1: %\documentstyle[12pt,aasms4]{article}
2: \documentclass[12 pt,preprint]{aastex}
3: %\documentstyle[aas2pp4]{article}
4: %\documentstyle[11pt,eqsecnum,aaspp4]{article}
5: %\documentstyle[12pt,amssym,aasms4]{article}
6: %\received{4 August 1988}
7: %\accepted{23 September 1988}
8: %\journalid{337}{15 January 1989}
9: %\articleid{11}{14}
10: %\slugcomment{DRAFT 09/05/01. For submission to ApJ.}
11: \shorttitle{3EG J2006-2321}
12: \shortauthors{Wallace et al.}
13: \begin{document}
14:
15: \title{An AGN Identification for 3EG J2006-2321}
16:
17: \author{P. M. Wallace}
18: \affil{Department of Physics, Astronomy, \& Geology, Berry College, Rome, GA 30149, USA\\
19: \rm{pwallace@berry.edu}}
20:
21: \author{J. P. Halpern}
22: \affil{Columbia Astrophysics Laboratory, Columbia University, New York, NY
23: 10027, USA\\
24: \rm{jules@astro.columbia.edu}}
25:
26: \author{A. M. Magalh\~{a}es}
27: \affil{Instituto de Astronomia, Geof\'{\i}sica, e Ci\^{e}ncias Atmosf\'{e}ricas, \\
28: Universidade de S\~{a}o Paulo, S\~{a}o Paulo - SP 01060-970, BRAZIL\\
29: \rm{mario@iagusp.usp.br}}
30:
31: \author{D. J. Thompson}
32: \affil{Code 661, NASA Goddard Space Flight Center, Greenbelt, MD 20771, USA\\
33: \rm{djt@egret.gsfc.nasa.gov}}
34:
35: \begin{abstract}
36:
37: We present a multiwavelength analysis of the high-energy gamma-ray source
38: 3EG J2006-2321 ($l=18^{\circ}\!.82$, $b=-26^{\circ}\!.26$). The flux
39: of this source above 100 MeV is shown to be variable on time scales of
40: days and months. Optical observations and careful examination of archived
41: radio data indicate that its most probable identification
42: is with PMN J2005--2310, a flat-spectrum radio quasar with a 5-GHz flux density
43: of 260 mJy. Study of the $V=19.3$ optical counterpart indicates a redshift
44: of 0.833 and variable linear polarization. No X-ray source has
45: been detected near the position of PMN J2005--2310, but an X-ray upper limit
46: is derived from ROSAT data. This upper limit provides for a
47: spectral energy distribution with global characteristics similar to those
48: of known gamma-ray blazars. Taken together, these data indicate that
49: 3EG J2006--2321, listed as unidentified in the 3rd EGRET Catalog, is a
50: member of the blazar class of AGN. The 5-GHz radio flux density of this
51: blazar is the lowest of the 68 EGRET-detected AGN. The fact that
52: EGRET has detected such a source has implications for unidentified EGRET
53: sources, particularly those at high latitudes ($|b|>30^{\circ}$), many of
54: which may be blazars.
55:
56: \end{abstract}
57:
58: \keywords{gamma rays: observations}
59: %\keywords{globular clusters,peanut clusters,bosons,bozos}
60:
61: \section{Introduction}
62:
63: Since the earliest days of gamma-ray astronomy, one of the foremost
64: questions in the field has been the identity of discrete sources.
65: In 1972-1973, SAS-2 was the first mission to detect radiation
66: from the Vela and Crab pulsars (Fichtel et al. 1975).
67: Launched in 1975, COS-B built on the
68: success of SAS-2 by detecting 25 point sources. However, only four of
69: those were identified (Swanenburg et al. 1981). As instruments became
70: more sophisticated, the number of identified and unidentified sources
71: has increased. The present list (Hartman et al. 1999, hereafter H99)
72: is comprised of the 271 sources detected by the Energetic Gamma-ray Experiment
73: Telescope (EGRET) on the late Compton Gamma-ray Observatory (CGRO) that
74: display significant flux above 100 MeV. Of these, 169 remain
75: unidentified. The 102 identified sources include a probable association
76: with the radio galaxy Cen A, a solar flare, and the LMC. The remaining
77: sources are pulsars (5), AGNs with low-confidence (27), and AGNs with
78: high-confidence
79: (66). The AGNs are blazars, typically flat-spectrum radio quasars (FSRQs) or
80: BL Lac objects. There is statistical evidence that
81: supernova remnants (Sturner \& Dermer 1995; Esposito et al. 1996), OB
82: associations (Kaaret \& Cottam 1996; Romero, Benaglia, \& Torres 1999),
83: and objects born in the Gould Belt (Gehrels et al. 2000) are also gamma-ray
84: emitters, but no single source type has been conclusively associated with any
85: of these classes.
86:
87: Due to the relatively poor angular resolution of even the best gamma-ray
88: telescopes, efforts to identify sources depend on factors other than
89: spatial coincidence. Pulsars are typically found in the Galactic plane while
90: most blazars are found at high latitude where the diffuse Galactic emission
91: does not overwhelm the source photons. Additionally, variability over hours
92: and days offers evidence against identification as a pulsar (Ramanamurthy
93: et al. 1995) while blazars are known to have strongly variable gamma-ray
94: flux (e.g. Mattox et al. 1997a). But latitude and variability studies
95: cannot be used alone to identify individual gamma-ray sources with high
96: confidence as pulsars or blazars; for pulsars, a clear pulse profile must
97: be detected, while for blazars, multiwavelength studies are necessary as
98: these sources have a distinct broad-band signature. They are variable
99: at many frequencies, feature flat radio spectra, show variable polarization
100: at radio and optical frequencies, display power-law spectra at X-ray and
101: gamma-ray energies, and have moderate to large redshifts.
102:
103: Recently there have been a number of papers published that describe
104: efforts to identify individual gamma-ray sources using multiwavelength
105: analyses (e.g., Mukherjee et al. 2000; Halpern et al. 2001; Mirabel
106: \& Halpern 2001; Reimer et al. 2001a). In the present paper a
107: multiwavelength study of 3EG J2006--2321 is presented, and the data
108: indicate that this source, listed as unidentified in H99, is a blazar.
109: In \S\S 2-5 the relevant data are presented; this is followed in
110: \S 6 by a short summary of the multifrequency data. The weak flux
111: density of the radio counterpart and its relevance to other
112: high-latitude unidentified EGRET sources is discussed in \S 7. In
113: \S 8 conclusions and suggestions for future work are summarized.
114:
115: \section{Gamma-ray Observations}
116:
117: The average flux displayed by 3EG J2006--2321 from 1991 April through
118: 1995 September (the time span covered by H99) is
119: $(7.3\pm2.7)\times10^{-8}$ photons~cm$^{-2}$~s$^{-1}$ in the energy range
120: $E>100$ MeV, making it a relatively faint
121: EGRET source. The position of 3EG J2006--2321 is $l=18^{\circ}\!.82$,
122: $b=-26^{\circ}\!.26$ and the mean radius of the 95\% confidence contour is
123: about $0^{\circ}\!.67$. The source is well-isolated; its closest gamma-ray
124: neighbor is more than $8^\circ$ away and its high Galactic latitude
125: ensures that the Galactic diffuse radiation does not interfere with the
126: source photons. The EGRET spectrum (from 30 MeV to 30 GeV) is
127: consistent with a power-law photon index $\Gamma = 2.47\pm0.44$.
128:
129: \begin{figure}[t!]
130: \epsscale{1.0}
131: \plotone{f1.eps}
132: \caption{Main plot: Flux history of 3EG J2006--2321 from 1991 April
133: to 1995 September, the time span covered by H99.
134: The first five points are averages over individual EGRET viewing
135: periods (VP's); each of the final two points represent averages over two
136: VP's; in no viewing periods after 1991 November
137: was the source detected. Inset plot: Light curve of 3EG J2006--2321
138: during VP 13.1, indicated by the third point on the main plot.
139: During this VP (1991 October 31-November 7) the source
140: was highly variable, with a peak flux nearly 24 times that of the average
141: over all VP's. The horizontal dashed lines indicate the mean flux for
142: the relevant time spans. \label{fig1}}
143: \end{figure}
144:
145: A light curve of this source from CGRO Phase 1 through Cycle 4 is shown
146: in Figure 1. The first detection by any telescope of 3EG J2006--2321 occured during
147: EGRET Viewing Period (VP) 5.0 (1991 July 12-26) when it was detected
148: with a flux of $(44.1\pm12.7)\times10^{-8}$ photons~cm$^{-2}$~s$^{-1}$ at
149: a significance of 4.4$\sigma$. During this time the source was
150: 28$^{\circ}\!.$6 from the instrument axis, and there is no evidence that
151: the source was variable on a $\sim1$-day time scale during this
152: observation. Three weeks later 3EG J2006--2321
153: was 13$^{\circ}$6 from the EGRET axis and was not detected. From 1991
154: October 31-November 7 (EGRET VP 13.1) the source was again
155: 13$^{\circ}\!.$6 off-axis and was detected with a flux of
156: $(32.7\pm8.7)\times10^{-8}$ photons~cm$^{-2}$~s$^{-1}$ at the 4.8$\sigma$
157: level. Although 3EG J2006--2321 was within 20$^{\circ}$ of the instrument
158: axis during six subsequent VP's, it was never again detected by EGRET.
159:
160: There is evidence that this source is variable over time scales of weeks.
161: McLaughlin et al. (1996) describe a statistic, called the \it{variability index}
162: \rm $V$. This statistic is used to calculate the probability that a given
163: light curve is consistent with an intrinsically nonvariable source and is
164: defined as follows. The $\chi^2$ for the light curve is
165: \begin{equation}
166: \chi^2=\sum_{i=1}^N \frac{(F-\bar{F}_i)^2}{\sigma_i^2}
167: \end{equation}
168: where $N$ is the number of observations, $F_i$ is the detected flux during the
169: \it{i}\rm th observation, $\bar{F}$ is the mean flux for the viewing period, and
170: $\sigma_i$ is the 1 $\sigma$ flux uncertainty of the \it{i}\rm th observation.
171: If $Q$ is the probability of obtaining a value of $\chi^2$ equal to or greater
172: than the empirical $\chi^2$ from an intrinsically nonvariable source, then
173: $V\equiv-\log Q$. Roughly speaking, if $V<1$ then the source
174: in question is not considered to be variable; if $V>1$ then
175: the source may be variable. The variability index for the long-term
176: light curve in Figure 1 is 2.9, corresponding to a probability of
177: $\sim0.0012$ that these data are produced by an intrinsically
178: nonvariable source.
179:
180: During VP 13.1, the source was variable on a $\sim1$ day time scale.
181: The gamma-ray observations of 3EG J2006--2321 during this VP
182: are discussed in detail in Wallace et al. (2000); the relevant information
183: will be summarized here. A light curve is shown in the inset of Figure 1.
184: The first two days are broken down into four 12-hour periods;
185: the remaining points represent full days. The peak flux, centered in MJD
186: 48560.25, is $(1.75\pm0.53)\times10^{-6}$ photons~cm$^{-2}$~s$^{-1}$; this 12-hour
187: detection has a significance of 5.4$\sigma$. The ratio of peak to average
188: flux for the VP is 5:1, and the ratio of peak flux to the overall EGRET
189: mean flux for 3EG J2006--2321 is nearly 24:1. It is very unlikely that the
190: source is not variable on this short time scale. Applying a $\chi^2$
191: test to the light curve yields a variability index of 3.2, corresponding
192: to a probability of 0.0006 that these data are produced by an intrinsically
193: nonvariable source.
194:
195: \section{Radio Observations}
196:
197: \begin{figure}[t!]
198: \epsscale{1.0}
199: \includegraphics[angle=90]{f2}
200: \caption{Grey-scale sky map of 3EG J2006--2321.
201: The vertical scale on the right indicates the gamma-ray flux in units
202: of $10^{-8}$ photons~cm$^{-2}$~s$^{-1}$.
203: The positions of 5-GHz radio
204: sources from the PMN survey are marked with triangles and numbered
205: as in Table 1.
206: %51. PMN J2003--2333, 2. PMN J2005--2310, 3. PMN
207: %5J2007--2335, 4. PMN J2008--2338, 5. PMN J2008--2305, 6. PMN J2009--2250.
208: \label{fig2}}
209: \end{figure}
210:
211: To find a radio counterpart to 3EG J2006--2321, we first search its error
212: circle for 5-GHz sources since gamma-ray blazars typically have
213: significant flux densities at this frequency (Mattox et al. 1997b).
214: At southern declinations, the Parkes-MIT-NRAO (PMN) survey
215: (Griffith \& Wright 1993) is the deepest at 5 GHz; we therefore look
216: to these sources. We expect that if 3EG J2006--2321 is indeed a blazar, it,
217: like other EGRET blazars, will be related to a flat-spectrum radio source,
218: with $-0.5\lesssim\alpha_r$ (where $F_{\nu}
219: \propto \nu^{\alpha_r}$). The NED database lists
220: six 5-GHz sources within the EGRET 99\% error contour; these sources
221: are listed in Table 1 with their 5-GHz flux densities, angular
222: separations from the gamma-ray position, and radio spectral indexes.
223: Figure 2 indicates their positions on a map of the gamma-ray intensity
224: of 3EG J2006--2321. These radio candidates are discussed below in order of
225: increasing right ascension.
226:
227: \it{1. PMN J2003--2333.} \rm This source has a 5-GHz flux density ($S_5$)
228: of 43 mJy, the weakest of all the candidate sources, and it is $42^{\prime}$
229: from the EGRET position, between the 95\% and 99\% confidence contours.
230: The nearest source to it in the
231: NRAO/VLA Sky Survey (NVSS; Condon et al. 1998) has a 1.4-GHz flux
232: density of 15.3 mJy, which, if identified with the PMN source
233: would give it a rising radio spectrum with an index
234: of 0.83. However, the beam size of the Parkes telescope is
235: $4^{\prime}\!.2$, and there are at least four NVSS sources within
236: $3^{\prime}$ of the PMN position having a combined flux density of 45 mJy
237: at 1.4 GHz. Given the weakness of all of these sources, and the
238: possibility of source confusion at 5 GHz, we do not consider this
239: to be a promising candidate.
240: %There is no known counterpart at other frequencies. This
241: %spectral index is beyond the flat-spectrum range, so this source is
242: %excluded from consideration.
243:
244: \it{2. PMN J2005--2310.} \rm This source is the brightest of the six, with a
245: 5-GHz flux density of 260 mJy. It is located well within the 50\%
246: confidence contour, at $11^{\prime}$ from the EGRET position. It appears
247: in the NVSS with a 1.4-GHz flux density of 302 mJy. We calculate
248: $\alpha_r=-0.12$ from these data. If we include the 365 MHz flux
249: density of 260 mJy from the Texas survey (Douglas et al. 1996) we
250: find that the radio spectrum is flatter, with $\alpha_r=0.002$.
251:
252: \it{3. PMN J2007--2335.} \rm This source has a 5-GHz flux density of
253: 141 mJy and is located near the 68\% confidence contour, at $23^{\prime}$
254: from the EGRET position. The nearest NVSS source has a flux density of 318 mJy,
255: giving it a spectral index of --0.65 between 1.4 and 4.85 (5)
256: GHz, below but near the threshold of -0.5 for flat sources. Its
257: lower-energy radio flux is relatively high; at 365 MHz it has a
258: flux density of 2.98 Jy. If this point is taken into account, its
259: spectral index falls to --1.2. There is also the possibility of
260: some confusion with another NVSS source $4^{\prime}\!.7$ away
261: that has a flux density of 177.5 mJy.
262: However, in light of its relative
263: brightness and near-flat spectrum, we do not yet exclude
264: it from our analysis.
265:
266: \it{4. PMN J2008--2338.} \rm This source has a 5-GHz flux density of 82
267: mJy and is $35^{\prime}\!.6$ from the EGRET position, just inside the 95\%
268: confidence contour. It has a 1.4-GHz counterpart with
269: flux density 275 mJy, giving it a steep spectrum with an index of
270: --0.97. It has a 993-mJy counterpart at 365 MHz; this does not
271: appreciably change $\alpha_r$. Only one EGRET blazar listed in H99,
272: 3EG J1937-1529, has a comparable radio spectral index (-0.96).
273: Therefore the steepness of the radio spectrum is not unheard of
274: among EGRET blazars, but it is rare. The combination of the steep
275: radio spectrum, weak $S_5$, and large separation from the gamma-ray
276: position leads us to exclude this source from consideration.
277:
278: \it{5. PMN J2008--2305.} \rm This source has a 5-GHz flux density of 47
279: mJy and is $41^{\prime}$ from the EGRET position, just inside the 95\% confidence
280: contour. It has a 1.4-GHz counterpart with
281: flux density 41 mJy, giving it a flat spectrum with an index of +0.05.
282: There are no known counterparts at other frequencies.
283:
284: \it{6. PMN J2009--2250.} \rm This source has a 5-GHz flux density of 49
285: mJy and is $55^{\prime}$ from the EGRET position, between the 95\% and 99\%
286: confidence contours. It has a 1.4-GHz counterpart
287: with flux density 97 mJy, giving it a spectral index of --0.55,
288: below but near the flat-spectrum threshold. If its 370-mJy
289: counterpart at 365 MHz is taken into account, the spectral index
290: falls to -0.78.
291:
292: After excluding the two weak steep-spectrum sources, we are left
293: with PMN J2005--2310, PMN J2007--2335, PMN J2008--2305, and PMN
294: J2009--2250 as candidates for association with 3EG J2006--2310.
295: For gamma-ray blazars, the 5-GHz flux density has been found to correlate
296: linearly with peak gamma-ray flux at the 99.998\% confidence level
297: (Mattox et al. 1997b). (It should be noted that this correlation is not
298: well-understood and it serves only as a coarse guide. See \S 8.) There
299: are only ten EGRET blazars with peak flux above $10^{-6}$
300: photons~cm$^{-2}$~s$^{-1}$, and they all have $S_5>1.0$ Jy. We therefore expect to
301: find a relatively bright 5-GHz counterpart to 3EG J2006--2321 with its peak
302: flux of $(1.75\pm0.53)\times10^{-6}$ photons~cm$^{-2}$~s$^{-1}$.
303: However, the brightest of these candidates, PMN J2005--2310, has
304: $S_5=260$ mJy. Two of the other candidates are far dimmer still; the
305: flux densities of PMN J2008--2305 and PMN J2009--2250 are $>5$ times
306: weaker than that of PMN J2005--2310. Therefore if we are to
307: identify 3EG J2006--2321 with any of the four remaining radio sources, we
308: see that PMN J2005--2310 is the most compelling candidate, followed by
309: PMN J2007--2335. Another point in favor of these two sources is
310: that their angular separations from the EGRET position are smaller
311: than for the weaker sources. We therefore regard PMN J2008--2305
312: and PMN J2009--2250 as highly improbable candidates and restrict
313: ourselves to considering only PMN J2005--2310 and PMN J2007--2335.
314:
315: \section{Optical Observations}
316:
317: \subsection{PMN J2007--2335}
318:
319: On 2001 June 29, six 300-s exposures of the field containing PMN
320: J2007--2335 were taken with a CCD on the 1.3m telescope of the
321: MDM Observatory. The combined $R$-band
322: image reveals a normal-looking galaxy at the position of the NVSS source,
323: (J2000) $20^{\rm h}07^{\rm m}25^{\rm s}\!.98,\
324: -23^{\circ}34^{\prime}35^{\prime\prime}\!.6$,
325: with $R\approx 19.8$. It is highly unlikely that such a galaxy is the
326: source of the high-energy gamma rays; therefore PMN J2007--2335 is
327: rejected as the counterpart of 3EG J2006--2321, leaving only
328: 3EG J2006--2321 as a viable candidate.
329:
330: \subsection{PMN J2005-2310}
331:
332: Three CCD images of 200-s exposure
333: in the $V$ band centered on the position of PMN J2005--2310 were taken with
334: the MDM Observatory 2.4m telescope on 2000 July 24.
335: The central $80^{\prime\prime} \times 80^{\prime\prime}$
336: of the combined image is shown in Figure 3.
337: A point-like optical counterpart with $V=19.3$ was found at
338: (J2000) $20^{\rm h}05^{\rm m}56^{\rm s}\!.59,\
339: -23^{\circ}10^{\prime}27^{\prime\prime}\!.0$, within $1^{\prime\prime}$ of
340: the NVSS position of PMN J2005--2310, consistent with their
341: combined astrometric uncertainty.
342: An object at this position and comparable brightness can be seen on
343: various Digitized Sky Survey plates.
344: Galactic extinction is moderate at these latitudes; the absorption in the $V$
345: band is $\approx 0.5$ mag \citep{sch98},
346: giving the source an intrinsic $V=18.8$.
347:
348: An optical spectrum of this object was obtained with the Goldcam
349: spectrometer on the 2.1m telescope of the Kitt Peak National Observatory (KPNO)
350: on 2000 June 3.
351: A single 3000-s exposure using grating 240 and a slit
352: width of $1^{\prime\prime}\!.9$ yielded a resolution of $\approx 5$ \AA.
353: The target was acquired by blind offset from a bright star $10^{\prime\prime}$
354: to the east (see Figure 3), and the slit was oriented at the parallactic
355: angle, in this case N-S, which is essential at such a southerly declination
356: in order to obtain accurate spectrophotometry.
357: The fully reduced spectrum is shown in Figure 4.
358: It features a single broad emission line at 5129 \AA\ which,
359: from the absence of other emission lines over the observed spectral range,
360: we identify as Mg \small{II} \normalsize $\lambda$2798 at $z=0.833$.
361: The Mg \small{II} \normalsize line has the unique property that there is a
362: range in redshift over which it is the only prominent quasar emission
363: line that falls in the region observed by the typical optical spectrograph
364: in first order, which covers at most a factor of 2 in wavelength. Reliable
365: redshifts are often achieved using Mg \small{II} \normalsize only. The neighboring broad emission
366: lines are [C \small{III}\normalsize] $\lambda$1909 and H$\gamma$ $\lambda$4340,
367: which at $z=0.833$
368: would be redshifted to 3499\AA\ and 7955\AA, respectively, both outside of
369: our observed spectral range. A narrow forbidden line that is sometimes seen
370: in quasars is [O~\small{II}\normalsize] $\lambda$3727, and there is a marginal detection of
371: this line at 6832\AA\ in our spectrum of PMN J2005--2310. The full width
372: at half maximum of the broad Mg~II line is 3400 km s$^{-1}$, and its rest-frame
373: equivalent width is 15.7 \AA, larger than that of BL Lac objects, but typical
374: of FSRQs that have been identified with EGRET sources. The continuum flux
375: is consistent with the $V$ magnitude measured from the CCD image obtained
376: 52 days later.
377:
378: \begin{figure}[ht!]
379: \epsscale{0.6}
380: \plotone{f3.eps}
381: \caption{Portion of a $V$-band image from the MDM 2.4m,
382: with the point-like optical counterpart of PMN J2005--2310 circled.
383: The field shown is $80^{\prime\prime} \times 80^{\prime\prime}$,
384: and the position of the $V = 19.3$ optical counterpart is
385: (J2000) $20^{\rm h}05^{\rm m}56^{\rm s}\!.59,\
386: -23^{\circ}10^{\prime}27^{\prime\prime}\!.0$, within $1^{\prime\prime}$
387: of the NVSS position. \label{fig3}}
388: \end{figure}
389:
390: \begin{center}
391: \begin{figure}[ht!]
392: \epsscale{0.6}
393: \includegraphics[width=4in,angle=270]{f4}
394: \caption{Spectrum of the optical counterpart of PMN J2005--2310
395: from the KPNO 2.1m. The single broad emission line is identified as Mg \small{II} \normalsize $\lambda 2798$
396: and there is a possible detection of the forbidden line [O \small{II}\normalsize] $\lambda$3727.\label{fig4}}
397: \end{figure}
398: \end{center}
399:
400: Strong and variable optical polarization is a characteristic blazar
401: signature \citep{pes97}, so optical polarimetry has been performed for
402: PMN J2005--2310. The source was observed with the IAGPOL
403: imaging polarimeter \citep{mag96} at the 61 cm IAG-USP telescope at
404: the Laborat\'orio Nacional de Astrof\'{\i}sca (LNA) on 2000 August~5
405: and on 2001 June~15. The polarimeter is a modification of the observatory's
406: direct CCD camera to allow for high-precision imaging polarimetry.
407: The first element in the beam is a 51mm diameter rotatable, achromatic
408: half--wave retarder followed by a Savart plate built by Opto
409: Eletr\^onica, S\~ao Paulo. One polarization modulation cycle is covered
410: for every $90\arcdeg$ rotation of the waveplate. The simultaneous
411: observations of the two beams allows observing under non-photometric
412: conditions at the same time that the sky polarization is practically
413: cancelled \citep{mag96}. Further details are given by \citet{kay99}.
414:
415: The log of observations is presented in Table~\ref{imagepol}.
416: CCD exposures were taken through the $V$ filter with the waveplate
417: rotated through 16 positions (2000 Aug 5) and 12 positions (2001 June 15)
418: $22.5\arcdeg$ apart. The exposure time at each position was 300\,s.
419: The instrumental Stokes parameters $Q$ and $U$ were then obtained, as
420: well as the theoretical (i.e., photon noise) and measurement errors.
421: The latter are estimated from the residuals of the observations at each
422: waveplate position angle (${{\psi}_i}$) with regards to the
423: expected cos ($4{{\psi}_i}$) curve and are quoted in Table~\ref{imagepol};
424: they are consistent with the photon noise errors \citep{mag84}.
425:
426: Due to the relative faintness (for the telescope) of PMN~J2005--2310,
427: care was taken to ensure that the estimates of sky values per pixel
428: were robust. We found that the {\it mode} option in the IRAF photometry
429: package gave consistent results for different annuli around the object.
430: The instrumental $Q$ and $U$ values were then
431: converted to the equatorial system from data of two polarized standard stars
432: \citep{tur90} obtained in each
433: night. The instrumental polarization was measured to be less than
434: 0.03\% from observations of unpolarized stars in the same night. We
435: have hence applied no such correction to our data.
436:
437: Table~\ref{imagepol} shows that on 2000 August 5, PMN~J2005--2310
438: was significantly polarized while that was not the case on 2001 June 15.
439: We conclude that PMN~J2005--2310 has a highly variable optical polarization,
440: a characteristic blazar signature, lending further credence to the association
441: of 3EG~J2006--2321 with PMN~J2005--2310.
442:
443: The imaging polarimetry includes data on targets angularly close to
444: PMN~J2005--2310, thereby providing the means to estimate the
445: interstellar (IS) polarization towards that general direction. We used
446: a package, {\it pccdpack}, specially written for the analysis of field
447: stars \citep{per00}. A sample of 25 field stars in the 2000 images and
448: 23 field stars in the 2001 images provided the weighted average polarization
449: values quoted in Table~\ref{imagepol}. The combined, weighted average for the
450: IS polarization is (1.47$\pm$0.01)\% @ 21.1$\arcdeg$. This
451: IS polarization estimate of about 1.5\% is entirely consistent
452: with the expected maximum percent IS polarization $P{_{max}} \le 9 E_{B-V}$
453: \citep{ser75}
454: and A$_V$=0.5 mag. Incidentally, this equatorial position angle
455: towards the direction of PMN~J2005-2310 corresponds to a Galactic position
456: angle $\theta_{G} = 91\arcdeg$, i.e., parallel to the Galactic plane,
457: confirming that we are measuring an IS component with the field stars.
458: The weighted IS polarization is much smaller than the 2000 August
459: observed polarization of PMN~J2005--2310 and hardly effects it; the intrinsic
460: polarization of the blazar on that date becomes (13.5$\pm$2.3)\% @
461: 61$\arcdeg$.
462:
463: \section{X-ray Observations}
464:
465: The region surrounding 3EG J2006--2321 has had little exposure to X-ray
466: instruments. The sole data come from the ROSAT All-Sky Survey (Voges
467: et al. 1999; Voges et al. 2000) during 1990 October 6-26; in these data (both the Bright and
468: Faint Source Catalogs) there is no source within $20^{\prime}$ of the position
469: of PMN J2005--2310. This places an upper limit on the X-ray flux of the
470: source in question. A typical dim source in this region has a PSPC count
471: rate of $\approx 2.4\times10^{-2}$ counts s$^{-1}$.
472: Assuming a power-law spectrum with photon index
473: with $\Gamma=2.0$ and Galactic $N_{\rm H} = 8.5 \times 10^{20}$~cm$^{-2}$,
474: this corresponds to an unabsorbed flux between 0.1 keV and 2.0
475: keV of $\approx 7 \times 10^{-13}$ ergs~cm$^{-2}$~s$^{-1}$, which we
476: adopt as an upper limit for PMN J2005--2310.
477:
478: \begin{figure}[h!]
479: \epsscale{1.0}
480: \plotone{f5.eps}
481: \caption{Spectral energy distribution for 3EG J2006--2321. The error
482: bars of the radio and optical points fall within the diamond markers,
483: and markers with arrows indicate upper limits. The X-ray upper limit
484: is derived from ROSAT All-Sky Survey data and the gamma-ray limits are
485: statistical.
486: The data shown were not taken simultaneously. As the gamma-ray data
487: are representative of 3EG J2006--2321 in a high state, the gamma-ray
488: contribution to the bolometric energy flux is probably exaggerated
489: in this SED. \label{fig5}}
490: \end{figure}
491:
492: \section{Summary of Multifrequency Data}
493:
494: As 3EG J2006-2321 is faint, well off the Galactic plane, and has no
495: conspicuous objects of astronomical interest (e.g., SNR's, OB
496: associations, X-ray sources) within its error circle, it
497: has had relatively little previous exposure to telescopes in any
498: frequency range. However, from the analysis of archived data in
499: the radio, X-ray, and gamma-ray regions, and from the results of
500: optical spectroscopy and polarimetry, it is evident that the source
501: is a member of the blazar class of AGN. Its gamma-ray flux is variable
502: on time scales of days and months. Its radio counterpart is a
503: flat-spectrum radio quasar (FSRQ) with a large redshift and strong,
504: variable optical polarization. Its X-ray upper limit makes its spectral
505: energy distribution (SED, Figure 5) consistent with a bimodal shape,
506: similar to SED's of known gamma-ray blazars (e.g., von Montigny
507: et al. 1995). Assuming a cosmology with $\Omega=1$ and $\Lambda=0$
508: and using the latest value of the Hubble constant $H_o=72$ km/s/Mpc
509: (Freedman et al. 2001), the peak gamma-ray isotropic
510: luminosity between 100 MeV and 10 GeV for 3EG J2006--2321
511: is $\sim4\times10^{47}$ ergs~s$^{-1}$, within the range of other flaring
512: gamma-ray AGN (e.g., McGlynn et al. 1997; Mattox et al. 1997a, 2001a).
513: Although there is evidence that the comological constant
514: is nonzero (Perlmutter et al. 1999), $\Lambda=0$
515: is assumed here only to normalize comparisons with previous
516: calculations.
517:
518:
519: \section{On the Low Flux Density of PMN J2005--2310 and
520: Implications for High-latitude Unidentified Sources}
521:
522: It may be argued that the present source is unusual among gamma-ray
523: blazars as its 5-GHz flux density is weaker than that of any other
524: EGRET AGN and much weaker than any other blazar with comparable
525: peak gamma-ray flux. [The EGRET blazars with the lowest
526: 5-GHz flux densities on record are 3EG J0743+5447 (272 mJy) and
527: 3EG J2158--3023 (407 mJy). The weakest $S_5$ among EGRET blazars
528: with peak high-energy gamma-ray flux $>10^{-6}$ photons~cm$^{-2}$~s$^{-1}$
529: is 1080 mJy, from PKS 1406--076, associated with 3EG J1409--0745.]
530: However, $S_5$ of PMN J2005--2310 is not anomalous among radio
531: counterparts of gamma-ray blazars; it is merely low. Figure 6 shows
532: the distribution of 5-GHz flux densities of EGRET blazars;
533: PMN J2005--2310 is shown in black, other sources with peak gamma-ray
534: flux $>10^{-6}$ photons~cm$^{-2}$~s$^{-1}$ are gray. It can be seen that while
535: the radio flux density of the present source is weak, it conforms to
536: the prevailing distributions and does not alter them in any significant
537: way. Additionally, Mattox et al. (1997b), as was mentioned in \S 3, have
538: shown that for confirmed EGRET blazars, the 5-GHz flux density and
539: the peak gamma-ray flux are correlated. Mirabel et al. (2000) indicate
540: that this correlation is not linear, but that there is a trend toward
541: low $S_5/[F(>100)$ MeV] with increasing peak gamma-ray flux, which suggests
542: that the identifications of Mattox et al. (2001b) are incomplete,
543: and that low-flux radio sources should be counterparts of unidentified
544: EGRET sources. The present source is in accordance with this trend;
545: it has the weakest 5-GHz flux density of any EGRET blazar, but has
546: displayed the seventh-highest peak gamma-ray flux value. We conclude that,
547: ultimately, 3EG J2006--2321 is not due any special attention for the weakness of
548: its radio counterpart.
549:
550: \begin{figure}[h!]
551: \epsscale{1.0}
552: \plotone{f6.eps}
553: \caption{Distribution of 5-GHz flux densities from EGRET blazars.
554: The bins are 0.25 Jy wide. The black rectangle represents 3EG J2006--2321
555: and the gray rectangles represent the 10 other blazars with peak
556: $E>100$ MeV flux $>10^{-6}$ photons~cm$^{-2}$~s$^{-1}$. There is an additional
557: source at 44.9 Jy; the peak $E>100$ MeV flux of this source is below
558: $10^{-6}$ photons~cm$^{-2}$~s$^{-1}$. The blazars used for this list are
559: the ``high-confidence'' AGN found in H99. The 2-Jy counterpart to
560: 3EG J2016+3657 (Mukherjee et al. 2000) is also included. \label{fig6}}
561: \end{figure}
562:
563: There are, however, implications of EGRET's ability to detect such a source.
564: Attention was first drawn to 3EG J2006--2321 when the EGRET database
565: was systematically searched for sources exhibiting
566: variability on short ($\sim1$ day) time scales (Wallace et al.
567: 2000). All the 3EG sources were examined across all VP's from
568: 1991 April to 1995 September, and 3EG J2006--2321 was the only
569: source to show new and compelling evidence of such variability.
570: All other instances of strong short-term variability uncovered by this
571: search had already been noted by investigators, connected with
572: previously-known AGN \footnote{The gamma-ray variability of 3EG
573: J0433+2910 led to its identification as a BL Lac object (Lundgren
574: et al. 1995; Wallace et al. 2001) but its flux variation was not on
575: a short time scale.}, and duly reported (e.g., Mukherjee et al. 1996;
576: Mattox et al. 1997a; McGlynn et al. 1997). The point to make about 3EG
577: J2006--2321 is not that its 5-GHz flux is so low, but that its
578: gamma-ray variability was strong enough for it to be picked out
579: by EGRET. There are probably other blazars with $S_5\lesssim500$
580: mJy near EGRET's threshold of detection, but no other happened to
581: flare so dramatically while in the EGRET field of view.
582: (Calculating the probability of detecting a flare
583: from such a near-threshold EGRET blazar is not practical.
584: Even if one assumed that all $|b|>30^{\circ}$ unidentified sources
585: are AGN whose variations were not detected in Wallace et al. (2000)
586: and that the population of known EGRET blazars is a good model,
587: the uncertainties inherent in such a calculation would render
588: the result meaningless.) Although a strongly variable
589: flux from a gamma-ray source indicates a probable blazar
590: identification, the reverse not true; that is, not all blazars
591: are indicated in the EGRET data by strongly variable flux
592: (McLaughlin et al. 1996). This point is especially relevant
593: for high-latitude sources, since they received, on average,
594: less exposure to EGRET than sources near the Galactic plane (H99),
595: and are less likely to be \it observed \rm to be variable.
596: We emphasize that the EGRET data do not rule out the possibility
597: that many unidentified sources with $|b|>30^{\circ}$ are blazars
598: similar to 3EG J2006--2321, with $S_5\lesssim500$ mJy.
599:
600: Of course, there are other possibilities. There is evidence that
601: a population of dim unidentified sources at low and middle latitudes
602: is associated with the Gould Belt (Gehrels et al. 2000), which reaches
603: a maximum of 30$^{\circ}$ above the Galactic center. There may be
604: some ``stragglers'' from the Belt at latitudes higher than this, but
605: even if this is the case the number of such sources must be small.
606: Galaxy clusters may be responsible for a very small number of
607: unidentified EGRET sources at high latitudes (Colafrancesco et al. 2001)
608: but any high-energy gamma-ray emission from such clusters
609: probably falls below EGRET's sensitivity (Reimer \& Sreekumar 2001b).
610: It is also possible that the Galactic halo houses
611: some of the persistent EGRET sources (Grenier et al. 2001), but to date
612: this is only a conjecture. It remains that the only identified
613: gamma-ray sources above 30 degrees Galactic latitude are blazars
614: with flat-spectrum radio counterparts, and we suggest that
615: there are others among the 30 unidentified sources with $|b|>30^{\circ}$.
616:
617: The present case brings to mind weaknesses in standard identification
618: methods that prevent these blazars from being identified. These
619: weaknesses cannot be easily overcome, as they are manifestations
620: of insufficient data and theory; however, it is important
621: to be reminded of them. It is common practice (Thompson et al. 1995; H99)
622: to use 5-GHz sources in the selection of radio counterparts.
623: Specifically, radio sources that are loud ($\gtrsim1$ Jy) and flat
624: ($-0.5\lesssim\alpha_r$) at this frequency have been
625: considered to be top candidates for association with EGRET AGN.
626: There are two weaknesses to this approach worth mentioning here.
627: First, \it{in general} \rm it does not easily work for dim
628: ($S_5\lesssim500$ mJy) radio sources, because the sky density
629: of such sources is high and source confusion becomes a problem,
630: especially for weak EGRET sources with large error circles.
631: In the present work we were forced to look at weak 5-GHz sources
632: that were not considered in the original identification process;
633: because of the low number of 5-GHz candidates we are able to
634: determine the appropriate counterpart. Also, as Bloom \&
635: Dale (2001) have noted, the use of this frequency to establish
636: the loud, flat-spectrum nature of radio counterparts to gamma-ray
637: AGN has no truly compelling physical justification but
638: is a necessity; there are no complete radio surveys at higher
639: frequencies with sensitivity below $\sim1$ Jy. Indeed, 3EG J0743+5447
640: ($S_5=272$ mJy) is representative of a small group of EGRET blazars
641: that are fairly dim and fairly flat at 5 GHz but have been found to
642: have brighter and flatter spectra extending beyond 200 GHz (Bloom et
643: al. 1997); 3EG J2006--2321 may be similar.
644:
645: Other, more sophisticated means of identifying EGRET sources with
646: flat-spectrum radio sources have been tried. In the most recent of these,
647: Mattox, Hartman, \& Reimer (2001b) searched for potential
648: radio counterparts to all
649: sources listed in H99, allowing for sources with arbitrarily low $S_5$.
650: Their approach is a quantitative one; for a given radio source and a
651: given EGRET source the probability of association is calculated, taking
652: into account certain properties of the radio source. Among these properties
653: are the 5-GHz flux density, the spectral index (when available) near 5
654: GHz, the angular separation from the EGRET position, and the
655: sky density of sources at least as bright and flat as the one in question.
656: The researchers list 46 blazar identifications with a ``high probability''
657: of being correct, and 37 additional radio associations that are
658: considered to be ``plausible''. The former list includes no unidentified
659: sources and the latter includes 15 unidentified sources, none of which
660: are in the region $|b|>30^{\circ}$. This does not mean that there are no
661: blazars among this group; by their generic calculation 3EG J2006--2321
662: does not meet the criteria for either list.\footnote{However, 3EG
663: J2006--2321 is included on the ``plausible'' list in light of its gamma-ray
664: variability, which is not factored into their standard calculation.} A
665: radio source must have a $\sim0.04$ probability of association
666: with the gamma-ray source to be included among the plausible associations.
667: The probability of association of PMN J2005--2310 with 3EG J2006--2321
668: was calculated to be 0.015, and the four ``high-probability'' blazars with
669: $S_5<1$ Jy have more than 10 times the probability of being associated
670: with their radio counterparts than does 3EG J2006--2321 with PMN J2005--2310.
671: As mentioned above, high-latitude sources are the most likely to be
672: indentifiable as blazars; it is worth noting that if any of these 30
673: sources are blazars, and it seems likely that \it some \rm should be,
674: the method of Mattox et al. (2001b) is not helpful for indicating them.
675: Stated another way, although this method may not prove that a particular
676: blazar is the true identification with high confidence, that does not mean
677: that such an identification is false with high confidence.
678:
679: The question should be asked: is it possible to make associations
680: of non-variable EGRET sources with weak 5-GHz radio counterparts?
681: If so, new search methods must be used. High-latitude sources
682: are the most likely to be identified as blazars and have
683: the easiest error circles to search because in the region
684: $|b|>30^{\circ}$ the sky densities of gamma-ray sources and
685: potential radio counterparts are low, and because in this region the
686: diffuse Galactic gamma radiation does not wash out source photons. Careful
687: searches for blazars within high-latitude EGRET error circles may be
688: helpful for pointing out likely identifications, and cases
689: in which no blazar can be found within the error circle would be interesting.
690: Such work has already begun; Bloom \& Dale (2001) have searched the
691: error circles of some high-latitude unidentified sources and are in
692: the process of monitoring optical counterparts of relevant flat-spectrum
693: radio sources. Their candidate radio counterparts have 5-GHz
694: flux densities between 33 mJy and 440 mJy. Optical spectroscopy
695: is planned for some of these sources. Similar work on other sources
696: is encouraged. The present work should embolden such investigations
697: as it provides additional evidence that EGRET has detected blazars with
698: radio properties similar to those for which they are searching.
699:
700: \section{Conclusion}
701:
702: From analysis of archived radio, X-ray, and gamma-ray data and our own
703: optical spectroscopy and polarimetry, we conclude that 3EG J2006-2321
704: is a member of the blazar class of AGN.
705: This identification is interesting because it is a reminder that EGRET
706: is capable of detecting blazars with $S_5$ on the order of a fourth of
707: a Jansky. The remaining EGRET unidentified sources most likely to be
708: identified as blazars are in the region $|b|>30^{\circ}$.
709: Further searches for possible radio and optical counterparts within
710: the error circles of the 30 unidentified sources in this region are
711: encouraged.
712:
713: While the present analysis is sufficient to identify 3EG J2006--2321,
714: little more can be said about the source, and no conclusions are reached
715: regarding beaming and radiation mechanisms of gamma-ray-bright AGN.
716: Complete and simultaneous multiwavelength observations are needed
717: to constrain blazar models. The Gamma-ray Large Area Space Telescope
718: (GLAST), scheduled for launch in 2006, is expected to uncover thousands
719: of gamma-ray blazars and other high-energy sources; however, in order
720: to realize a full return of GLAST science, these sources must be
721: observed not just in gamma rays, but across the electromagnetic spectrum.
722:
723: \acknowledgments
724:
725: PMW gratefully acknowledges support from the NASA/ASEE Summer Faculty
726: Fellowship Program and from an AAS/NASA Small Research Grant (2000).
727: The authors thank O. Reimer for his helpful comments and sustained interest
728: in this project. AMM is thankful for support from S\~ao Paulo State
729: funding agency FAPESP (97/11299-2) and CNPq, and is grateful to Rocio
730: Melgarejo and Fernando Nascimento da Silva for help with data gathering
731: and to Antonio Pereyra for help with \it{pccdpack}. \rm This research made
732: use of the NASA/IPAC Extragalactic Database (NED) which is operated by
733: the Jet Propulsion Laboratory/California Institute of Technology under
734: contract with NASA.
735:
736: \begin{center}
737: \begin{deluxetable}{ccccc}
738: \footnotesize
739: \tablecaption{5-GHz Sources Within the 3EG J2006--2321 99\% Confidence Contour.
740: \label{tbl-1}}
741: \tablewidth{0pt}
742: \tablehead{\colhead{Source No.}&
743: \colhead{PMN Coordinate Name} &\colhead{$S_5$ (mJy)} & \colhead{Angular Separation
744: (arcmin)} & \colhead{$\alpha_r\tablenotemark{a}$}
745: }
746: \startdata
747: 1. & J2003--2333 & $43\pm11$ & 41.7 & ... \\
748: 2. & J2005--2310 & $260\pm17$ & 11.0 & --0.12 \\
749: 3. & J2007--2335 & $141\pm13$ & 22.9 & --0.65 \\
750: 4. & J2008--2338 & $82\pm11$ & 35.6 & --0.97 \\
751: 5. & J2008--2305 & $47\pm11$ & 40.8 & +0.05 \\
752: 6. & J2009--2250 & $49\pm11$ & 55.3 & --0.55 \\
753: \enddata
754: \tablenotetext{a}{The radio spectral index $\alpha_r$ (where $F_{\nu}
755: \propto \nu^{\alpha_r}$) is calculated from the flux
756: densities at 1.4 and 4.85 (5) GHz}
757: \end{deluxetable}
758: \end{center}
759:
760: \begin{deluxetable}{llcccc}
761: \tablecaption{ PMN~J2005--2310 Field Imaging Polarimetry Observations}
762: \tablewidth{0pt}
763: \tablehead{
764: \colhead{Object} &
765: \colhead{Date} &
766: \colhead{Exp} &
767: \colhead{P} &
768: \colhead{$\sigma_P$} &
769: \colhead{$\theta$}
770: \\[.2ex]
771: & \colhead{(UT)} &
772: \colhead{(min)} & \colhead{\%} & \colhead{\%} &
773: \colhead{(deg)} %\\
774: }
775: \startdata
776: PMN~J2005--2310 & 2000 Aug 5 & $16\times5$ & 13.8 & 2.3 & 58 \\
777: & 2001 Jun 15 & $12\times5$ & 4.8 & 3.4 & 35 \\
778: Field Stars & 2000 Aug 5 & $16\times5$ & 1.465 & 0.013 & 20.4 \\
779: & 2001 Jun 15 & $12\times5$ & 1.492 & 0.016 & 22.1 \\
780: \enddata
781: \label{imagepol}
782: \end{deluxetable}
783:
784: \newpage
785:
786: \begin{thebibliography}{}
787:
788: \bibitem[Bloom, S. D. \& Dale, D. A. (2001)]{Blo01} Bloom, S. D. \& Dale, D. A. 2001, in AIP Conf. Proc. 587, Gamma 2001, eds.
789: S. Ritz, N. Gehrels, \& C. R. Schrader (Melville, NY: AIP), 329
790:
791: \bibitem[Colafrancesco et al. (2001)]{Col01} Colafrancesco, S. 2001, in AIP Conf. Proc. 587, Gamma 2001, eds.
792: S. Ritz, N. Gehrels, \& C. R. Schrader (Melville, NY: AIP), 427
793:
794: \bibitem[Condon, J. J. et al. (1998)]{Con98} Condon, J. J., Cotton, W. D., Greisen, E. W., Yin, Q. F., Perley, R. A., Taylor, G. B., \& Broderick, J. J. 1998, AJ, 115, 1693
795:
796: \bibitem[Douglas, J. N. et al. (1996)]{Dou96} Douglas, J. N., Bash, F. N., Bozyan, F. A., Torrence, G. W., \& Wolfe, C. 1996, AJ, 111, 1945
797:
798: \bibitem[Esposito et al. (1996)]{Esp96} Esposito, J. A., Hunter, S. D., Kanbach, G., \& Sreekumar 1996, ApJ, 461, 820
799:
800: \bibitem[Fichtel, C. E. et al. (1975)]{Fic75} Fichtel, C. E., Hartman, R. C., Kniffen, D. A., Thompson, D. J., Ogelman, H., Ozel, M. E., Turner, T., \& Bignami, G. F. 1975, ApJ, 198, 163
801:
802: \bibitem[Freedman, W. L. et al. (2001)]{Fre01} Freedman, W. L. et al. 2001, ApJ, 553, 47
803:
804: \bibitem[Gehrels et al. (2000)]{Geh00} Gehrels, N., Macomb, D. J., Bertsch, D. L., Thompson, D. J., \& Hartman, R. C. 2000, Nature, 404, 363
805:
806: \bibitem[Grenier, I. et al. (2001)]{Gre01} Grenier, I. et al. 2001, in AIP Conf. Proc. 587, Gamma 2001, eds.
807: S. Ritz, N. Gehrels, \& C. R. Schrader (Melville, NY: AIP), 649
808:
809: \bibitem[Griffith, M. R. \& Wright, A. E. (1993)]{Gri93} Griffith, M. R. \& Wright, A. E. 1993, AJ, 105, 1666
810:
811: \bibitem[Halpern et al. (2001)]{Hal01} Halpern, J. P., Camilo, F., Gotthelf, E. V., Helfand, D. J., Kramer, M., Lyne, A. G., Leighly, K. M., \& Eracleous, M. 2001, \apj, 552, L125
812:
813: \bibitem[Hartman et al. (1999)]{Har99} Hartman, R. C. et al. 1999, \apjs, 123, 79
814:
815: \bibitem[Kaaret, P. \& Cottam, J. 1995]{Kaa95} Kaaret, P. \& Cottam, J. 1995, ApJ, 462L, 35
816:
817: \bibitem[Kay et al. (1999)]{kay99} Kay, L.E., Magalh\~aes, A. M., Elizalde, F., \& Rodrigues, C.V. 1999,
818: \apj, 518, 210
819:
820: \bibitem[Lundgren et al. (1995)]{Lun95} Lundgren, S. C. et al. 1995, IAU Circular 6258
821:
822: \bibitem[Magalh\~aes, Benedetti \& Roland(1984)]{mag84} Magalh\~aes,
823: A. M., Benedetti, E., \& Roland, E. 1984, \pasp, 96, 384
824:
825: \bibitem[Magalh\~aes et al.(1996)]{mag96} Magalh\~aes, A. M., Rodrigues, C. V., Margoniner, V. E., \& Pereyra, A. 1996, ASP Conf.
826: Ser. 97 (W. G. Roberge and D. C. B. Whittet eds.), p. 118
827:
828: \bibitem[Mattox et al. (1997a)]{Mat97a} Mattox, J. R., Wagner, S. J., Malkan, M., McGlynn, T. A., Schachter, J. F., Grove, J. E., Johnson, W. N., \& Kurfess, J. D. 1997a, \apj, 476, 692
829:
830: \bibitem[Mattox et al. (1997b)]{Mat97b} Mattox, J. R., Schachter, J. Molnar, L., Hartman, R. C., \& Patnaik, A. R. 1997b, \apj, 481, 95
831:
832: \bibitem[Mattox et al. (2001a)]{Mat01b} Mattox, J. R., Hallum, J. C., Marscher, A. P., Jorstad, S. G., Waltman, E. B., Ter\"{a}sranta, H., Aller, H. D., \& Aller, M. F. 2001a, \apj, 549, 906
833:
834: \bibitem[Mattox et al. (2001b)]{Mat01a} Mattox, J. R., Hartman, R. C., \& Reimer, O. 2001b, ApJS, 135, 155
835:
836: \bibitem[Mirabel et al. (2000)]{Mir00} Mirabel, N., Halpern. J. P., Eracleous, M., \& Becker, R. H. 2000, ApJ, 541, 180
837:
838: \bibitem[Mirabel \& Halpern (2001)]{Mir01} Mirabel, N. \& Halpern, J. P. 2001, \apj, 547, L137
839:
840: \bibitem[McGlynn et al. (1997)]{McG97} McGlynn, T. A. et al. 1997, \apj, 481, 625
841:
842: \bibitem[McLaughlin et al. (1996)]{McL96} McLaughlin, M. A., Mattox, J. R., Cordes, J. M., \& Thompson, D. J. 1996, \apj, 473, 763
843:
844: \bibitem[Mukherjee et al. (1996)]{Muk96} Mukherjee, R. et al. 1996, \apj, 470, 831
845:
846: \bibitem[Mukherjee et al. (2001)]{Muk00} Mukherjee, R., Gotthelf, E. V., Halpern, J. P., \& Tavani, M. 2000, \apj, 542, 740
847:
848: \bibitem[Pereyra(2000)]{per00} Pereyra, A. 2000, PhD Thesis, IAG-USP
849:
850: \bibitem[Perlmutter et al. (1999)]{Per97} Perlmutter, S. et al. 1999, \apj, 517, 565
851:
852: \bibitem[e.g., Pesce et al.(1997)]{pes97} Pesce, J. E. et al. 1997, \apj, 486, 770
853:
854: \bibitem[Ramanamurthy et al. (1997)]{Ram95} Ramanamurthy, P. V. et al. 1995, \apj, 450, 791
855:
856: \bibitem[Reimer et al. (2001a)]{Rei01a} Reimer, O., Brazier, K. T. S., Carrami\~{n}ana, A., Kanbach, G., Nolan, P. L., and Thompson, D. J. 2001a, MNRAS 324, 772
857:
858: \bibitem[Reimer et al. (2001b)]{Rei01b} Reimer, O. \& Sreekumar, P. 2001b, in AIP Conf. Proc. 587, Gamma 2001, eds.
859: S. Ritz, N. Gehrels, \& C. R. Schrader (Melville, NY: AIP), 422
860:
861: \bibitem[Romero et al. (1999)]{Rom99} Romero, G. E., Benaglia, P., \& Torres, D. F. 1999, A\&A, 348, 868
862:
863: \bibitem[Schlegel et al.(1998)]{sch98} Schlegel, D. J., Finkbeiner, D. P.,
864: \& Davis, M. 1998, \apj, 500, 525
865:
866: \bibitem[Serkowski et al.(1975)]{ser75} Serkowski, K., Mathewson, D.S. \& Ford, V.L., \apj, 196, 261
867:
868: \bibitem[Sturner \& Dermer (1995)]{Stu95} Sturner, S. J. \& Dermer, C. D. 1995, A\&A, 293L, 17
869:
870: \bibitem[Voges et al. (1999)]{Vog99} Voges, W. et al. 1999, A\&A, 349, 389
871:
872: \bibitem[Voges et al. (2000)]{Vog00} Voges, W. et al. 2000, IAU Circular 7432
873:
874: \bibitem[von Montigny et al. (1995)]{von95} von Montigny et al. 1995, ApJ, 440, 525
875:
876: \bibitem[Thompson et al. (1995)]{Tho95} Thompson, D. J. et al. 1995, ApJS, 86, 629
877:
878: \bibitem[Turnshek et al.(1990)]{tur90} Turnshek, D. A., Bohlin, R. C., Williamson II, R. L., Lupie,
879: O. L., Koornneef, J., \& Morgan, D. H. 1990, \aj, 99, 1243
880:
881: \bibitem[Wallace et al. (2000)]{Wal00} Wallace, P. M., Griffis, N. J., Bertsch, D. L., Hartman, R. C., Thompson, D. J., Kniffen, D. A., \& Bloom, S. D. 2000, \apj, 540, 184
882:
883: \bibitem[Wallace et al. (2001)]{Wal01} Wallace, P. M., Eracleous, M., Foreman, J. V., Halpern, J. P., Reimer, O., \& Thompson, D. J. 2001, in AIP Conf. Proc. 587, Gamma 2001, eds. S. Ritz, N. Gehrels, \& C. R. Schrader (Melville, NY: AIP), 319
884:
885:
886: \end{thebibliography}
887:
888: \end{document}
889: