1: \documentclass[twocolumn]{emulateapj}
2:
3: %\documentclass[manuscript]{aastex}
4:
5: \bibliographystyle{apj}
6:
7: \usepackage{epsfig}
8: \usepackage{amsmath}
9: \usepackage{natbib}
10:
11: %internal short cuts
12:
13: \begin{document}
14:
15: \title{Quasar Microlensing: when compact masses mimic smooth matter}
16: \author{Geraint F. Lewis \& Rodrigo Gil-Merino}
17: \email{gfl@physics.usyd.edu.au}
18: \email{rodrigo@physics.usyd.edu.au}
19: \affil{Institute of Astronomy, School of Physics, A28,
20: University of Sydney, NSW 2006, Australia}
21:
22:
23: \begin{abstract}
24: The magnification induced by gravitational microlensing is sensitive
25: to the size of a source relative to the Einstein radius, the natural
26: microlensing scale length. This paper investigates the effect of
27: source size in the case where the microlensing masses are distributed
28: with a bimodal mass function, with solar mass stars representing the
29: normal stellar masses, and smaller masses (down to $8.5\times
30: 10^{-5}$M$_\odot$) representing a dark matter component. It is found
31: that there exists a critical regime where the dark matter is initially
32: seen as individual compact masses, but with an increasing source size
33: the compact dark matter acts as a smooth mass component. This study
34: reveals that interpretation of microlensing light curves, especially
35: claims of small mass dark matter lenses embedded in an overall stellar
36: population, must consider the important influence of the size of the
37: source.
38: \end{abstract}
39:
40: \keywords{gravitational lensing - microlensing - dark matter halos}
41:
42:
43: \maketitle
44:
45: %Section heading
46: \section{Introduction}
47: As the light from a distant source shines through a foreground galaxy,
48: the gravitational lensing effect of the individual stars within the
49: galaxy can induce rapid fluctuations in the apparent brightness of the
50: source \citep{1986ApJ...301..503P}. Since the detection of the first
51: lensing induced variability, seen in the quadruply imaged quasar
52: Q2237+0305 \citep{1989AJ.....98.1989I,1991AJ....102...34C}, there has
53: been intense theoretical study of this {\it gravitational
54: microlensing}, revealing that the degree of brightness fluctuations is
55: strongly dependent upon the size of the source being lensed
56: \citep{1986A&A...166...36K, 1990ApJ...358L..33W, 1991AJ....102..864W}
57: and the mass function of the lensing objects
58: \citep{1996MNRAS.283..225L,2001MNRAS.320...21W}, potentially providing
59: a useful probe of both. Further studies have scrutinized the influence
60: of source size and shape \citep{1991A&A...250...62R,
61: 1993A&A...278L...5R, 1997A&A...325..877R, 2003ApJ...594..449A,
62: 2005ApJ...628..594M}, and more recent programs with dedicated
63: monitoring programs \citep{2000ApJ...529...88W, 2002ApJ...572..729A,
64: 2003MNRAS.346..415U} have finally resulted in high quality
65: microlensing light curves that can be compared directly with
66: theoretical models~\citep{2000MNRAS.318.1120W,2004ApJ...605...58K}.
67:
68: \begin{figure*}
69: \centerline{ \psfig{figure=fig1.ps,angle=270,width=6.8in}}
70: \caption[]{Examples of the magnifications maps utilized in this paper.
71: Each map is 12 Einstein radii on a side with a normalized surface mass
72: density of $\sigma=0.2$ and global shear of $\gamma=0.5$. The
73: lower-right panel presents the case where half of the surface mass
74: density is in the form of solar mass objects and the rest in the form
75: of smoothly distributed matter. In the additional panels, this smooth
76: matter component has been replaced with compact mass objects with
77: solar and then sub-solar masses; the individual masses of the
78: sub-solar population are noted in the lower left-hand corner of each
79: panel. Note that in each panel, the solar mass stars retain their
80: positions in the lensing plane.
81: \label{fig1}}
82: \end{figure*}
83:
84: More recently, gravitational microlensing has been proposed as an
85: explanation of the anomalous flux ratios seen in multiply imaged
86: quasars, with the addition of smoothly distributed dark matter
87: broadening the width of the magnification probability distribution
88: \citep{2002ApJ...580..685S}. This result appears somewhat surprising
89: given a conjecture that has existed for a number of years in
90: microlensing studies, namely that the shape of a magnification
91: probability distribution should be independent of the mass function of
92: the microlensing masses
93: \citep{1992ApJ...386...19W,1996MNRAS.283..225L}; the smooth matter
94: limit is approached as the mass of the individual objects is
95: decreased, but the surface mass density remains the same. Clearly, the
96: significant differences seen in the magnification probability
97: distributions with smooth and compact matter violates this conjecture,
98: a position recently confirmed by \citet{2004ApJ...613...77S}.
99:
100: The fall of the microlensing conjecture therefore implies that there
101: is a fundamental difference in the action of compact masses as opposed
102: to smoothly distributed mass. But can compact masses appear as smooth
103: matter to a source? It is well known that increasing the size of a
104: microlensing source smooths out the variability of the light curve,
105: but does this imply that to a large source, a population of small
106: masses appears as smooth matter? While this situation was not
107: examined in the work of \citet{2004ApJ...613...77S}, it existence
108: means that the breakdown of the microlensing conjecture is further
109: blurred by the size of the source under consideration. Studying this
110: regime, therefore, is the goal of this paper. In
111: Section~\ref{approach}, the numerical approach is outlined, while
112: Section~\ref{analysis} discusses the influence of a range of source
113: sizes and compares the resulting properties of the magnification maps.
114: In Sections~\ref{implications} and \ref{conclusions}, the implications
115: of this study and the conclusions are respectively presented.
116:
117: \section{Approach}\label{approach}
118: In studying the influence of gravitational microlensing, this study
119: employs the backwards ray tracing algorithm of
120: \citet{1986A&A...166...36K} and \citet{1992ApJ...386...19W}, whereby a
121: large number of rays are fired through a field of microlensing masses
122: and collected to form a magnification map over the source plane. The
123: important length scale for gravitational microlensing is the Einstein
124: radius in the source plane, given by
125: \[
126: \eta_o = \sqrt{ \frac{4 G M}{c^2} \frac{D_{os} D_{ol}}{D_{ls}} }
127: \]
128: where $D_{ij}$ are angular diameter distances between the observer
129: ($o$), lens ($l$) and source ($s$) respectively and typically this
130: scale is $\eta_o\sim 0.04$pc for cosmological circumstances
131: \citep[e.g.][]{1998ApJ...501..478L}. As well as the mass function of
132: the microlensing bodies, other parameters need to be specified,
133: including the normalized surface mass density ($\sigma$) which may be
134: composed of compact and smooth masses, and a shear term ($\gamma$)
135: which describes the global asymmetric gravitational lensing influence
136: of the overall galactic/intracluster mass distribution. Given the
137: potential parameter space, a comprehensive study of potential
138: microlensing configurations is beyond the scope of this paper. As a
139: first step, a representative model was chosen with a total surface
140: mass density of $\sigma=0.2$ and global shear of of $\gamma=0.5$. Each
141: simulation covers an area of $12^2$ Einstein radii for a solar mass
142: star and consists of $2048^2$ pixels.
143:
144: \begin{figure*}
145: \centerline{ \psfig{figure=fig2.ps,angle=270,width=6.7in}}
146: \caption[]{An example of the procedure undertaken in this paper. The
147: left-most column contains a magnification map formed when half the
148: surface mass density is in solar mass objects, with the rest smoothly
149: distributed. The centre panel presents the same circumstance, except
150: the smooth matter component has been replaced with compact objects
151: each with a mass of $8.5\times10^{-5}{\rm M_\odot}$. Following these
152: are light curves for the smooth matter (red) and compact mass (green)
153: cases, taken across the centre of the maps (indicated by the lines on
154: the magnification maps). Note that the light curves are in magnitude
155: and the range on the y-axis from +3 to -5 magnitudes about the
156: theoretically expected mean. This scale has been omitted for clarity.
157: The right hand column presents the magnification probability
158: distributions. The top maps represent a single pixel source, while
159: moving downwards the source size is increased. As it does, the small
160: scale caustics structure induced by the smaller masses is washed out
161: and the maps, light curves and magnification distributions become the
162: same. While the top panel is unconvolved, the scale-radius of the
163: Gaussian kernel in the subsequent panels are $5.9\times10^{-3}$,
164: $1.8\times10^{-2}$, $5.7\times10^{-2}$ and $1.8\times10^{-1}$ Einstein
165: radii respectively.
166: \label{fig2}}
167: \end{figure*}
168:
169: A fiducial model was chosen in which half the surface mass density is
170: in the form of solar mass stars, while the remaining mass is in the
171: form of smoothly distributed matter. In addition to this, 43
172: additional simulations were undertaken. Again, half the mass was in
173: the form of solar mass stars, while the remainder is also in the form
174: of compact objects, but with progressively smaller masses, with each
175: subsequent dark matter mass being 80\% of the proceeding value. The
176: limit of this procedure was compact dark matter masses of
177: $8.5\times10^{-5}$M$_\odot$, resulting in a total number of
178: microlensing masses in the simulation exceeding $4.2\times10^6$; it
179: was found that this represented the memory limit of available
180: computational hardware, while greatly increasing the calculation time
181: of the magnification maps, and hence no smaller masses were
182: considered. It is important to note that the locations of the solar
183: mass stars are identical in all simulations.
184:
185: Figure~\ref{fig1} presents examples of the results of this approach,
186: with the upper-left panel representing the magnification maps for the
187: case where half the matter is in the form of smooth matter. For the
188: additional panels in this figure, this smooth matter component has
189: been replaced with compact objects, beginning with $1$M$_\odot$ and
190: continuing down to $1.1\times10^{-4}$M$_\odot$. There is a significant
191: difference between the caustic structures apparent in the
192: magnification map for the smooth matter case and that for all the
193: lensing bodies being of 1M$_\odot$. However, as the mass of the
194: compact objects replacing the smooth matter component is reduced,
195: there is a very apparent evolution in the form of the caustic
196: structure; maps appear to possess caustics structure on two distinct
197: scales, corresponding to the two microlensing masses under
198: consideration. For the magnification map with the smallest masses,
199: the large scale caustic structure becomes very similar to that of the
200: case with smooth matter. However, it is also apparent that there is
201: significant small scale structure distributed throughout the map. As
202: will be seen in the next section, while the magnification maps in
203: smooth matter and small mass cases appear similar, the statistical
204: properties, especially the magnification probability distributions,
205: differ markedly.
206:
207: \section{Analysis}\label{analysis}
208: To determine the microlensing properties of a particular extended
209: source, the magnification maps must be convolved with the appropriate
210: surface brightness distribution. For the purposes of this study, the
211: source is assumed to possess a Gaussian profile with a characteristic
212: radius. As the source size is increased, the resulting magnification
213: map and magnification probability distribution are compared with those
214: of the solar mass plus smooth matter component.
215:
216: Figure~\ref{fig2} illustrates this procedure for the case where the
217: compact dark matter component have individual masses of
218: $8.5\times10^{-5}$M$_\odot$, the smallest masses that are
219: computationally viable at this time. The upper panels present the
220: unconvolved magnification maps for the smooth dark matter (left-hand)
221: and compact dark matter (right-hand)\footnote{It should be remembered
222: that these maps, by virtue of the pixelated nature, are implicitly
223: convolved with a source on the scale of a single pixel.}. The maps
224: possess similar large scale caustic structure, but clearly the
225: presence of small masses instead of smooth dark matter has introduced
226: significant substructure into the caustic network. Over-plotted on
227: these magnification maps are the resultant light curves of a source
228: passing horizontally across the middle of the two maps and these
229: present quite dramatic behaviour. This is reflected in the
230: magnification probability distributions which are markedly different.
231: Progressing down the page these magnification maps are convolved with
232: larger and larger sources. As the source size increases, the
233: substructure in the magnification map with compact dark matter begins
234: to wash out. This is further reflected in the subsequent light curves,
235: with the degree of variability decreasing. Interestingly, as the
236: source size is increased, the light curves for the compact dark matter
237: and correspondingly convolved smooth matter cases become very similar,
238: again a fact seen in the magnification probability distributions which
239: too become identical. With even larger smoothing, the magnification
240: maps remain identical and hence the compact matter will effectively
241: appear as a smooth dark matter component.
242:
243: \begin{figure}
244: \centerline{ \psfig{figure=fig3.ps,angle=270,width=3.5in}}
245: \caption[]{The solid line in this figure denotes the Einstein radius
246: for the smaller mass microlensing simulations, whereas the points
247: present the radius of the Gaussian kernel at which the difference in
248: the binned, cumulative magnification patterns differed by less than
249: $f=0.05$ (circles) and $f=0.10$ (squares); note the relation shows an
250: linear trend in this logarithmic space. The discrepant points are due
251: to ``noise'' in the binned distribution preventing the fit from
252: reaching the fitting criteria. The arrows in this figure denote the
253: magnification maps which did not achieve this criterion and the
254: magnification distributions are too different to allow an equivalence
255: between the two maps even with smoothing on the scales considered.
256: \label{fig3}}
257: \end{figure}
258:
259: To examine this further, the entire sample was subjected to the same
260: convolution procedure, starting with small Gaussian sources and
261: increasing the radius of the kernel until the magnification
262: probability distributions became similar. In defining this threshold,
263: the binned cumulative magnification distributions were compared, with
264: the difference critical convolution radius defined as being that at
265: which
266: \begin{equation}
267: \mbox{Max } \left( 2 \frac{\left| D^1_i - D^2_i \right|} {\left( D^1_i
268: + D^2_i \right)} \right) < f
269: \end{equation}
270: where $D^1$ and $D^2$ are the cumulative probability distributions;
271: such cumulative distributions ease differences in the differential
272: magnification probability distributions due to noise. The factor $f$
273: represents the fractional difference between the two magnification
274: distribution and for the purposes of this study we choose $f=0.05$ and
275: $f=0.10$ as representative values. Of course, the choice of the value
276: of $f$ is rather arbitrary and other choices could be made. Hence,
277: the functional form employed here should be viewed as illustrative
278: rather than definitive. The results of this procedure are presented
279: in Figure~\ref{fig3}; here the solid line represents the Einstein
280: radius of the smaller mass component, with the points denoting the
281: critical radius as defined by the above criterion. This figure shows
282: that there is a clear trend in this statistic with the critical
283: Gaussian radius increasing in step with the Einstein radius. For
284: $f=0.05$ and mass scales from $\sim10^{-4}$M$_\odot$ up to
285: $\sim$0.1M$_\odot$, the ratio of these two quantities is (remarkably)
286: constant at $\sim4$ (with a couple of discrepant points due to noise
287: spikes in the cumulative distribution delaying convergence), although
288: this ratio drops to $\sim1.1$ for $f=0.10$. Note, the arrows denote
289: masses at which smoothing on the scale of less than 5 Einstein radii
290: failed to result in convergence between the two magnification
291: distributions. A reexamination of Figure~\ref{fig1} illustrates that
292: at this high mass end, the caustic structure introduced by the compact
293: masses is similar in scale to the overall caustic network and it is
294: understandable that this structure has not been smoothed out on the
295: scales under consideration.
296:
297: Returning to Figure~\ref{fig2} it is clear that the third and fourth
298: row of panels straddle this critical region, with magnification
299: distributions for the smaller Gaussian source appearing quite
300: different for the smooth matter and compact cases. Considering the
301: lower panel, and hence larger source, the distributions in both
302: scenarios have become very similar. Notice in this case, even when
303: smoothed with the larger source, the resulting magnification
304: distributions possess significant structure which is reflected in
305: quite dramatic events in the microlensing light curve. Hence the
306: origin of the similarity in the distributions is not due to all of the
307: magnification structure being smoothed out of the microlensing maps.
308: For cases where the difference in mass between the compact dark matter
309: and the solar mass stars is less, then a lot of the structure in the
310: magnification maps is smoothed out, with the resulting light curves
311: possessing very little structure.
312:
313: \subsection{Further simulations}\label{further}
314: So far this paper has considered a single combination of microlensing
315: optical depth and shear, but how general are the results uncovered
316: thus far? In addressing this question, three more sets of
317: microlensing parameters were employed, namely $(\sigma,\gamma) =
318: (0.2,0.0)$, $(\sigma,\gamma) = (0.2,0.2)$ and $(\sigma,\gamma) =
319: (0.6,0.6)$. As with the previous simulations, the source region under
320: consideration covered 12$^2$ Einstein radii for a solar mass star and
321: consisted of $2048^2$ pixels. Again, in each case, half of the optical
322: depth comprised of solar mass stars while the other half was in the
323: form of smooth matter or compact objects, of mass either
324: $3.5\times10^{-2}{\rm M_\odot}$ or $1.2\times10^{-3}{\rm M_\odot}$.
325:
326: Rather than simply repeating the analysis presented in
327: Section~\ref{analysis}, a slightly different approach was undertaken.
328: As revealed in Fig.~\ref{fig1}, as the masses of the compact objects
329: decreases, the structure in the magnification map tends to that of the
330: smooth matter case, with the smaller masses providing strong, but
331: localized, perturbations, and it is these perturbations which are
332: washed out by convolving with a large enough source; note, however, if
333: the masses of the compact objects are large enough, they produce gross
334: changes into the magnification map such that convolving cannot make
335: the magnification identical to the smooth matter case. Hence, given
336: the small masses considered in this second set of simulations, we
337: would expect that the structures in the compact matter and smooth
338: matter magnification maps should become similar, if convolved with a
339: large enough source.
340:
341: \begin{figure}
342: \centerline{ \psfig{figure=fig4.ps,angle=270,width=3.2in}}
343: \caption[]{The variation in the subtracted convolved magnification
344: maps, as discussed in Section~\ref{further}. The differing line styles
345: correspond to the three parameter sets employed (as denoted by the
346: key) where as the lighter lines represent the compact masses of
347: $3.5\times10^{-2}{\rm M_\odot}$, whereas the heavier lines are for
348: masses of $1.2\times10^{-3}{\rm M_\odot}$. Note that the cases with
349: $\sigma=0.2$ possess similar structure and overlap in this figure.
350: \label{fig4}}
351: \end{figure}
352:
353: The analysis procedure again consisted of convolving the smooth matter
354: and compact matter magnification maps with a Gaussian profile. A
355: residual map was constructed by converting the smoothed magnification
356: maps to magnitudes and then subtracting them. These residuals possess
357: a Gaussian-like profile, centred upon zero, and so a residual
358: dispersion, $\sigma_{Mag}$ was calculated. The results of this
359: procedure are presented in Figure~\ref{fig4}, in which this residual
360: dispersion is plotted against the smoothing radius, for the
361: simulations under consideration. The trend is as expected, with the
362: residual map showing large variations, with $\sigma_{Mag}\sim1$ at
363: small smoothing radii for $(\sigma,\gamma)=(0.6,0.6)$, but as the
364: smoothing radius is increased, $\sigma_{Mag}$ falls, and hence the
365: convolved magnification maps present similar structure. Also,
366: $\sigma_{Mag}$ falls faster for the smaller compact masses, as their
367: magnification structure gets washed out at smaller radii.
368:
369: Interestingly, the curves for the larger mass compact objects in
370: Fig.~\ref{fig4} clearly show a plateau-like structure, remaining
371: constant over a range of smoothing radii, and then beginning to
372: decrease. Furthermore, the similarity of the curves for both masses
373: considered suggests that the smaller mass component possesses a
374: similar plateau structure, existing in the regime below the minimum
375: smoothing radius we considered. Hence, this again reveals that there
376: appears to be a critical radius below which the influence of the
377: compact matter changes character, effectively representing a smooth
378: mass component. A systematic study of this critical radius for a range
379: of microlensing parameters, and the influence of physical size and
380: structure of the emission regions in quasars, will be the subject of a
381: forthcoming contribution.
382:
383: \section{Implications}\label{implications}
384: One immediate implication of the results presented in this paper is
385: that if the dark matter is in the form of relatively small compact
386: masses, then the resulting statistical properties of a microlensing
387: light curve will be dependent upon the size of the source under
388: consideration. It is important to note that this is somewhat different
389: to the usual size dependence of microlensing statistics where there is
390: a straightforward smoothing of the light curve as the source size is
391: increased, with a resultant narrowing of the magnification probability
392: distribution. Rather, a small source and large source, where relative
393: sizes of small and large can be determined from the relationship
394: presented in Figure~\ref{fig3}, will be subject to significantly
395: different magnification patterns and corresponding statistics.
396:
397: The effect influences the question of how well we can reveal the
398: nature of the dark matter component, if present. In realistic
399: gravitational microlensing scenarios, the distribution of matter in
400: the lensing galaxy will not be a clean two-mass component population,
401: but rather a certain mass function (or even a combination of mass
402: functions). Depending upon the source size for a given system, this
403: mass function will suffer a cutoff point below which observations will
404: not be able to distinguish between smooth and granular matter. This
405: does not mean that small lens masses and large sources combination
406: with a single mass component cannot produce microlensing imprints.
407: Similarly means that for large enough source sizes, sub-Jupiter masses
408: cannot be detected unambiguously and that claims involving planetary
409: and sub-planetary populations that produce light curves fluctuations
410: should be thoroughly justified \citep[i.e.][]{2003ApJ...594...97C}.
411: For example, from Fig.~\ref{fig3}, for $f=0.05$, we see that for
412: microlenses masses of $\sim$0.04~M$_\odot$, there are still source
413: sizes ($\sim$0.7~ER) that make such a low-mass population to appear as
414: a smooth distributed matter component. Considering the corresponding
415: physical length scale in microlensed quasars corresponds to
416: $\sim0.04$pc \citep{1998ApJ...501..478L}, it is too large to represent
417: the optical/UV emitting region of quasars. However, for more
418: realistic quasar accretion disk sizes of $<$0.1~ER, ``secondary''
419: populations with masses $<$10$^{-3}$~M$_\odot$ will be completely
420: smoothed out. Since the trend in Fig.~\ref{fig3} is quite constant
421: for both considered values of $f$, this approach would also be useful
422: in putting limits to source sizes, though a better understanding of
423: the role of the mass populations is needed.
424:
425: To some extent, it seems that our results are in contradiction with
426: those obtained by Schechter et al. (2004). In that work, the authors
427: concluded that ``the magnification probability distribution for two
428: disparate components is not \emph{exactly} that of a single component
429: and a smooth component''. The apparent contradiction vanishes when it
430: is realised that Schechter et al. (2004) did not account for the
431: source size effect (or, in other words, they assumed the same pixel
432: size source in their study). Indeed, this is an important issue,
433: because observationally the source effect will be always present.
434: This means that the additional structure in the magnification maps
435: induced by the low-mass component, seen by Schechter et al. (2004), is
436: again blurred out by the source. So, the dependence of the
437: magnification probability on the higher-order moments of the mass
438: distributions is no longer true when introducing the source effect (at
439: least for certain regions of the mass distribution function).
440:
441: \section{Conclusions}\label{conclusions}
442: This paper has presented a study of the influence of source size on
443: the properties of gravitational microlensing with a bimodal
444: distribution of lensing masses. Given the computational limitations,
445: the study focused upon a set of macrolensing parameters, but
446: considered the influence of compact dark matter on a range of mass
447: scales. It was found that small sources resolved the caustic network
448: produced by the smaller masses, but as the source size was increased
449: this fine scale network becomes washed out. When compared to the
450: scenario where the smaller masses are replaced with an equivalent
451: quantity of smooth matter, it is seen that the light curves of the
452: smallest sources in both circumstances are quite different. As the
453: source size is increased, the light curves, and corresponding
454: magnification probability distributions, in each case become
455: identical, indicating that there is a critical source scale where
456: compact dark matter behaves, in a gravitational lensing sense, like
457: smoothly distributed matter. This critical source scale is dependent
458: upon the ratio of the stellar to dark matter masses, with a relatively
459: linear trend in log-log space (i.e. source sizes smaller than the
460: circles in Figure~\ref{fig3} are small enough to resolve the compact
461: dark matter, and we can expect to see their signature in the
462: microlensing light curve, whereas for source size greater than the
463: circles, this signature will be smoothed out).
464:
465: The results found in this paper point to another problem concerning
466: magnification probabilities. To what extent magnification
467: distributions are showing valuable information in \emph{real}
468: gravitational lensed systems? In other words, can we characterize
469: systems according to their magnification distributions? To answer
470: these questions we would need to explore a large range of combinations
471: of $\sigma$ and $\gamma$, together with different mass distribution
472: functions; while computationally expensive, we are currently planning
473: the first stages of such an exploration. However, it seems clear that
474: when the source effect is taken into account, the information
475: available from the magnification distributions can be rather poor. Of
476: course, the magnification distributions are only the zeroth level
477: statistic that can be used to examine the influence of gravitational
478: microlensing, but clearly the results presented in this contribution
479: similarly influence higher order temporal statistics that are applied
480: to microlensing light curves.
481:
482: In closing, this contribution returns to the question of the
483: microlensing conjecture that was discussed earlier. The fall of the
484: conjecture implies that the mass function of the compact objects was
485: potentially amenable to microlensing observations as it is now
486: apparent that this will imprint a signature into any microlensing
487: light curves. However, the issue has been clouded further with the
488: results presented in this paper illustrating that the size of a source
489: being of vital importance when addressing the question of how well the
490: properties of an underlying mass function can be determined. While
491: introducing a mass function expands the potential parameter space
492: enormously, it is important to understand the influence of the source
493: on the values of microlensing statistics and hence future
494: contributions will investigate its role in more general microlensing
495: scenarios.
496:
497: % References
498: \section*{Acknowledgments}
499: GFL thanks the Selby Trustees for the 2004 Selby Research Award which
500: part funded this project. The anonymous referee is thanked for
501: comments which improved the paper.
502:
503: \begin{thebibliography}{DUM}
504:
505: \bibitem[Agol(2003)]{2003ApJ...594..449A} Agol, E.\ 2003, \apj, 594, 449
506:
507: \bibitem[Alcalde et al. (2002)]{2002ApJ...572..729A} Alcalde et al.\ 2002,
508: \apj, 572, 729
509:
510: \bibitem[Colley \& Schild(2003)]{2003ApJ...594...97C} Colley, W.~N., \&
511: Schild, R.~E.\ 2003, \apj, 594, 97
512:
513: \bibitem[Corrigan et al.(1991)]{1991AJ....102...34C} Corrigan, R.~T., et
514: al.\ 1991, \aj, 102, 34
515:
516: %\bibitem[Granot et al.(2003)]{2003ApJ...583..575G} Granot, J., Schechter,
517: %P.~L., \& Wambsganss, J.\ 2003, \apj, 583, 575
518:
519: \bibitem[Irwin et al.(1989)]{1989AJ.....98.1989I} Irwin, M.~J., Webster,
520: R.~L., Hewett, P.~C., Corrigan, R.~T., \& Jedrzejewski, R.~I.\ 1989, \aj,
521: 98, 1989
522:
523: \bibitem[Kayser et al.(1986)]{1986A&A...166...36K} Kayser, R.,
524: Refsdal, S., \& Stabell, R.\ 1986, \aap, 166, 36
525:
526: \bibitem[Kochanek(2004)]{2004ApJ...605...58K} Kochanek, C.~S.\ 2004, \apj,
527: 605, 58
528:
529: \bibitem[Lewis \& Irwin(1996)]{1996MNRAS.283..225L} Lewis, G.~F., \& Irwin,
530: M.~J.\ 1996, \mnras, 283, 225
531:
532: \bibitem[Lewis \& Ibata(1998)]{1998ApJ...501..478L} Lewis, G.~F., \& Ibata,
533: R.~A.\ 1998, \apj, 501, 478
534:
535: \bibitem[Mortonson et al.(2005)]{2005ApJ...628..594M} Mortonson, M.~J.,
536: Schechter, P.~L., \& Wambsganss, J.\ 2005, \apj, 628, 594
537:
538: \bibitem[Paczynski(1986)]{1986ApJ...301..503P} Paczynski, B.\ 1986, \apj,
539: 301, 503
540:
541: \bibitem[Refsdal \& Stabell(1991)]{1991A&A...250...62R} Refsdal, S., \&
542: Stabell, R.\ 1991, \aap, 250, 62
543:
544: \bibitem[Refsdal \& Stabell(1993)]{1993A&A...278L...5R} Refsdal, S., \&
545: Stabell, R.\ 1993, \aap, 278, L5
546:
547: \bibitem[Refsdal \& Stabell(1997)]{1997A&A...325..877R} Refsdal, S., \&
548: Stabell, R.\ 1997, \aap, 325, 877
549:
550: \bibitem[Schechter et al.(2004)]{2004ApJ...613...77S} Schechter, P.~L.,
551: Wambsganss, J., \& Lewis, G.~F.\ 2004, \apj, 613, 77
552:
553: \bibitem[Schechter \& Wambsganss(2002)]{2002ApJ...580..685S} Schechter,
554: P.~L., \& Wambsganss, J.\ 2002, \apj, 580, 685
555:
556: \bibitem[Ull{\' a}n et al.(2003)]{2003MNRAS.346..415U} Ull{\' a}n, A., et
557: al.\ 2003, \mnras, 346, 415
558:
559: \bibitem[Wambsganss(1992)]{1992ApJ...386...19W} Wambsganss, J.\ 1992, \apj,
560: 386, 19
561:
562: \bibitem[Wambsganss \& Paczynski(1991)]{1991AJ....102..864W} Wambsganss,
563: J., \& Paczynski, B.\ 1991, \aj, 102, 864
564:
565: \bibitem[Wambsganss et al.(1990)]{1990ApJ...358L..33W} Wambsganss, J.,
566: Schneider, P., \& Paczynski, B.\ 1990, \apjl, 358, L33
567:
568: \bibitem[Wo{\' z}niak et al.(2000)]{2000ApJ...529...88W} Wo{\' z}niak,
569: P.~R., Alard, C., Udalski, A., Szyma{\' n}ski, M., Kubiak, M., Pietrzy{\'
570: n}ski, G., \& Zebru{\' n}, K.\ 2000, \apj, 529, 88
571:
572: \bibitem[Wyithe et al.(2000)]{2000MNRAS.318.1120W} Wyithe, J.~S.~B.,
573: Webster, R.~L., \& Turner, E.~L.\ 2000, \mnras, 318, 1120
574:
575: \bibitem[Wyithe \& Turner(2001)]{2001MNRAS.320...21W} Wyithe, J.~S.~B., \&
576: Turner, E.~L.\ 2001, \mnras, 320, 21
577:
578: \bibitem[Wyithe et al.(2002)]{2002MNRAS.331.1041W} Wyithe, J.~S.~B., Agol,
579: E., \& Fluke, C.~J.\ 2002, \mnras, 331, 1041
580:
581:
582: \end{thebibliography}
583: \end{document}
584:
585: