1: % For ApJ submission -----------------------------------------
2: %\documentclass[12pt,preprint]{aastex}
3: %\def\baselinestretch{1.3}
4:
5: \documentclass{emulateapj}
6: \usepackage{apjfonts}
7:
8:
9: %==== CUSTOMIZED LATEX MACROS ========================================
10:
11: \newcommand{\vect}[1]{\ensuremath{\mbox{\boldmath $#1$}}}
12: \newcommand{\deltavec}{\mbox{\boldmath $\delta$}}
13:
14: % Equation align
15: \def\eqalign#1{\null\,\vcenter{\openup\jot
16: \ialign{\strut\hfil$\displaystyle{##}$&$
17: \displaystyle{{}##}$\hfil \crcr#1\crcr}}\,}
18:
19: \newcommand{\xivec}{\mbox{\boldmath $\xi$}}
20: \newcommand{\rhovec}{\mbox{\boldmath $\rho$}}
21: \newcommand{\zetavec}{\mbox{\boldmath $\zeta$}}
22:
23:
24: %=======================================================================
25:
26: \lefthead{HAN}
27: \righthead{PLANET VS BINARY}
28:
29: \begin{document}
30: \title{Distinguishing between Planetary and Binary Interpretations of Microlensing
31: Central Perturbations under Severe Finite-Source Effect}
32:
33:
34: \author{Cheongho Han}
35: \affil{Program of Brain Korea 21, Department of Physics,
36: Chungbuk National University, Chongju 361-763, Korea;
37: cheongho@astroph.chungbuk.ac.kr}
38:
39:
40:
41:
42: \submitted{Submitted to The Astrophysical Journal}
43:
44: \begin{abstract}
45: In the current strategy of microlensing planet searches focusing
46: on high-magnification events, wide and close binaries pose
47: important sources of contamination that imitates planetary
48: signals. For the purpose of finding systematic differences,
49: we compare the patterns of central perturbations induced by
50: a planet and a binary companion under severe finite-source
51: effect. We find that the most prominent difference shows up
52: in the morphology of the edge features with negative excess
53: that appear at the edge of the circle with its center located
54: at the caustic center and a radius equivalent to the source
55: radius. It is found that the feature induced a binary companion
56: forms a complete annulus, while the feature induced by a planet
57: appears as several arc segments. This difference provides a
58: useful diagnostic for immediate discrimination of a planet-induced
59: perturbation from that induced by a binary companion,
60: where the absence of a well-developed dip in the residual
61: from the single-lensing light curve at both or either of the moments of the
62: caustic center's entrance into and exit from the source
63: star surface indicates that the perturbation is
64: produced by a planetary companion.
65: We find that that
66: this difference is basically caused by the difference between
67: the shapes of the central caustics induced by the two different
68: types of companions.
69: \end{abstract}
70:
71:
72: \keywords{gravitational lensing -- planets and satellites: general}
73:
74: % ==================================================================
75:
76:
77: \section{Introduction}
78:
79: Microlensing planets are being discovered at an accelerating rate
80: and the total number of detections now reaches 8 \citep{bond04,
81: udalski05, beaulieu06, gould06, gaudi08, dong08, bennett08}.
82: The microlensing signal of a planet is a brief perturbation to
83: the smooth standard light curve of the primary-induced lensing
84: event occurring on a background star. To achieve the monitoring
85: frequency required to detect short-duration planetary signals,
86: current planetary lensing searches are being conducted by using
87: a combination of survey and follow-up observations, where alerts
88: of ongoing lensing events are issued by the survey observations
89: (OGLE: Udalski et al.\ 1994, MOA: Bond et al.\ 2001) and the
90: alerted events are intensively monitored by the follow-up
91: observations (Micro-FUN: Dong et al.\ 2006-, PLANET: Kubas et
92: al.\ 2008). However, the number of telescopes available for
93: follow-up observations is far less for intensive monitoring of
94: all alerted events \citep{dominik08} and thus observations are
95: focused on events which can maximize the planet detection
96: probability with follow-up observations. Currently, the prime
97: targets of follow-up observations are high-magnification events
98: %\citep{albrow00, bond02, rattenbury02, yoo04, abe04, han07}
99: for which the source trajectories always pass close to the
100: perturbation region around the central caustic induced by a
101: planet and thus planet detection efficiency is intrinsically
102: high \citep{griest98}.
103:
104:
105: Although high-magnification events yield high sensitivity to
106: planets, interpretation of the observed perturbation often
107: suffers from several potential degeneracies. There are two
108: major degeneracies causing this complication. The first
109: well-known wide/close degeneracy arises due to the fact that
110: the perturbation induced by a planet with a projected separation
111: in units of the Einstein ring, $s$, is very similar to the
112: perturbation induced by a planet with a separation $1/s$
113: \citep{dominik99, an05, chung05}.\footnote{We note that the
114: wide/close degeneracy occurs not only for planetary events
115: but also for binary events in general \citep{dominik96}.}
116: The other degeneracy arises due to the fact that a central
117: perturbation of a high-magnification event can also be produced
118: by a very close ($s\ll 1$) or a very wide ($s\gg 1$) binary with
119: roughly equal mass components. Hereafter, we refer the latter
120: degeneracy as the `planet/binary' degeneracy. In the sense that
121: the planet/binary degeneracy causes indeterminacy of both of
122: the mass ratio and the separation between the lens components,
123: while the wide/close degeneracy causes ambiguity only in the
124: separation, the planet/binary degeneracy poses more serious
125: problem in the interpretation of an observed perturbation.
126: Fortunately, the perturbations induced by a planet and a binary
127: are intrinsically different and thus it would be possible to
128: determine whether the perturbation is caused by a planet or a
129: binary. However, distinguishing between the two interpretations
130: usually requires detailed modelling which demands time-consuming
131: search for a solution in the vast space of many parameters.
132: Therefore, a simple diagnostic that can resolve the planet/binary
133: degeneracy would be very helpful not only for the prompt interpretation
134: of an observed perturbation but also for the establishment of the
135: observational setup optimizing the coverage of the characteristic
136: features helping to resolve the degeneracy. \citet{han08a} provided
137: such a diagnostic but this diagnostic applies to a subset of light
138: curves with double-peak features.
139:
140:
141:
142: Recently, \cite{dong08} reported a planet detected from the
143: analysis of a new type of high-magnification event where the
144: angular extent of the perturbation region induced by the planet
145: is significantly smaller than the angular size of the source
146: and thus finite-source effect is very severe. The main
147: perturbation features of this event are double spikes in the
148: residuals, which are approximately centered at the times when
149: the lens enters and exits the source. In 2008 season, several
150: more such events were detected (A.\ Gould 2008, P.\ Fouque
151: 2008, private communication), implying that events with these
152: features might be common. From detailed investigation,
153: \citet{han08b} found that spike features commonly appear for
154: planetary-lensing events affected by severe finite-source effect
155: and thus they can be used for the diagnosis of the existence
156: of a planet. However, they noted that such features can also
157: be produced by a wide or a close binary companion and thus the
158: existence of the feature does not necessarily confirm the
159: planetary interpretation of the perturbation.
160:
161:
162: In this paper, we investigate systematic differences between
163: the patterns of central perturbation induced by a planet and
164: a binary companion. From this, we identify a diagnostic that
165: can be used to immediately distinguish between the planetary
166: and binary interpretations. We also investigate the origin of
167: the difference.
168:
169:
170:
171:
172:
173: \section{Perturbation Pattern}
174:
175:
176: The pattern of central perturbations is basically determined
177: by the caustic shape. The shapes of the caustics induced by a
178: planet and a binary companion are intrinsically different and
179: thus the resulting patterns of the perturbations produced by
180: the two types of lens systems will be different.
181:
182:
183: For a wide binary system, the central caustic has a hypocycloid
184: shape with four cusps. When the horizontal and vertical widths
185: of the caustic are measured as the separations between the two
186: confronting cusps located on and off the binary axis, respectively,
187: the ratio of the vertical to horizontal widths is represented by
188: \citep{han05}
189: \begin{equation}
190: R_b\sim \left( {1-\gamma\over 1+\gamma}\right)^{1/2};\qquad
191: \gamma={q\over (1+q)s^2},
192: \label{eq1}
193: \end{equation}
194: where $\gamma$ represents the shear exerted by the companion.
195: As the binary separation increases, the shear decreases and
196: the width ratio becomes $R_b\rightarrow 1$, implying that the
197: horizontal and the vertical widths are nearly identical for a
198: wide binary. In addition, the shape of the caustic is symmetric
199: with respect to both of the binary axis and the axis normal to
200: the binary axis. The caustic induced by a close binary is
201: approximately identical to that of the wide binary with a
202: separation of $s^{-1}$.
203:
204:
205: In contrast to the central caustic induced by a binary
206: companion, the central caustic induced by a planet has an
207: elongated wedge-like shape. The width ratio of the planetary
208: central caustic defined in a similar way as that of the binary
209: case is represented by \citep{chung05}
210: \begin{equation}
211: R_p \sim {(s-s^{-1})|\sin^3\phi| \over
212: (s+s^{-1}-2\cos\phi)^2},
213: \label{eq2}
214: \end{equation}
215: where
216: $\cos\phi=(3/4)(s+s^{-1}) \{ 1-[32/9(s+s^{-1})^2]^{1/2}\}$.
217: For the range of planetary separations where the size of
218: the central caustic is not negligible (see \citet{han09}),
219: the width ratio of the central caustic is substantially
220: smaller than unity, implying that the caustic is elongated
221: along the planet-primary axis. In addition, three of the
222: four cusps of the caustic lean toward the primary direction.
223: As a result, the caustic is {\it not} symmetric with respect
224: to the axis normal to the planet-primary axis.
225:
226:
227: Another important factor that affects the perturbation pattern
228: is the effect of the finite size of a source star. The lensing
229: magnification affected by the finite-source effect corresponds
230: to the magnification averaged over the source star surface.
231: %\begin{equation}
232: %A={
233: %\int_0^{\rho_\star} I(r)A'(|{\bf r}-{\bf r}_0|) r dr
234: %\over
235: %\int_0^{\rho_\star} I(r) r dr
236: %},
237: %\label{eq4}
238: %\end{equation}
239: %where ${\bf r}_0$ is the displacement vector of the source center
240: %with respect to the lens, ${\bf r}$ is the vector to a position
241: %on the source star surface from the center of the source star,
242: %$I(r)$ represents the surface brightness profile of the source
243: %star, $A'$ is the point-source magnification, and $\rho_\star$
244: %represents the source radius normalized by the Einstein radius.
245: As a result, the signal of the companion is smeared out by the
246: finite-source effect and the degree of the effect depends on
247: the ratio of the caustic size to the size of the source star.
248:
249:
250:
251:
252:
253:
254: \subsection{Excess Maps}
255:
256: To see the difference between the patterns of central perturbations
257: induced by a planet and a binary companion under severe finite-source
258: effect, we construct maps of magnification excess around the central
259: caustics of the individual lens systems. The magnification excess
260: is defined as
261: \begin{equation}
262: \epsilon={A-A_0\over A_0},
263: \label{eq3}
264: \end{equation}
265: where $A$ and $A_0$ represent the lensing magnifications with
266: and without the companion, respectively. For the computation
267: of the magnification, we use a ray-shooting based algorithm
268: developed by \citet{dong06}. The algorithm is optimized for
269: high-magnification events and saves computation time by limiting
270: the range of ray shooting on the image plane to a narrow annulus
271: around the Einstein ring. We take the finite-source effect into
272: consideration by modelling the source brightness profile as
273: \begin{equation}
274: {I(\theta)\over I_0}=
275: 1-\Gamma\left( 1-{3\over 2}\cos\theta\right)
276: -\Lambda \left( 1-{5\over 4}\cos^{1/2}\theta\right),
277: \label{eq4}
278: \end{equation}
279: where $\theta$ is the angle between the normal direction to the
280: source-star surface and the line of sight. We adopt a linear
281: and a square-root limb-darkening coefficients of $(\Gamma,
282: \Lambda)=(-0.46,1.11)$.
283:
284:
285: Figure~\ref{fig:one} and \ref{fig:two} show the constructed
286: excess-pattern maps represented in color scales. In each
287: figure, the map in the upper left panel is constructed by
288: considering the finite-source effect while the map in the
289: lower left panel is for a point source. The abscissa of
290: the map is parallel with the axis connecting the two lens
291: components and the center is located at the caustic center,
292: that is located at the position with an offset from the position
293: of the primary of
294: \begin{equation}
295: \deltavec =
296: \cases{
297: {\bf s}^{-1}q/(1+q) & for $s>1$, \cr
298: -{\bf s}[(1+q)^{-1}-1] & for $s<1$, \cr
299: }
300: \label{eq5}
301: \end{equation}
302: where the sign is positive when the offset vector is directed
303: toward the companion and vise versa. The companion is
304: located on the right and all lengths are normalized by the
305: source radius. Colors are chosen such that the regions with
306: brown and blue-tone colors represent the areas where the
307: magnification is higher ($\epsilon > 0$) and lower ($\epsilon
308: < 0$) than the single-lensing magnification without the companion,
309: respectively. For each tone, the color scale becomes darker at
310: the excess levels of $|\epsilon|=2\%$, 4\%, 8\%, 16\%, and 32\%,
311: respectively.
312:
313:
314:
315: % Figure 1 ----------------------------------------------------
316: \begin{figure}[t]
317: \epsscale{1.2}
318: \plotone{f1.eps}
319: \caption{\label{fig:one}
320: Color-scale maps of magnification excess around the central
321: caustic induced by a planet. The map in the upper left panel
322: is constructed by considering the finite-source effect while
323: the map in the lower left panel is for a point source. The
324: abscissa of the map is parallel with the axis connecting the
325: two lens components and the center is located at the caustic
326: center. In each map, the planet is located on the right and
327: all lengths are normalized by the source radius $\rho_*$.
328: Colors are chosen such that the regions with brown and blue-tone
329: colors represent the areas where the magnification is higher
330: ($\epsilon > 0$) and lower ($\epsilon < 0$) than the single-lensing
331: magnification without the companion, respectively. For each
332: tone, the color scale becomes darker at the excess levels of
333: $|\epsilon| =2\%$, 4\%, 8\%, 16\%, and 32\%, respectively.
334: Each of the encircled maps on the right side shows the excess
335: pattern enclosed by the source star at the time when the center
336: of the source star is located at the position marked by `x'
337: with a corresponding number. The straight lines with arrows
338: represent the source trajectories where the residual of the
339: light curves of the resulting events from that of a single-lensing
340: event are presented in Fig.~\ref{fig:three}.
341: }\end{figure}
342: % -------------------------------------------------------------
343:
344:
345:
346: % Figure 2 ----------------------------------------------------
347: \begin{figure}[t]
348: \epsscale{1.2}
349: \plotone{f2.eps}
350: \caption{\label{fig:two}
351: Color-scale maps of magnification excess around the central
352: caustic induced by the companion of a wide-separation binary.
353: Notations are same in Fig.~\ref{fig:one}.
354: }\end{figure}
355: % -------------------------------------------------------------
356:
357:
358:
359:
360: From the maps, one finds several features that commonly appear
361: in the perturbation patterns of both planetary and binary cases.
362: The first feature is a region of little excess inside a circle
363: with its center located at the caustic center and a radius
364: corresponding to that of the source star (dashed circle in each
365: map). The other feature is the perturbation regions that appear
366: at the edge of the circle with either positive or negative excess.
367: \citet{han08b} pointed out that the very small excess inside the
368: circle is caused by the cancellation of the positive and negative
369: excesses by the finite-source effect. They also indicated that
370: the feature at the edge of the circle (hereafter we refer this
371: feature as `edge feature') is formed by the break of the balance
372: between the positive and negative excesses due to a partial
373: coverage of the strong excess region around the caustic by the
374: source star. These features in combination result in a distinctive
375: signal of a companion in the residual of a light curve, that is
376: characterized by a spike of either positive or negative excess at
377: the moment when the center of the caustic enters or exits the source
378: star and a flat residual region between the spikes. See Fig.\ 3 of
379: \citet{han08b}.
380:
381:
382:
383: \subsection{Diagnosis}
384:
385: Along with the similarities, we find that there also exist
386: systematic differences between the patterns of the perturbations
387: induced by a planet and a binary companion. The most prominent
388: difference shows up in the morphology of the edge features with
389: {\it negative} excess. It is found that the feature induced a
390: binary companion forms a complete annulus, while the feature
391: induced by a planet appears as several arc segments.
392: We find that this trend holds for nearly all planetary
393: systems with mass ratios $q\lesssim 10^{-2}$ and separations
394: where the planet detection efficiency is important.
395: See Fig.\ 1 of \citet{han09}.
396:
397:
398: To search for the origin of these differences, we produce
399: additional maps showing the region enclosed by the source star
400: at the time when the source star is located at various positions
401: of the annulus where the negative edge features systematically
402: occur. These maps are presented on the right side of
403: Figure~\ref{fig:one} and \ref{fig:two}. From close examination
404: of the perturbation pattern, it is found that the morphological
405: differences between the planetary and binary cases are basically
406: caused by the difference in the shape of the caustics. The shape
407: of the caustic induced by the planet lacks symmetry. As a result,
408: the location of the source position where the overall excess is
409: negative depends on the orientation of the source position with
410: respect to the caustic. For example, when the source is located
411: at the position on the annulus toward the direction of the
412: sharp-pointed cusp marked as `1', the number of strong-negative
413: excess regions within the area encompassed by the source star
414: is three (two just outside the fold caustics and the other
415: inside of the caustic) while there exists only a single
416: strong-positive excess region (the one extending from the cusp),
417: resulting in overall negative excess. By contrast, when the
418: source is located at the opposite position marked as `4', there
419: are equal numbers of positive and negative excess regions of
420: three, respectively, and thus the resulting excess is very small.
421: In contrast to the shape of the planet-induced caustic, the shape
422: of the binary-induced caustic is more symmetric and thus the
423: regions of negative excess appear in all regions along the annulus.
424:
425:
426:
427:
428: % Figure 3 ----------------------------------------------------
429: \begin{figure}[t]
430: \epsscale{1.2}
431: \plotone{f3.eps}
432: \caption{\label{fig:three}
433: Residuals of lensing light curves from a single-lensing
434: one for example planetary and wide-separation binary events.
435: Note that for the case of the binary, well-developed negative
436: spikes occur at both moments when the center of the caustic
437: enters and exits the source star. On the other hand, a
438: negative spike may {\it not} occur at one (middle panel) or
439: both of the moments (upper panel). The source trajectories
440: responsible for the individual events are marked in
441: Fig.~\ref{fig:one} and \ref{fig:two} with the corresponding
442: roman numbers. Time is normalized by the duration required
443: for the lens to cross the source radius, $t_c$.
444: }\end{figure}
445: % -------------------------------------------------------------
446:
447:
448:
449:
450: The systematic difference in the morphology of the perturbation
451: patterns enables one to distinguish between the perturbations
452: caused by planetary and binary companions. Due to the continuity
453: of the negative-edge feature in the excess pattern, the perturbation
454: in the light curve of a lensing event induced by a wide-separation
455: binary will have well-developed negative spikes in the residual at
456: both moments when the caustic center enters and exits the source
457: star. On the other hand, the edge feature induced by a planet is
458: split into segments and thus the resulting light curve will not
459: often have a negative spike in the residual at both or either of
460: the moments of entrance and exit. Therefore, the absence of the
461: double negative-spike feature can be used as a simple diagnostic
462: to immediately distinguish the planetary interpretation from the
463: binary interpretation.
464:
465:
466: In Figure~\ref{fig:three}, we present residuals of lensing light
467: curves of several example planetary (solid curve) and binary
468: events (dotted curve), where the source trajectories responsible
469: for the individual events are marked in Figure\ref{fig:one} and
470: \ref{fig:two} with the corresponding roman numbers. For the
471: binary case, there always exist two well-developed negative
472: spikes. For a planetary case, on the other hand, there can be
473: no (upper panel) or only a single (middle panel) spike. Setting
474: a critical value of the magnification excess to define the term
475: "well-developed spike" is not easy due to the variation of the
476: perturbation pattern combined with the difficulty in the conversion
477: of the observed flux into magnification. Nevertheless, the term
478: can be defined in the observational point of view as {\it a
479: well-resolved dip in the residual observed with high enough precision}.
480: Under this definition, a spike with deviation $\gtrsim 3\%$ can be
481: well resolved considering that the photometric precision of the
482: current follow-up observations reaches $\sim 1\%$ at the peaks of
483: high-magnification events. We note that a planet can produce
484: perturbations with double negative-spike features as shown in the
485: lower panel of Figure~\ref{fig:three}. Therefore, the existence
486: of double negative spikes does not necessarily confirm that the
487: perturbation is caused by a binary companion. In this case, the
488: proposed diagnostic cannot be used and detailed modelling is
489: required to distinguish between the two interpretations.
490:
491:
492: \section{Conclusion}
493:
494: We compared the patterns of central perturbations induced by
495: a planet and a binary companion under severe finite-source
496: effect. We found that the most prominent difference shows
497: up in the morphology of the edge features with negative excess,
498: where the feature induced a binary companion forms a complete
499: annulus, while the feature induced by a planet appears as
500: several arc segments. This difference provides a useful
501: diagnostic for immediate discrimination of a planet-induced
502: perturbation from that induced by a binary companion, where
503: the absence of a well-developed dip in the residual from the
504: single-lensing light curve at both or either of the moments
505: of the caustic center's entrance into or exit from the source
506: star surface indicates that the perturbation is produced by a
507: planetary companion. We found that that this difference is
508: basically caused by the difference between the shapes of the
509: central caustics induced by the two different types of companions.
510:
511:
512:
513:
514:
515:
516: \acknowledgments
517: This work was supported by the Science Research Center (SRC)
518: program of Korea Science and Engineering Foundation (KOSEF).
519:
520:
521:
522: \begin{thebibliography}{99}
523: \frenchspacing
524:
525: %\bibitem[Abe et al.(2004)]{abe04}
526: %Abe, F., et al.\ 2004, Science, 305, 1264
527:
528: %\bibitem[Albrow et al.(2000)]{albrow00}
529: %Albrow, M.\ D., et al.\ 2000, \apj, 535, 176
530:
531: \bibitem[An(2005)]{an05}
532: An, J.\ H.\ 2005, \mnras, 356, 1409
533:
534: \bibitem[Bennett et al.(2008)]{bennett08}
535: Bennett, D.\ P., et al.\ 2008, \apj, 684, 663
536:
537: %\bibitem[Bond et al.(2002)]{bond02}
538: %Bond, I.\ A., et al.\ 2002, \mnras, 333, 71
539:
540: \bibitem[Bond et al.(2004)]{bond04}
541: Bond, I.\ A., et al.\ 2004, \apj, 606, L155
542:
543: \bibitem[Beaulieu et al.(2006)]{beaulieu06}
544: Beaulieu, J.-P., et al.\ 2006, Nature, 439, 437
545:
546: \bibitem[Chung et al.(2005)]{chung05}
547: Chung, S.-J., et al.\ 2005, \apj, 630, 535
548:
549: \bibitem[Dominik(1999)]{dominik99}
550: Dominik, M.\ 1999, \aap, 349, 108
551:
552: \bibitem[Dominik \& Hirshfeld(1996)]{dominik96}
553: Dominik, M., \& Hirshfeld\ 1996, \aap, 313, 841
554:
555: \bibitem[Dominik et al.(2008)]{dominik08}
556: Dominik, M., et al.\ 2008, Astronomische Nachrichen, 329, 248
557:
558: \bibitem[Dong et al.(2006)]{dong06}
559: Dong, S., et al.\ 2006, \apj, 642, 842
560:
561: \bibitem[Dong et al.(2008)]{dong08}
562: Dong, S., et al.\ 2008, \apj, submitted (arXiv:0809.2997)
563:
564: \bibitem[Gaudi et al.(2008)]{gaudi08}
565: Gaudi, B.\ S., et al.\ 2008, Science, 319, 927
566:
567: \bibitem[Gould et al.(2006)]{gould06}
568: Gould, A., et al.\ 2006, \apj, 644, L37
569:
570: \bibitem[Griest \& Safizadeh(1998)]{griest98}
571: Griest, K., \& Safizadeh, N.\ 1998, \apj, 500, 37
572:
573: %\bibitem[Han(2007)]{han07}
574: %Han, C.\ 2007, \apj, 661, 1202
575:
576: \bibitem[Han(2009)]{han09}
577: Han, C.\ 2009, \apj, 691, 000
578:
579: \bibitem[Han \& Kim(2008)]{han08b}
580: Han, C., \& Kim, D.\ 2008, \apj, submitted (arXiv:0809.4292)
581:
582: \bibitem[Han \& Gaudi(2008)]{han08a}
583: Han, C., \& Gaudi, B.\ S.\ 2008, \apj, 689, 53
584:
585: \bibitem[Han et al.(2005)]{han05}
586: Han, C., Gaudi, B.\ S., An, J.\ H., \& Gould, A.\ 2005, \apj, 618, 962
587:
588: %\bibitem[Rattenbury et al.(2002)]{rattenbury02}
589: %Rattenbury, N.\ J., Bond, I.\ A., Skuljan, J., \& Yock, P.\ C.\ M.\ 2002, \mnras, 335, 159
590:
591: \bibitem[Udalski et al.(2005)]{udalski05}
592: Udalski, A., et al.\ 2005, \apj, 629, L109
593:
594: %\bibitem[Yoo et al.(2004)]{yoo04}
595: %Yoo, J., et al.\ 2004, \apj, 616, 1204
596:
597:
598: \end{thebibliography}
599:
600: \end{document}
601:
602:
603:
604:
605:
606:
607: