1: \documentclass{emulateapj}
2: %\documentclass[12pt,preprint]{aastex}
3:
4: \newcommand{\LT}{L-~$\&$~T-}
5:
6: \newcommand{\myemail}{Russell.Ryanjr@asu.edu}
7: \newcommand{\super}[1]{$^{#1}$}
8: \newcommand{\sub}[1]{$_{#1}$}
9: \newcommand{\tab}[1]{Table~\ref{#1}}
10: \newcommand{\fig}[1]{Figure~\ref{#1}}
11:
12: \newcommand{\lon}{l^{{\rm II}}}
13: \newcommand{\lat}{b^{{\rm II}}}
14: \newcommand{\lb}{\ensuremath{(\lon,\lat)}}
15: \newcommand{\nfields}{\ensuremath{15}}
16: \newcommand{\nltds}{\ensuremath{28}}
17: \newcommand{\minz}{\ensuremath{350}}
18: \newcommand{\errz}{\ensuremath{50}}
19: \newcommand{\finz}{\ensuremath{\minz\!\pm\!\errz}}
20: \newcommand{\rscl}{\ensuremath{2100}}
21: \newcommand{\blim}{\ensuremath{15}}
22: \newcommand{\absb}{\ensuremath{|\lat|\!\lesssim\!\blim\degr}}
23: \newcommand{\iz}{\ensuremath{(i'\!-\!z')}}
24: \newcommand{\gi}{\ensuremath{(g'\!-\!i')}}
25: \newcommand{\zj}{\ensuremath{(z'-J)}}
26:
27: \shorttitle{Galactic \LT Dwarfs}
28: \shortauthors{Ryan et al.}
29:
30: \begin{document}
31: \title{Constraining the Distribution of \LT Dwarfs in the
32: Galaxy\footnote{Based on observations made with the NASA/ESA Hubble
33: Space Telescope, obtained from the Data Archive at the Space Telescope
34: Science Institute, which is operated by the Association of
35: Universities for Research in Astronomy, Inc., under NASA contract NAS
36: 5-26555.}}
37:
38: \author{R. E. Ryan Jr., N. P. Hathi, S. H. Cohen \& R. A. Windhorst\altaffilmark{2}}
39: \affil{Department of Physics and Astronomy, Arizona State University, Tempe, AZ 85281}
40: \email{\myemail}
41:
42: \begin{abstract}
43:
44: We estimate the thin disk scale height of the Galactic population of
45: \LT dwarfs based on star counts from $\nfields$ deep parallel fields
46: from the Hubble Space Telescope. From these observations, we have
47: identified $\nltds$ candidate \LT dwarfs based on their \iz\ color and
48: morphology. By comparing these star counts to a simple Galactic
49: model, we estimate the scale height to be $\finz$~pc that is
50: consistent with the increase in vertical scale with decreasing stellar
51: mass and is independent of reddening, color-magnitude limits, and
52: other Galactic parameters. With this refined measure, we predict that
53: less than $10^{9}$~M$_{\odot}$ of the Milky Way can be in the form \LT
54: dwarfs, and confirm that high-latitude, $z\!\simeq\!6$ galaxy surveys
55: which use the $i'$-band dropout technique are 97-100\% free of \LT
56: dwarf interlopers.
57:
58: \end{abstract}
59:
60: \keywords{stars: low-mass, brown dwarfs --- Galaxy: structure --- Galaxy: stellar content}
61:
62:
63: \section{Introduction} \label{introduction}
64:
65: The method of counting stars to infer the shape and size of the Galaxy
66: is as old as astronomy itself. Earliest efforts using this technique
67: were famously flawed as they often relied on insufficient data or
68: wholly incorrect assumptions \citep[eg.][]{herschel,kapteyn22}. Aided
69: by advanced technology, \citet[B\&S;][]{bahc80,bahc81} demonstrated
70: that the true power of star counts is realized when they are compared
71: to simulations of the fundamental equation of stellar statistics
72: \citep{vonsee}. The B\&S method relies heavily on the assumed
73: luminosity functions and density distributions, and has been a
74: standard method for many subsequent studies.
75:
76: The Galaxy is traditionally characterized by having a Population~I
77: disk and Population~II spheroid. In a series of studies of the
78: exponential disk, \citet{gil83} and \citet{gil84} established the need
79: for a thick and thin disk whose scale heights are inversely
80: proportional to the masses of the studied stars \citep[see Table~1
81: in][]{siegel02}. The standard description of the Galactic halo is a
82: de Vaucouleur or power-law profile, while the Besan\c{c}on flattened
83: spheroid with $c/a\!\approx\!0.5-0.8$
84: \citep{bahc84,robin00,larsen03,robin03} being the currently favored
85: parameters. Thorough discussions of star counts and their relevance
86: to Galactic structure are given in the Annual Reviews by
87: \citet{bahc86}, \citet{gil89}, and \citet{majewski}.
88:
89: Many of the Galactic models and the majority of the literature examine
90: on relatively luminous dwarf and/or giant stars and rarely address
91: sub-stellar objects. The discovery of the first extra-solar,
92: sub-stellar object, Gliese~229B \citep{naka95,opp95} motivated the
93: creation of the L and T spectral classes. With surface temperatures
94: ranging from 750--2200~K \citep{burg99}, the \LT dwarfs can contaminate
95: searches for $z\!\simeq\!6$, $i'$-band dropout objects \citep{yan03}
96: by mimicking the extremely red broad-band colors. This effect has
97: remained largely unquantified due to insufficient knowledge of the \LT
98: dwarf IMF, Galactic distribution, and local number density. Previous
99: work on their IMF and local number density
100: \citep{reid99,chab01,chab02,liu02} has suffered from limited
101: statistics. With the deep imaging of \nfields\ Hubble Space Telescope
102: (HST) Advanced Camera for Surveys (ACS) parallel fields, this study
103: increases the number of {\it faint} dwarfs by adding $\nltds$ to the
104: growing list. Little work has been done on the scale height of \LT
105: dwarfs; \citet{liu02} and \citet{grapes} estimated 100--400~pc as
106: based on a single object or a 3--4 objects in a single field.
107: Therefore the primary goal of this work is to estimate the scale
108: height of the \LT dwarf population by comparing the surface densities
109: from the ACS parallel fields to the Galactic structure models.
110:
111: \section{Observations} \label{observations}
112: The \LT dwarf candidates were selected from $\nfields$ HST/ACS
113: parallel fields covering a broad range in Galactic latitudes and
114: longitudes (see \tab{thetable}). All ACS fields have at least three
115: independent exposures in F775W (SDSS-$i'$) and F850LP (SDSS-$z'$) with
116: a total exposure time of 2--10~ks per bandpass. All fields are
117: $\geq$90\% complete at $z'$(AB){$\simeq$}26.0~mag \citep{yan04a}. We
118: adopt the AB magnitude system \citep{oke83}.
119:
120: After combining the individual ACS frames into final stacks using the
121: PyRAF-based script {\it multidrizzle} \citep{multi}, the SExtractor
122: package \citep{sex} was used in double-input mode to perform the
123: matched-aperture photometry. The F850LP stack was used to define the
124: optimal apertures for the flux measurements in both stacks. For
125: source detection, we used a 5$\times$5 Gaussian smoothing kernel with
126: a FWHM of 2.0 pixels, which is approximately the same as the FWHM of
127: the ACS point-spread function (PSF) on both image stacks.
128: We used total magnitudes (corresponding to the {\it MAG\_AUTO} option
129: in SExtractor) for the photometry and adopted the zero points
130: published in HST ACS Instrument Science Report \citep{HSTzero}.
131:
132: As a cursory selection, all objects with \iz$>$1.3~mag and
133: $z'\!<$26.0~mag were considered. \fig{colormag} is the
134: color-magnitude diagram for all point sources detected in all
135: $\nfields$~fields. Dashed lines indicate the imposed color-magnitude
136: limits and the large, filled stars represent the candidate \LT dwarfs.
137: For the five fields for which the F475W (SDSS-$g'$) band was
138: available, we required candidates to have \gi$>$0~mag. Objects near
139: the detector edges were not considered, yielding an effective area of
140: $\sim$9~arcmin$^2$ per ACS field. Extended objects were eliminated
141: from the analysis by using the FWHM parameter of SExtractor. In
142: \fig{elephant}, we plot the FWHM as a function of apparent magnitude
143: for each object in Field~1 as small dots. The locus of points at
144: FWHM$\sim\!0\farcs13$ and $z'\!<\!26$~mag (hereafter the ``stellar
145: locus''), represented as asterisks, are the unresolved objects with
146: the minimum possible FWHM. In addition to the above color-magnitude
147: criterion, we required all \LT dwarf candidates to lie within this
148: locus. Sources of contamination to these criteria are discussed in
149: \S~\ref{contamination}.
150:
151: \begin{figure}
152: \epsscale{.85}%fraction of original size....
153: \plotone{f1.ps}
154: \caption{Color-magnitude diagram for all point sources detected in our
155: $\nfields$~ACS fields. The small dots represent all objects that met
156: the stellar morphology classification, the solid stars are the
157: $\nltds$ \LT dwarf candidates, and the dashed lines represent the
158: imposed color-magnitude limits. Many point sources associated with
159: diffraction spikes, field edges, and spurious detections (objects
160: smaller than the PSF, likely residual cosmic rays) brighter than
161: $z'\!=\!26$~mag were manually removed. The similar points fainter
162: than $z'\!=\!26$~mag were not removed from this figure. The sample of
163: stellar candidates, becomes incomplete for
164: $z'\!\geq\!25$~mag.}\label{colormag}
165: \end{figure}
166:
167: \begin{figure}
168: \epsscale{1.0}%fraction of original size....
169: \plotone{f2.ps}
170: \caption{The observed FWHM for all objects in Field~1 is plotted as a
171: function of apparent $z'$ magnitude. The small dots, asterisks, and
172: filled stars represent all objects, the point sources used to define
173: the ``stellar-locus,'' and the objects selected as candidate \LT
174: dwarfs. Clearly many objects were recognized as point sources,
175: however only two met the \iz\ color criterion.}\label{elephant}
176: \end{figure}
177:
178: \input{tab1.tex}
179:
180: Typically Galactic structure studies examine star counts
181: from one or many shallow fields with large surveyed areas
182: \citep[eg.][]{siegel02,larsen03}. Thus the data of the 2MASS
183: \citep{burg99,kirk99}, DENIS \citep{delf99}, and/or SDSS
184: \citep{stra99,tsve00,hawl02} are natural choices to study the Galactic
185: distribution of the \LT dwarfs. These surveys have significantly more
186: detection area than our HST dataset and their \LT dwarfs are typically
187: closer to the Sun than $\sim\!300$~pc or 1~disk scale height.
188: Therefore to avoid {\it extrapolating} the vertical scale height
189: beyond this distance, we chiefly analyzed the HST dataset where all
190: sample stars are likely more distant than 1~$e$-folding length.
191:
192: \section{The Simple Galactic Model}\label{model}
193:
194: The Galactic structure models were made by distributing
195: 10\super{10}~points according to an exponential disk that was
196: motivated by the light profiles of edge-on galaxies \citep{degrijs97}:
197: $n(r,\theta,z)=n_0\,\exp{\left(\frac{R_0-r}{r_l}\right)}\exp{\left(\frac{Z_0-z}{z_h}\right)}$,
198: where $r_l$=$\rscl$~pc is the radial scale length found by
199: \citet{porc98}, $n_0$=0.12~pc$^{-3}$ is the local space density \LT
200: dwarfs taken from \citet{chab02}, and $R_0$=8~kpc and $Z_0$=15~pc
201: \citep{yama92} are the solar radius and height, respectively. The
202: vertical scale height, $z_h$, is the free parameter and found by
203: minimizing the squared difference between the number counts from the
204: model and the HST data. Altering the assumed coordinates of the Sun
205: and the radial scale length have little effect on the vertical scale
206: height estimate. To generate absolute magnitudes, we adopted the
207: $J$-band luminosity function of \citet{cruz03} and the \zj\ colors of
208: \citet{hawl02} over the appropriate range of spectral type.
209:
210: The effects of interstellar extinction were included into the model
211: using the COBE dust maps of \citet{schl98} in two ways. First, we
212: assumed that each point was located beyond the dust, establishing a
213: lower bound on the model counts. Alternatively, an upper bound is
214: reached by assuming that the Galaxy has {\it no} dust whatsoever.
215: Since an overwhelming majority of the dust is localized to \absb\ (see
216: \fig{modelgal}) and only two of our observed ACS fields are in this
217: range, either approach yielded the same result within
218: the uncertainties, therefore we adopted the third method for
219: simplicity. \fig{modelgal} is a representative realization the model
220: with a scale height of \minz\ pc, the with locations of the $\nfields$
221: observed fields are indicated with plus signs.
222:
223:
224: \begin{figure}
225: \epsscale{1.0}%fraction of original size....
226: \plotone{f3.ps}
227: \caption{A sample realization of the Monte Carlo simulation
228: with $10^{10}$ random points with the best-fit vertical scale height of
229: \minz\ pc. The over-plotted crosses with numbers represent each of the
230: observed HST/ACS field in \tab{thetable}. The effects of the
231: \citet{schl98} dust maps are readily apparent when comparing the upper (no
232: extinction) and lower (extinction) panels.}\label{modelgal}
233: \end{figure}
234:
235:
236: The canonical disk/spheroid Galaxy likely has additional components
237: \citep{bahc86}, the models used here did not contain any contribution
238: from the Galactic bulge or a two-component disk \citep{gil83,gil84}
239: for the following reasons. First, a bulge distribution was not
240: modeled since its radius is $\sim$1--2~kpc or
241: $\sim$7$\degr$--14$\degr$, and every field is well beyond 14$\degr$ of
242: the Galactic center, hence we do not require a bulge component in the
243: model. Second, this sample contains only $\nltds$ \LT dwarfs which
244: are likely within $\sim$1000~pc (based on the luminosity function) of
245: the Sun. Since the thick disk has a scale height of $\gtrsim$1000~pc,
246: we expect the star counts to be dominated by a {\it single} disk
247: population. Moreover, With only $\nltds$ candidates the models and
248: analyses must remain simple and straightforward.
249:
250:
251: \section{Analysis} \label{analysis}
252:
253: Despite this work utilizing the largest dataset of \LT dwarfs compiled
254: from HST observations, the star counts remain very sparse requiring a
255: simple analysis scheme. Using the grid of Monte Carlo models
256: described in \S~\ref{model}, we sought the scale height which
257: minimizes the squared difference between the integrated star counts of
258: the model and those from the HST/ACS dataset. For the fields where no
259: \LT dwarf candidates were found, we assumed an upper limit of one
260: object (per field) could have been detected and perform this
261: minimization technique simultaneously on all $\nfields$ fields. This
262: procedure yielded a vertical scale height of \finz~pc. In the upper
263: panel of \fig{surfres}, we plot the modeled surface density averaged
264: over Galactic longitude as a function of Galactic latitude as computed
265: from the model with a scale height of \minz~pc, with the HST data
266: points from \tab{thetable} over-plotted for comparison. The residuals
267: in the lower panel clearly demonstrate that the model with a scale
268: height of \minz~pc reproduces the HST star counts for
269: $\left|\lat\right|\!\geq\!15\degr$ where dust extinction is minimal.
270:
271:
272: \begin{figure}
273: \epsscale{1.0}
274: \plotone{f4.ps}
275: \caption{{\bf Top panel:} model surface density as a function of
276: Galactic latitude.
277: Here we have averaged over all longitudes with a
278: 2.5$\sigma$-clipping to better handle the sparse statistics at high
279: latitudes.
280: The dotted and full lines indicate models with and without
281: the extinction corrections, respectively.
282: The data from the \nfields\ ACS fields are plotted for
283: comparison as filled circles and downward arrows as an upper limit
284: when either zero or one object was detected. {\bf Bottom panel:} the
285: residuals from the upper panel as a function of Galactic latitude.
286: Clearly the data where the dust corrections are large (ie. \absb) are
287: the most deviant. The model used in both panels has a vertical scale
288: height of \minz\ pc. The two fields outlying fields at \absb\ are
289: discussed in \S~\ref{contamination}. }\label{surfres}
290: \end{figure}
291:
292: \subsection{Sources of Contamination} \label{contamination}
293: While all \LT dwarf candidates were systematically found by
294: color and FWHM criteria, each was visually confirmed as a point
295: source. However, the color-magnitude rules outlined in
296: \S~\ref{observations} potentially find three classes of contaminates:
297:
298: (1) The primary motivation of this study is to reliably correct the
299: $z\!\simeq\!6$, $i'$-band dropout galaxy surveys for interloping \LT
300: dwarfs \citep{yan02,yan03,yan04a,yan04b}. Since our method is similar
301: to the $i'$-band dropout technique, we expect possible contamination
302: to the star counts from the $z\!\simeq\!6$ galaxies. In a recent
303: study of the Hubble Ultra-Deep Field (HUDF), \citet{yan04b} find 108
304: $i'$-band dropouts using the \iz$>$1.3~mag color selection. Their
305: sample has a median $z'$-band magnitude of 28.5~mag, and contains only
306: three objects brighter than our limit of $z'\!=$26.0~mag. Each of
307: these three objects is considerably extended and could not be mistaken
308: for a point-source. Since the majority of unresolved $z\!\simeq\!6$
309: galaxies will be $\sim$2.5~mag too faint, we conclude their
310: contamination in our sample must be negligible.
311:
312: (2) Another known source of possible contaminates comes from dusty,
313: elliptical galaxies with redshifts $1.0\!\leq\!z\!\leq\!1.5$, whose
314: 4000~$\mbox{\AA}$ break occurs between the $i'$ and $z'$ bands
315: \citep{yan03}. With a typical color of \iz$\sim$1.0~mag, many of
316: these objects would appear too blue in the absence of extreme internal
317: reddening. Moreover the visual identification confirms only
318: point-like objects populate the ``stellar locus'' in the lower left of
319: \fig{elephant}. Hence, extended elliptical galaxies could not grossly
320: corrupt our sample.
321:
322: (3) In addition to the above extragalactic sources, we anticipated
323: contamination from galactic M-dwarfs for two different reasons.
324: First, the color criterion of \iz$>\!1.3$~mag was primarily motivated
325: by the $z\!\simeq\!6$ galaxy surveys and is $\sim$0.5~mag too blue
326: \citep{hawl02} to have included only \LT dwarfs. When we repeated the
327: above analysis for \iz$\geq$1.8~mag \citep{hawl02} the star counts
328: were reduced by $\sim$50\% and the inferred vertical scale height was
329: $300\!\pm\!100$~pc. While there is a significant contribution from
330: late M-dwarfs, the vertical scale height was unaffected by the 0.5~mag
331: color difference. Second, an appreciably reddened M-dwarf could have
332: an \iz\ color of an unreddened L- or T-dwarf. This scenario requires
333: considerable reddening, only the two fields with \absb\ have $E\iz
334: \!\gtrsim\!0.1$~mag. This effect could account these fields lying
335: more than $1\sigma$ above the best-fit line in \fig{surfres}, however
336: we cannot be certain without more broad-band filters or spectroscopy.
337: We investigated this effect's contribution by removing the two suspect
338: fields and repeated the analysis. While this procedure reduced the
339: observed star counts by $\sim$50\%, it resulted in a vertical scale
340: height of 360$\pm$180~pc. Without further observations it is
341: difficult to definitively remove highly reddened M-dwarfs, however
342: their contribution should not grossly affect our main goal.
343:
344: \section{Discussion} \label{discussion}
345:
346: Using a suite of Monte Carlo simulations and $\nfields$ HST/ACS
347: parallel fields, we find a vertical scale height of $\finz$~pc for the
348: \LT dwarf population based on $\nltds$ faint candidates. This
349: estimated scale height is consistent with the known trend of
350: increasing scale height with decreasing stellar mass, independent of
351: reddening, color selections, and other Galactic parameters and is
352: within the uncertainties of and is largely a refinement of previous
353: work \citep{liu02,grapes}. Using our value of the scale height, and
354: the parameters given in \S~\ref{model}, we predict a total of
355: $\sim\!10^{11}$ \LT dwarfs and a total mass of
356: $\lesssim10^{9}$~M$_{\odot}$ in the Milky Way.
357:
358: This improved understanding of the \LT dwarf Galactic distribution
359: will aide high-redshift surveys in better estimating the contamination
360: of \LT dwarfs in their samples. In the recent HUDF pointed at
361: $(\alpha,\delta)$=(3$^{\mathrm h}$32$^{\mathrm
362: m}$39$\fs$0,--27\degr47'29$\farcs$1) and a depth of $z'\!\sim$29~mag,
363: we predict $\gtrsim$2 \LT dwarfs in its $\sim$11~arcmin$^{2}$ field of
364: view, which has been confirmed by \citet{grapes} who have
365: spectroscopically identified three \LT dwarfs. We confirm that
366: Galactic \LT dwarfs cannot significantly corrupt the $z\!\simeq\!6$
367: surveys in high-latitude fields (the HUDF for example), however
368: low-latitude fields will find a modest number of interloping \LT
369: dwarfs. With only $\nltds$ candidates from $\nfields$ fields, our
370: statistics remain sparse and ideally require further observations.
371:
372: \acknowledgments This work was funded by the ASU NASA Space Grant.
373: The authors thank Dave Burstein and Neill Reid for their helpful
374: discussions on Galactic structure and \LT dwarfs. The authors are
375: greatly appreciative for the referee's insightful and useful comments.
376: We wish to dedicate this work to the memory of Dr. John Bahcall.
377:
378: Facilities: \facility{HST(ACS)}
379:
380:
381:
382: \begin{thebibliography}{}
383: \bibitem[Bahcall(1986)]{bahc86} Bahcall, J. N. 1986, ARA\&A, 24, 577
384: \bibitem[Bahcall \& Soneira(1980)]{bahc80} Bahcall, J. N. \& Soneira, R. M. 1980, \apjs, 44, 73
385: \bibitem[Bahcall \& Soneira(1981)]{bahc81} Bahcall, J. N. \& Soneira, R. M. 1981, \apj, 246, 122
386: \bibitem[Bahcall \& Soneira(1984)]{bahc84} Bahcall, J. N. \& Soneira, R. M. 1984, \apjs, 55, 67
387: \bibitem[Bertin \& Arnouts(1996)]{sex} Bertin, E. \& Arnouts, S. 1996, A\&AS, 117, 393
388: \bibitem[Burgasser et al.(2003)]{burg03} Burgasser, A., Kirkpatrick, D., Burrows, A., Liebert, J., Reid. N., Gizis, J., McGovern, M., \& Prato, L. 2003, \apj, 592, 1186
389: \bibitem[Burgasser(1999)]{burg99} Burgasser, A. J., et al. 1999, \apj, 522, L65
390: \bibitem[Chabrier(2001)]{chab01} Chabrier, G. 2001, \apj, 554, 1274
391: \bibitem[Chabrier(2002)]{chab02} Chabrier, G. 2002, \apj, 567, 304
392: \bibitem[Cruz et al.(2003)]{cruz03} Cruz, K. C., Reid, I. N., Liebert, J., Kirkpatrik, J. D., Lowrance, P. J. 2003, \aj, 126, 2421
393: \bibitem[de Grijs, Peletier, \& van der Kruit(1996)]{degrijs97} de Grijs, R., Peletier, R., \& van der Kruit, P. 1997, A\&A, 327, 966
394: \bibitem[Delfosse(1999)]{delf99} Delfosse, X., Tinney, C. G., Forveille, T., Epchtein, N., Borsenberger, J., Fouqu\'{e}, P., Kimeswenger, S., \& Tip\`{e}ne, D. 1999, \aaps, 135, 41
395: \bibitem[De Marchi et al.(2004)]{HSTzero} De Marchi, G. et al. 2004, Detector Quantum Efficiency and Photometric Zero Points of the ACS, Instrument Science Report, (STScI, Baltimore)
396: \bibitem[Gilmore(1984)]{gil84} Gilmore, G. 1984, \mnras, 207, 223
397: \bibitem[Gilmore \& Reid(1983)]{gil83} Gilmore, G \& Reid, N. 1983, \mnras, 202, 1025
398: \bibitem[Gilmore, Wyse, \& Kuijken(1989)]{gil89} Gilmore, G., Wyse, R., \& Kuijken, K. 1989, ARA\&A, 27, 555
399: \bibitem[Hawley et al.(2002)]{hawl02} Hawley, S. et al. 2002,\apj, 123, 3409
400: \bibitem[Herschel(1785)]{herschel} Herschel W. 1785, RSPT, 75, 213
401: \bibitem[Kapteyn(1922)]{kapteyn22} Kapteyn, J. C. 1922, \apj, 55, 302
402: \bibitem[Kirkpatrick et al.(1999)]{kirk99} Kirkpatrick, J. D., et al. 1999, \apj, 519, 802
403: \bibitem[Kirkpatrick(2000)]{kirk00} Kirkpatrick, J. D. 2000, \aj, 120, 447
404: \bibitem[Koekemoer et al.(2002)]{multi} Koekemoer, A. M., Fruchter, A. S., Hook, R., \& Hack, W. 2002 ``{\it The 2002 HST Calibration Workshop}'' Space Telescope Science Institute, 2002
405: \bibitem[Larsen \& Humphreys(2003)]{larsen03} Larsen, J. A. \& Humphreys, R. M. 2003, \aj, 125,1958
406: \bibitem[Liu et al.(2002)]{liu02} Liu, M., Wainscoat, R., Mart\'{i}n, E. L., Barris, B., \& Tonry, J. 2002, \apj, 568, L107
407: \bibitem[Majewski(1993)]{majewski} Majewski, S. R. 1993, ARA\&A, 31, 575
408: \bibitem[Nakajima et al.(1995)]{naka95} Nakajima, T., Oppenheimer, B. R., Kulkarni, S. R., Golimowski, D. A., Matthews, K., \& Durrance, S. T. 1995, Nature, 378, 463
409: \bibitem[Oke \& Gunn(1983)]{oke83} Oke, J. B. \& Gunn, J. E. 1983, \apj, 266,713
410: \bibitem[Oppenheimer et al.(1995)]{opp95} Oppenheimer, B. R., Kulkarni, S. R., Matthews, K., \& Nakajima, T. 1995, Science, 270, 1478
411: \bibitem[Porcel et al(1998)]{porc98} Porcel, C., Garz\'{o}n, F., Jimen\'{e}nez-Vicente, J., \& Battaner, E. 1998, A\&A, 330, 136
412: \bibitem[Pirzkal et al.(2005)]{grapes} Pirzkal N., et al. 2005, \apj, 622, 319
413: \bibitem[Reid \& Majewski(1993)]{reid93} Reid, I. N. \& Majewski, S. R. 1993, \apj, 409, 635
414: \bibitem[Reid(2003)]{reid03} Reid, I. N. 2003, \aj, 126, 2449
415: \bibitem[Reid et al.(1999)]{reid99} Reid, I. N., et al. 1999, \apj, 521, 613
416: \bibitem[Robin, Reyl\'{e}, \& Cr\'{e}z\'{e}(2000)]{robin00} Robin, A., Reyl\'{e}, C., \& Cr\'{e}z\'{e} M. 2000, A\&A 359, 103
417: \bibitem[Robin et al.(2003)]{robin03} Robin, A., Reyl\'{e}, C., Derri\`{e}re, S., \& Picaud, S. 2003, A\&A, 409, 523
418: \bibitem[Schlegel et al.(1998)]{schl98} Schlegel, D. J., Finkbeiner, D. P., \& Davis, M. 1998, \aj, 500, 525
419: \bibitem[Siegel et al.(2002)]{siegel02} Siegel, M. H., Majewski, S. R., Reid, I. N., \& Thompson, I. B. 2002, \apj, 578, 151
420: \bibitem[Strauss et al.(1999)]{stra99} Strauss, M. A., et al. 1999, \apj, 522, L61
421: \bibitem[Tsvetanov et al.(2000)]{tsve00} Tsvetanov, Z. I., et al. 2000, \apj, 531, L61
422: \bibitem[von Seeliger(1898)]{vonsee} von Seeliger, H. 1898, Abh. Bayerische Akad. Wiss., Math.-Phys. KI, 19, 564
423: \bibitem[Yan \& Windhorst(2004a)]{yan04a} Yan, H. \& Windhorst, R. A. 2004a, \apj, 600, L1
424: \bibitem[Yan \& Windhorst(2004b)]{yan04b} Yan, H. \& Windhorst, R. A. 2004b, \apj, 612, L93
425: \bibitem[Yan, Windhorst, \& Cohen(2003)]{yan03} Yan, H., Windhorst, R. A., \& Cohen S. H. 2003, \apj, 585, L93
426: \bibitem[Yan et al.(2002)]{yan02} Yan, H. et al. 2002, \apj, 580, 725
427: \bibitem[Yamagata \& Yoshii(1992)]{yama92} Yamagata, T., \& Yoshii, Y. 1992, \aj, 103, 117
428: \end{thebibliography}
429:
430: \end{document}
431: