1: %\documentstyle[aasms4,natbib]{article}

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

3: \def\kms{{\rm km\,s^{-1}}}

4: \def\kpc{{\rm kpc}}

5: \def\erf{{\rm erf}}

6: \def\masyr{{\rm mas}\,{\rm yr}^{-1}}

7: \def\lim{{\rm lim}}

8: \def\det{{\rm det}}

9: \def\obs{{\rm obs}}

10: \def\nom{{\rm nom}}

11: \def\true{{\rm true}}

12: \def\mod{{\rm mod}}

13: \def\rpm{{\rm rpm}}

14: \def\bv{{\bf v}}

15: \newcommand{\bdv}[1]{\mbox{\boldmath$#1$}}

16: \def\e{{\rm E}}

17: \def\bpi{{\bdv{\pi}}}

18: \def\bvt{{\bdv{\tilde v}}}

19: \def\ret{{\tilde r_\E}}

20: \def\rehat{{\hat r_\E}}

21: \def\rel{{\rm rel}}

22: \def\pirel{\pi_{\rm rel}}

23: \def\au{{\rm AU}}

24: \def\mbr{{M_{\rm brk}}}

25: \def\bmurel{{\bdv{\mu}_{\rm rel}}}

26: \def\bmu{{\bdv{\mu}}}

27: \def\btheta{{\bdv{\theta}}}

28: \def\kms{{\rm km}\,{\rm s}^{-1}}

29: 

30: \begin{document}

31: 

32: \title{Probing MACHOs Toward the Galactic Bulge}

33: 

34: \author{Andrew Gould}

35: \affil{Department of Astronomy, The Ohio State University,

36: 140 W.\ 18th Ave., Columbus, OH 43210}

37: \authoremail

38: %\email

39: {gould@astronomy.ohio-state.edu}

40: 

41: \singlespace

42: 

43: \begin{abstract}

44: 

45: If the massive compact halo object (MACHO) fraction of the 

46: Galactic dark halo is $f\sim 20\%$ as suggested by some microlensing

47: experiments, then about $1.2\%$ of lensing events toward the

48: Galactic bulge are due to MACHOs.  For the 40\% of these that lie

49: nearby ($D_l<4\,\kpc$), measurement of their distance $D_l$ would

50: distinguish them from bulge lenses, while measurement of their

51: transverse velocity $\bv_l$ would distinguish them from disk lenses.

52: Hence, it would be possible to identify about $0.5\%(f/20\%)$ of

53: all events as due to MACHOs.  I show that a planned experiment

54: using the {\it Space Interferometry Mission (SIM PlanetQuest)}

55: could thereby detect 1 or 2 such events.  This is at the margin

56: of what is required because of a small, but non-negligible background

57: from spheroid stars.

58: 

59: \end{abstract}

60: \keywords{dark matter -- galaxies: stellar content --

61: gravitational lensing -- instrumentation: interferometers}

62: %\clearpage

63: %\newpage

64:  

65: \section{Introduction

66: \label{sec:intro}}

67: 

68: Following the suggestion of \citet{pac86}, the MACHO \citep{alcock93}

69: and EROS \citep{aubourg93} collaborations began searching for dark matter

70: in the form of massive compact halo objects (MACHOs) by microlensing

71: observations toward the Large Magellanic Cloud (LMC).  This target

72: seemed ideal because of the small column of known populations of

73: stars compared to the

74: huge volume of space that would be home to the putative MACHOs.

75: The microlensing optical depth due to known stars was estimated

76: to be $\tau_{\rm MW}^{\rm LMC}=8\times 10^{-9}$ for the Milky Way 

77: disk \citep{gbf97}

78: and $\tau_{\rm LMC}^{\rm LMC}= 1\times 10^{-8}$ for the LMC itself 

79: \citep{gould95b}.

80: By contrast, if the dark halo were completely composed of MACHOs,

81: their optical depth would be of order,

82: \begin{equation}

83: \tau^{\rm LMC}_{\rm halo}\sim {v_{\rm rot}^2\over c^2} = 5\times 10^{-7},

84: \label{eqn:taulmchalo}

85: \end{equation}

86: roughly 25 times higher.  Here, $v_{\rm rot}=220\,\kms$ is the Milky

87: Way rotation speed.  Hence, when the experiments began, it seemed

88: as though even a crude measurement of $\tau$ would unambiguously

89: determine whether the halo was composed of MACHOs.

90: 

91: A decade later, the situation is far less clear than was anticipated.  MACHO 

92: \citep{alcock00} found $\tau\sim 1\times 10^{-7}$, roughly the root-mean-square

93: of the results expected from MACHOs and stars.  They interpreted this

94: to mean that the halo was 20\% composed of MACHOs and estimated the typical

95: mass to be $M\sim 0.4\,M_\odot$.  On the other hand, the EROS collaboration

96: \citep{afonso03a,tisserand05} found an upper limit for the optical depth

97: due to MACHOs of 5\% of the full-halo value.

98: 

99: One option for resolving this conflict is to explore other lines of sight.

100: \citet{crotts92} and \citet{baillon93} advocated M31, and several

101: collaborations, including AGAPE \citep{ansari99}, Columbia-VATT 

102: \citep{uglesich04}, MEGA \citep{dejong04}, NainiTal \citep{joshi05}

103: POINT-AGAPE \citep{auriere01}, SLOTT-AGAPE \citep{calchi03}, and

104: WeCAPP \citep{riffeser03}, have pursued this suggestion.  In many

105: ways this is substantially more challenging than the observations

106: toward the LMC, simply because M31 is 15 times farther away and

107: hence the sources are substantially fainter.  Events are now being

108: reported from these experiments, and their implications for dark matter

109: should be available soon.

110: 

111: The microlensing target field that has been monitored the most intensively

112: is the Galactic bulge.  Originally proposed by \citet{pac91} and

113: \citet{griest91}, major surveys have been carried out by the

114: OGLE \citep{udalski93,udalski03}, DUO \citep{alard95}, MACHO \citep{popow05},

115: EROS \citep{afonso03b}, and MOA \citep{abe04} collaborations.

116: The primary motivation of both proposals was to probe for disk dark matter

117: and other exotic objects such as a large population of Jupiters.

118: \citet{griest91} does mention that if the halo is composed of MACHOs, then

119: these will give rise to an optical depth 

120: $\tau_{\rm halo}^{\rm bulge}=1.3\times 10^{-7}$, but since this is 4 times

121: smaller than the predicted optical depth due to disk stars

122: $\tau_{\rm disk}^{\rm bulge}=5.1\times 10^{-7}$, there did not appear

123: to be any way to isolate the MACHO events.

124: 

125: Bulge microlensing observations have been enormously fruitful.

126: \citet{kiraga94} showed that the optical depth due to bulge self-lensing

127: was even greater than that due to disk stars.  The high event rate

128: encouraged searches for lensing anomalies due to planetary 

129: companions of the lenses \citep{mao91,gould92,rhie00,albrow01b,gaudi02,abe04},

130: which has now yielded the first firm microlensing planet detection

131: \citep{bond04}.  Bulge microlensing has enabled the 

132: first microlens mass measurement \citep{an02} and the

133: probing of bulge-star atmospheres with $\mu$as resolution both photometrically

134: \citep{alcock97,albrow99,albrow00,fields03} and spectroscopically

135: \citep{castro01,albrow01a,cassan04}.

136: 

137: Here I show that bulge microlensing can also be used to probe for 

138: halo dark matter (MACHOs) in the inner Galaxy.  This seems absurd at

139: first sight because the observed optical depth, 

140: $\tau_{\rm bulge}^{\rm obs}\sim 2\times 10^{-6}$, is about 15 times

141: higher than the rate predicted by \citet{griest91},

142: $\tau_{\rm bulge}^{\rm halo}\sim 1.3\times 10^{-7}$, even assuming that

143: the dark halo were completely composed of MACHOs.  However, the microlensing

144: experiments toward the LMC seem to imply that this fraction is no larger

145: than 20\%, which means that only about 1\% of Galactic bulge microlensing

146: would be due to halo objects.  How would one identify these halo

147: microlensing events within the barrage of microlensing by ordinary

148: bulge and disk stars?

149: 

150: \section{Needle in Haystack

151: \label{sec:haystack}}

152: 

153: Halo lenses are distinguished from disk lenses by the their transverse

154: velocity $\bv_l$ 

155: relative to the Sun, and from bulge lenses by their distance

156: from the Sun, $D_l$ (or equivalently, their absolute parallax $\pi_l$).

157: Hence, to reliably identify the nearby, fast MACHOs, one must

158: reliably measure $\bv_l$ and $\pi_l$.  Since the MACHOs are by definition

159: ``dark'' matter, direct observations of the lens cannot be employed

160: in making these determinations, as they were for example for MACHO-LMC-5

161: \citep{alcock01,dck04,gould04a,gba04}.  Instead, these quantities must

162: be derived entirely from observations of the source during and after the

163: microlensing events.

164: 

165: \subsection{Observables

166: \label{sec:observables}}

167: 

168: These two quantities can be expressed in terms of microlensing observables

169: by (e.g., \citealt{gould00}),

170: \begin{equation}

171: \pi_l = \pirel + \pi_s,\qquad \pirel =\pi_\e\theta_\e

172: \label{eqn:pil}

173: \end{equation}

174: and

175: \begin{equation}

176: \bv_l = {\bmu_\rel + \bmu_s\over \pirel + \pi_s}\au,\qquad

177: \bmu_\rel = {\btheta_\e\over t_\e}.

178: \label{eqn:bvl}

179: \end{equation}

180: Here, $\pi_l$, $\pi_s$ $\bmu_l$, $\bmu_s$ are the absolute parallaxes and

181: proper motions of the lens and source, $\pi_\rel=\pi_l-\pi_s$ and

182: $\bmu_\rel=\bmu_l-\bmu_s$ are the lens-source relative parallax and

183: proper motion, $\theta_\e$ is the angular Einstein radius, $t_\e$ is

184: the Einstein timescale, and $\pi_\e$ is the microlens parallax

185: (i.e., the inverse of the projected Einstein radius, 

186: $\pi_\e = \au/\tilde r_\e$).  The direction of $\btheta_\e$ is that of

187: the lens-source relative proper motion.

188: 

189: In brief, to determine $\pi_l$ and $\bv_l$, one must measure five observables,

190: two 2-vectors ($\bmu_s$ and $\btheta_\e$) and three scalars 

191: ($\pi_s, \pi_\e,$ and $t_\e$).

192: 

193: \subsection{Parameter Measurement

194: \label{sec:parms}}

195: 

196: Two of these five parameters ($\pi_s$ and $\bmu_s$) are

197: related solely to the source, while the remaining three ($\pi_\e$, $t_\e$,

198: and $\btheta_\e$) are microlensing-event parameters.  Of these three, only

199: one $(t_\e)$ is routinely measured during microlensing events.  The other

200: two are higher order parameters.  While there are a variety of 

201: methods to measure $\pi_\e$ and $\theta_\e$ (see \citealt{gould01}),

202: these generally apply to only a small fraction of events.  There

203: are only two events (out of almost 3000 discovered) for which

204: both parameters have been measured from microlensing data alone 

205: \citep{an02,kubas05}, and both of these were binary lenses.

206: 

207: The only known way to {\it routinely} determine $\btheta_\e$

208: is by high-precision astrometric measurements of the microlensing

209: event \citep{HNP95,MY95,Wa95,Pa98,BSV98}.  The centroid of the microlensed

210: images deviates from the source position by an amount and direction

211: that yields both components of $\btheta_\e$.

212: 

213: The only known way to {\it routinely} determine $\pi_\e$ is to

214: make photometric measurements of the event from two locations separated

215: by of order $\tilde r_\e$ \citep{Re66,gould94}.  The difference in the

216: event parameters then yields both the size of $\tilde r_\e$ and the 

217: direction of motion (the latter potentially confirming the direction

218: extracted from $\btheta_\e$).

219: Since $\tilde r_\e \sim O(\au)$, in practice this means placing a satellite

220: in solar orbit.  Although there is a four-fold ambiguity in the determination

221: of $\pi_\e$,

222: this can be resolved by higher-order effects \citep{gould95a}.  Moreover,

223: measurement of the direction of $\btheta_\e$ also helps resolve this

224: degeneracy.

225: 

226: 

227: \subsection{{\it SIM PlanetQuest} Measurements

228: \label{sec:sim}}

229: 

230: 

231: \citet{gs99} showed that the 

232: {\it Space Interferometry Mission (SIM PlanetQuest)} combined with

233: ground-based photometry,  could determine both

234: of these parameters with good $\sim 3\%$ precision with about 5 hours total 

235: obsrvation time for bright $(I\sim 15)$

236: events having typical lens parameters.  

237: Moreover, they showed that the same observations would

238: also yield good measurements of $\pi_s$ and $\bmu_s$.  Hence, {\it SIM}

239: (combined with ground-based photometry) could measure all the required

240: quantities for about 200 events with about 1000 hours of observing time.

241: Indeed, a {\it SIM} Key Project has been awarded 1200 hours of observation

242: time to carry out such observations.  The main objective of this project

243: is to measure the bulge mass function but the same observations could

244: cull out the handful of halo events that could be present in the same

245: sample.

246: 

247: {\it SIM} has been descoped since \citet{gs99} made their analysis.

248: The new performance is not precisely known but it is likely that the

249: precision will degrade to something like $\sim 5\%$ for $\pi_\e$ and

250: $\sim 10\%$ for $\theta_\e$ for the canonical events considered by

251: \citet{gs99}.  Moreover, it is unlikely that 200 $I=15$ events will

252: be found during the 5-year primary {\it SIM} mission, and using

253: fainter sources (e.g., $I=16.5$) would further degrade the

254: precision by a factor 2.  Nevertheless, as I show below, this

255: precision would be quite adequate for distinguishing halo lenses.

256: 

257: \section{Background from Spheroid/Bulge Stars

258: \label{sec:spheroid}}

259: 

260: Halo lenses could produce events anywhere along the line of sight

261: from the Sun to the bulge and, assuming an isothermal halo model

262: with core radius $a=5\,\kpc$, the density

263: \begin{equation}

264: \rho_{\rm halo} = {v_c^2\over 4\pi G(R^2 + a^2)}

265: \label{eqn:rhoh}

266: \end{equation}

267: rises all the way to the Galactic center.  Here $R$ is Galactocentric

268: distance.  The optical depth per unit path length along a line of

269: sight toward the Galactic center therefore also rises almost all

270: the way in,

271: \begin{equation}

272: {d\tau_{\rm halo}\over d D_l} = {\rho_{\rm halo}\over M}\,

273: {\pi r_\e^2}  = {v_c^2\over c^2}\,{f\over R_0}

274: \,{x(1-x)\over (a/R_0)^2+(1-x)^2}.

275: \label{eqn:tauhalo}

276: \end{equation}

277: Here, $r_\e=(4GM D_{l} D_{s}/c^2 D_{os})^{1/2}$ is the Einstein

278: radius, $D_l$ and $D_s$ are the source and lens distances, $D_{ls}=D_s-D_l$,

279: $f$ is the fraction of the halo in the form of MACHOs, $R_0=8\,\kpc$ is the

280: Solar Galactocentric distance, $x\equiv D_l/R_0$, 

281: and $M$ (which cancels out) is the mass of the lens.

282: 

283: However, in the inner Galaxy, these halo lenses are completely 

284: submerged in the background of bulge lenses, and since they

285: have similar kinematics, there is no way to reliably distinguish them.

286: It is only out closer to the Sun, where the spheroidal population

287: (here usually called ``spheroid'' or ``stellar halo''), thins out that

288: one may hope to separate the two populations.

289: Even here, there is some possibility of contamination.  The

290: local spheroid density is only about 1\% of the dark halo, but

291: if $f\sim 20\%$ as \citet{alcock00} suggest, then MACHOs are only

292: 20 times more common than spheroid stars locally.  Moreover, 

293: as one approaches the Galactic center, the spheroid density

294: grows substantially more rapidly than does the dark halo.  To make

295: a quantitative comparison, I adopt 

296: \begin{equation}

297: \rho_{\rm spheroid} = 1\times 10^{-4}{M_\odot\over \rm pc^3}

298: \biggl({R\over R_0}\biggr)^{-3.2}.

299: \label{eqn:rhospheroid}

300: \end{equation}

301: After accounting for observed stars and extrapolating down to

302: brown dwarfs and up to the progenitors of remnants, \citet{gfb98} 

303: estimate $6.4\times 10^{-5}M_\odot\,\rm pc^{-3}$.  

304: However, both \citet{dahn95} and

305: \citet{gould03} find substantially more low-luminosity ($M_V>8$) stars

306: than did \citet{gfb98} in their more local sample

307: (see Fig.~2 from \citealt{gould04b}), so I have adjusted

308: their estimate upward.

309: The power-law slope is measured by several techniques (\citealt{gfb98} and

310: references therein).  

311: 

312: Figure \ref{fig:opdep} shows the optical depth per unit distance due 

313: to spheroid stars and to 

314: putative MACHOs under the assumption that $f=20\%$. 

315: It shows that even with a MACHO fraction of 20\%, the halo dominates

316: the spheroid from $R=R_0$ to  

317: $R=4\,\kpc$, which latter is about the limit to which the

318: local spheroid density profile can be reliably extrapolated.  However,

319: this domination is not overwhelming: at $R=4\,\kpc$ it is only a

320: factor of 5 and even at $R=7\,\kpc$ (where the halo optical depth

321: has fallen by a factor 5) the halo only dominates by a factor 10.

322: This means that 2 or 3 halo lenses would have to be identified to

323: constitute a reliable ``MACHO detection''.  Otherwise, there would

324: be a significant possibility that spheroid lenses were responsible.

325: 

326: Since one must restrict attention to $D_l<4\,\kpc$, the total available

327: halo optical depth is reduced by a factor $0.4$ relative to the

328: $1.3\times 10^{-7}$ calculated by \citet{griest91}.  If we further assume

329: $f=20\%$, the available halo optical depth is further reduced to $10^{-8}$,

330: about 0.5\% of the observed optical depth of 

331: $\tau\sim 2\times 10^{-6}$ \citep{afonso03b,popow05,sumi05}.  

332: Hence, assuming for the

333: moment that the event rates are in proportion to the optical depths,

334: roughly 200 measurements would be required to identify a single halo lens.

335: Thus, if the {\it SIM} mission were extended from 5 to 10 years (as is

336: currently envisioned) then one might expect to find about 2 halo lenses.

337: As noted above this is just at the margin of a viable detection.

338: 

339: \section{Practical Considerations

340: \label{sec:practical}}

341: 

342: \subsection{Event Timescales

343: \label{sec:timescale}}

344: 

345: The Einstein timescales of these halo events are given by,

346: \begin{equation}

347: t_\e = 15\,{\rm day}

348: \biggl[{M\over 0.4\,M_\odot}\,{x(1-x)\over 0.25}\biggr]^{1/2}

349: \biggl({v_\perp\over 300\,\kms}\biggr)^{-1},

350: \label{eqn:teeval}

351: \end{equation}

352: where $v_\perp$ is the transverse lens velocity relative to the

353: observer-source line of sight.

354: Thus, for the mass range advocated by \citet{alcock00}, the

355: typical event timescales will be fairly short.  This is important

356: because the event must be identified and alerted to the satellite

357: well before peak in order to measure $\pi_\e$ \citep{gs99}.

358: Hence, a fairly aggressive posture is required to keep the halo

359: events in the sample.

360: 

361: However, the fact that these halo events are somewhat shorter than 

362: typical bulge events means that they are also more frequent than

363: would be indicated by their optical depth alone.  That is, the

364: event rate $\Gamma\propto \tau/t_\e$, so the rate is inversely

365: proportional to the timescale.  Hence, the shorter timescales

366: enhances the viability of a given experiment relative to what

367: was discussed in \S~\ref{sec:spheroid}, provided that not too

368: many halo events are lost because they are too short.

369: 

370: \subsection{Signal-to-Noise Ratios

371: \label{sec:sn}}

372: 

373: The two microlensing parameters being measured are related to the

374: underlying physical parameters by,

375: \begin{equation}

376: \pi_\e = \sqrt{\pi_\rel \over \kappa M},\qquad

377: \theta_\e = \sqrt{\kappa M\pi_\rel},

378: \label{eqn:piethetae}

379: \end{equation}

380: where $\kappa = 4G/\au c^2 \sim 8.1\,{\rm mas}\,M^{-1}_\odot$.

381: Hence, for fixed $M$, both $\pi_\e$ and $\theta_\e$ 

382: are proportional to $\pi_\rel^{1/2}$.

383: Since the absolute errors in these two quantities are approximately

384: independent of their size, this means that the fractional errors

385: decline as $\pi_\rel^{-1/2}$.  The basic experiment is designed

386: for typical bulge-bulge lensing, in which the lenses are of order

387: $M\sim 0.5\,M_\odot$ and the relative parallaxes are 

388: $\pi_\rel\sim \au/7\,\kpc - \au/9\,\kpc =31\,\mu$as. By contrast,

389: since $D_l\leq 4\,\kpc$, the halo-lens relative parallaxes are

390: $\pi_\rel>125\,\mu$as.

391: Hence, if a halo event is successfully monitored, both $\pi_\e$ and

392: $\theta_\e$ will be measured substantially more accurately than for

393: typical events.

394: 

395: \acknowledgments

396: This work was supported by grant AST 02-01266 from the NSF and by

397: JPL contract 1226901.

398: 

399: \clearpage

400: 

401: \begin{thebibliography}{}

402: 

403: \bibitem[Abe et al.(2004)]{abe04} Abe, F. et al. 2004, Science, 305, 1264

404: 

405: \bibitem[Afonso et al.(2003a)]{afonso03a} Afonso, C. et al. 2003a, \aap, 400, 

406: 951

407: 

408: \bibitem[Afonso et al.(2003b)]{afonso03b} Afonso, C. et al. 2003b, \aap, 404, 

409: 145 %bulge

410: 

411: \bibitem[Alard et al.(1995)]{alard95} Alard, C., Mao, S., \& Guibert, J. 1995

412: \aap, 300, L17

413: 

414: \bibitem[Albrow et al.(1999)]{albrow99} Albrow, M.D. et al. 1999, 

415: \apj, 512, 672

416: 

417: \bibitem[Albrow et al.(2000)]{albrow00} Albrow, M.D. et al. 2000, 

418: \apj, 534, 894

419: 

420: \bibitem[Albrow et al.(2001a)]{albrow01a} Albrow, M.D. et al. 2001a, 

421: \apj, 550, L173

422: 

423: \bibitem[Albrow et al.(2001b)]{albrow01b} Albrow, M.D. et al. 2001b, 

424: \apj, 556, L113

425: 

426: \bibitem[Alcock et al.(1993)]{alcock93} Alcock, C. et al. 1993, Nature, 365, 621

427: 

428: \bibitem[Alcock et al.(1997)]{alcock97} Alcock, C. et al. 1997, \apj, 491, 436

429: 

430: \bibitem[Alcock et al.(2000)]{alcock00} Alcock, C. et al. 2000, \apj, 542, 281

431: 

432: \bibitem[Alcock et al.(2001)]{alcock01} Alcock, C. et al. 2001, Nature, 414, 617

433: 

434: \bibitem[An et al.(2003)]{an02} An, J.H. et al. 2002, \apj, 549, 759

435: 

436: \bibitem[Ansari et al.(1999)]{ansari99} Ansari, R., et al. 1999, \aap, 344, L49

437: 

438: \bibitem[Aubourg et al.(1993)]{aubourg93} Aubourg, E. et al. 1993, 

439: Nature, 365, 623

440: 

441: \bibitem[Auri\`ere et al.(2001)]{auriere01} 

442: Auri\`ere, M., et al. 2001, \aap, 553, L137

443: 

444: \bibitem[Baillon et al.(1993)]{baillon93} Baillon, P., Bouquet, A.,

445: Giraud-H\'eraud, Y. \& Kaplan, J. 1993, \aap, 277, 1

446: 

447: \bibitem[Boden et al.(1998)]{BSV98}

448: Boden A. F., Shao M., \& Van Buren D. 1998, \apj, 502, 538

449: 

450: \bibitem[Bond(2004)]{bond04} Bond, I.A. et al. 2004, \apj, 606, L155

451: 

452: \bibitem[Calchi Novati(2003)]{calchi03} Calchi Novati, S. 2003, \aap, 405, 851

453: 

454: \bibitem[Castro et al.(2001)]{castro01} Castro, S., Pogge, R.W., Rich, R.M.,

455: DePoy, D.L., \& Gould, A. 2001, \apj, 548, L197

456: 

457: \bibitem[Cassan et al.(2004)]{cassan04} Cassan, A. et al. 2004, \aap, 419 L1

458: 

459: \bibitem[Crotts(1992)]{crotts92} Crotts, A.P.S. 1992, \apj, 399, L43

460: 

461: \bibitem[Dahn et al.(1995)]{dahn95} Dahn, C.C., Liebert, J.W., Harris, H., 

462: \& Guetter, H.C.\ 1995, 

463: p.\ 239, An ESO Workshop on: the Bottom of the Main Sequence and Beyond,

464: C.G.\ Tinney ed.\ (Heidelberg: Springer)

465: 

466: \bibitem[Drake et al.(2004)]{dck04}

467: Drake, A.J., Cook, K.H., \& Keller, S.C. 2004, \apj, 607, L29

468: 

469: \bibitem[de Jong et al.(2004)]{dejong04} de Jong et al. 2004, \aap, 417, 461

470: 

471: \bibitem[Joshi et al.(2005)]{joshi05} Joshi, Y.C., Pandey, A.K.,

472: Narasimha, D., \& Sagar, R. 2005, \aap, in press (astroph/0412550)

473: 

474: \bibitem[Fields et al.(2003)]{fields03} Fields, D.L. et al. 2003, 

475: \apj, 596, 1305

476: 

477: \bibitem[Gaudi et al.(2002)]{gaudi02} Gaudi, B.S. et al. 2002, \apj, 566, 463

478: 

479: \bibitem[Gould(1994)]{gould94}

480: Gould, A. 1994, \apjl, 421, L75

481: 

482: \bibitem[Gould(1995a)]{gould95a}

483: Gould, A. 1995a, \apjl, 441, L21

484: 

485: \bibitem[Gould(1995b)]{gould95b} Gould, A. 1995b, \apj, 441, 77

486: 

487: \bibitem[Gould(2000)]{gould00} Gould, A. 2000, \apj, 542, 785

488: 

489: \bibitem[Gould(2001)]{gould01} Gould, A. 2001, \pasp, 113, 903

490: 

491: \bibitem[Gould(2003)]{gould03} Gould, A., 2003, \apj, 583, 765

492: 

493: \bibitem[Gould(2004a)]{gould04a} Gould, A., 2004a, \apj, 606, 319

494: 

495: \bibitem[Gould(2004b)]{gould04b} Gould, A., 2004b, \apj, 607, 653

496: 

497: \bibitem[Gould et al.(2004)]{gba04}

498: Gould, A., Bennett, D.P., \& Alves, D.R., \apj, 614, 404

499: 

500: \bibitem[Gould et al.(1997)]{gbf97} Gould, A., Bahcall, J.N., \&

501: Flynn, C. 1997, \apj, 482, 913

502: 

503: \bibitem[Gould et al.(1998)]{gfb98} Gould, A., 

504: Flynn, C. Bahcall, \& J.N. 1998, \apj, 503, 798

505: 

506: \bibitem[Gould \& Loeb(1992)]{gould92} Gould, A. \& Loeb, A. 1992, 

507: \apj, 396, 104

508: 

509: \bibitem[Gould \& Salim(1999)]{gs99} Gould, A. \& Salim, S. 1999

510: \apj, 524, 794

511: 

512: \bibitem[Griest et al.(1991)]{griest91} Griest, K. et al. 1991, \apj, 372, L79

513: 

514: \bibitem[H{\o}g et al.(1995)]{HNP95}

515: H{\o}g, E., Novikov, I. D., \& Polanarev, A. G. 1995, \aap, 294, 287

516: 

517: \bibitem[Kiraga \& Paczy\'nski(1994)]{kiraga94} 

518: Kiraga, M. \& Paczy\'nski, B. 1994, \apj, 430, L101

519: 

520: \bibitem[Kubas et al.(2005)]{kubas05} Kubas, D. et al. 2005, \aap, in press

521: (astroph/0502018)

522: 

523: \bibitem[Mao \& Paczy\'nski(1991)]{mao91} Mao, S. \& Paczy\'nski, B. 1991, \apj, 

524: 374, 37

525: 

526: \bibitem[Miyamoto \& Yoshii(1995)]{MY95}

527: Miyamoto,  M., \& Yoshii, Y. 1995, \aj, 110, 1427

528: 

529: \bibitem[Paczy\'nski(1986)]{pac86} Paczy\'nski, B. 1986, \apj, 304, 1

530: 

531: \bibitem[Paczy\'nski(1991)]{pac91} Paczy\'nski, B. 1991, \apj, 371, L63

532: 

533: \bibitem[Paczy\'nski(1998)]{Pa98}

534: Paczy\'nski, B. 1998, \apjl, 494, L23

535: 

536: \bibitem[Popowski(2005)]{popow05} Popowski, P. et al. 2005, \apj, submitted 

537: (astroph/0410319)

538: 

539: \bibitem[Refsdal(1966)]{Re66}

540: Refsdal, S. 1966, \mnras, 134, 315

541: 

542: \bibitem[Riffeser(2003)]{riffeser03} Riffeser, A., Fliri, J., Bender, R.,

543: Seitz, S., \& G\"ossl, C.A. 2003, \apj, 599, L17

544: 

545: \bibitem[Rhie et al.(2000)]{rhie00} Rhie, S.H. et al. 2000, \apj, 533, 378

546: 

547: \bibitem[Sumi et al.(2005)]{sumi05} Sumi, T. et al. 2005, \apj, submitted

548: (astroph/0502363)

549: 

550: \bibitem[Tisserand \& Milsztajn(2005)]{tisserand05} 

551: Tisserand, P. \& Milsztajn, A.

552: 2005, Proceedings of the 5th Rencontres du Vietnam ``New Views on the

553: Universe'', in press, astroph/0501584

554: 

555: \bibitem[Udalski et al.(1993)]{udalski93} Udalski, A., Szymanski, M., 

556: Kaluzny, J., Kubiak, M., Krzeminski, W., Mateo, M., Preston, G.W., 

557: Paczy\'nski, B. 2003, Acta Astron., 43, 289 

558: 

559: \bibitem[Udalski(2003)]{udalski03} Udalski, A. 2003, Acta Astron., 53, 291 

560: 

561: \bibitem[Uglesich et al.(2004)]{uglesich04} Uglesich, R.R. et al. 

562: 2004, \apj, 612, 877

563: 

564: \bibitem[Walker(1995)]{Wa95}

565: Walker, M. A. 1995, \apj, 453, 37

566: 

567: 

568: \end{thebibliography}

569: \clearpage

570: 

571: \begin{figure}

572: \plotone{f1.ps}

573: \caption{\label{fig:opdep}

574: Optical Depth per unit path length $d\tau/dD_l$ as a function of

575: distance from the Galactic center for a source near the Galactic

576: center.  The halo (assuming a $f=20\%$ MACHO fraction) and the spheroid

577: are shown by {\it solid} and {\it dashed} curves, respectively.  

578: For $f=20\%$, spheroid stars are a 20\% background at $R=4\,\kpc$ and

579: a 10\% background at $R=7\,\kpc$, which implies that 2 or 3 halo lenses

580: must be identified at $R>4\,\kpc$ for a reliable halo ``detection''.

581: Inside $R<4\,\kpc$ the spheroid continues to grow (and also transforms

582: into the bulge), making the identification of halo lenses less secure.

583: Hence, the experiment should be restricted to $R>4\,\kpc$, where the

584: total optical depth is $\tau = 5\times 10^{-8}f$.

585: }\end{figure}

586: 

587: \end{document}

588: