1:
2: % For ApJ submission -----------------------------------------
3: %\documentclass[12pt,preprint]{/home/cheongho/LatexStyle/AAS5.02/aastex}
4: %\documentclass[12pt, preprint]{aastex}
5: %\def\baselinestretch
6:
7: \documentclass{emulateapj}
8: \usepackage{apjfonts}
9: \lefthead{HAN} \righthead{LENS-LIGHT ANALYSIS METHOD}
10:
11: %==== CUSTOMIZED LATEX MACROS ========================================
12: % Italic bold font
13: \def\dslash{\mathbin{/\mkern-4mu/}}
14:
15: \newcommand{\dvec}{\mbox{\boldmath $d$}}
16: \newcommand{\rvec}{\mbox{\boldmath $r$}}
17: \newcommand{\uvec}{\mbox{\boldmath $u$}}
18: \newcommand{\vvec}{\mbox{\boldmath $v$}}
19: \newcommand{\xvec}{\mbox{\boldmath $x$}}
20: \newcommand{\yvec}{\mbox{\boldmath $y$}}
21: \newcommand{\zvec}{\mbox{\bf z}}
22: \newcommand{\zetavec}{\mbox{\boldmath $\zeta$}}
23: \newcommand{\svec}{\bold s}
24:
25: \newcommand{\te}{t_{\rm E}}
26: \newcommand{\re}{r_{\rm E}}
27: \newcommand{\rh}{r_{\rm H}}
28: \newcommand{\retilde}{\tilde{r}_{\rm E}}
29: \newcommand{\thetae}{\theta_{\rm E}}
30:
31: % Equation align
32: \def\eqalign#1{\null\,\vcenter{\openup\jot
33: \ialign{\strut\hfil$\displaystyle{##}$&$
34: \displaystyle{{}##}$\hfil \crcr#1\crcr}}\,}
35:
36: %===================================================================
37:
38:
39: \begin{document}
40: \title{On the Feasibility of Characterizing Lens Stars in
41: Future Space-Based Microlensing Surveys}
42:
43: % ==================================================================
44:
45: \author{Cheongho Han}
46: \affil{Department of Physics, Institute for Basic Science
47: Research, Chungbuk National University, Chongju 361-763, Korea;
48: cheongho@astroph.chungbuk.ac.kr}
49:
50:
51: %\submitted{Submitted to The Astrophysical Journal}
52:
53:
54:
55: \begin{abstract}
56:
57: If a light-emitting star is responsible for a gravitational microlensing
58: event, the lens can be characterized by analyzing the blended light from
59: the lens. In this paper, we investigate the feasibility of characterizing
60: lenses by using this method in future space-based lensing surveys. To
61: judge the feasibility of the method, we estimate the portions of events
62: whose blended flux $F_b$ can be firmly noticed and most of it can be
63: attributed to the lens by carrying out detailed simulations of Galactic
64: bulge lensing events considering various blending sources, including
65: the lens, background stars, and binary companions to the lens and source.
66: From this, it is estimated that among the events to be detected from a
67: survey using a 1 m space telescope, $\sim 27\%$ will have blending
68: fractions of $F_b/F\geq 10\%$ and the blended flux of half of these
69: events will be contaminated more than $(F_b-F_L)/F_b=20\%$ by the flux
70: from blending sources other than the lens, implying the contamination
71: of the blended flux will be substantial. Although the contamination by
72: the background stars can be reduced by using an instrument with a higher
73: resolution, it is estimated that the blended flux of more than 1/3 of
74: events will still be contaminated (mostly by binary companions) even
75: using a telescope equivalent to the {\it Hubble Space Telescope}, assuming
76: 50\% binary frequency. We, therefore, conclude that caution and
77: consideration of the blending contaminants are required in applying the
78: lens-light analysis method.
79:
80: \end{abstract}
81:
82: \keywords{gravitatinal lensing}
83:
84: \section{Introduction}
85:
86: Proposed by \citet{paczynski86}, initiated by the MACHO
87: \citep{alcock93, alcock95}, EROS \citep{aubourg93}, and OGLE
88: \citep{udalski94} collaborations, and succeeded by the OGLE-II
89: \citep{wozniak01}, OGLE-III \citep{udalski03}, and MOA \citep{bond01}
90: collaborations, the Galactic microlensing experiments have been and
91: are searching for microlensing events toward the Galactic bulge and
92: Magellanic Cloud fields. Current experiments are routinely detecting
93: more than 500 events every season and the total number of detected
94: events now exceeds 2000. Except for several dozens of events, most of
95: these events have been detected toward the Galactic bulge field.
96:
97:
98: The light curve of a microlensing event is represented by
99: \begin{equation}
100: F = AF_0 + F_b;\qquad A={u^2+2\over u(u^2+4)},
101: \label{eq1}
102: \end{equation}
103: where $F_0$ is the baseline flux of the source star, $F_b$ is the
104: blended flux, $A$ represents the lensing magnification, and $u$ is
105: the lens-source separation normalized by the angular Einstein radius
106: $\thetae$. When the lensing event is observed, one can measure
107: three observables related to the physical parameters of the lens,
108: namely the Einstein timescale $\te$, the angular Einstein radius
109: $\thetae$, and the Einstein ring radius projected onto the observer
110: plane $\retilde$. These observables are related to the underlying
111: physical lens parameters of the mass $M$, relative lens-source
112: parallax $\pi_{\rm rel}={\rm AU}\ (D_{\rm L}^{-1}- D_{\rm S}^{-1})$,
113: and proper motion $\mu_{\rm rel}$, by \citep{gould00b}
114: \begin{equation}
115: \te = {\thetae\over \mu_{\rm rel}},\ \ \
116: \thetae=\sqrt{4GM\pi_{\rm rel} \over c^2\ {\rm AU}},\ \ \
117: \retilde = \sqrt{4GM\ {\rm AU} \over c^2 \pi_{\rm rel}}.
118: \label{eq2}
119: \end{equation}
120: Once all these observables are measured, the lens mass is
121: uniquely determined by
122: \begin{equation}
123: M=\left( {c^2 \over 4G} \right)\retilde \thetae.
124: \label{eq3}
125: \end{equation}
126: Among these three observables, however, only $\te$ is routinely
127: measurable from the lensing light curve and $\thetae$ and $\retilde$
128: can be measured only in some favorable cases. As a result, of the
129: roughly 2000 microlensing events detected to date, there have been
130: only about a dozen events for which $\thetae$ has been measured and
131: another dozen for which $\retilde$ has been measured. Moreover,
132: simultaneous measurements of $\thetae$ and $\retilde$ are very
133: rare, and so far there are only three events, MACHO-LMC-5,
134: OGLE-2000-BLG-5, and OGLE-2003-BLG-238, for which the microlens
135: mass and distance have been reliably determined \citep{alcock01,
136: an02, jiang04, gould04}.
137:
138:
139: If a light-emitting star is responsible for a microlensing event, the
140: lens can be characterized by another method. In this case, the light
141: from the lens star contributes to the observed flux and thus one can
142: characterize the lens star by analyzing the flux from the lens
143: \citep{kamionkowski95, buchalter96, mao98}: `lens-light analysis'
144: method. If realized, this method can be applied to a large number of
145: events because it is believed that a majority of events detected toward
146: the Galactic bulge field are caused by stars \citep{han03}. Currently,
147: this method is not being used due to two major observational difficulties.
148: First, because of the high number density of source stars in the Galactic
149: bulge field combined with the poor resolution of the ground-based observation,
150: current lensing surveys can effectively monitor only bright resolved stars,
151: and thus the relative contribution of the lens flux to the total
152: observed flux is small. Second, even if the source star is relatively
153: faint, most blended flux comes from nearby background stars, and thus
154: the flux from the lens has a meager contribution to the total blended
155: flux.
156:
157:
158: However, the situation might be different in future space-based
159: microlensing surveys, such as the {\it Galactic Exoplanet Survey Telescope}
160: ({\it GEST}) mission \citep{bennett02}, whose concept was later
161: succeeded by {\it Microlensing Planet Finder} ({\it MPF}) mission
162: \citep{bennett04}. The proposed mission plans to monitor faint
163: main-sequence stars to optimize the detections of terrestrial planetary
164: signals by minimizing finite-source effect. In addition, events to be
165: observed by the mission would suffer from much less blending caused by
166: background stars thanks to the high resolution provided by space
167: observation. Then, the proposed space-based microlensing survey might
168: be able to detect enough light from the lens stars for a significant
169: fraction of the event sample. By measuring the apparent magnitude and
170: color of the lens star, then, it would be possible to determine the
171: spectral type of the lens star and thus can estimate the approximate
172: lens mass and distance.
173:
174:
175: The feasibility of the lens-light analysis method depends greatly on
176: whether the flux from the lens accounts for not only a substantial
177: fraction of the observed flux but also most of the blended flux. If the
178: blended flux is seriously contaminated by the flux from blending sources
179: other than the lens, analysis of the blended light would be difficult or
180: might lead to wrong characterization of the lens star. Despite the high
181: resolution of the space mission, accidental overlap of the source star
182: image with those of background stars will be unavoidable for a fraction
183: of events. In addition, binary companions to either the source star or
184: the lens can also contaminate the blended flux. Therefore, it is
185: important to evaluate the degree of contamination to the blended flux
186: by these possible contaminants. In this paper, we judge the feasibility
187: of the lens-light analysis method in future space-based microlensing
188: surveys by estimating the fractions of events where the blended flux
189: can be firmly noticed and the most of the blended flux is attributed
190: to the lens star.
191:
192:
193:
194:
195:
196:
197: \section{Microlensing Event Simulations}
198:
199: To estimate the fraction of events whose blended flux can be noticed
200: and attributed mostly to the lens star, we carry out detailed simulations
201: of Galactic bulge microlensing events considering blending caused by
202: various sources, including the lens star, background stars (denoted as
203: `bs'), and the companions to the source and lens (denoted as `S2' and
204: `L2', respectively).
205:
206:
207: The simulations of lensing events are proceeded as follows.
208: \begin{enumerate}
209: \item
210: %{\bf stellar density of the field:}
211: We assume that the survey is conducted toward the Baade's Window with
212: the central coordinates of $(l,b)=(1^\circ,-4^\circ)$. Based on the
213: luminosity function (LF) of \citet{holtzman98} constructed using the
214: {\it Hubble Space Telescope} ({\it HST}), we normalize the stellar number
215: density to $1.2\times 10^4/{\rm arcmin}^2$ for stars up to $M_V=12.25$,
216: which corresponds to the brightness of a M5 star. Blending by background
217: stars is taken into consideration by assuming that the images of stars
218: cannot be resolved if the separation between stars is less than the
219: the angular resolution, $\theta_{\rm res}$. To see the dependence on
220: the angular resolution, we test two cases of $\theta_{\rm res}=
221: 0''\hskip-2pt .1$ and $0''\hskip-2pt .25$, which correspond to the
222: resolutions of the {\it HST} (with an aperture of ${\cal D}=2.4\ {\rm m}$)
223: and {\it MPF} (with ${\cal D}=1.0\ {\rm m}$), respectively.
224:
225:
226: \item
227: %{\bf density and velocity models:}
228: For the density and velocity distributions of the lens and source,
229: we adopt the barred bulge model of \citet{han03}. In the model, the
230: bulge mass distribution is scaled by the deprojected infrared light
231: density profile of \citet{dwek95}, specifically model G2 with
232: $R_{\rm max}=5$ kpc in their Table 1. The velocity distribution of
233: the bulge is deduced from the tensor virial theorem and the resulting
234: distribution of the lens-source transverse velocity, $\vvec$, is
235: listed in Table 1 of \citet{han03}, specifically non-rotating barred
236: bulge model. In our simulations, we consider only bulge self-lensing
237: events because they account for the majority of Galactic bulge events
238: \citep{zhao96}.
239:
240: \item
241: %{\bf source brightness:}
242: The absolute magnitude of the source star is assigned based on the
243: binary-corrected $V$-band LF of stars in the solar neighborhood listed
244: in \citet{allen00}.\footnote{For the source brightness, it might be
245: thought that using the LF of Galactic bulge stars obtained from the
246: {\it HST} observation would be a better choice. However, this LF is
247: not binary-corrected because the companions to source stars cannot be
248: resolved even with the {\it HST}. Since the purpose of this paper is
249: investigating the effect of contaminants to the blended flux and the
250: binary companion to the source star is an important contaminant candidate,
251: we use binary-corrected LF of solar-neighborhood stars instead of that
252: of binary-corrected LF of Galactic bulge stars.} We assume that the
253: brightness range of the source star to be monitored by the survey is
254: $2.7\lesssim M_V\lesssim 7.5$, which corresponds to early F to late
255: K-type main-sequence stars. The apparent source magnitude is determined
256: considering the distance to the source star and extinction. The extinction
257: is determined such that the source star flux decreases exponentially with
258: the increase of the dust column density, where the dust column density is
259: computed based on an exponential dust distribution model with a scale
260: height of $h_z=120\ {\rm pc}$, i.e.\ $\propto \exp(-|z|/h_z)$. We
261: normalize the amount of extinction such a way that that $A_V=1.28$
262: for a star located at $D_{\rm S}=8\ {\rm kpc}$ following the measurement
263: of \citet{holtzman98}. The brightness of the companion to the source
264: star is assigned under the assumption that the companion follows the
265: same LF of the primary star \citep{duquennoy91}. Since the star to be
266: monitored by the survey will be selected based on the combined flux of
267: the primary and companion, there is no lower limit of the companion
268: star brightness. We note, however, that the terms `primary' does not
269: designate the brighter component of the binary, but it indicates the
270: star participating in the lensing magnification, although the primary
271: source star is brighter than the companion in most cases.
272:
273:
274: \item
275: %{\bf lens mass and brightness:}
276: We assign the lens mass based on the mass function (MF) of
277: \citet{gould00a}. The model MF is composed of stars following a double
278: power-law distribution, brown dwarfs (BDs), and stellar remnants of
279: white dwarfs (WDs), neutron stars (NSs), and black holes (BHs), where
280: the mass fractions of the individual lens components are
281: \begin{equation}
282: {\rm stars}:{\rm BD}:{\rm WD}:{\rm NS}:{\rm BH}=62:7:22:6:3.
283: \label{eq4}
284: \end{equation}
285: Once the mass of the lens is chosen and it is turned out to be a
286: stellar lens, its brightness is assigned based on the mass by using
287: the mass-$M_V$ relation listed in \citet{allen00}. If the lens is
288: a remnant, on the other hand, it is assumed to be dark. The brightness
289: of the companion to the lens is assigned by the same way as the primary
290: lens. Similar to the notations for the source stars, the primary lens
291: designates the binary component that is involved with the lensing process.
292: The apparent magnitude of the lens is determined by the same way as that
293: of the source star considering the distance and extinction.
294:
295:
296: \item
297: Once the lens and source are chosen, the rate of the event associated
298: with the lens and source is computed by
299: \begin{equation}
300: \Gamma \propto \rho(D_{\rm S})D_{\rm S}^2 \rho(D_{\rm L}) \sigma v
301: u_{0,{\rm th}},
302: \label{eq5}
303: \end{equation}
304: where $\rho(D)$ is the matter density along the line of sight, the
305: factor $D_{\rm S}^2$ is included to account for the increase of the
306: number of source stars with the increase of $D_{\rm S}$, $\sigma$
307: represents the lensing cross-section corresponding to the diameter
308: of the Einstein ring, i.e.\ $\sigma=2\re=2D_{\rm L}\thetae$, and
309: $u_{0,{\rm th}}$ represents the threshold lens-source separation
310: (normalized by $\thetae$) required for the event detection. Due
311: to blending, the threshold magnification for the event detection
312: becomes higher. The increased threshold magnification is
313: \begin{equation}
314: {A'}_{\rm th}=A_{\rm th}\left( 1+{F_b\over F_{\rm S}} \right)-
315: {F_b\over F_{\rm S}},
316: \label{eq6}
317: \end{equation}
318: where $A_{\rm th}$ is the threshold magnification without blending.
319: Then, the threshold lens-source separation corresponding to the
320: increased threshold magnification \citep{distefano95, han99} is
321: \begin{equation}
322: {u}_{0,{\rm th}} = \left[ {2\over \left( 1-{A'}_{\rm th}^{-2}
323: \right)^{1/2}} -2 \right]^{1/2}.
324: \label{eq7}
325: \end{equation}
326: In our simulations, we set $A'_{\rm th}=3/\sqrt{5}$ following the
327: conventional threshold magnification. Since the future lensing
328: survey will be carried out with a high monitoring frequency, we
329: assume that there is no bias in event detections depending on
330: the event timescale.
331:
332: \end{enumerate}
333:
334:
335:
336:
337: \section{Blending and Contamination to the Blended Flux}
338:
339: We produce a large number ($10^5$) of events from the simulations
340: following the procedure described in the previous section and
341: compute the relative event rates of the individual events. Based
342: on these events and their rates, we then construct the distribution
343: of the blended light fraction $F_b/F$. In order to inspect the
344: contamination of the blended flux by the blending sources other
345: than the lens star, we also construct the distribution of the
346: fraction of the lens flux $F_{\rm L}$ out of the total blended
347: flux, $F_{\rm L}/F_b$, among the events for which blending is
348: firmly detected. We assume that blended light can be noticed
349: if $F_b/F\geq 0.1$. As candidates of the blending contaminants,
350: we consider background stars, and companions to the source and lens,
351: i.e.\ $F_b=F_{\rm L}+F_{\rm bs}+F_{\rm S2}+ F_{\rm L2}$, where
352: $F_{\rm bs}$, $F_{\rm S2}$, and $F_{\rm L2}$ represent the fluxes
353: from the background stars and companions to the source and lens,
354: respectively.
355:
356:
357:
358: % Figure 1 --------------------------------------------------------------
359: \begin{figure}[th]
360: \epsscale{1.2}
361: \plotone{f1.eps}
362: \caption{\label{fig:one}
363: The portion of the Galactic bulge events where the blended light
364: fraction, $F_b/F$, is greater than a certain value $x$.
365: }\end{figure}
366:
367:
368:
369: % Figure 2--------------------------------------------------------------
370: \begin{figure}[th]
371: \epsscale{1.2}
372: \plotone{f2.eps}
373: \caption{\label{fig:two}
374: The portion
375: of the Galactic bulge events where the lens light fraction out of the
376: total blended flux, $F_{\rm L}/F_b$, is greater than a certain value
377: $x$ among the blended events with $F_b/F\geq 10\%$.
378: }\end{figure}
379:
380:
381:
382:
383:
384:
385: In Figure~\ref{fig:one}, we present the distribution of the portion of
386: events where the blended light fraction is greater than a certain value
387: $x$, ${\cal F}( F_b/F\geq x)$. In Figure~\ref{fig:two}, we also present
388: the distribution of the portion of events where the lens light fraction
389: out of the total blended flux is greater than a certain value $x$ among
390: the blended events with $F_b/F\geq 0.1$, ${\cal F}( F_{\rm L} /F_b\geq
391: x\ {\rm and}\ F_b/F\geq 0.1)$. We test two binary frequencies of
392: $f_{\rm bi}=0.5$ and $0.0$, where the latter corresponds to the case
393: that the background star is the only blending contaminant, i.e.\ $F_b=
394: F_{\rm L}+F_{\rm bs}$. To see the blending contribution only by the
395: lens star, we also construct the distribution ${\cal F}(F_b/F\geq x)$
396: under the condition that $\theta_{\rm res} =0''\hskip-2pt.01$ and
397: $f_{\rm bi}=0\%$, i.e.\ $F_b=F_{\rm L}$ (pure lens blending case).
398: In Table~\ref{table1}, we summarize the results of blending and
399: contamination of the blended flux.
400:
401:
402:
403: % Table 1 ---------------------------------------------------------------
404: \begin{deluxetable*}{crcccl}
405: \tablecaption{Blended Event Portions \label{table1}}
406: \tablewidth{0pt}
407: \tablehead{
408: \multicolumn{2}{c}{cases} &
409: \multicolumn{2}{c}{portions} &
410: \multicolumn{2}{c}{comments} \\
411: \colhead{$\theta_{\rm res}$} &
412: \colhead{$f_{\rm bi}$} &
413: \colhead{${\cal F}(F_b/F\geq 0.1)$} &
414: \colhead{${\cal F}(F_{\rm L}/F_b\geq 0.8\ \&\ F_b/F\geq 0.1)$} &
415: \colhead{instrument} &
416: \colhead{blending sources}}
417: \startdata
418: $0''\hskip-2pt .25$ & 50\% & 27\% & 53\% & {\it MPF} & L, bs, S2, L2 \\
419: -- & 0\% & 17\% & 78\% & -- & L, bs \\
420: $0''\hskip-2pt .10$ & 50\% & 23\% & 63\% & {\it HST} & L, bs, S2, L2 \\
421: -- & 0\% & 13\% & 97\% & -- & L, bs \\
422: $0''\hskip-2pt .01$ & 0\% & 12\% & 100\% & & L \\
423: \enddata
424: \tablecomments{
425: The portion of events where the blended light fraction is greater
426: than 10\%, ${\cal F}(F_b/F\geq 0.1)$, and the portion of events
427: where the lens light fraction out of the total blended flux is
428: greater than 80\% among the blended events with $F_b/F\geq 0.1$,
429: ${\cal F} (F_{\rm L}/F_b\geq 0.8\ \&\ F_b/F\geq 0.1)$. The values
430: are estimated under various angular resolutions $\theta_{\rm res}$
431: of the instrument and assumptions of the binary frequency $f_{\rm bi}$.
432: }
433: \end{deluxetable*}
434:
435:
436:
437: From Table~\ref{table1} and Figure~\ref{fig:one} and \ref{fig:two},
438: we find the following results.
439: \begin{enumerate}
440: \item
441: From the distribution of ${\cal F}( F_b/F\geq x)$ for the
442: pure-lensing blending case ($\theta_{\rm res}=0''\hskip-2pt.01$
443: and $f_{\rm bi}=0\%$), one finds that the lens does contribute
444: to the observed flux for a substantial portion of events. We
445: find that the flux from the lens star will contribute to the
446: observed flux more than $F_b/F\geq 0.1$ for $\sim 12\%$ of events.
447:
448: \item
449: However, we find that the contribution to the blended flux by sources
450: other than the lens star is not negligible as well. Even not considering
451: binary companions ($f_{\rm bi}=0\%$), we estimate that the blended flux
452: will be contaminated by the flux from background stars by more than
453: $(F_b-F_{\rm L})/F_b=20\%$ for nearly a quarter of events with
454: $F_b/F\geq 0.1$ to be detected from the survey using a 1 m space
455: telescope.
456:
457: \item
458: Although the contamination by the background stars can be reduced by
459: using an instrument with a higher resolution, the contamination by the
460: binary companions to the source and lens is unavoidable. When these
461: additional contaminants are considered, we estimate that the portion
462: of events that will suffer from relatively less severe blending
463: ($F_{\rm L}/ F_{b}\geq 0.8$) will be only 53\% and 63\% for surveys
464: using instruments equivalent to the {\it MPF} and {\it HST}, respectively.
465:
466: \end{enumerate}
467:
468:
469:
470: \section{Conclusion}
471:
472: We investigated the feasibility of characterizing lens stars from the
473: analysis of the blended lens flux in the future space-based lensing
474: surveys. We judged this feasibility by estimating the portions of
475: events whose blended flux can be firmly noticed and most of the blended
476: flux can be attributed to the lens star based on detailed simulations
477: of bulge lensing events considering various blending sources, including
478: the lens, background stars, and the binary companions to the lens and
479: source. From this estimation, we found that although the flux from
480: the lens star will contribute to the observed flux for a substantial
481: fraction of events, the contamination of the blended flux by the light
482: from the blending sources other than the lens star will not be negligible.
483: Considering these blending contaminants, we estimated that that among
484: the events to be detected from a survey using a 1 m space telescope,
485: $\sim 27\%$ will have blending fractions of $F_b/F\geq 10\%$ and the
486: blended flux of half of these events will be contaminated more than
487: $(F_b-F_L)/F_b=20\%$ by the flux from blending sources other than the
488: lens. Although the contamination by the background stars can be reduced
489: by using an instrument with a higher resolution, it is estimated that
490: the blended flux of more than 1/3 of events will still be contaminated
491: (mostly by binary companions) even using a telescope equivalent to
492: {\it HST}, assuming 50\% binary frequency. We, therefore, conclude
493: that caution and consideration of the blending contaminants are required
494: in applying the lens-light analysis method
495:
496:
497:
498:
499: \acknowledgments
500: This work was supported by the Astrophysical Research Center for the
501: Structure and Evolution of the Cosmos (ARCSEC") of Korea Science and
502: Engineering Foundation (KOSEF) through Science Research Program
503: (SRC) program.
504:
505:
506:
507: \begin{thebibliography}{99}
508: \frenchspacing
509: \bibitem[An et al.(2002)]{an02}
510: An, J.H., et al.\ 2002, \apj, 572, 521
511:
512: \bibitem[Alcock et al.(1993)]{alcock93}
513: Alcock, C., et al.\ 1993, Nature, 365, 621
514:
515: \bibitem[Alcock et al.(1995)]{alcock95}
516: Alcock, C., et al.\ 1995, \apj, 445, 133
517:
518: \bibitem[Alcock et al.(2001)]{alcock01}
519: Alcock, C., et al.\ 2001, Nature, 414, 617
520:
521: \bibitem[Allen(2000)]{allen00}
522: Allen, C.W.\ 2000, Astrophysical Quantities, eds. A.\ N.\ Cox
523: (The Athlon Press, London)
524:
525: \bibitem[Aubourg et al(1993)]{aubourg93}
526: Aubourg, E., et al.\ 1993, Nature, 365, 623
527:
528: \bibitem[Bennett \& Rhie(2002)]{bennett02}
529: Bennett, D.P., \& Rhie, S.H.\ 2002, \apj, 574, 985
530:
531: \bibitem[Bennett et al.(2004)]{bennett04}
532: Bennett, D.P., et al.\ 2004, BAAS 205 \#11.26
533:
534: \bibitem[Bond et al.(2001)]{bond01}
535: Bond, I., et al.\ 2001, \mnras, 327, 868
536:
537: \bibitem[Buchalter, Kamionkowski, \& Rich(1996)]{buchalter96}
538: Buchalter, A., Kamionkowski, M., \& Rich, R.M.\ 1996, \apj, 469, 676
539:
540: \bibitem[Di Stefaano \& Esin(1995)]{distefano95}
541: Di Stefano, R., \& Esin, A.A.\ 1995, \apj, 448, L1
542:
543: \bibitem[Duquennoy \& Mayor(1991)]{duquennoy91}
544: Duquennoy, A., \& Mayor, M.\ 1991, \aap, 248, 485
545:
546: \bibitem[Dwek et al.(1995)]{dwek95}
547: Dwek, E., et al.\ 1995, \aj, 445, 716
548:
549: \bibitem[Gould(2000a)]{gould00a}
550: Gould, A.\ 2000, \apj, 539, 928
551:
552: \bibitem[Gould(2000b)]{gould00b}
553: Gould, A.\ 2000, \apj, 542, 785
554:
555: \bibitem[Gould(2004)]{gould04}
556: Gould, A.\ 2004, \apj, 606, 319
557:
558: \bibitem[Han(1999)]{han99}
559: Han, C.\ 1999, \mnras, 309, 373
560:
561: \bibitem[Han \& Gould(2003)]{han03}
562: Han, C., \& Gould, A.\ 2003, \apj, 592, 172
563:
564: \bibitem[Holtzman et al.(1998)]{holtzman98}
565: Holtzman, J.A., Watson, A.M., Baum, W.A., Grillmair, C.J., Groth, E.J.,
566: Light, R.M., Lynds, R., \& O'Neil, E.J.\ 1998, \aj, 115, 1946
567:
568: \bibitem[Jiang et al.(2004)]{jiang04}
569: Jiang, G., et al.\ 2004, \apj, 617, 1307
570:
571: \bibitem[Kaminokowski(1995)]{kamionkowski95}
572: Kamionkowski, M.\ 1995, \apj, 442, L9
573:
574: \bibitem[Mao, Reetz, \& Lennon(1998)]{mao98}
575: Mao, S., Reetz, J., \& Lennon, D.J.\ 1998, \aap, 338, 56
576:
577: \bibitem[Paczy\'nski(1986)]{paczynski86}
578: Paczy\'nski, B.\ 1986, \apj, 304, 1
579:
580: \bibitem[Udalski et al.(1994)]{udalski94}
581: Udalski, A., et al.\ 1994, Acta Astron., 44, 165
582:
583: \bibitem[Udalski(2003)]{udalski03}
584: Udalski, A.\ 2003, Acta Astron., 53, 291
585:
586: \bibitem[Wo\'zniak et al.(2001)]{wozniak01}
587: Wo\'zniak, P.R., Udalski, A., Szyma\'nski M., Kubiak, M.,
588: Pietrzy\'nski, G., Soszy\'nski, I., \& \.Zebru\'n, K.\ 2001,
589: Acta Astron., 51, 175
590:
591: \bibitem[Zhao, Rich, \& Spergel(1996)]{zhao96}
592: Zhao, H., Rich, R.M., \& Spergel, D.N.\ 1996, \mnras, 282, 175
593:
594:
595: \end{thebibliography}
596:
597: \end{document}
598: