0803.2044/ms.tex
1: %\documentclass[12pt,preprint]{aastex}
2: \documentclass{emulateapj} 
3: \shorttitle{Circumbinary disk around CoKu Tau/4}
4: \shortauthors{Ireland and Kraus}
5: \begin{document}
6: 
7: \title{The Disk Around CoKu Tau/4: Circumbinary, not Transitional
8: \altaffilmark{1} }
9: 
10: \author{Ireland, M.J.}
11: \affil{Division of Geological and Planetary Sciences, California Institute of
12:   Technology, Pasadena, CA 91125}
13: \email{mireland@gps.caltech.edu}
14: \and
15: \author{Kraus, A.L.}
16: \affil{Division of Physics, Mathematics and Astronomy, California Institute of
17:   Technology, Pasadena, CA 91125}
18: 
19: \altaffiltext{1}{Data presented herein were obtained at the
20: W.M. Keck Observatory, which is operated as a scientific partnership
21: among the California Institute of Technology, the University of
22: California and the National Aeronautics and Space Administration. The
23: Observatory was made possible by the generous financial support of the
24: W.M. Keck Foundation. }
25: 
26: \begin{abstract}
27: CoKu Tau/4 has been labeled as one of the very few known transition
28: disk objects: disks around young stars that have their inner disks cleared of
29: dust, arguably due to planetary formation. We report aperture-masking
30: interferometry and adaptive optics 
31: imaging observations showing that CoKu Tau/4 is in fact a near-equal
32: binary star of projected separation $\sim$53\,mas ($\sim$8\,AU). The spectral 
33: energy distribution of the disk is then naturally explained by inner
34: truncation of the disk through gravitational interactions with the
35: binary star system. We discuss the possibility that such ``unseen''
36: binary companions could cause other circumbinary disks to be labeled as
37: transitional.
38: \end{abstract}
39: \keywords{stars: low-mass, brown dwarfs}
40: 
41: \section{Introduction}
42: 
43: The lifetime of disks around young stars sets the time-scale for 
44: giant planet formation. Most young stars have disks that are optically
45: thick at all wavelengths or have no observable disk at all, while a
46: small subset of young stars are in an intermediate state, thought to
47: be in ``transition'' \citep{Skrutskie90}. Many of these intermediate
48: objects have optically thick ``cold disks'' as seen by large excesses
49: at $\sim$30\,$\mu$m, but much smaller or non-existent near- and mid-
50: infrared excesses \citep{Brown07}.
51: 
52: %For the vast
53: %majority of objects, the presence of an inner disk, seen by a
54: %near-infrared excess, correlates with the the spectroscopic
55: %classification of a classical T~Tauri star, while the lack of an
56: %optically-thick inner disk means an object is a weak-lined T~Tauri star
57: %\citep[e.g.][]{Hartmann05}. 
58: 
59: One model of this transition process is
60: one where the inner disk is cleared of material, leaving an
61: inner hole and low accretion rates, followed $\sim10^5$ years later by
62: a clearing of the outer disk. This inner hole could be caused by natural
63: processes of grain growth and/or photo-evaporation
64: \citep{Alexander06}, or truncation of the disk due to dynamical 
65: interactions with companions \citep{Lin79}.
66: Although the most exciting interpretation for the development of gaps
67: is arguably truncation due to planetary formation, it is also possible 
68: to dynamically truncate disks due to stellar companions 
69: \citep[e.g][]{Jensen97,Beust05,White05}. An example of
70: this is CS~Cha, which was announced as a transitional disk
71: \citep{Espaillat07} shortly after its discovery as a circumbinary disk
72: \citep{Guenther07}. Some authors \cite[e.g.][]{Furlan07} have
73: considered the term ``transitional'' to be an observational term that
74: means the spectral signature of a cleared inner disk
75: rather than an evolutionary term that implicitly applies to single stars. 
76: We use the term ``cold disk'' for this purpose.
77: 
78: CoKu Tau/4 is a weak-lined M1.5 T~Tauri star \citep[H-$\alpha$ equivalent
79:   width 1.8\,$\AA$, ][]{Cohen79} in the Taurus star-forming region
80: (1$\sim$2\,Myr, $\sim$145\,pc).
81: CoKu Tau/4 was discovered to have a large $\sim$20-30\,$\mu$m excess
82: and no excess at wavelengths $<8$\,$\mu$m through {\em Spitzer Space
83:   Telescope} Infrared Spectrograph observations \citep{Forrest04}. The
84: disk has been modeled as having an inner hole of radius $\sim$10\,AU
85: \citep{DAlessio05}, with 
86: suggestions that this hole is due to a giant planet \citep{Quillen04}.
87: 
88: In this paper, we describe near-infrared aperture-masking
89: interferometry and imaging observations that demonstrate that CoKu
90: Tau/4 is a near-equal mass binary star system. We show that the
91: predicted inner hole size from dynamical models is comparable to, but larger
92: than, that predicted from radiative transfer models. 
93: Finally, we discuss whether other so-called ``transitional'' disks could 
94: be circumbinary disks. We conclude that for candidates in Taurus, much of the 
95: mass ratio-separation space for stellar companions can by ruled out by 
96: existing observations, but that definitively ruling out the possibility
97: of binarity is generally difficult for individual so-called ``transitional'' 
98: disks. 
99: 
100: \section{Observations}
101: \label{sectObservations}
102: 
103: CoKu Tau/4 was observed with the NIRC2 camera behind Adaptive Optics
104: (AO) at the Keck II telescope on 2007 Nov 23 as
105: part of an ongoing aperture-masking survey of nearby young
106: star-forming associations. Aperture-masking
107: interferometry \citep[e.g.][]{Tuthill00} is a well established
108: technique for achieving the full diffraction limit of a single
109: telescope, recently applied to observations behind adaptive optics
110: systems \citep[e.g.][]{Lloyd06,Kraus08a}. A 9-hole mask was placed in a
111: filter wheel near a pupil plane in the NIRC2 camera, enabling
112: interference fringes on 36 baselines to be simultaneously recorded on
113: the camera's imaging array. The observations of CoKu Tau/4 consisted
114: of two image sets taken through a K' filter,
115: each with eight 20 second exposures, calibrated by 
116: two interleaved image sets of CX~Tau. The airmass of CoKu Tau/4
117: observations varied from 1.28 to 1.38, while the airmass of CX Tau (6
118: degrees from CoKu Tau/4) varied from 1.32 to 1.45 over the 45 minutes
119: it took to complete the observations. From the Fourier transforms of the target
120: and calibrator's images, squared-visibilities and closure-phases were
121: extracted. The squared visibilities of CoKu Tau/4 were calibrated by
122: dividing by the squared visibilities of the calibrator, and the
123: closure-phases were calibrated by subtracting the closure-phases of
124: the calibrator (which are non-zero due to instrumental effects). These
125: observations were made 
126: in poor seeing conditions (uncorrected seeing 1-2\arcsec in
127: K-band), so our observations only achieved typical sensitivities of
128: $\sim$50:1 at the diffraction limit, less than the $\sim$200:1 we
129: achieved in Upper Sco \citep{Kraus08a}. These high sensitivities are
130: primarily due to the non-redundant nature of the aperture-mask, which
131: means that calibration is independent of wavefront structure on scales
132: larger than a single sub-aperture ($\sim$1\,m).
133: 
134: \begin{figure}
135:  \plotone{f1.eps}
136:  \caption{Calibrated squared-visibilities for CoKu~Tau~4, plotted
137:    against baseline projected along a position angle of 306 degrees. 
138:    A binary model with contrast 1.25 and separation 54\,mas is
139:    over-plotted. The dashed line shows the visibility expected for a
140:    single star.}
141:  \label{figV2}
142: \end{figure}
143: 
144: \begin{figure}
145:  \plotone{f2.eps}
146:  \caption{Measured closure phases for CoKu~Tau~4 as a function of
147:    modeled closure phases, assuming that CoKu~Tau~4 is a 1.25:1 binary
148:    with separation 54\,mas at a position angle of 306 degrees. The
149:    dashed line shows the zero closure-phase expected for a single star.}
150:  \label{figCP}
151: \end{figure}
152: 
153: \begin{figure}
154:  \plotone{f3.eps}
155:  \caption{Top line: Contour plots of the original AO image, only
156:    the A component (best-fit B solution subtracted), and only the B component
157:    (best-fit A solution subtracted) in the Br$\gamma$ filter,
158:    demonstrating the PSF fitting technique. Bottom line: AO images in
159:    the K$_{cont}$, H$_{cont}$ and J$_{cont}$ filters, clearly showing
160:    the binary despite the low Strehl. The contours are drawn at 10\%
161:    to 90\% in intervals of 10\%. } 
162:  \label{figIm}
163: \end{figure}
164: 
165: \begin{deluxetable}{lcrrr}
166: \tabletypesize{\scriptsize}
167: \tablewidth{0pt}
168: \tablecaption{Direct Imaging Observations}
169: \tablehead{\colhead{Filter} & \colhead{Wavelength} &
170: \colhead{$\rho$} & \colhead{PA} & \colhead{$\Delta$}
171: \\
172: \colhead{} & \colhead{Range ($\mu$m)} &
173: \colhead{(mas)} & \colhead{(deg)} & \colhead{(mag)}}
174: \startdata
175: \multicolumn{5}{l}{Aperture Masking}\\ %320s integration
176: $K'$     &1.948-2.299 & $54.1\pm0.3$ & $306.7\pm 0.4$ & $0.23\pm0.01$ \\
177: \multicolumn{5}{l}{PSF Fitting} \\
178: $H_{cont}$&1.569-1.592 & $53.6\pm0.7$ & $306.3\pm0.3$ & $0.24\pm0.04$ \\
179: $Br\gamma$&2.152-2.185 & $53.3\pm0.2$ & $307.7\pm0.7$ &$0.25\pm0.01$ \\
180: $K_{cont}$&2.256-2.285 & $53.6\pm0.2$ & $307.1\pm0.2$ &$0.21\pm0.01$ \\
181: \multicolumn{5}{l}{Bispectrum Analysis} \\
182: $J_{cont}$ &1.203-1.223 & $51.7\pm0.9$  & $305.8\pm0.6$ & $0.23\pm0.06$  \\
183: $H_{cont}$ &1.569-1.592 & $53.3\pm0.2$  & $306.9\pm0.2$ & $0.21\pm0.02$ \\
184: $Br\gamma$ &2.152-2.185 & $53.9\pm0.2$  & $306.4\pm0.2$  & $0.19\pm0.01$ \\
185: $K_{cont}$ &2.256-2.285 & $53.6\pm0.2$  & $306.0\pm0.2$ & $0.20\pm0.01$ \\
186: \enddata
187: \label{tabImaging}
188: \end{deluxetable}
189: 
190: We found that only a binary model provided
191: a good fit to the aperture-masking data, with a near-equal binary well
192: within our detection limits. The calibrated squared visibilities and
193: closure phases are shown in 
194: Figures~\ref{figV2} and \ref{figCP}, demonstrating that a binary
195: solution provides an excellent fit to the data. The binary solution
196: based on the aperture-masking data is given in Table~\ref{tabImaging}. 
197: %a $54.1\pm0.3$\,mas separation
198: %at a position angle of $306.7\pm 0.4$ and a contrast of
199: %$0.23\pm0.01$\,magnitudes.  
200: 
201: Preliminary analysis of the masking data indicated that the binary
202: companion's separation and flux ratio would allow it to be resolved
203: with direct imaging, so we obtained regular AO images
204: on the following night under slightly better observing conditions
205: (uncorrected seeing $\sim$1\arcsec in K-band) at an airmass of 1.06.   
206: These observations were obtained with NIRC2 in the narrow
207: camera mode ($\sim$10 mas pix$^{-1}$), and since the target is very
208: bright in the near-infrared, we used the narrow-band filters
209: $J_{cont}$, $H_{cont}$, $K_{cont}$, and $Br\gamma$.  Total exposure
210: times were 40\,s in all filters except for $J_{cont}$, where twice as
211: many images were taken due to the very low Strehl, for a total
212: exposure time of 80\,s. We show the images in
213: Figure~\ref{figIm}. Despite the low Strehl ratio and the elongation of
214: the speckle halo due to a dominant wind direction, the binary is
215: clearly resolved in all images.
216: 
217: We extracted photometry and astrometry from the direct AO imaging 
218: of these sources using the
219: PSF reconstruction technique described in \citet{Kraus07b},
220: which iteratively fits a template PSF to the primary and then
221: subtracts the secondary to fit an improved estimate of the
222: primary. This routine was implemented using the ALLSTAR routine in the
223: IRAF package DAOPHOT (Stetson 1987), and based on previous experience
224: with NIRC2 imaging, we chose to model the PSF using a Gaussian core
225: with Lorentzian wings. This method is illustrated in
226: Figure~\ref{figIm}. In addition to this method, we used a recursive-phase
227: algorithm \citep[e.g][]{Lohmann83} to reconstruct Fourier Phase from  
228: bispectrum phase, and then fitted a binary model to the reconstructed phase of
229: the image. 
230: In Table~\ref{tabImaging}, we summarize the relative astrometry
231: and photometry that we measured from our direct imaging observations. 
232: The consistency in separation and position angle at the 1\% level in H and 
233: K bands and the 5\% level in J band despite very different PSFs and diffraction
234: effects demonstrates the reliability of our results. However, the
235: small reported statistical errors in Table~\ref{tabImaging} are not
236: entirely consistent internally, requiring additional
237: systematic errors to be added. Therefore, we assign a weighted mean
238: solution of separation $53.6\pm0.5$\,mas and a position angle of
239: $306.4\pm0.6$\,degrees. The contrast ratio in J, H and K  is
240: consistent with there being no significant difference in reddening
241: between the components. 
242: 
243: %From Adam:
244: 
245: %Kcont
246: %Sep = 53.1 +/- 0.1 mas
247: %PA = 307.1 +/- 0.2 deg
248: %deltmag = 0.21 +/- 0.01 mag
249: %Ktot = 8.656
250: %K_a = 9.31
251: %K_b = 9.52
252: %
253: %Brgamma
254: %Sep = 52.8 +/- 0.2 mas
255: %PA = 307.7 +/- 0.7 deg
256: %deltamag = 0.25 +/- 0.01 mag
257: %Ktot = 8.656
258: %K_a = 9.29
259: %K_b = 9.54
260: %
261: %Hcont
262: %Sep = 53.1 +/- 0.7 mas
263: %PA = 306.3 +/- 0.3 deg
264: %deltamag = 0.24 +/- 0.04 mag
265: %Htot = 9.077
266: %H_a = 9.72
267: %H_b = 9.97
268: %
269: %Colors:
270: %
271: %(H-K)_a ~ 0.42
272: %(H-K)_b ~ 0.43
273: 
274: %\begin{deluxetable*}{lccrrrrrrrr}
275: %\tabletypesize{\scriptsize}
276: %\tablewidth{0pt}
277: %\tablecaption{Direct Imaging Observations. [ADAM: please help me
278: %    construct a pretty deluxetable header for the two analyses]}
279: %\tablehead{\colhead{Filter} & \colhead{Wavelength} & \colhead{$T_{int}$} &
280: %\colhead{$\rho_{\rm IM}$} & \colhead{PA$_{\rm IM}$} & \colhead{$\Delta$$m_{\rm IM}$} &
281: %\colhead{$\rho_{\rm BS}$} & \colhead{PA$_{\rm BS}$} & \colhead{$\Delta$$m_{\rm BS}$} &
282: %\colhead{$m_A$} & \colhead{$m_B$}
283: %\\
284: %\colhead{} & \colhead{Range ($\mu$m)} & \colhead{(sec)} &
285: %\colhead{(mas)} & \colhead{(deg)} & \colhead{(mag)} &
286: %\colhead{(mas)} & \colhead{(deg)} & \colhead{(mag)} &
287: %\colhead{(mag)} & \colhead{(mag)} 
288: %}
289: %\startdat%a
290: %$J_{cont}$&1.2033-1.2231 &80& - & - & - & 
291: % $51.7\pm0.5$ & $305.8\pm0.4$ & $0.23\pm0.06$ &- & - \\
292: %$H_{cont}$&1.5688-1.5920 &40& $53.6\pm0.7$ & $306.3\pm0.3$ & $0.24\pm0.04$ &
293: % $53.3\pm0.14$ & $306.92\pm0.14$ & $0.21\pm0.02$ &9.72 & 9.97 \\
294: %$Br\gamma$&2.1523-2.1849 &40& $53.3\pm0.2$ & $307.7\pm0.7$ &$0.25\pm0.01$ &
295: % $53.9\pm0.1$  & $306.4\pm0.1$  & $0.19\pm0.01$ &9.29 & 9.54 \\
296: %$K_{cont}$&2.2558-2.2854 &40& $53.55\pm0.1$ & $307.1\pm0.2$&$0.21\pm0.01$ &
297: % $53.62\pm0.08$& $306.03\pm0.08$ & $0.20\pm0.01$ &9.31 & 9.52 \\
298: %\enddata
299: %\label{tabImaging}
300: %\end{deluxetable*}
301: 
302: \section{System Properties}
303: 
304: \subsection{Stellar and Binary Properties}
305: \label{sectProperties}
306: 
307: Both components have near infrared colors that are consistent with other young
308: Taurus members of spectral type $\sim$M1-M2, suggesting that they have
309: similar temperatures and that neither has a significant $K$ band
310: excess due to a hot inner disk. In order to model the stars and the
311: disk, it is important to know the bolometric 
312: luminosity of the system, and hence the reddening.
313: To estimate the reddening, we used the following data: intrinsic colors
314: reported in \citet{Bessell88} and \citet{Bessell90} interpolating
315: between giants (log$(g) \sim 1$) and dwarfs (log$(g) \sim 5$) at an
316: assumed log$(g)$ of 3.7, JHK photometry from 2MASS,  V-band photometry
317: averaged between \citet{Cohen79} and \citet{Hanson04} with an assumed
318: 0.1\,mag overall uncertainty, and an M1.5 spectral classification 
319: from \citet{Cohen79} with an assumed 1 sub-class uncertainty. We arrive at
320: $E_{V-K}=2.3\pm0.3$ and $E_{J-K}=0.58\pm0.04$.  The ratio $E_{V-K}/E_{J-K} =
321: 4.0\pm0.6$ is significantly lower than the 5.6 typical of Taurus
322: \citep{Whittet01}, suggesting a reddening law more typical of a dense
323: core. Based on $E_{V-K}$, the reddening $A_V$ is $2.5\pm0.3$ using the
324: relationship $A_V \approx 1.1 E_{V-K}$ from \citet{Whittet01}, but given
325: the uncertainties in the reddening law, $A_V=3$ as used by
326: \citet{DAlessio05} is certainly consistent. 
327: %NB: A small K-band excess doesn't make sense, because even with
328: %optically-thin dust, a ~0.1 mag K-band excess gives ~1 mag at the
329: %short end of IRS
330: 
331: %NB2: Kirkpatrick (1991) has a good spectral atlas of M stars.
332: 
333: %Using Lynne's Reddening table: We arrive at a
334: %reddening of $A_V=2.59\pm0.12$ from $V-K$, and  $A_V=3.56\pm0.22$ from
335: %J, H and K photometry alone. An uncertainty of 1 sub-class in spectral
336: %type corresponds to an additional reddening uncertainty of
337: %$\sim$0.25 for $V-K$ based reddening, with a negligible difference to
338: %the reddening based on $J$, $H$ and $K$ magnitudes. We do not have a
339: %way to reconcile these numbers, but note that \citet{Flaherty07} find
340: %that in dense clouds typical of star forming regions, near infrared
341: %reddening laws differ substantially from those based on interstellar
342: %dust.
343: 
344: %NB, if we were to B-band photometry from \citet{Hanson04}, then that
345: %is consistent with an _unreddened_ star!
346: 
347: Absolute stellar properties like
348: masses can be difficult to estimate because pre-main-sequence
349: evolutionary models are not well-calibrated, but based on the
350: prescription described in \citet{Kraus07a}, both
351: components are likely to have masses of $\sim$0.5-0.6
352: $M_{\sun}$.  Using a 145\,pc distance to Taurus, the system has an
353: absolute K magnitude of 2.55, which implies absolute K magnitudes of
354: 3.2 and 3.4 for the components. According to the models of
355: \citet{Baraffe98}, assuming an effective temperature of $\sim$3600\,K
356: applicable to M1.5 stars \citep{Luhman99}, the stars are 4\,MYr old
357: and are of mass $\sim$0.52 and $\sim$0.62\,$M_\sun$. If the effective
358: temperature is closer to 3500\,K, the stars would be of slightly lower
359: mass and $<3$\,Myr old, consistent with the canonical age of Taurus. 
360: The observational scatter in the Taurus HR diagram
361: \citep[e.g][]{Luhman04} is consistent with CoKu Tau/4 being either
362: single as assumed by previous authors or an equal brightness double.
363: The mass ratio, however, should be more robust than the masses or age
364: (as it is only weakly dependent on reddening), so we can assert with
365: greater confidence that $M_B/M_A=$0.85$\pm$0.05.
366: 
367: The minimum possible semi-major axis ($a$) for the orbit is
368: the apparent separation divided by (1+$e$), which is 7.8/(1+$e$)\,AU
369: based on an assumed distance 
370: of 145\,pc for Taurus, giving an orbital period $>$10 years. 
371: Based on considerations of dynamical disk
372: truncation in Section~\ref{sectModel}, it is likely that the system
373: was observed near maximum elongation.
374: 
375: \subsection{The Circumbinary Disk}
376: \label{sectModel}
377: 
378: %NB Roche Lobe really has nothing to do with this, at only 0.764a. 
379: %2/(2r-a) + 2/(2r+a) - r^2/a^3 = 4/a
380: %If instead, emission were spread over a range
381: %of radii (e.g. in a flared disk with a curved wall), one would expect
382: %this to drive the inner-most radius of emission inwards. Furthermore,
383: %\citet{DAlessio05} find an insignificant contribution from dust inside
384: %the inner wall to the mid-infrared spectrum. We should therefore
385: %consider their radii to be model-dependent outer limits of the disk
386: %truncation radius. 
387: 
388: We expect the inner radius of the CoKu Tau/4 disk to be set by tidal
389: truncation. To simplify our discussion, we will assume that the
390: disk is roughly co-planar with the binary and (as described in
391: Section~\ref{sectProperties}) that the binary components have similar
392: masses. In this geometry, \citet{Artymowicz94} calculate that the 3:1
393: orbital resonance at $R_D\sim2.08a$ is opened for all but improbably
394: small eccentricities ($e\ga0.03$). \citet{Beust05} predict a
395: gap size of $>2.6a$ for eccentricities $>0.1$ for GG~Tau, which has a
396: similar mass ratio to CoKu Tau/4 (and find this not to be large enough
397: for that system). For very large eccentricities ($e\ga0.6$), larger
398: resonances open and predicted gap sizes from \citet{Artymowicz94} are
399: in excess of $R_D\sim3a$ but are viscosity-dependent. The closer
400: near-equal spectroscopic binary HD~98800B \citep{Akeson07} is an
401: example of a system with high eccentricity that should have opened
402: large resonances in the disk: but the radiative transfer modeling has
403: too many uncertainties to clearly say which resonance has opened. 
404: 
405: %NB St 34 is near equal but old, and is Hartmann05b
406: 
407: %The term of order $e^1$ for
408: %the 4:1 resonance vanishes for equal-mass binaries
409: %\citep{Artymowicz94}, so we will not consider this here. Note,
410: %however, that the opening of this and higer order resonances will be
411: %dependent on disk viscosity considerations, and for the binary GG~Tau,
412: %which has a similar mass-ratio to CoKu Tau/4, \citet{Beust05} predict
413: %a gap size of $>2.6a$ for eccentricities $>0.1$. 
414: %See equation 24 in Artymowicz. There is a (1-2\mu) term.
415: 
416: Based on the minimum semi-major axis of 7.8/(1+$e$)\,AU from
417: Section~\ref{sectObservations}, we arrive at minimum disk
418: truncation radius of $\sim$16\,AU if CoKu Tau/4 is a low ($e\sim0.1$)
419: eccentricity binary, or $\sim$13\,AU if CoKu Tau/4 is a high ($e\ga0.6$)
420: eccentricity binary. These numbers are
421: discrepant from the 9-12\,AU truncation radius predicted by the
422: detailed spectral modeling of \citet{DAlessio05}, who examined
423: radiative transfer models consisting of a geometrically thin wall and
424: a range of dust types. We will examine several ways to resolve this
425: discrepancy in turn: the differing radiation pattern of a binary
426: versus a single star, errors in the stellar luminosity, and the
427: possibility of a substantially different radiative transfer model.
428: 
429: %Note that this discrepancy has the opposite sign to that
430: %for the well-known circumbinary disk around 
431: %GG~Tau \citep{Beust05}, where the disk appears too large given the
432: %binary separation. 
433: 
434: A binary star has a different radiation pattern to a single star. In
435: particular, on a circular ring of radius $R_d$ centered on the binary center
436: of mass, flux is greatest at points on this ring aligned with the
437: axis of the binary. For an 8\,AU binary (each component 4\,AU from the
438: center of mass) and a 14\,AU ring, this geometry means that the flux
439: from the binary is between 0.92 and 1.28 times the flux from a single
440: star with the same total luminosity. We will assume that the
441: $\sim$145\,K modeled by \citet{DAlessio05} represents the maximum
442: temperature seen in the ring and this temperature determines
443: $L_S/R_D^2$. Here $L_S$ is the total stellar luminosity and $R_D$ is
444: the dust radius from the center of mass. Therefore, a 28\% increase in
445: flux $F$ could represent a $\sim$13\% increase in the ring radius from
446: radiative transfer modeling. The temperature contrast around the ring
447: for optically thin dust with opacity $\kappa_\lambda \propto
448: \lambda^{-1}$goes as $F^{1/5}$, meaning that the temperature around
449: the ring ranges from $\sim$136\,K to 145\,K. Gray dust would vary from
450: $\sim$133\,K to 145\,K. Although this contrast is only moderate, it
451: may present an additional challenge in detailed spectral fitting, as
452: the models of \citet{DAlessio05} required emission from a single
453: temperature only. 
454: 
455: %so the apparently excellent fit of the
456: %CoKu~Tau/4 spectrum with a dust at a single radius in \citet{DAlessio05} is
457: %not as compelling evidence for a single optically-thin dust
458: %temperature as it initially might seem.
459: 
460: % 0.8 power of luminosity. 1.28 versus 0.92.
461: %So a factor of 1.3 between coolest and warmest parts to the disk.
462: 
463: %A 14\,AU radius ring that is co-planar with this
464: %same binary in turn receives a 10\% increase in flux over a
465: %single star of equivalent luminosity. This effect would cause the disk
466: %radii of \citet{DAlessio05} to be underestimated by $\sim$5\%.
467: 
468: One way to reconcile the gap radius from radiative transfer modeling
469: with that from
470: dynamical considerations would be if the luminosity of CoKu Tau/4 is
471: significantly underestimated. If we keep the dust temperature and
472: therefore $L_S/R_D^2$ constant, increasing $L_S$ increases the model
473: $R_D$. An uncertainty in distance due to
474: the depth of Taurus \cite[e.g.][]{Torres07} can not reconcile this
475: discrepancy, as both $R_D$ determined from radiative transfer
476: modeling and the minimum $R_D$ determined by dynamical
477: considerations both increase linearly with the distance estimate.
478: However, an uncertainty in the reddening of CoKu Tau/4 translates to a
479: luminosity uncertainty. The probable reddening calculated in
480: Section~\ref{sectProperties} is lower than that used by
481: \citet{DAlessio05}, but uncertainties are considerable, and it may be
482: possible that a larger bolometric reddening could increase the gap
483: radius derived from radiative transfer modeling.
484: 
485: %2/(14^2 + 4^2) versus 1/10^2 + 1/18^2
486: 
487: %!!! I did say 60% what is the real number...? The infrared reddening
488: % would produce a 1 mag V-band excess, and a 2.4 mag B-band excess
489: %... which is significantly different from the value in
490: %\citet{Cohen79} of $0.94\pm0.29$.
491: 
492: In order to independently estimate the dust radius from radiative
493: transfer modeling and appreciate possible further complexities, we
494: will examine a very simple model. This 
495: model contains small grains of amorphous olivine (i.e. the limit where
496: the grain radius is much smaller than the blackbody emission peak),
497: with optical constants from \cite{Dorschner95}. We assign a dust temperature
498: $T_D$=145\,K and a stellar luminosity of 0.61\,$L_\sun$ as in 
499: \citet{DAlessio05}, and a stellar temperature of an M1.5 star of
500: $T_S=$3600\,K \citep{Luhman99}. The stellar radius for each of the two
501: stars is 3.7\,$R_\sun$, equivalent to a single star of radius
502: 5.2\,$R_\sun$. We assume a blackbody spectrum for
503: the star\footnote{Assuming instead a NextGen model spectrum from
504:   \citet{Hauschildt99} at 3600\,K and log($g$)=4.0 reduces $R_D$ by
505:   2\,AU, but has far too much TiO absorption when compared to an M1.5
506:   spectrum}. The dust is assumed optically thin (which could represent
507: the outer layers of an optically-thick disk), so that the dust radius
508: is then given by:
509: %3600K from Adam's Coma Ber paper.
510: 
511: %This is also in calc_redenning.script
512: \begin{equation}
513:   R_D = R_S \sqrt{\frac{\int \kappa_\lambda B_\lambda(T_S) d\lambda}
514:     {2 \int \kappa_\lambda B_\lambda(T_D) d\lambda} }
515: \end{equation}
516: 
517: In this simple model, the dust radius is $R_D=$15\,AU, roughly consistent with
518: disk truncation radii from dynamical considerations. With the many
519: possible parameters to tweak in a more complete model, and the
520: possibility of the optically-thin dust being additionally heated by
521: surrounding dust, increasing $T_D$ and driving the
522: model radius outwards, it appears likely that a detailed radiative
523: transfer model could indeed be made to match the dynamical truncation
524: radius. The edge in the spectrum of disk
525: emission from CoKu Tau/4 at $\sim$9\,$\mu$m coincides with a sharp
526: rise in silicate emissivity, meaning that the interplay between
527: temperature, extinction, geometry and dust composition is all the more
528: complex for this source. We hope that this preliminary discussion will
529: motivate other authors to examine possible models in more detail.
530: 
531: %This statement is conditional on the orbit having low
532: %eccentricity. For example HD 98800...
533: %OR At first glance, CoKu Tau/4 appears unique amongst circumbinary disks
534: %in that the dust radius is roughly equal to or smaller than that
535: %derived from dynamical modelling. It is certainly....
536: 
537: %Such a ring, 4\,AU in height, contains ~1% of the mass estimated from
538: %mm observations.
539: 
540: % Therefore, model derived radii are too
541: %small when compared to dynamical considerations, and we will examine
542: %several possible ways to reconcile these numbers in turn.
543: 
544: %inally, we should ask: could some other source of energy be
545: %heating the disk? The obvious potential source is viscous heating as
546: %the binary transfers angular momentum to the disk, losing energy. The
547: %total amount of energy available is of order $G M_* M_D / R_D$, which
548: %could potentially be emitted over a timescale as short as $\sim10^5$
549: %years. Given the mass of the disk is estimated at
550: %0.0005\,$M_\sun$\citep{Andrews05}, this luminosity is only
551: %$\sim10^{-4}$L$_\sun$. This is small compared to the luminosity of
552: %the central stars as intercepted by the disk, estimated at
553: %$\sim10^{-1}$L$_\sun$ from the system's geometry
554: %\citep{DAlessio05}. Therefore, we will not consider viscous heating
555: %further here. 
556: %A better way to do this: \Omega^2 \sim GM^3R, and
557: % E \sim G^2 M^5 / \Omega^2
558: %Then we put \DeltaE = dE/d\Omega \Delta \Omega. We regard \Delta
559: %\Omega as the change in angular momentum due to the disk, and get a
560: % factor of (M_D/M_*)^0.5 difference, meaning that the 10^-4 has to be
561: % multiplied by 0.022.
562: %M_* = 1.989e30
563: %M_D = 0.0005 M_*
564: %R_D = 14AU = 2.1e12 m
565: %G = 6.67e-11
566: %L_sun = 3.827e26
567: %T = 3.16e12
568: 
569: %Therefore, we find that the derived disk truncation radius of
570: %$>13$\,AU is only marginally consistent with that derived from
571: %modelling the spectral energy distribution, as it requires CoKu Tau/4
572: %to be a low-eccentricity binary, observed at maximum
573: %apparent separation and with an extinction $A_V>3.0$. 
574: 
575: \section{Discussion and Conclusions}
576: \label{sectConclusion}
577: 
578: The circumbinary nature of the CoKu Tau/4 disk begs the question: are
579: other so-called transition disks also likely circumbinary in nature?
580: The other well-known cold disks in Taurus, 
581: DM~Tau, LkCa~15, UX~Tau and GM~Aur, are also generally assumed to be
582: single stars.  Although all these systems differ from one another,
583: CoKu Tau/4 is perhaps most unique because of its very low accretion
584: rate. Our aperture-masking
585: observations of these systems (paper in preparation) can so far eliminate
586: stellar (mass ratio $q>0.1$) companions over a 20-160\,mas projected separation
587: range, but can say little about the possibility of closer
588: companions. Accurate ($\sigma \la1$\,km\,s$^{-1}$)  radial velocity
589: monitoring of these systems is 
590: required to determine if any of them harbor a close ($\la$4\,AU)
591: companion. Radial velocity monitoring is most applicable
592: to DM~Tau, where the cleared inner hole is expected to have only a
593: 3\,AU radius \citep{Calvet05}.
594: \citet{Simon00} used resolved CO 2-1 observations of disks around
595: several T~Tauri stars to measure their masses directly. These
596: measurements would have resulted in the total mass of the stellar
597: primaries and any unresolved companions. The total masses of
598: $0.84\pm0.05$, $0.97\pm0.03$ and $0.55\pm0.03$ for GM~Aur, LkCa~15 and
599: DM~Tau respectively provide convincing evidence that these systems do
600: not harbor companions with mass ratios $q>0.3$.  
601: 
602: In summary, CoKu Tau/4
603: is a binary star system with a current projected separation of $\sim$7.8\,AU,
604: and a mass ratio near unity. The disk surrounding
605: CoKu Tau/4 is therefore a circumbinary disk, with the disk inner edge
606: set by dynamical truncation. The dynamical disk truncation radius of
607: $\ga13-16$\,AU is inconsistent with the disk radius derived from previous
608: spectral modeling, but it is likely that a different choice of
609: reddening law and correction, dust composition and geometry will
610: resolve this small discrepancy. We suggest 
611: %that most binary systems with separations
612: %similar to CoKu Tau/4 may be free of optically-thick disks, and
613: that systematic radial velocity monitoring and aperture masking
614: surveys are required to determine if other
615: so-called transition disks are in fact multiple systems. 
616: 
617: \acknowledgements
618: 
619: We gratefully acknowledge helpful discussions with Lynne Hillenbrand,
620: Klaus Pontoppidan, Geoffrey Blake, Gregory Herczeg and Colette
621: Salyk, and acknowledge the assistance of Thierry Forveille in pointing
622: out an oversight in an early version of the manuscript. M.I. would like to
623: acknowledge 
624: Michelson Fellowship support from the Michelson Science Center and the
625: NASA Navigator Program. A.L.K. is supported by a NASA/Origins grant to
626: L. Hillenbrand. 
627: The authors wish to recognize and acknowledge
628: the very significant cultural role and reverence that the summit of
629: Mauna Kea has always had within the indigenous Hawaiian community.  We
630: are most fortunate to have the opportunity to conduct observations
631: from this mountain. 
632: 
633: Facilities: \facility{Keck:II (NIRC2)}
634: 
635: \bibliography{../mireland}
636: \bibliographystyle{apj}
637: 
638: \begin{thebibliography}{35}
639: \expandafter\ifx\csname natexlab\endcsname\relax\def\natexlab#1{#1}\fi
640: 
641: \bibitem[{{Akeson} {et~al.}(2007){Akeson}, {Rice}, {Boden}, {Sargent},
642:   {Carpenter}, \& {Bryden}}]{Akeson07}
643: {Akeson}, R.~L., {Rice}, W.~K.~M., {Boden}, A.~F., {Sargent}, A.~I.,
644:   {Carpenter}, J.~M., \& {Bryden}, G. 2007, \apj, 670, 1240
645: 
646: \bibitem[{{Alexander} {et~al.}(2006){Alexander}, {Clarke}, \&
647:   {Pringle}}]{Alexander06}
648: {Alexander}, R.~D., {Clarke}, C.~J., \& {Pringle}, J.~E. 2006, \mnras, 369, 229
649: 
650: \bibitem[{{Artymowicz} \& {Lubow}(1994)}]{Artymowicz94}
651: {Artymowicz}, P., \& {Lubow}, S.~H. 1994, \apj, 421, 651
652: 
653: \bibitem[{{Baraffe} {et~al.}(1998){Baraffe}, {Chabrier}, {Allard}, \&
654:   {Hauschildt}}]{Baraffe98}
655: {Baraffe}, I., {Chabrier}, G., {Allard}, F., \& {Hauschildt}, P.~H. 1998, \aap,
656:   337, 403
657: 
658: \bibitem[{{Bessell}(1990)}]{Bessell90}
659: {Bessell}, M.~S. 1990, \pasp, 102, 1181
660: 
661: \bibitem[{{Bessell} \& {Brett}(1988)}]{Bessell88}
662: {Bessell}, M.~S., \& {Brett}, J.~M. 1988, \pasp, 100, 1134
663: 
664: \bibitem[{{Beust} \& {Dutrey}(2005)}]{Beust05}
665: {Beust}, H., \& {Dutrey}, A. 2005, \aap, 439, 585
666: 
667: \bibitem[{{Brown} {et~al.}(2007){Brown} , {Blake}, {Dullemond}, {Mer{\'{\i}}n},
668:   {Augereau}, {Boogert}, {Evans}, {Geers}, {Lahuis}, {Kessler-Silacci},
669:   {Pontoppidan}, \& {van Dishoeck}}]{Brown07}
670: {Brown}, J.~M. et al., 2007, \apjl, 664, L107
671: %{Blake}, G.~A., {Dullemond}, C.~P., {Mer{\'{\i}}n}, B.,
672: %  {Augereau}, J.~C., {Boogert}, A.~C.~A., {Evans}, II, N.~J., {Geers}, V.~C.,
673: %  {Lahuis}, F., {Kessler-Silacci}, J.~E., {Pontoppidan}, K.~M., \& {van
674: %  Dishoeck}, E.~F. 
675: 
676: \bibitem[{{Calvet} {et~al.}(2005){Calvet}, {D'Alessio}, {Watson},
677:   {Franco-Hern{\'a}ndez}, {Furlan}, {Green}, {Sutter}, {Forrest}, {Hartmann},
678:   {Uchida}, {Keller}, {Sargent}, {Najita}, {Herter}, {Barry}, \&
679:   {Hall}}]{Calvet05}
680: {Calvet}, N. et al., 2005, \apjl, 630, L185
681: %{D'Alessio}, P., {Watson}, D.~M., {Franco-Hern{\'a}ndez}, R.,
682: %  {Furlan}, E., {Green}, J., {Sutter}, P.~M., {Forrest}, W.~J., {Hartmann}, L.,
683: %  {Uchida}, K.~I., {Keller}, L.~D., {Sargent}, B., {Najita}, J., {Herter},
684: %  T.~L., {Barry}, D.~J., \& {Hall}, P. 
685: 
686: \bibitem[{{Cohen} \& {Kuhi}(1979)}]{Cohen79}
687: {Cohen}, M., \& {Kuhi}, L.~V. 1979, \apjs, 41, 743
688: 
689: \bibitem[{{D'Alessio} {et~al.}(2005){D'Alessio}, {Hartmann}, {Calvet},
690:   {Franco-Hern{\'a}ndez}, {Forrest}, {Sargent}, {Furlan}, {Uchida}, {Green},
691:   {Watson}, {Chen}, {Kemper}, {Sloan}, \& {Najita}}]{DAlessio05}
692: {D'Alessio}, P. et al., 2005, \apj, 621, 461
693: % {Hartmann}, L., {Calvet}, N., {Franco-Hern{\'a}ndez}, R.,
694: %  {Forrest}, W.~J., {Sargent}, B., {Furlan}, E., {Uchida}, K., {Green}, J.~D.,
695: %  {Watson}, D.~M., {Chen}, C.~H., {Kemper}, F., {Sloan}, G.~C., \& {Najita}, J.
696:   
697: \bibitem[{{Dorschner} {et~al.}(1995){Dorschner}, {Begemann}, {Henning},
698:   {Jaeger}, \& {Mutschke}}]{Dorschner95}
699: {Dorschner}, J., {Begemann}, B., {Henning}, T., {Jaeger}, C., \& {Mutschke}, H.
700:   1995, \aap, 300, 503
701: 
702: \bibitem[{{Espaillat} {et~al.}(2007){Espaillat}, {Calvet}, {D'Alessio},
703:   {Bergin}, {Hartmann}, {Watson}, {Furlan}, {Najita}, {Forrest}, {McClure},
704:   {Sargent}, {Bohac}, \& {Harrold}}]{Espaillat07}
705: {Espaillat}, C. et al., 2007, \apjl, 664, L111
706:  %{Calvet}, N., {D'Alessio}, P., {Bergin}, E., {Hartmann}, L.,
707:  % {Watson}, D., {Furlan}, E., {Najita}, J., {Forrest}, W., {McClure}, M.,
708:  % {Sargent}, B., {Bohac}, C., \& {Harrold}, S.~T. 
709: 
710: \bibitem[{{Forrest} {et~al.}(2004){Forrest}, {Sargent}, {Furlan}, {D'Alessio},
711:   {Calvet}, {Hartmann}, {Uchida}, {Green}, {Watson}, {Chen}, {Kemper},
712:   {Keller}, {Sloan}, {Herter}, {Brandl}, {Houck}, {Barry}, {Hall}, {Morris},
713:   {Najita}, \& {Myers}}]{Forrest04}
714: {Forrest}, W.~J. et al., 2004, \apjs, 154, 443
715: %{Sargent}, B., {Furlan}, E., {D'Alessio}, P., {Calvet}, N.,
716: %  {Hartmann}, L., {Uchida}, K.~I., {Green}, J.~D., {Watson}, D.~M., {Chen},
717: %  C.~H., {Kemper}, F., {Keller}, L.~D., {Sloan}, G.~C., {Herter}, T.~L.,
718: %  {Brandl}, B.~R., {Houck}, J.~R., {Barry}, D.~J., {Hall}, P., {Morris}, P.~W.,
719: %  {Najita}, J., \& {Myers}, P.~C. 
720: 
721: \bibitem[{{Furlan} {et~al.}(2007){Furlan}, {Sargent}, {Calvet}, {Forrest},
722:   {D'Alessio}, {Hartmann}, {Watson}, {Green}, {Najita}, \& {Chen}}]{Furlan07}
723: {Furlan}, E. et al., 2007, \apj, 664, 1176
724: %, {Sargent}, B., {Calvet}, N., {Forrest}, W.~J., {D'Alessio}, P.,
725: %  {Hartmann}, L., {Watson}, D.~M., {Green}, J.~D., {Najita}, J., \& {Chen},
726: %  C.~H. 
727: 
728: \bibitem[{{Guenther} {et~al.}(2007){Guenther}, {Esposito}, {Mundt}, {Covino},
729:   {Alcal{\'a}}, {Cusano}, \& {Stecklum}}]{Guenther07}
730: {Guenther}, E.~W., {Esposito}, M., {Mundt}, R., {Covino}, E., {Alcal{\'a}},
731:   J.~M., {Cusano}, F., \& {Stecklum}, B. 2007, \aap, 467, 1147
732: 
733: \bibitem[{{Hanson} {et~al.}(2004){Hanson}, {Klemola}, {Jones}, \&
734:   {Monet}}]{Hanson04}
735: {Hanson}, R.~B., {Klemola}, A.~R., {Jones}, B.~F., \& {Monet}, D.~G. 2004, \aj,
736:   128, 1430
737: 
738: \bibitem[{{Hartmann} {et~al.}(2005){Hartmann}, {Megeath}, {Allen}, {Luhman},
739:   {Calvet}, {D'Alessio}, {Franco-Hernandez}, \& {Fazio}}]{Hartmann05}
740: {Hartmann}, L., {Megeath}, S.~T., {Allen}, L., {Luhman}, K., {Calvet}, N.,
741:   {D'Alessio}, P., {Franco-Hernandez}, R., \& {Fazio}, G. 2005, \apj, 629, 881
742: 
743: \bibitem[{{Hauschildt} {et~al.}(1999){Hauschildt}, {Allard}, \&
744:   {Baron}}]{Hauschildt99}
745: {Hauschildt}, P.~H., {Allard}, F., \& {Baron}, E. 1999, \apj, 512, 377
746: 
747: \bibitem[{{Jensen} \& {Mathieu}(1997)}]{Jensen97}
748: {Jensen}, E.~L.~N., \& {Mathieu}, R.~D. 1997, \aj, 114, 301
749: 
750: \bibitem[{{Kraus} {et~al.}(2008){Kraus}, {Ireland}, {Martinache}, \&
751:   {Lloyd}}]{Kraus08a}
752: {Kraus}, A., {Ireland}, M., {Martinache}, F., \& {Lloyd}, J. 2008, \apj,
753:   Submitted
754: 
755: \bibitem[{{Kraus} \& {Hillenbrand}(2007{\natexlab{a}})}]{Kraus07a}
756: {Kraus}, A.~L., \& {Hillenbrand}, L.~A. 2007{\natexlab{a}}, \apj, 662, 413
757: 
758: \bibitem[{{Kraus} \& {Hillenbrand}(2007{\natexlab{b}})}]{Kraus07b}
759: ---. 2007{\natexlab{b}}, \apj, 664, 1167
760: 
761: \bibitem[{{Lin} \& {Papaloizou}(1979)}]{Lin79}
762: {Lin}, D.~N.~C., \& {Papaloizou}, J. 1979, \mnras, 186, 799
763: 
764: \bibitem[{{Lloyd} {et~al.}(2006){Lloyd}, {Martinache}, {Ireland}, {Monnier},
765:   {Pravdo}, {Shaklan}, \& {Tuthill}}]{Lloyd06}
766: {Lloyd}, J.~P., {Martinache}, F., {Ireland}, M.~J., {Monnier}, J.~D., {Pravdo},
767:   S.~H., {Shaklan}, S.~B., \& {Tuthill}, P.~G. 2006, \apj, 650, L131
768: 
769: \bibitem[{{Lohmann} {et~al.}(1983){Lohmann}, {Weigelt}, \&
770:   {Wirnitzer}}]{Lohmann83}
771: {Lohmann}, A.~W., {Weigelt}, G., \& {Wirnitzer}, B. 1983, \ao, 22, 4028
772: 
773: \bibitem[{{Luhman}(1999)}]{Luhman99}
774: {Luhman}, K.~L. 1999, \apj, 525, 466
775: 
776: \bibitem[{{Luhman}(2004)}]{Luhman04}
777: ---. 2004, \apj, 617, 1216
778: 
779: \bibitem[{{Quillen} {et~al.}(2004){Quillen}, {Blackman}, {Frank}, \&
780:   {Varni{\`e}re}}]{Quillen04}
781: {Quillen}, A.~C., {Blackman}, E.~G., {Frank}, A., \& {Varni{\`e}re}, P. 2004,
782:   \apjl, 612, L137
783: 
784: \bibitem[{{Simon} {et~al.}(2000){Simon}, {Dutrey}, \& {Guilloteau}}]{Simon00}
785: {Simon}, M., {Dutrey}, A., \& {Guilloteau}, S. 2000, \apj, 545, 1034
786: 
787: \bibitem[{{Skrutskie} {et~al.}(1990){Skrutskie}, {Dutkevitch}, {Strom},
788:   {Edwards}, {Strom}, \& {Shure}}]{Skrutskie90}
789: {Skrutskie}, M.~F., {Dutkevitch}, D., {Strom}, S.~E., {Edwards}, S., {Strom},
790:   K.~M., \& {Shure}, M.~A. 1990, \aj, 99, 1187
791: 
792: \bibitem[{{Torres} {et~al.}(2007){Torres}, {Loinard}, {Mioduszewski}, \&
793:   {Rodr{\'{\i}}guez}}]{Torres07}
794: {Torres}, R.~M., {Loinard}, L., {Mioduszewski}, A.~J., \& {Rodr{\'{\i}}guez},
795:   L.~F. 2007, \apj, 671, 1813
796: 
797: \bibitem[{{Tuthill} {et~al.}(2000){Tuthill}, {Monnier}, {Danchi}, {Wishnow}, \&
798:   {Haniff}}]{Tuthill00}
799: {Tuthill}, P.~G., {Monnier}, J.~D., {Danchi}, W.~C., {Wishnow}, E.~H., \&
800:   {Haniff}, C.~A. 2000, \pasp, 112, 555
801: 
802: \bibitem[{White} \& {Hillenbrand}(2005)]{White05}
803: {White}, R.~J., \& Hillenbrand, L.~A., 2005, \apj, 621, L65
804: 
805: \bibitem[{{Whittet} {et~al.}(2001){Whittet}, {Gerakines}, {Hough}, \&
806:   {Shenoy}}]{Whittet01}
807: {Whittet}, D.~C.~B., {Gerakines}, P.~A., {Hough}, J.~H., \& {Shenoy}, S.~S.
808:   2001, \apj, 547, 872
809: 
810: \end{thebibliography}
811: 
812: \end{document}
813: