer79} Emerson D.T., Klein U., Haslam C.G.T.

1222:    1979, \aap, 76, 92

1223: 	

1224: \bibitem[Fesen et al.(1992)]{fese92} Fesen, R.A., Martin, C.L., Shull, J.M.,

1225:    1992, \apj, 399, 599

1226: 

1227: \bibitem[Foster \& MacWilliams(2006)]{fost05} Foster, T., MacWilliams, J.

1228:    2006, \apj, in press (May, 2006, \apj, 642)

1229: 

1230: \bibitem[F\"urst \& Reich (1986)]{furs86} F\"urst, E., Reich, W. 

1231:    1986, \aap 163, 185

1232:    

1233: %\bibitem[Gray et al.(1999)]{gray99} Gray A.D., Landecker T.L., Dewdney P.E.,

1234: %   et al., 1999, \apj, 514, 221

1235: 

1236: \bibitem[Green(1987)]{gree87} Green D.A., 1987, \mnras, 225, L11

1237: 

1238: \bibitem[Green(1994)]{gree94} Green, D.A., 1994, \apjs, 90, 871

1239: 

1240: \bibitem[Green(2004)]{gree04} Green D.A., 2004, Bulletin

1241:    of the Astronomical Society of India, 32, 335

1242:    (also available on the World Wide Web at 

1243:    http://www.mrao.cam.ac.uk/surveys/snrs)

1244: 

1245: \bibitem[Gregory \& Condon(1991)]{greg91} Gregory P.C., Condon J.J.,

1246: 1991, \apjs, 75, 1011

1247: 

1248: \bibitem[Gregory et al.(1996)]{greg96} Gregory P.C., Scott W.K.,

1249: Douglas K., Condon J.J., 1996, \apjs, 103, 427

1250: 

1251: \bibitem[Haslam et al.(1982)]{hasl82} Haslam C.G.T., Stoffel H., Salter

1252:    C.J., Wilson W.E., 1982, \aaps 47, 1

1253: 

1254: \bibitem[Hill(1967)]{hill67} Hill E.R., 1967, AuJPh, 29, 29

1255: 

1256: \bibitem[Kothes et al.(2004)]{koth04} Kothes R., Landecker T.L.,

1257:    Wolleben M., 2004, \apj, 607, 855

1258: 

1259: \bibitem[Kothes et al.(2006a)]{koth06a} Kothes, R., Reich, W., 

1260:    Uyan{\i}ker, B., 2006a, \apj, 638, 225

1261:    

1262: \bibitem[Kothes et al.(2006b)]{koth06b} Kothes, R., Fedotov, K., Foster,

1263:    T.J., Uyan{\i}ker, B., 2006b, \aap, 457, 1081

1264: 

1265: \bibitem[Kovalenko et al.(1994)]{kova94} Kovalenko A.V., Pynzar A.V.,

1266:    Udal'tsov V.A., 1994, Astr. Rep., 38, 95

1267: 

1268: \bibitem[Landecker \& Caswell(1983)]{land83} Landecker, T.L., Caswell,

1269:    J.L., 1983, \aj, 88, 1810

1270: 

1271: \bibitem[Landecker et al.(1990)]{land90} Landecker T.L., Clutton-Brock M,

1272: Purton C.R., 1990, \aap, 232, 207

1273: 

1274: %\bibitem[Landecker et al.(2001)]{land01} Landecker T.L., Uyan{\i}ker B.,

1275: %   Kothes R., 2001, AAS, 199, \#58.07

1276: 

1277: %\bibitem[Landecker et al.(1993)]{land93} Landecker T.L., Higgs L.A.,

1278: %   Wendker H.J., 1993, \aap, 276, 522

1279: 

1280: \bibitem[Langston et al.(1990)]{lang90} Langston G.I., Heflin M.B.,

1281: Conner S.R., Lehar J., Carilli C.L., Burke B.F., 1990, \apjs, 72, 621

1282: 

1283: \bibitem[Lyne et al.(1996)]{lyne96} Lyne A.G., Pritchard R.S.,

1284:    Graham-Smith F., Camillo F., 1996, Nature 381, 497

1285: 

1286: \bibitem[Lyne et al.(1988)]{lyne88} Lyne A.G., Pritchard R.S., Smith F.G.,

1287:    1988, \mnras, 233, 667

1288:    

1289: \bibitem[Lyne \& Smith(1989)]{lyne89} Lyne A.G., Smith F.G., 1989, 

1290: \mnras, 237, 533

1291: 

1292: \bibitem[Manchester et al.(2005)]{manc05} Manchester, R. N., Hobbs, G. B., 

1293:    Teoh, A., Hobbs, M., \aj, 129, 1993

1294: 

1295: \bibitem[Morsi \& Reich(1986)]{mors86} Morsi H.W., Reich W., 

1296:    1986, \aap, 163, 313

1297: 

1298: \bibitem[Morsi \& Reich(1987)]{mors87} Morsi H.W., Reich W., 

1299:    1987, \aap, 69, 533

1300: 

1301: \bibitem[Pacini \& Salvati(1973)]{paci73} Pacini F., Salvati M.,

1302:    1973, \apj, 186, 249

1303: 

1304: \bibitem[Petre et al.(2002)]{petr02} Petre, R., Kuntz, K.D., Shelton, R.L.,

1305: 2002, \apj, 579, 404

1306: 

1307: %\bibitem[Possenti et al.(2002)]{poss02} Possenti A., Cerutti R.,

1308: %   Colpi M., Mereghetti S., 2002, \aap, 387, 993

1309: 

1310: \bibitem[Reich et al.(1984)]{reic84} Reich, W., F\"urst, E., Sofue, Y.,

1311:     1984, \aap, 133, L4

1312: 

1313: \bibitem[Reich et al.(1990)]{reic90} Reich W., F\"urst E., Reich P., Reif

1314:     K., 1990, \aaps, 85, 633

1315: 	

1316: \bibitem[Reich et al.(1992)]{reic92} Reich W., F\"urst E., 

1317:    Arnal E.M., 1992, \aap, 256, 214

1318: 

1319: \bibitem[Rengelink et al.(1997)]{reng97} Rengelink R.B., Tang Y., de

1320: Bruyn A.G., Miley G.K., Bremer M.N., Roettgering H.J.A., Bremer

1321: M.A.R., 1997, \aaps, 124, 259

1322: 

1323: \bibitem[Salter et al.(1989)]{salt89} Salter C.J., Reynolds S.P., Hogg D.E.,

1324:     Payne J.M., Rhodes P.J., 1989, \apj, 338, 171

1325:     

1326: \bibitem[Simard-Normandin \& Kronberg(1979)]{sima79} Simard-Normandin M., 

1327: Kronberg P.P., 1979, Nature, 279, 115

1328: 	

1329: \bibitem[Sokoloff et al.(1998)]{soko98} Sokoloff, D.D., Bykov, A.A.,

1330: Shukurov, A., Berkhuijsen, E.M., Beck, R., \& Poezd, A.D., 1998,

1331: MNRAS, 299, 189

1332: 

1333: \bibitem[Stephenson \& Green(2002)]{step02} Stephenson, F.R., 

1334:    Green, D.A., 2002, Historical Supernovae and their Remnants, 

1335:    Oxford University Press

1336: 

1337: \bibitem[Strom \& Greidanus(1992)]{stro92} Strom, R.G., Greidanus, H.,

1338: 1992, Nature, 358, 654

1339: 

1340: \bibitem[Taylor \& Cordes(1993)]{tayl93} Taylor J.H., Cordes J.M.,

1341:    1993, \apj, 411, 674

1342: 

1343: \bibitem[Taylor et al.(2003)]{tayl03} Taylor A.R., Gibson S.J., Peracaula,

1344:    M. et al.,

1345: %      Martin, P.G., Landecker, T.L., Brunt, C., Dewdney, P.E., Dougherty,

1346: %      S.M., Durand, D.,

1347: %      Gibson, S.J., Gray, A.D., Higgs, L.A., Kerton, C.R., Knee, L.B.G.,

1348: %      Kothes, R., Peracaula, M., Purton, C.R., Uyan{\i}ker, B., Wallace,

1349: %      B.J., Willis, A.G.,

1350:    2003, \aj, 124, 3145

1351: 

1352: \bibitem[Uyan{\i}ker et al.(2003)]{uyan03} Uyan{\i}ker B., Landecker T.L.,

1353:    Gray A.D., Kothes R., 2003, \apj, 585, 785

1354: 	

1355: \bibitem[Velusamy et al.(1989)]{velu89} Velusamy T., Becker R.H.,

1356: Goss W.M., Helfand D.J., 1989, J. Astrophys. Astr., 10, 161

1357: 

1358: \bibitem[Weiler \& Panagia(1980)]{weil80} Weiler K.W., Panagia N.,

1359:    1980, \aap, 90, 269

1360: 

1361: \bibitem[Willis(1973)]{will73} Willis A.G., 1973, \aap, 26, 237

1362: 

1363: \bibitem[Woltjer et al.(1997)]{wolt97} Woltjer L., Salvati M., 

1364:    Pacini F., Bandiera R., 1997, \aap, 325, 295

1365: 

1366: \bibitem[Xiao et al.(2008)]{xiao08} Xiao L., F\"urst E., Reich W., Han J.L.,

1367: 2008, \aap, 482, 783

1368: 

1369: \end{thebibliography}

1370: %%%%%%%%%%%%%%%%%%%%%%%%

1371: \clearpage

1372: 

1373: 

1374: \begin{deluxetable}{lllll}

1375:    \tablewidth{0pc}

1376:    \tablecolumns{5}

1377:    \tablecaption{Observation characteristics\label{obspara}}

1378:    \tablehead{{Frequency [MHz]} & {408} & {1420} 

1379:      & {4850} & {10550}}

1380:    \startdata

1381: Telescope & DRAO & DRAO & Effelsberg & Effelsberg \\

1382: Observation Date & January & January & February & January \\

1383:      &         &         &          & to July \\

1384:      & 2002 & 2002 & 1996  & 1990  \\

1385: Half-power beam ($'$) & $2.9 \times 5.95$ & $0.82 \times 1.75$ & 2.45 & 2.45$^a$ \\

1386:                       & (R.A. $\times$~Dec.) & (R.A. $\times$~Dec.) & & \\

1387: rms noise, $I$, [mJy/beam] & 7 & 0.4 & 0.5 & $<1.5^b$ \\

1388: rms noise, $PI$, [mJy/beam] & & 0.4 & 0.3 & 0.4 \\

1389: Calibrators, $I$, (Flux Density, Jy) & 3C48 (38.9)  & 3C48 (15.7) & 

1390: 3C286 (7.5) & 3C286 (4.5) \\

1391:                                     & 3C147 (48.0) & 3C147 (22.0) & & \\

1392:                                     & 3C295 (54.0) & 3C295 (22.1) & & \\

1393: Calibrators $Q, U$ &  & 3C286 & 3C286 & 3C286 \\

1394: Linear Pol. [\%] &  & 9.25 & 11.3 & 11.7 \\

1395: Pol. Angle [$\degr$] &  & 33 & 33 & 33\\

1396: Coverages$^c$ & & & 14 & 5 \\

1397: \enddata

1398: \\$^a$ Original resolution 1\farcm15. Smoothed to 2\farcm45 for

1399: presentation.

1400: 

1401: $^b$ Estimated. There is no region in the image entirely free of emission.

1402: 

1403: $^c$ Effelsberg observations only.

1404: \end{deluxetable}

1405: 

1406: \clearpage

1407: 

1408: \begin{deluxetable}{cccccc}

1409:    \tablewidth{0pc}

1410:    \tablecolumns{6}

1411:    \tablecaption{Integrated flux densities of DA\,495 \label{fluxes}}

1412:    \tablehead{

1413: \colhead{Frequency [MHz]} & \colhead{Beam} & \colhead{$S_{meas}$ [Jy]} & 

1414: \colhead{$S_{cs-r}$ [Jy]} & \colhead{$S_{HII-r}$ [Jy]} & 

1415: \colhead{reference}}

1416:    \startdata

1417: 34.5  & $26' \times~44'$                  & $<35.0$       & $<$24.8        &

1418: & 1 \\

1419: 74 & $1\farcm33 \times 1\farcm33$   &               & $<$14          & 

1420: &  5 \\

1421: 83    & $0.2^{\circ} \times~5.2^{\circ}$  & $18 \pm~5$    & 13.0 $\pm$~5   &

1422: & 2 \\

1423: 111   & $0.15^{\circ} \times~3.9^{\circ}$ & $17 \pm~5$    & 13.0 $\pm$~5   &

1424: & 2 \\

1425: 318   & $14' \times~16'$                  & $9.7 \pm~2.2$ & 8.1 $\pm$~2.2  &

1426: & 3 \\

1427: 327   & 1\farcm1 $\times$~0\farcm6        &               & $7.8 \pm~1.0$  &

1428: & 4 \\

1429: 408   & 2\farcm8 $\times$~5\farcm7        &               & $6.5 \pm~0.6$  &

1430: & 5 \\

1431: 430   & 8\farcm4 $\times$~9\farcm6        & $8.7 \pm~4.9$ & $7.5 \pm~4.9$  &

1432: & 3 \\

1433: 610.5 & $16' \times~18'$                  & $7.0 \pm~1.5$ & $6.0 \pm~1.5$  &

1434: & 6, 7 \\

1435: 1420  & 0\farcm8 $\times$~1\farcm6        &               & $4.0 \pm~0.2$  &

1436: $3.6 \pm~0.3$ & 5 \\

1437: 1665  & $22'$                             & $4.4 \pm~0.5$ & $4.0 \pm~0.5$  & 

1438: $3.6 \pm~0.5$ & 7 \\

1439: 2695  & 4\farcm3                          &               & $2.7 \pm~0.2$  &

1440: $2.3 \pm~0.3$ & 8 \\

1441: 2735  & 4\farcm9 $\times$~5\farcm7        & $2.8 \pm~0.4$ & $2.7 \pm~0.4$  &

1442: $2.3 \pm~0.4$ & 7 \\

1443: 4850  & 2\farcm45                         &               & $1.59 \pm~0.1$ &

1444: $1.2 \pm~0.1$ & 5 \\

1445: 10550 & 1\farcm15                         &               & $1.12 \pm~0.1$ &

1446: $0.76 \pm~0.18$ & 5 \\

1447:    \enddata

1448: \\Notes: $S_{meas}$ is the measured integrated flux density of

1449: DA\,495. $S_{cs-r}$ is the measured flux density minus the total flux

1450: density of the three brightest compact sources (see text)

1451: appropriately scaled with frequency. For the new data presented in

1452: this paper, and for other high-resolution measurements, this is the

1453: integrated flux density of DA\,495 without the compact

1454: sources. $S_{HII-r}$ is the integrated flux density minus an allowance

1455: for the \ion{H}{2} region which overlaps the SNR (see text).

1456: 

1457: References: (1) \citet{dwar90}, (2) \citet{kova94}, (3) \citet{dick75}, 

1458: (4) \citet{velu89}, (5) present work, (6) \citet{dick71}, 

1459: (7) \citet{will73}, (8) \citet{reic90}

1460: \end{deluxetable}

1461: 

1462: \clearpage

1463: 

1464: \begin{deluxetable}{ccccc}

1465:    \tablewidth{0pc}

1466:    \tablecolumns{5}

1467:    \tablecaption{Flux densities of compact sources [mJy]\label{sourcefd}}

1468:    \tablehead{

1469: \colhead{Frequency, MHz} & \colhead{Source 1} & \colhead{Source 2} & 

1470: \colhead{Source 3} & Reference }  

1471:    \startdata

1472:  74   & 1470 $\pm$~180 & 1200 $\pm$~400 & 900 

1473: $\pm$~300 & 3 \& 10 \\

1474: 330   & 613 $\pm$~50 & 711 $\pm$~70 & 328 $\pm$~30 & 1 \\

1475: 365   & 557 $\pm$~67 & 739 $\pm$~52 & 294 $\pm$~28 & 2 \\

1476: 408   & 650 $\pm$~80 & 504 $\pm$~50 & 200 $\pm$~40 & 3 \\

1477: 1420  & 139 $\pm$~5  & 280 $\pm$~8  &  93 $\pm$~5  & 4 \\

1478: 1420  & 131 $\pm$~5  & 244 $\pm$~25 &  80 $\pm$~6  & 5 \\

1479: 1420  & 134 $\pm$~8  & 288 $\pm$~15 &  83 $\pm$~6  & 3 \\

1480: 2695  &              & 160 $\pm$~32 &              & 6 \\

1481: 4850  &  50 $\pm$~10 &              &              & 7 \\

1482: 4850  &              & 102 $\pm$~15 &              & 8 \\

1483: 4850  &              &  89 $\pm$~10 &              & 9 \\

1484: 4850  &  58 $\pm$~10 & 108 $\pm$~5  &  36 $\pm$~4  & 3 \\

1485: 10550 & 15 $\pm$~4  &  60 $\pm$~10 &  17 $\pm$~5  & 3 \\

1486:    \enddata

1487: \\References: (1) WENSS \citep{reng97}, (2) Texas \citep{doug96}, 

1488: (3) present work, (4) NVSS \citep{cond98}, (5) \citet{land83}, 

1489: (6) \citet{reic90}, (7) MIT-GB \citep{lang90}, (8) 87GB \citep{greg91},

1490: (9) GB6 \citep{greg96}, (10) VLSS \citep{cohe07}

1491: \end{deluxetable}

1492: 

1493: \begin{deluxetable}{cccc}

1494:    \tablewidth{0pc}

1495:    \tablecolumns{4}

1496:    \tablecaption{Positions and spectra of compact sources\label{sourceps}}

1497:    \tablehead{

1498:  & \colhead{Source 1} & \colhead{Source 2} & \colhead{Source 3}}  

1499:    \startdata

1500: RA(2000)  & 19$^h$ 51$^m$ 52.1$^s$ & 19$^h$ 53$^m$ 3.1$^s$ &

1501: 19$^h$ 52$^m$ 40.8$^s$ \\

1502: Dec(2000) & $29^{\circ}~26'~7''$ & $29^{\circ}~10'~36''$ &

1503: $29^{\circ}~35'~47''$ \\

1504: A         & 5.05 $\pm$~0.09 & 4.69 $\pm$~0.07 & 4.58 

1505: $\pm$~0.11 \\

1506: B         & $-$0.92 $\pm$~0.03 & $-$0.72 $\pm$~0.02 &  $-$0.83 

1507: $\pm$~0.04 \\

1508:    \enddata

1509: \\log(S[mJy]) = A + B~log($\nu$[MHz])

1510: \end{deluxetable}

1511: 

1512: \begin{deluxetable}{ccc}

1513:    \tablewidth{0pc}

1514:    \tablecolumns{3}

1515:    \tablecaption{Integrated polarization properties\label{intpol}}

1516:    \tablehead{

1517: Frequency [MHz] & \colhead{Polarized Intensity [Jy]} &

1518: \colhead{Fractional Polarization [\%]}}

1519:    \startdata

1520: 1420  & $0.44 \pm~ 0.02$ & $12 \pm~2$ \\

1521: 2695  & $0.59 \pm~ 0.05$ & $26 \pm~4$ \\

1522: 4850  & $0.29 \pm~ 0.03$ & $24 \pm~4$ \\

1523: 10550 & $0.18 \pm~ 0.02$ & $24 \pm~6$ \\

1524:    \enddata

1525: \\Fractional polarization is computed using integrated flux densities

1526: corrected for compact sources and the superimposed \ion{H}{2}

1527: region. See Table~\ref{fluxes} 

1528: \end{deluxetable}

1529: 

1530: \begin{deluxetable}{lccc}

1531:    \tablewidth{0pc}

1532:    \tablecolumns{4}

1533:    \tablecaption{Breaks in the spectra of pulsar wind nebulae

1534: \label{break_freq}}

1535:    \tablehead{

1536: \colhead{SNR} & \colhead{Break Frequency} & \colhead{injected/cooling$^a$} &

1537: \colhead{reference}}

1538:    \startdata

1539: Crab Nebula & 40~keV~\& 1000~\AA & {\it{i}} & \citet{wolt97} \\

1540: Crab Nebula & 14000 GHz & {\it{c}} & \citet{stro92} \\

1541: W44 & 8000 GHz & {\it{c}} & \citet{petr02} \\

1542: Vela X & 100 GHz & {\it{c}} & \citet{weil80} \\

1543: G29.7$-$0.3 & 55 GHz & {\it{i}} & \citet{bock05} \\

1544: 3C~58 & 50 GHz & {\it{i}} & \citet{salt89,gree94} \\

1545: G21.5$-$0.9 & $\sim 30$~GHz & {\it ?} & \citet{salt89} \\

1546: G16.7+0.1 & 26 GHz & {\it{i}} & \citet{bock05} \\

1547: CTB 87 & 10 GHz & {\it{c}} & \citet{mors87} \\

1548: G27.8$+$0.6 & 5-10~GHz & {\it{c}} & \citet{reic84} \\

1549: G106.3+2.7 & 4.5 GHz & {\it{c}} & \citet{koth06a} \\

1550: DA 495 & 1.3 GHz & {\it{c}} & this paper \\ 

1551:    \enddata

1552: \\ $^a$ {\it{i}} denotes spectral break due to a break in the injected electron

1553: spectrum; {\it{c}} denotes spectral break due to synchrotron cooling

1554: \end{deluxetable}

1555: 

1556: \begin{deluxetable}{lcccccc}

1557:    \tablewidth{0pc}

1558:    \tablecolumns{7}

1559:    \tablecaption{Present day and birth properties of the six pulsars for which 

1560:    a reliable age can be found.

1561: For a description of individual parameters see Section 4.2.

1562: \label{pulsars}}

1563:    \tablehead{

1564: \colhead{SNR} & \colhead{Pulsar} & \colhead{$\tau$ [yr]} &

1565: \colhead{$\dot{E}$ [erg/s]} & \colhead{$\tau_0$ [yr]} &

1566: \colhead{$\dot{E}_0$ [erg/s]} & \colhead{$t$ [yr]}}

1567:    \startdata

1568: Crab Nebula & B0531+21     & 1270  & $4.4\times 10^{38}$ &

1569: 320         & $1.0\times 10^{40}$ & 950 \\

1570: 3C~58       & J0205+6449   & 5370  & $2.7\times 10^{37}$ &

1571: 4550        & $3.8\times 10^{37}$ & 820 \\

1572: G11.2$-$0.3 & J1811$-$1925 & 23300 & $6.4\times 10^{36}$ &

1573: 21680       & $7.4\times 10^{36}$ & 1620 \\

1574: Vela X ($\beta = 2.0$) & B0833-45     & 11300 & $6.9\times 10^{36}$ &

1575: 110         & $7.3\times 10^{40}$ & 11200 \\

1576: Vela X ($\beta = 6.0$) & B0833-45     & 11300 & $6.9\times 10^{36}$ &

1577: 3000         & $2.0\times 10^{40}$ & 8300 \\

1578: G106.3+2.7  & J2229+6114   & 10460 & $2.2\times 10^{37}$ &

1579: 6560        & $5.6\times 10^{37}$ & 3900 \\

1580: W~44        & B1853+01     & 20300 & $4.3\times 10^{35}$ &

1581: 13300       & $1.0\times 10^{36}$ & 7000 \\

1582:    \enddata

1583: \end{deluxetable}

1584: 

1585: \begin{deluxetable}{lccccc}

1586:    \tablewidth{0pc}

1587:    \tablecolumns{6}

1588:    \tablecaption{Properties of DA 495 extrapolated from birth

1589:    properties of the six pulsars in Table~\ref{pulsars}.

1590: \label{DA495_pulsar}}

1591:    \tablehead{

1592: \colhead{Basis} & \colhead{$t_{\rm DA\,495}$ [yr]} & \colhead{$\tau_{\rm DA\,495}$ 

1593: [yr]} & \colhead{$B_{\rm req}$ [mG]} & \colhead{$E_{\rm tot}$ [erg]} &

1594: \colhead{$B_{\rm max}$ [mG]}}

1595:    \startdata

1596: Crab Nebula ($\beta = 2.0$) & 100900 & 101200 & 0.45 & $1.0\times 10^{50}$ & 0.98 \\

1597: Crab Nebula ($\beta = 2.3$) & 47600 & 47900 & 0.74 & $7.8\times 10^{49}$ & 0.86 \\

1598: 3C~58       & 84100  & 88700  & 0.51 & $5.2\times 10^{48}$ & 0.22 \\

1599: G11.2$-$0.3 & 165000 & 186000 & 0.32 & $4.5\times 10^{48}$ & 0.22 \\

1600: Vela X ($\beta = 2.0$)  & 93900  & 94000  & 0.47 & $2.5\times 10^{50}$ & 1.56 \\

1601: Vela X ($\beta = 6.0$)  & 20000  & 23000  & 1.32 & $3.8\times 10^{50}$ & 1.91 \\

1602: G106.3+2.7  & 149000  & 155000 & 0.35 & $1.1\times 10^{49}$ & 0.32 \\

1603: W~44        & 28800  & 42100  & 1.03 & $2.3\times 10^{47}$ & 0.05 \\

1604:    \enddata

1605: \end{deluxetable}

1606: \clearpage

1607: %%Figure 1%%

1608: \begin{figure*}

1609:   \centerline{\includegraphics[bb = 47 62 509 701,width=13cm,clip]{f1.ps}}

1610: %  \includegraphics[width=16cm,clip]{DA\,495_tot_int.eps}

1611:    \caption{Total intensity maps of DA\,495 at five radio

1612:    frequencies, and the corresponding IRAS 60$\mu$m image.

1613:    The position of the neutron star is shown in each panel by the white star.

1614:    Contour levels are at: 10~K ($12\sigma$) to 100~K in steps of 10~K

1615:    at 408~MHz (1~K = 8.5~mJy/beam); 0.5~K ($9\sigma$) to 5.0~K in steps of 

1616:    0.5~K

1617:    at 1420 MHz (1~K = 8.5~mJy/beam); 0.1~K ($5\sigma$) to 0.9~K in steps of 

1618:    0.1~K

1619:    at 2695 MHz (1~K = 400~mJy/beam); 5~mJy/beam ($20\sigma$) to 55~mJy/beam in

1620:    steps of 5~mJy/beam at 4850 MHz (1~mJy/beam = 2.4~mK); 

1621:    5~mJy/beam ($4\sigma$) to 

1622:    40~mJy/beam in steps of 5~mJy/beam at 10550 MHz (1~mJy/beam = 0.5~mK);

1623:    20 MJy/sr to 28 MJy/sr in steps of 1 MJy/sr at 60$\mu$m.}

1624:    \label{tp}

1625: \end{figure*}

1626: 

1627: %%Figure 2%%

1628: \begin{figure}

1629: \centerline{\includegraphics[bb = 45 32 540 505,width=15cm,clip]{f2.ps}}

1630:  %  \includegraphics[width=16cm,clip]{DA\,495_nth_th.eps}

1631:    \caption{Separation of the total power emission at 1420~MHz 

1632:    into a non-thermal component (${\alpha}={-0.85}$ -- shown as black 

1633:    contours and shading) and a thermal component (${\alpha}={-0.1}$ -- white

1634:    contours). Black contours are from 30~mJy/beam ($10\sigma$) to 170~mJy/beam 

1635:    in steps of

1636:    20~mJy/beam. White contours are from, 10~mJy/beam ($3\sigma$) to 40~mJy/beam in 

1637:    steps of 5~mJy/beam.

1638:     The angular resolution is 2\farcm45 (see Section 3.2).}

1639:    \label{nthth}

1640: \end{figure}

1641: 

1642: %%Figure 3%%

1643: \begin{figure*}

1644: \centerline{\includegraphics[bb = 30 185 585 495,width=15cm,clip]{f3.ps}}

1645: %   \includegraphics[width=16cm,clip]{DA\,495_int_spectrum.eps}

1646:    \caption{Integrated radio continuum spectrum of DA\,495. Flux densities

1647:    taken from the literature are displayed in gray and those

1648:    determined in this publication in black. The left panel shows a

1649:    spectrum constructed from the measured flux densities of

1650:    Table~\ref{fluxes}. The indicated line, corresponding to

1651:    ${\alpha}={-0.59}$, is a fit to all plotted points. The right panel

1652:    shows the spectrum after allowance for the contribution of compact

1653:    sources and the superimposed \ion{H}{2} region. The solid lines

1654:    corresponding to ${\alpha}={-0.45}$ and ${\alpha}={-0.87}$ are fits

1655:    to points below and above 1~GHz respectively (see Section 3.2). The dashed

1656:    line corresponds to ${\alpha}={-0.59}$, the fit from the left panel.

1657:    The lengths of the arrows denoting the upper limits are not intended to

1658:    indicate error estimates.}

1659:    \label{intspec}

1660: \end{figure*}

1661: 

1662: %%Figure 4%%

1663: \begin{figure*}

1664: \centerline{\includegraphics[bb = 45 60 511 505,width=15cm,clip]{f4.ps}}

1665: %   \includegraphics[width=16cm,clip]{DA\,495_poln_images.eps}

1666:    \caption{Polarized intensity of the emission from DA\,495 and its 

1667:    surroundings at four frequencies, shown by contours and grayscale.

1668:    The orientation of the E-field is shown by the superimposed vectors.

1669:    The location of the pulsar is indicated in each panel 

1670:    by the star.

1671:    Contour levels are at:

1672:    1.5~mJy/beam ($4\sigma$) to 12~mJy/beam at 1420~MHz; 

1673:    6~mJy/beam ($2\sigma$) to 42~mJy/beam

1674:    in steps of 6~mJy/beam at 2695~MHz; 

1675:    1.2~mJy/beam ($6\sigma$) to 9.6~mJy/beam

1676:    in steps of 1.2~mJy/beam at 4850~MHz; 

1677:    0.8~mJy/beam ($2\sigma$) to 5.6~mJy/beam

1678:    in steps of 0.8~mJy/beam at 10550 MHz.}

1679:    \label{pi}

1680: \end{figure*}

1681: 

1682: %%Figure 5%%

1683: \begin{figure}

1684: \centerline{\includegraphics[bb = 30 60 555 560,width=15cm,clip]{f5.ps}}

1685:    \caption{Upper panels: Maps of rotation measure between 4850~MHz

1686:    and 10550~MHz, at the resolution of the 2695~MHz map, $4\farcm3$

1687:    (left) and at a resolution of $2\farcm45$ (right). Black contours

1688:    represent rotation measure, with values 140, 180, 220, 260, 300,

1689:    340, 380, and 420~rad/m$^2$. White contours indicate polarized

1690:    intensity at 10550~MHz at the appropriate resolution. Lower panels:

1691:    Left: Map of rotation measure between the four frequency bands of

1692:    the CGPS at $2\farcm45$ (left). Contour levels are the same as in

1693:    the other RM maps. Right: Map of polarized intensity at 4850~MHz

1694:    shown by contours and grayscale, with levels 1.2 mJy/beam to 9.6

1695:    mJy/beam in steps of 1.2 mJy/beam. Overlaid white vectors indicate

1696:    the intrinsic magnetic field directions calculated using the RM map

1697:    at the top right.}

1698:    \label{rmpib}

1699: \end{figure}

1700: 

1701: %%Figure 6%%

1702: \begin{figure}

1703:   \centerline{\includegraphics[bb = 45 30 540 505,width=15cm,clip]{f6.ps}}

1704:    \caption{Map of total intensity at 1420~MHz made by merging

1705:    the CGPS 1420~MHz map with a map taken from the NVSS (see Section 4.2).

1706:    The resolution is about $45\arcsec$. Contour levels are at

1707:    0.8 ($8\sigma$), 1.6, 2.4, 3.2, 4.0, 4.8, and 5.6\,K. Overlaid white 

1708:    contours indicate the polarized intensity at 4850\,MHz with 

1709:    levels at 7, 10, 13, 16, and 19\,mJy/beam.}

1710:    \label{tppi}

1711: \end{figure}

1712: 

1713: %%Figure 7%%

1714: \begin{figure*}

1715: \centerline{\includegraphics[bb = 60 50 565 760,angle=-90,width=16cm,clip]{f7.ps}}

1716:    \caption{Foreground rotation measure for all pulsars within $15\degr$ 

1717:    of DA\,495 as a function of distance from us in kpc. 

1718:    See Section 4.4.1 for source of data.}

1719:    \label{pulsarrm}

1720: \end{figure*}

1721: 

1722: \begin{figure*}

1723: \centerline{\includegraphics[bb = 45 110 480 510,width=16cm,clip]{f8.ps}}

1724:    \caption{Sketch of a possible magnetic field configuration inside DA\,495,

1725:    which contains a dipole and a toroidal component. Dashed contours represent

1726:    the polarized emission as observed at 4850~MHz.}

1727:    \label{magsketch}

1728: \end{figure*}

1729: 

1730: %%Figure 8%%

1731: \begin{figure*}

1732: \centerline{\includegraphics[bb = 55 45 530 515,width=16cm,clip]{f9.ps}}

1733:    \caption{Sketch of the dipole field configuration adopted as a model

1734:    of one component of the magnetic field in DA\,495. The field lines

1735:    have been assumed to have a dipole configuration that is stretched

1736:    vertically by a factor of 2.5. The direction of the magnetic 

1737:    field $\vec{\rm B}$ is indicated. The spin axis of the pulsar

1738:    should be aligned with the dipole axis. The viewing angle $\theta$

1739:    is defined as shown. In Section 4.4.4 we consider this configuration

1740:    viewed from the right.}

1741:    \label{dipole}

1742: \end{figure*}

1743: 

1744: %%Figure 9%%

1745: \begin{figure*}

1746: \centerline{\includegraphics[bb = 80 115 825 460,width=16cm,clip]{f10.ps}}

1747:    \caption{The required electron density $n_e$ as a function of viewing angle

1748:    $\theta$. The dashed lines in both panels correspond to the line of sight

1749:    through the center of the dipole (Fig.~\ref{dipole}), the lower line

1750:    for the observed rotation measure of +170~rad/m$^2$ and the upper

1751:    line for +200~rad/m$^2$, which is the observed value corrected for

1752:    the foreground RM. The solid lines represent the line of sight through

1753:    the dipole at a distance of 1.2~pc from the center, looking through

1754:    the top of the dipole (left) and the bottom of it (right) as displayed

1755:    in Fig.~\ref{dipole}. The three

1756:    lines were calculated from the top for rotation measures of +350, +300,

1757:    and +250~rad/m$^2$. The small dotted rectangles delineate the range of

1758:    parameter values for which the model can provide a reasonable match

1759:    to observed values of RM.}

1760:    \label{rmcalc}

1761: \end{figure*}

1762: 

1763: %%Figure 10%%

1764: \begin{figure*}

1765: \centerline{\includegraphics[bb = 80 115 820 485,width=16cm,clip]{f11.ps}}

1766:    \caption{Orientation of the dipole field inside DA\,495, as determined

1767:    in Sections 4.3 and 4.4. On the left the projection of the dipole field

1768:    onto the plane of the sky is shown. On the right the line of sight

1769:    orientation is indicated with view from the right hand side.}

1770:    \label{oridip}

1771: \end{figure*}

1772: \end{document}

1773: