1: \documentclass{emulateapj}
2: \shorttitle{ROTSE-III Observations of GRB 030329}
3: \shortauthors{Smith et al.}
4: \newcommand{\PSbox}[3]{\mbox{\special{psfile=#1}\hspace{#2}\rule{0pt}{#3}}}
5:
6: \begin{document}
7:
8: \slugcomment{Accepted to Astrophysical Journal Letters, 4 Sept 2003}
9:
10: \title{ROTSE-III Observations of the Early Afterglow From GRB 030329}
11:
12: \author{Smith, D. A.\altaffilmark{1}, Rykoff, E. S.\altaffilmark{1}, Akerlof,
13: C. W.\altaffilmark{1}, Ashley, M. C. B.\altaffilmark{2}, Bizyaev,
14: D.\altaffilmark{4,6}, McKay, T. A.\altaffilmark{1}, Mukadum,
15: A.\altaffilmark{5}, Phillips, A.\altaffilmark{2}, Quimby,
16: R.\altaffilmark{5}, Schaefer, B.\altaffilmark{5}, Sullivan,
17: D.\altaffilmark{7}, Swan, H. F.\altaffilmark{1}, Vestrand,
18: W. T.\altaffilmark{3}, Wheeler, J. C.\altaffilmark{5}, and Wren,
19: J.\altaffilmark{3}}
20:
21: \email{donaldas@umich.edu}
22:
23: \altaffiltext{1}{University of Michigan, 2477 Randall Laboratory,
24: 500 E. University Ave., Ann Arbor, MI, 48104}
25: \altaffiltext{2}{School of Physics, Department of Astrophysics and Optics, University of New South Wales, Sydney, NSW 2052, Australia}
26: \altaffiltext{3}{Los Alamos National Laboratory, NIS-2 MS D436, Los Alamos, NM 87545}
27: \altaffiltext{4}{Department of Physics, University of Texas at El Paso, El Paso, Texas, USA}
28: \altaffiltext{5}{Department of Astronomy, University of Texas, Austin, TX 78712}
29: \altaffiltext{6}{Sternberg Astronomical Institute, Moscow, Russia}
30: \altaffiltext{7}{School of Chemical \& Physical Sciences, Victoria University of Wellington, PO Box 600, Wellington, NZ}
31:
32: \begin{abstract}
33:
34: Using two identical telescopes at widely separated longitudes, the ROTSE-III
35: network observed decaying emission from the remarkably bright afterglow of
36: GRB~030329. In this report we present observations covering 56\% of the period
37: from 1.5--47~hours after the burst. We find that the light curve is piecewise
38: consistent with a powerlaw decay. When the ROTSE-III data are combined with
39: data reported by other groups, there is evidence for five breaks within the
40: first 20~hours after the burst. Between two of those breaks, observations from
41: 15.9--17.1~h after the burst at 1-s time resolution with McDonald Observatory's
42: 2.1-m telescope reveal no evidence for fluctuations or deviations from a simple
43: power law. Multiple breaks may indicate complex structure in the jet. There
44: are also two unambiguous episodes at 23 and 45~hours after the burst where the
45: intensity becomes consistent with a constant for several hours, perhaps
46: indicating multiple injections of energy into the GRB/afterglow system.
47:
48: \end{abstract}
49: \keywords{Gamma-rays: bursts}
50:
51: \section{Introduction}
52:
53: The ROTSE program is dedicated to recording optical observations of Gamma-Ray
54: Bursts (GRBs) within seconds of their detection at high energies. Although the
55: study of late-time afterglows has revolutionized our understanding of GRBs,
56: comparatively little is known about the nature and diversity in the prompt
57: optical and early afterglow emission~\citep{piran99}. To date the ROTSE-I
58: observations of GRB~990123 remain the only optical detections of a GRB while
59: the burst was still bright in gamma-rays~\citep{abbbb99}. ROTSE-I and LOTIS
60: observations of many bursts demonstrate that prompt emission is not, in
61: general, well represented by extrapolation of late time decay
62: curves~\citep{kabbb01,ppwab99}. More sensitive instruments are required to
63: study prompt emission from typical bursts.
64:
65: In preparation for HETE-2~\citep{vvdjl99} and the Swift
66: satellites~\citep{gehre00} which can distribute rapid localizations of GRBs
67: with arcminute accuracy, the ROTSE team developed a new set of 0.45-m
68: telescopes. The ROTSE-III instruments use fast optics (f/1.9) to yield a
69: $1\arcdeg.85 \times 1\arcdeg.85$ field of view over a four megapixel CCD
70: camera. Four of these robotically controlled, automatically scheduled
71: instruments are being erected around the globe, both to increase the amount of
72: sky available for instant viewing and to enable around-the-clock coverage of
73: GRB afterglow behavior~\citep{akmrs03}. At the time of this writing, the first
74: two instruments, labeled ROTSE-IIIa and ROTSE-IIIb, are operating at Siding
75: Spring Observatory in New South Wales, Australia, and McDonald Observatory at
76: Ft. Davis, Texas, USA, respectively. The installation of ROTSE-IIIc is nearing
77: completion in Namibia, and ROTSE-IIId is slated to arrive in Turkey in the Fall
78: of 2003.
79:
80: Within a week of the first two ROTSE-III instruments becoming fully
81: operational, the network was tested by the discovery with HETE-2 of GRB~030329.
82: Spectroscopic observations found the source to lie at a redshift of
83: $z=0.168$~\citep{gpekv03}, which makes it the closest GRB to the Earth yet to
84: be measured, with the possible exception of the anomalously faint
85: GRB~980425~\citep{gvpka99}. This event provided the first direct spectral
86: evidence for a supernovae associated with a high redshift
87: GRB~\citep{smgmb03,hsmfw03,kdwmn03}. We report here on the light curve of
88: GRB~030329 for the first two days of observations by ROTSE-IIIa \& b.
89:
90: \begin{figure*}
91: \PSbox{figure1.eps hoffset=-10 hscale=75 vscale=75 voffset=10}{6.0in}{3.9in}
92: \figcaption{Decay of GRB 030329 over the first two days of observations.
93: Magnitudes are calibrated to $V_{\rm ROTSE}$ as described in the text.
94: Diamonds indicate measurements made with ROTSE-IIIa and triangles those made
95: with ROTSE-IIIb. Intervals when the source was above the horizon at each site
96: are indicated by horizontal bars at the top of the graph.\label{fig:ful}}
97: \end{figure*}
98:
99: \section{Observations and Analysis}
100:
101: GRB 030329 was first observed by the instruments on board the HETE-2 satellite
102: at 11:37:14.67 UTC. Ground analysis of the SXC data produced a localization
103: that was reported in a Gamma-Ray Burst Coordinate Network (GCN) Notice 73
104: minutes after the burst~\citep{vcdvm03}. Although the burst's celestial
105: location was above the Australian horizon at the time, the ROTSE-IIIa
106: connection to the GCN was temporarily experiencing sporadic outages. It was
107: therefore impossible for the ROTSE automated scheduler daemon to respond to the
108: alert. However, electronic mail was also distributed to ROTSE team members,
109: and a manual response by the ROTSE-IIIa telescope was initiated over the
110: Internet within 15~minutes of the alert's distribution, 88~minutes after the
111: burst itself.
112:
113: Clouds and intermittent rain interfered with observations and delayed analysis
114: of the images, but once~\citet{petpr03} and~\citet{torii03} identified the
115: optical counterpart, we were able to extract a light curve from the setting
116: source. Our earliest useful images began 92~minutes after the burst
117: event~\citep{ryksm03}. At 4.8~h after the burst, the air mass to the source
118: grew too large for further observation.
119:
120: \begin{figure}
121: \PSbox{figure2.eps hoffset=0 hscale=50 vscale=45 voffset=-25}{3.0in}{4in}
122: \figcaption{Expanded views of two segments of the ROTSE-III light curve for
123: GRB~030329. Diamonds indicate measurements made with ROTSE-IIIa and triangles
124: those made with ROTSE-IIIb. The top panel shows the interval from 14 to 30~h
125: after the burst, while the bottom panel covers the time from 38 to 48~h. These
126: panels also overlay best-fit power law functions onto the observations. The
127: data in the top panel are clearly inconsistent with a single power law decay.
128: Parameters for these functions are given in Table~\protect{\ref{tab:pars}} The
129: burst intensity levels off at $\sim23$~h after the burst, maintaining a
130: constant intensity for the rest of the Australian night, but begins decaying
131: again in the second night of Texas observations, as shown in the bottom panel,
132: before entering a second flattening episode.
133: \label{fig:sub}}
134: \end{figure}
135:
136:
137: ROTSE-IIIb (in Texas) began observations 14.8~h after the burst, and followed
138: the source for a further 8.5~h. Once the source had risen again in Australia,
139: contemporaneous observation from both ROTSE-III instruments was possible for
140: 1.8~h. The ROTSE-III automatic scheduler normally plans observations of fading
141: GRB afterglows to occur at logarithmically increasing intervals, leading to
142: gaps in coverage at the start of each night's observing. As it became clear
143: how bright this source was, we overrode the default schedule to instruct the
144: instruments to look at nothing else as long as GRB~030329 was above the local
145: horizon.
146:
147: The ROTSE-III images were bias-subtracted and flat-fielded in the standard way.
148: For images recorded later than 10~h after the burst, we co-added sets of ten
149: frames before deriving a light curve. We applied SExtractor~\citep{berar96}
150: with aperture photometry to these co-added images to detect objects and derive
151: their intensities. We compared these intensities to the USNO~A2.0 catalog to
152: derive an approximate R-band magnitude zero-point offset for the field.
153: Zero-point offsets among the ROTSE images are determined through application of
154: a relative photometry procedure. Absolute calibration to a V-magnitude
155: equivalent (dubbed $V_{\rm ROTSE}$) is derived by comparison of several nearby
156: stable stars to the $UBVR_cI_c$ photometry of~\citet{h2023}.
157:
158: The offset between unfiltered $V_{\rm ROTSE}$ observations and standard V
159: magnitudes is a function of color. Without color information for the early-time
160: afterglow, we have not applied any color correction to these data. However,
161: from 6.0--14.4~h after the burst, during observations reported
162: by~\citet{bspdt03}, the afterglow color is rather blue, with $V-R\approx0.3$.
163: This color remained stable through a break in the lightcurve. If the afterglow
164: maintained this color throughout the ROTSE-III observations, it would be
165: appropriate to further adjust the $V_{\rm ROTSE}$ magnitudes reported here to
166: be approximately 0.5 magnitudes brighter.
167:
168: GRB~030329 was also observed with a time resolution of 1-second using a CCD
169: time-series photometer mounted at the prime focus position of the McDonald
170: Observatory 2.1-m telescope. The photometer ``Argos''~(Nather \&~Mukadam, in
171: preparation) operates in frame transfer mode, and thus allows rapid exposures
172: as short as one second. The time-series photometry started 15.87 hours after
173: the burst trigger, and continued for 4290 seconds in essentially photometric
174: conditions. No filter was used, so the spectral response of the instrument was
175: dominated by the combination of a typical CCD sensitivity and the Earth's
176: atmosphere: the transmission is about 70\% in the wavelength range 560--640~nm,
177: and greater than 50\% in the wavelength range 450--800~nm.
178:
179: Typical ADU count rates for the optical transient were 17,000 per second, while
180: the sky contributed about 10,500 counts per second in the chosen software
181: aperture. There were two comparison stars in the 2.8~arcminute field of view
182: of the instrument, and sky-subtracted light curves for each object were
183: obtained by using an optimized circular software aperture. A corrected light
184: curve for the target object was then generated by dividing its light curve to
185: the sum of the two comparison star light curves (which have been normalized to
186: unity).
187:
188: \section{Results}\label{sec:results}
189:
190: The full light curve from both ROTSE-IIIa \& b is shown on a logarithmic scale
191: in the top panel of Figure~\ref{fig:ful}. Figure~\ref{fig:sub} shows two
192: segments of the light curve on expanded scales. In this section, we describe
193: the behavior of the afterglow during each observing sequence.
194:
195: The poor observing conditions during the first observing run (from 1.5 to 4.8~h
196: after the burst) introduce large systematic errors into the photometry, but
197: there is no evidence for any significant deviations from a power law decay with
198: an index of 0.88$\pm$0.05, a value not atypical for early GRB afterglows. The
199: earliest detections reveal an unfiltered $V_{\rm ROTSE}$~magnitude of 12.55,
200: over 100 times brighter than any other afterglow to date to be observed at this
201: epoch~\citep{pfkps03}, consistent with the very low redshift for the object.
202: The latest points in this sequence, recorded after an hour during which
203: ROTSE-IIIa automatically ceased operation due to a light rain, are also
204: consistent with the same decay curve.
205:
206: When ROTSE-IIIb commenced observations at 14.8~h, the intensity of GRB~030329
207: was decaying much more rapidly, consistent with a power law index of
208: 1.85$\pm$0.07 and strongly inconsistent with the earlier curve observed with
209: ROTSE-IIIa. If we assume a single, sharp transition from one decay mode to the
210: other, these data predict a transition time of $\sim$12--14~h after the burst.
211: Over the course of the 8.5~h observing sequence with ROTSE-IIIb on 30 Mar 2003,
212: the decay slope steepened further, demanding the introduction of a third power
213: law with decay index of 2.25$\pm$0.08. The best-fit transition time to this
214: decay slope is at 19.0~h after the burst.
215:
216: The interval during which GRB~030329 was observed by the 2.1-m coincided with
217: the early part of this decay phase, before the break at 19~h. The corrected
218: light curve at 1-s resolution shows the optical transient to be fading
219: smoothly. The fitted power law decline has an index of 1.98$\pm$0.02. This
220: power law fits well to both the beginning and ending 500-second time intervals
221: with the same normalization to within 0.2\%, thus indicating that there were no
222: departures from the overall power law decline. The largest deviation from the
223: best fit power law was one point that was low by only $4.4\sigma$. Analysis of
224: binned light curves reveals no significant fluctuations from the power law on
225: any longer time scales. The computed discrete Fourier transform did not reveal
226: any power out to the Nyquist frequency of 0.5~Hz up to a limit of 2~millimag
227: (i.e., 0.2\% change in intensity). In all, we found no deviations from a
228: simple power law decline.
229:
230: During the interval when both ROTSE-III instruments could observe the source
231: simultaneously, our calibration procedure yielded consistent results. However,
232: ROTSE-IIIa subsequently observed the source intensity to level off and remain
233: constant for at least three hours. When the source again became visible to
234: ROTSE-IIIb in Texas, it was found to still have the same intensity as when last
235: seen by ROTSE-IIIa. It shortly commenced a rapid decay phase, consistent with
236: a single power law of index 2.34$\pm$0.08. This decay rate is consistent with
237: the decay slope observed at the end of the previous night. The decay phase
238: lasted for approximately six hours. The source then entered a second
239: flattening phase that was still continuing when observations ceased for the
240: night two hours later\footnote{The ROTSE-III light curve is stored in
241: electronic form at http://www.rotse.net/transients/grb030329}.
242:
243: The decay curve for GRB~030329 over its first two days is clearly inconsistent
244: with a single power-law function. We find that at least four power laws plus
245: two intervals of constant intensity are necessary to adequately reproduce the
246: behavior of the light curve. Three of these functions are plotted as broken
247: lines in Figure~\ref{fig:sub}. The decay indices as well as the intervals over
248: which each power-law segment applies are given in Table~\ref{tab:pars}.
249:
250: \section{Discussion}
251:
252: The world-wide distribution of the ROTSE-III network allows for unprecedented
253: coverage of the early decay curve of GRB afterglows, and in the case of
254: GRB~030329, the afterglow was above the horizon for 56\% of the first 47~h
255: after the event. Since the ROTSE-III network was only half complete at the
256: time of GRB~030329, there are still large gaps in the derived light curve when
257: the source was not visible to either telescope. However, at least two groups
258: have published observations of the afterglow's behavior during the time on the
259: first day when neither ROTSE-IIIa nor ROTSE-IIIb could observe it.
260:
261: \citet{ukiym03} observed GRB~030329 with a variety of telescopes from 1.2~h to
262: $\sim14$~h after the event, and they report three breaks in the light curve
263: during this time: from 0.74$\pm$0.02 to 0.95$\pm$0.01 at $t=2.04$~h, to
264: 0.65$\pm$0.04 at $t=3.9$~h, and again to 1.16$\pm$0.01 at $t=5.4$~h. This
265: characterization is consistent with the ROTSE-IIIa data, which have neither
266: sufficient coverage nor small enough errors to distinguish between this and the
267: best-fit single power law reported above in Section~\ref{sec:results}.
268: \citet{bspdt03} present observations taken with the 1.5~m Russian-Turkish
269: Telescope shortly after the time of this last transition and give a slope of
270: 1.19$\pm$0.01 from 6.2~h to 14.4~h after the burst, after which the slope
271: changes to $\sim1.9$, which is consistent with the ROTSE-IIIb observations at
272: 14.8~h. The Japanese and Turkish programs, therefore, neatly fill in the gap
273: between ROTSE-III observations, and the joint data set provides evidence for
274: five changes in slope decay index within the first 20~hours after the burst.
275:
276: \begin{deluxetable}{lll}
277: \tablecaption{Parameters of Best-Fit Power-Law Decay Curve\label{tab:pars}}
278: \tablehead{
279: \colhead{Start Time (h)} & \colhead{Stop Time (h)} & \colhead{Index}
280: }
281: \startdata
282: 1.5 & 4.8 & 0.87$\pm$0.03 \\
283: 14.8 & 19.0 & 1.85$\pm$0.06 \\
284: 19.0 & 23.0 & 2.25$\pm$0.08 \\
285: 23.0 & 26.0 & $<$0.4 \\
286: 39.5 & 45.0 & 2.34$\pm$0.08 \\
287: 45.0 & 46.5 & $<$0.4
288: \enddata
289: \end{deluxetable}
290:
291: No GRB afterglow has been observed nearly as intensely as GRB~030329, and for
292: most other bursts only a single steepening in the light curve decay has been
293: reported. These breaks have been taken as evidence for deceleration of a
294: relativistic jet: as the jet decelerates, the beaming angle must increase, and
295: when the beaming angle exceeds the jet angle, the intensity of visible light is
296: expected to decrease more rapidly~\citep{rhoad97,sapih99,psf03}. Multiple
297: breaks within a single decay curve are more difficult to interpret in this
298: paradigm, but \citet{weiji02}, for example, have suggested that they may result
299: from off-axis viewing geometry or perhaps non-uniform structure within the jet.
300: A detailed analysis of what geometry five decay breaks might demand is beyond
301: the scope of this paper, but the data clearly are not consistent with the
302: simplest models and demand caution when trying to interpret $t\sim0.5$~d as
303: {\it the} jet break~\citep{gnp03,bspdt03,pfkps03}.
304:
305: The other striking feature of the afterglow is the deviation from its rapid
306: decay, beginning at $\sim23$~h after the burst. In the ROTSE-III observations,
307: the intensity levels off and remains constant for several hours. A second
308: decay phase begins $\sim39$~h after the burst. It is unlikely that this
309: episode is related to the emergence of a supernova in the light curve, as the
310: first spectral evidence for a supernova associated with GRB~030329 emerged only
311: after $\sim7$~d~\citep{smgmb03}. At one day after the event, the optical
312: intensity should be dominated by the burst afterglow. Furthermore, the light
313: curve flattens yet again at $\sim44$~h. \citet{weiji02} argue that geometric
314: factors could produce a flattening of the light curve at about this time, but
315: it is hard to understand the repeating nature of the phenomenon within that
316: paradigm. On the other hand, \citet{gnp03} have interpreted these episodes as
317: refreshed shocks produced by further ejection events from the central engine
318: catching up with the decelerating blast wave. In this interpretation, the
319: central engine remains an active part of the GRB process well into the
320: afterglow phase, further underscoring the need for early and thorough
321: observations of the entire burst event if the complex physical interactions are
322: ever to be teased out.
323:
324: \acknowledgments
325:
326: This work has been supported by NASA grants NAG 5-5281 and F006794, NSF grant
327: AST 01-19685, the Michigan Space Grant Consortium, the Australian Research
328: Council, the University of New South Wales, the University of Texas, and the
329: University of Michigan. DAS is supported by an NSF Astronomy and Astrophysics
330: Postdoctoral Fellowship under award AST 01-05221. Work performed at LANL is
331: supported by NASA SR\&T through Department of Energy (DOE) contract
332: W-7405-ENG-36 and through internal LDRD funding.
333:
334: \newcommand{\noopsort}[1]{} \newcommand{\printfirst}[2]{#1}
335: \newcommand{\singleletter}[1]{#1} \newcommand{\switchargs}[2]{#2#1}
336: \begin{thebibliography}{}
337:
338: \bibitem[\protect\astroncite{{Akerlof} {\rm et~al.\/}}{1999}]{abbbb99}
339: {Akerlof}, C., et~al.
340: \newblock 1999, \nat, 398, 400
341:
342: \bibitem[\protect\astroncite{{Akerlof} {\rm et~al.\/}}{2003}]{akmrs03}
343: {Akerlof}, C., et~al.
344: \newblock 2003, \pasp, 115, 132
345:
346: \bibitem[\protect\astroncite{{Bertin} \& {Arnouts}}{1996}]{berar96}
347: {Bertin}, E. \& {Arnouts}, S.
348: \newblock 1996, \aaps, 117, 393
349:
350: \bibitem[\protect\astroncite{{Burenin} {\rm et~al.\/}}{2003}]{bspdt03}
351: {Burenin}, R., et~al.
352: \newblock 2003, Astronomy Letters, 9, 1, astro-ph/0306137
353:
354: \bibitem[\protect\astroncite{{Galama} {\rm et~al.\/}}{1999}]{gvpka99}
355: {Galama}, T.~J., et~al.
356: \newblock 1999, \aaps, 138, 465
357:
358: \bibitem[\protect\astroncite{{Gehrels}}{2000}]{gehre00}
359: {Gehrels}, N.~A. 2000, , in Proc. SPIE Vol. 4140, p. 42-49, X-Ray and Gamma-Ray
360: Instrumentation for Astronomy XI, Kathryn A. Flanagan; Oswald H. Siegmund;
361: Eds., 42
362:
363: \bibitem[\protect\astroncite{{Granot}, {Nakar}, \& {Piran}}{2003}]{gnp03}
364: {Granot}, J., {Nakar}, E., \& {Piran}, T.
365: \newblock 2003, astro-ph/0304563
366:
367: \bibitem[\protect\astroncite{Greiner {\rm et~al.\/}}{2003}]{gpekv03}
368: Greiner, J., Peimbert, M., Estaban, C., Kaufer, A., Vreeswijk, P., Jaunsen, A.,
369: Smoke, J., Klose, S., \& Reimer, O.
370: \newblock 2003, GCN Circ. No. 2020
371:
372: \bibitem[\protect\astroncite{Henden}{2003}]{h2023}
373: Henden, A.
374: \newblock 2003, GCN Circ. No. 2023
375:
376: \bibitem[\protect\astroncite{{Hjorth} {\rm et~al.\/}}{2003}]{hsmfw03}
377: {Hjorth}, J. et~al.
378: \newblock 2003, \nat, 423, 847
379:
380: \bibitem[\protect\astroncite{{Kawabata} {\rm et~al.\/}}{2003}]{kdwmn03}
381: {Kawabata}, K.~S. et~al.
382: \newblock 2003, \apjl, in press, astro-ph/0306155
383:
384: \bibitem[\protect\astroncite{{Kehoe} {\rm et~al.\/}}{2001}]{kabbb01}
385: {Kehoe}, R., et~al.
386: \newblock 2001, \apjl, 554, L159
387:
388: \bibitem[\protect\astroncite{{Park} {\rm et~al.\/}}{1999}]{ppwab99}
389: {Park}, H.~S., et~al.
390: \newblock 1999, \aaps, 138, 577
391:
392: \bibitem[\protect\astroncite{{Perna}, {Sari}, \& {Frail}}{2003}]{psf03}
393: {Perna}, R., {Sari}, R., \& {Frail}, D.
394: \newblock 2003, \apj, astro-ph/0305145
395:
396: \bibitem[\protect\astroncite{Peterson \& Price}{2003}]{petpr03}
397: Peterson, B. \& Price, P.
398: \newblock 2003, GCN Circ. No. 1985
399:
400: \bibitem[\protect\astroncite{Piran}{1999}]{piran99}
401: Piran, T.
402: \newblock 1999, \physrep, 314(6), 575
403:
404: \bibitem[\protect\astroncite{{Price} {\rm et~al.\/}}{2003}]{pfkps03}
405: {Price}, P., et~al.
406: \newblock 2003, \nat, 423, 844
407:
408: \bibitem[\protect\astroncite{{Rhoads}}{1997}]{rhoad97}
409: {Rhoads}, J.~E.
410: \newblock 1997, \apjl, 487, L1
411:
412: \bibitem[\protect\astroncite{Rykoff \& Smith}{2003}]{ryksm03}
413: Rykoff, E.~S. \& Smith, D.~A.
414: \newblock 2003, GCN Circ. No. 1995
415:
416: \bibitem[\protect\astroncite{{Sari}, {Piran}, \& {Halpern}}{1999}]{sapih99}
417: {Sari}, R., {Piran}, T., \& {Halpern}, J.~P.
418: \newblock 1999, \apjl, 519, L17
419:
420: \bibitem[\protect\astroncite{{Stanek} {\rm et~al.\/}}{2003}]{smgmb03}
421: {Stanek}, K.~Z., et~al.
422: \newblock 2003, \apjl, 591, L17
423:
424: \bibitem[\protect\astroncite{Torii}{2003}]{torii03}
425: Torii, K.
426: \newblock 2003, GCN Circ. No. 1986
427:
428: \bibitem[\protect\astroncite{{Uemura} {\rm et~al.\/}}{2003}]{ukiym03}
429: {Uemura}, M., {Kato}, T., {Ishioka}, R., {Yamaoka}, H., {Monard}, B., {Nogami},
430: D., {Maehara}, H., {Sugie}, A., \& {Takahashi}, S.
431: \newblock 2003, \nat, astro-ph/0306396
432:
433: \bibitem[\protect\astroncite{{Vanderspek} {\rm et~al.\/}}{2003}]{vcdvm03}
434: {Vanderspek}, R., et~al.
435: \newblock 2003, GCN Circ. No. 1997
436:
437: \bibitem[\protect\astroncite{{Vanderspek} {\rm et~al.\/}}{1999}]{vvdjl99}
438: {Vanderspek}, R., {Villase{\~ n}or}, J., {Doty}, J., {Jernigan}, J.~G.,
439: {Levine}, A., {Monnelly}, G., \& {Ricker}, G.~R.
440: \newblock 1999, \aaps, 138, 565
441:
442: \bibitem[\protect\astroncite{{Wei} \& {Jin}}{2003}]{weiji02}
443: {Wei}, D.~M. \& {Jin}, Z.~P.
444: \newblock 2003, \aap, 400, 415
445:
446: \end{thebibliography}
447:
448:
449: \end{document}
450: