0710.2127/ms.tex
1: %\documentclass[12pt,preprint]{aastex}
2: %\documentclass[10pt, letterpaper, onecolumn]{emulateapj}
3: \documentclass[10pt, letterpaper]{emulateapj}
4:  
5: %\usepackage{graphicx}
6: %\usepackage{epstopdf} 
7:  
8: \begin{document}
9:  
10: \title{The orbits of the quadruple star system 88 Tau A from PHASES differential astrometry and radial velocity}
11: 
12: \author{Benjamin F.~Lane\altaffilmark{1}, Matthew W.~Muterspaugh\altaffilmark{2,3}, 
13: Francis C. Fekel\altaffilmark{4}, Michael Williamson\altaffilmark{4}, Stanley Browne\altaffilmark{2}, Maciej Konacki\altaffilmark{5}, 
14: Bernard F.~Burke\altaffilmark{1}, M.~M.~Colavita\altaffilmark{6}, S.~R.~Kulkarni\altaffilmark{7}, M.~Shao\altaffilmark{6}
15: }
16: \altaffiltext{1}{MIT Kavli Institute for Astrophysics and Space Research, MIT Department of Physics, 70 Vassar Street, Cambridge, MA 02139}
17: \altaffiltext{2}{University of California, Space Sciences Laboratory, 7 Gauss Way, Berkeley, CA 94720-7450}
18: \altaffiltext{3}{Townes Fellow}
19: \altaffiltext{4}{Center of Excellence in Information Systems, Tennessee State University, 3500 John A. Merritt Blvd, Box 9501, Nashville, TN 37209}
20: \altaffiltext{5}{Nicolaus Copernicus Astronomical Center, Polish Academy of Sciences, Rabianska 8, 87-100 Torun, Poland}
21: \altaffiltext{6}{Jet Propulsion Laboratory, California Institute of Technology, 4800 Oak Grove Dr., Pasadena, CA 91109}
22: \altaffiltext{7}{Division of Physics, Mathematics and Astronomy, 105-24, California Institute of Technology, Pasadena, CA 91125}
23: 
24: \begin{abstract} 
25: We have used high precision differential astrometry from the Palomar High-precision Astrometric Search for Exoplanet Systems (PHASES)
26: project and radial velocity measurements covering a time-span of 20 years to determine the 
27: orbital parameters of the 88 Tau A system. 88 Tau is a complex hierarchical multiple system comprising a total of six stars; we have studied
28: the brightest 4, consisting of two short-period pairs orbiting each other with an $\sim$18-year period.  We present 
29: the first orbital solution for one of the short-period pairs, and determine the masses of the components and 
30: distance to the system to the level of a few percent. In addition, our astrometric measurements allow us to make the 
31: first determination of  the mutual inclinations of the orbits. We find that the sub-systems are not coplanar.
32: \end{abstract}
33: 
34: \keywords{techniques:interferometric--star:88 Tau}
35:  
36: \section{Introduction}
37: 
38: 88 Tau (HD 29140, HR 1458, HIP 21402) is a bright 
39: ($m_V=4.25, m_K=3.69\pm 0.25$; Skrutskie et al. 2006\nocite{2mass}), nearby
40: ($\sim 50$pc) hierarchical sextuple stellar system \citep{msc}. The A
41: component contains a pair of systems (designated Aa
42: and Ab) in an $\sim18$-year \citep{balega99}
43:  orbit that has been resolved by speckle interferometry \citep{mac87}.
44:  The Aa component is a known spectroscopic binary system 
45:  ($P\sim 3.57$-day), with a composite
46: spectral type of A5m \citep{c69}. In previous work it had been noted \citep{bc88} 
47: that the A system is likely complex, with possibly as many as 5 components. 
48: \citet{balega99} noted a discrepancy between the total estimated 
49: mass of this system based on photometry and spectral types, and 
50: the total mass derived from the visual orbit and {\em Hipparcos}
51: parallax.  In this work we have determined that, like the Aa component, the Ab 
52: component is a double-lined binary; this newly-resolved binary has a period of 7.89 
53: days. Finally, there is a common-proper-motion companion, labeled B, located 
54: $\sim$69 arcseconds away from the A system;  it, too, is
55: known to be a binary \citep{tg01}.  For clarity we provide a 
56: schematic of this complex system in Figure \ref{fig1}.
57: 
58: %\clearpage
59: \begin{figure}
60: \figurenum{1}
61: %%\includegraphics[scale=0.5,angle=90 ]{f1.eps}
62: %\epsscale{0.25}
63: \plotone{f1.eps}
64: \figcaption{A schematic diagram of the 88 Tau system. 
65: \label{fig1}}
66: 
67: \end{figure}
68: %\clearpage
69: There are several reasons why multiple stellar systems such as 88 Tau
70: merit attention: first, binary orbits make it possible to measure accurate
71: stellar masses and distances, while the larger number of
72: presumably co-eval stars allows one to impose the additional
73: constraint that any given model must accurately match all of the
74: stars. This approach has proven particularly fruitful when applied 
75: to another famous hierarchical sextuple system: Castor\citep{tr02}.
76: Second, as outlined in \citet{st02}, the relative orientations
77: of the orbital angular momenta allow one to constrain the properties
78: of the cloud from which the stars are thought to have formed, as well as the
79: subsequent dynamical decay process. Despite their value, observational
80: problems have limited the number of triple or higher-order systems
81: with accurately measured orbits to fewer than 10. Given their hierarchical nature it is
82: often the case that either the close system is unresolvable or the
83: outer system has an impractically long orbital period. 
84: 
85: With the advent of long-baseline stellar interferometry, and more recently 
86: phase-referenced long-baseline interferometric astrometry \citep{lm04}
87: capable of 10--20 $\mu$-arcsecond astrometric precision between 
88: pairs of stars with separations in the range 0.05--1 arcsecond, it
89: has become possible to resolve the orbital motion of several 
90: interesting multiple systems \citep{kapPeg,v819Her}. Here we 
91: report on astrometric and radial velocity measurements of the 88 Tau A
92: system, which allow us to constrain the orbits of the 3.57-day, 7.89-day and 
93: 18-year components with improved precision, and for the first time 
94: provide a relative orientation of the orbits as well as component masses. 
95: 
96: Astrometric measurements were made with the Palomar Testbed
97: Interferometer \citep{colavita99} as part of the Palomar
98: High-precision Astrometric Search for Exoplanet Systems (PHASES)
99: program \citep{limits}.  The Palomar Testbed Interferometer is located
100: on Palomar Mountain near San Diego, CA. It was
101: developed by the Jet Propulsion Laboratory, California Institute of
102: Technology for NASA as a testbed for interferometric techniques
103: applicable to the Keck Interferometer and the
104: Space Interferometry Mission (SIM). It operates in the J (1.2 $\mu$m), H
105: (1.6 $\mu$m), and K (2.2 $\mu$m) bands and combines starlight from two out of
106: three available 40 cm apertures. The apertures form a triangle with 86
107: and 110 m baselines.
108: 
109: 
110: \section{Observations \& Models}
111: 
112: \subsection{PHASES Astrometry}
113: 88 Tau A was successfuly observed with PTI on 29 nights in 2004--2007
114: with the use of the phase-referenced fringe-scanning mode \citep{lm04} developed
115: for high-precision astrometry; the data were reduced with the
116: algorithms described therein, as well as with the modifications
117: described in \citet{delEqu}. 
118: 
119: The obtained differential astrometry is listed in Table
120: \ref{tab:data}.  Note that the astrometry on any single night is
121: essentially that of a single-baseline interferometer, yielding a very
122: small error in the direction aligned with the baseline, but limited to
123: the effect of Earth-rotation synthesis in the perpendicular direction.
124: The median minor-axis formal uncertainty is 10 $\mu$arcseconds, while
125: the median major-axis uncertainty is 312 $\mu$arcseconds.  To
126: properly weight the data set when doing a combined fit with previous
127: astrometry and radial velocity data, we fit an orbital model to the
128: PHASES astrometry by itself, and rescaled the formal uncertainties so
129: as to yield a reduced $\chi^2$ of unity; the resulting scale factor
130: was 2.5, indicating a substantial amount of excess scatter beyond the
131: internal error estimates. We do not believe this scatter to be due to
132: the effect of starspots, given that the {\it Hipparcos} photometry
133: of this system indicates a scatter of no more than 5 mmag; the resulting maximum
134: starspot-induced astrometric noise would be $\sim 4 \mu$arcseconds
135: \citep{v819Her}.  We have however identified possible instrumental
136: sources of this systematic error and developed methods for reducing
137: it, see \cite{muOri}. Nevertheless, the existing astrometry is
138: sufficient to detect astrometric motion induced by the short-period subsystems. 
139: 
140: %\clearpage
141: \begin{deluxetable*}{ccccccccccccl}
142: %%\label{tab:data}
143: %\rotate
144: \tabletypesize{\footnotesize}
145: \tablecaption{PHASES Astrometric data for 88 Tau A\label{tab:data}}
146: \tablewidth{0pt}
147: \tablehead{
148: HJD-2400000.5 &  $\Delta$RA$\cos(\delta)$ & $\Delta$Dec & $\sigma_{min}$ & $\sigma_{maj}$ & $\phi_e$ & $\sigma_{RA}$ & $\sigma_{Dec}$ & $\frac{\sigma^2_{RA,Dec}}{\sigma_{RA}\sigma_{Dec}} $ & N & ADC & Align & Rate \\
149:  & (mas) & (mas) & ($\mu$as) & ($\mu$as) & (deg) & ($\mu$as) & ($\mu$as) & & & & &(Hz)
150: }
151: \startdata 
152: 52979.34108 & -32.1751 & -101.5913 & 20.6 & 477.9 & 163.71 & 458.8 & 135.5 & -0.98733 & 2353  & 0 & 0 &100  \\
153: 53034.13365 & -40.3179 & -91.5425 & 14.0 & 311.9 & 152.78 & 277.4 & 143.2 & -0.99394 & 2951  & 0 & 0 &100 \\
154: 53250.50169 & -65.6825 & -53.5438 & 40.2 & 1775.4 & 147.24 & 1493.2 & 961.3 & -0.99876 & 1027 & 0 & 0 &100  \\
155: 53271.46723 & -70.0370 & -48.2688 & 16.6 & 669.7 & 150.15 & 580.9 & 333.7 & -0.99836 & 4005 & 0 & 0 &100\\
156: 53291.40191 & -71.5816 & -44.8765 & 88.3 & 1759.4 & 148.37 & 1498.7 & 925.8 & -0.99371 & 408  & 0 & 0 &100\\
157: 53294.47661 & -71.0249 & -44.6002 & 32.3 & 1700.7 & 164.09 & 1635.6 & 467.2 & -0.99742 & 1355   & 0 & 0 &100\\
158: 53312.38113 & -74.1285 & -40.9475 & 7.8 & 68.0 & 154.15 & 61.3 & 30.5 & -0.95879 & 8150 & 0 & 0 &100\\
159: 53320.33408 & -74.5082 & -39.6052 & 40.3 & 2076.1 & 150.23 & 1802.1 & 1031.5 & -0.99899 & 1838  & 0 & 0 &100 \\
160: 53340.29304 & -77.4848 & -35.5088 & 18.2 & 215.1 & 152.98 & 191.8 & 99.1 & -0.97862 & 3602   & 0 & 0 &100\\
161: 53341.28228 & -77.5104 & -35.2746 & 17.2 & 551.5 & 150.70 & 481.0 & 270.3 & -0.99734 & 3630   & 0 & 0 &100\\
162: 53605.52764 & -104.7426 & 15.6166 & 18.3 & 956.5 & 147.64 & 808.1 & 512.1 & -0.99911 & 2855   & 0 & 0 &100\\
163: 53606.51559 & -106.5080 & 16.8975 & 15.7 & 972.4 & 146.14 & 807.5 & 542.0 & -0.99939 & 3205   & 0 & 0 &100\\
164: 53614.50698 & -105.6799 & 17.1892 & 25.4 & 773.7 & 147.94 & 655.8 & 411.2 & -0.99735 & 1716   & 0 & 0 &100\\
165: 53687.37864 & -112.2999 & 31.2082 & 24.7 & 227.4 & 31.80 & 193.7 & 121.7 & 0.97114 & 3260  & 0 & 0 &100 \\
166: 53711.29715 & -114.6779 & 35.8970 & 80.7 & 880.5 & 29.41 & 768.1 & 438.0 & 0.97743 & 833   & 1 & 0 &100\\
167: 53712.28437 & -114.0278 & 36.0764 & 41.2 & 326.1 & 27.87 & 288.9 & 156.8 & 0.95491 & 2889   & 1 & 0 &100\\
168: 53789.14419 & -120.0382 & 51.3801 & 59.4 & 3107.9 & 40.17 & 2375.3 & 2005.2 & 0.99925 & 761  & 1 & 0 &100\\
169: 53790.13775 & -119.0373 & 52.0709 & 77.5 & 2663.3 & 39.39 & 2058.8 & 1691.3 & 0.99824 & 624  & 1 & 0 & 100 \\
170: 54030.46119 & -135.6408 & 94.2582 & 50.4 & 1514.6 & 163.75 & 1454.1 & 426.6 & -0.99240 & 1268  & 1 & 1 & 50 \\
171: 54055.38631 & -136.2430 & 98.5452 & 48.0 & 870.1 & 161.74 & 826.4 & 276.5 & -0.98318 & 1679   & 1 & 1 & 50\\
172: 54061.37247 & -136.6968 & 99.4972 & 16.3 & 302.1 & 163.08 & 289.0 & 89.3 & -0.98165 & 4216& 1 & 1 &50\\
173: 54075.33188 & -136.8574 & 101.6204 & 13.6 & 250.1 & 162.18 & 238.2 & 77.7 & -0.98300 & 5902 & 1 & 1 &50\\
174: 54083.30564 & -137.8524 & 103.1506 & 31.6 & 688.5 & 161.39 & 652.6 & 221.8 & -0.98863 & 3845  & 1 & 1 &50\\
175: 54084.31595 & -136.7985 & 103.1084 & 63.9 & 1741.5 & 163.11 & 1666.5 & 509.6 & -0.99139 & 1182  & 1 & 1 &50 \\
176: 54103.25456 & -137.6742 & 106.3338 & 42.7 & 1011.9 & 162.34 & 964.3 & 309.6 & -0.98949 & 743   & 1 & 1 &50\\
177: 54138.15348 & -139.3256 & 112.2494 & 28.6 & 457.9 & 161.38 & 434.1 & 148.7 & -0.97916 & 3142  & 1 & 1 &50 \\
178: \enddata
179: \tablecomments{All quantities are in the ICRS 2000.0 reference frame. The uncertainty values 
180: presented in these data have been scaled by a factor of 2.5 over the formal (internal) uncertainties 
181: for each night.  Column 6 ($\phi_e$) is the angle between the major axis of the uncertainty 
182: ellipse and the right ascension axis, measured from increasing differential right ascension 
183: through increasing differential declination. N is the number of scans obtained in a night; each scan 
184: typically represents 0.5--1 second of integration. ADC indicates that the observations made use of the 
185: automatic dispersion compensator. Rate indicates the tracking rate of the fringe tracker used to stabilize the
186: fringe phase during measurement. Align indicates whether or not the automatic alignment 
187: system was used to stabilize the system pupil.  }
188: \end{deluxetable*} 
189: %\clearpage
190: 
191: \subsection{Previous Astrometry}
192: 
193: In addition to our astrometry, 88 Tau A has been followed by a number of
194: observers with speckle-interferometric techniques. We use 20
195: observations tabulated in the {\em 4th Catalog of 
196: Interferometric Measurements of Binary Stars\footnote{\tt http://ad.usno.navy.mil/wds/int4.html} } \citep{int4}
197: to further constrain our fit. Although of somewhat lower precision, the considerable
198: time-baseline (including observations dating from 1985) help constrain
199: the parameters of the wide orbit. In many cases the published
200: astrometry lacks uncertainties, and we therefore assigned a 
201: plausible initial uncertainty of 3 mas in separation and 2 degrees in 
202: position angle to these points (we used published uncertainties for the 
203: points where such were available). We then performed a least-squares fit of a single Keplerian
204: orbital model (corresponding to the Aa-Ab orbit - the subsystems are far too small
205: to be detected by these data), and scaled all of the uncertainties
206: so as to yield a reduced $\chi^2$ of unity.  We find the average
207: uncertainty in separation to be 5 milli-arcseconds, and the average
208: position-angle uncertainty to be 3 degrees.
209: 
210: \subsection{Spectroscopic observations and reductions}
211: 
212: %\clearpage
213: \begin{deluxetable*}{lcccccccc}
214: %%\label{tab:dataknpo}
215: \tabletypesize{\scriptsize}
216: \tablecaption{KPNO Radial Velocity data for 88 Tau A\label{tab:dataknpo}}
217: \tablewidth{0pt}
218: \tablehead{
219: HJD-2400000.5 &  $V_{Aa1}$& Weight\tablenotemark{a} & $V_{Aa2}$ & Weight\tablenotemark{b} & $V_{Ab1}$ & Weight\tablenotemark{c} & $V_{Ab2}$ & Weight\tablenotemark{d} \\
220:                               &   (${\rm km\,s^{-1}}$) & & (${\rm km\,s^{-1}}$) & & (${\rm km\,s^{-1}}$) & & (${\rm km \,s^{-1}}$) & 
221: }
222: \startdata 
223:  45718.327 & 91.4 & 1.0  & -66.5 & 1.0 & - & - & - & - \\
224:  46388.389 & -48.7 & 1.0  & 141.6 & 1.0 & 40.1 & 1.0 & 1.1 & 1.0 \\
225:  46390.212 & 99.8 & 1.0  & -85.2 & 1.0 & -5.8 & 1.0 & - & - \\
226:  46718.473 & 96.5 & 1.0  & -93.7 & 1.0 & 59.8 & 0.4 & -4.5 & 1.0 \\
227:  46720.487 & -51.6 & 1.0  & 139.5 & 1.0 & - & - & - & - \\
228:  47152.190 & -45.3 & 1.0  & 123.6 & 1.0 & 57.9 & 1.0 & - & - \\
229:  47245.112 & -50.2 & 1.0  & 131.5 & 1.0 & - & - & - & - \\
230:  47456.264 & -52.2 & 1.0  & 129.7 & 1.0 & 9.3 & 1.0 & - & - \\
231:  47556.098 & -57.7 & 1.0  & 135.8 & 1.0 & - & - & - & - \\
232:  47624.100 & -53.5 & 1.0  & 125.7 & 1.0 & - & - & - & - \\
233:  47626.145 & 66.5 & 1.0  & -57.7 & 1.0 & - & - & 4.3 & 1.0 \\
234:  47627.123 & -49.3 & 1.0  & 118.9 & 1.0 & - & - & - & - \\
235:  48345.123 & -60.5 & 1.0  & 133.2 & 1.0 & - & - & - & - \\
236:  48347.103 & 96.5 & 1.0  & -102.4 & 1.0 & 9.4 & 1.0 & 69.9 & 0.4 \\
237:  48356.100 & -58.6 & 1.0  & 128.6 & 1.0 & 10.0 & 1.0 & 59.1 & 1.0 \\
238:  48505.518 & -46.5 & 1.0  & 106.1 & 1.0 & 14.0 & 1.0 & 53.0 & 1.0 \\
239:  48573.407 & -49.1 & 1.0  & 112.7 & 1.0 & 42.5 & 0.4 & 26.2 & 0.4 \\
240:  48604.284 & 90.2 & 1.0  & -98.1 & 1.0 & 57.7 & 0.4 & 15.4 & 1.0 \\
241:  48607.306 & 75.8 & 1.0  & -74.8 & 1.0 & 3.0 & 1.0 & - & - \\
242:  48913.412 & -43.0 & 1.0  & 105.6 & 1.0 & 19.0 & 1.0 & 50.9 & 1.0 \\
243:  48916.375 & -55.5 & 1.0  & 125.2 & 1.0 & - & - & - & - \\
244:  49246.472 & 76.6 & 1.0  & -67.1 & 1.0 & 6.6 & 0.4 & - & - \\
245:  49248.525 & -56.1 & 1.0  & 131.7 & 1.0 & 44.0 & 1.0 & 15.3 & 1.0 \\
246:  49250.537 & 101.6 & 1.0  & -93.6 & 1.0 & 58.9 & 0.4 & 1.7 & 1.0 \\
247:  49302.418 & -53.7 & 1.0  & 132.4 & 1.0 & 12.9 & 1.0 & 51.9 & 1.0 \\
248:  49307.374 & 91.3 & 1.0  & -87.2 & 1.0 & 23.0 & 0.4 & 34.6 & 0.4 \\
249:  49618.436 & 94.0 & 1.0  & -92.1 & 1.0 & - & - & - & - \\
250:  49622.425 & 60.6 & 1.0  & -41.1 & 1.0 & - & - & - & - \\
251:  49677.381 & -50.0 & 1.0  & 132.6 & 1.0 & - & - & - & - \\
252:  49971.535 & 93.8 & 1.0  & -78.7 & 1.0 & -9.8 & 1.0 & 53.5 & 0.4 \\
253:  49973.434 & -48.2 & 1.0  & 138.8 & 1.0 & - & - & - & - \\
254:  49973.535 & -51.1 & 1.0  & 142.9 & 1.0 & 17.0 & 0.4 & 26.0 & 0.4 \\
255:  50364.509 & 102.5 & 1.0  & -85.3 & 1.0 & 5.8 & 0.4 & 30.0 & 0.4 \\
256:  50366.371 & -48.5 & 1.0  & 146.0 & 1.0 & - & - & - & - \\
257:  50400.392 & 100.7 & 1.0  & -88.4 & 1.0 & 38.2 & 1.0 & -1.3 & 1.0 \\
258:  50404.270 & 91.1 & 1.0  & -66.5 & 1.0 & - & - & - & - \\
259:  50721.423 & 93.8 & 1.0  & -67.1 & 1.0 & -9.9 & 1.0 & 38.0 & 0.4 \\
260:  50721.501 & 96.1 & 1.0  & -77.1 & 1.0 & -11.3 & 1.0 & 40.2 & 1.0 \\
261:  50755.460 & -46.1 & 1.0  & 139.7 & 1.0 & 38.8 & 1.0 & -11.3 & 0.4 \\
262:  50757.384 & 104.0 & 1.0  & -86.7 & 1.0 & 34.0 & 1.0 & -2.7 & 1.0 \\
263:  50832.293 & 104.7 & 1.0  & -82.2 & 1.0 & -9.7 & 1.0 & 44.2 & 1.0 \\
264:  50833.156 & 57.7 & 1.0  & - & - & - & - & - & - \\
265:  51088.384 & -7.6 & 1.0  & 92.2 & 1.0 & - & - & - & - \\
266:  51089.481 & 106.0 & 1.0  & -85.2 & 1.0 & - & - & - & - \\
267:  51091.427 & -49.1 & 1.0  & 149.9 & 1.0 & -18.8 & 0.4 & 47.3 & 1.0 \\
268:  51093.343 & 103.8 & 1.0  & -79.4 & 1.0 & 3.6 & 0.4 & 22.0 & 0.4 \\
269:  51473.377 & -51.8 & 1.0  & 147.2 & 1.0 & 43.5 & 1.0 & -18.3 & 0.4 \\
270:  51475.281 & 107.5 & 1.0  & -85.7 & 1.0 & - & - & - & - \\
271:  51475.362 & 107.8 & 1.0  & -85.7 & 1.0 & 22.1 & 0.4 & 4.1 & 0.4 \\
272:  51803.489 & 95.8 & 1.0  & -66.7 & 1.0 & - & - & - & - \\
273:  51805.463 & -49.1 & 1.0  & 147.2 & 1.0 & 41.6 & 1.0 & -21.2 & 0.4 \\
274:  51807.463 & 107.0 & 1.0  & -86.9 & 1.0 & 5.1 & 0.4 & 22.7 & 0.4 \\
275:  52016.110 & -46.8 & 1.0  & 142.6 & 1.0 & - & - & - & - \\
276:  52180.462 & -49.0 & 1.0  & 146.2 & 1.0 & -14.7 & 1.0 & 43.8 & 1.0 \\
277:  52182.481 & 104.0 & 1.0  & -86.2 & 1.0 & 28.2 & 0.4 & 0.4 & 0.4 \\
278:  52327.104 & -48.8 & 1.0  & 145.5 & 1.0 & 35.9 & 1.0 & -0.9 & 0.4 \\
279:  52329.148 & 89.8 & 1.0  & -64.0 & 1.0 & -8.0 & 0.4 & 38.2 & 0.4 \\
280:  52537.494 & -50.0 & 1.0  & 137.7 & 1.0 & 35.0 & 1.0 & -4.0 & 0.4 \\
281:  52539.468 & 106.0 & 1.0  & -89.1 & 1.0 & 42.9 & 1.0 & -7.0 & 1.0 \\
282:  52541.378 & -46.3 & 1.0  & 144.8 & 1.0 & - & - & - & - \\
283:  52541.473 & -45.5 & 1.0  & 139.3 & 1.0 & 1.5 & 0.4 & 38.9 & 1.0 \\
284:  52705.198 & -45.8 & 1.0  & 126.6 & 1.0 & 48.5 & 1.0 & - & - \\
285:  52707.190 & 101.6 & 1.0  & -88.8 & 1.0 & - & - & - & - \\
286:  52709.127 & -54.6 & 1.0  & 144.6 & 1.0 & - & - & 51.1 & 1.0 \\
287:  52903.438 & 93.6 & 1.0  & -75.7 & 1.0 & - & - & - & - \\
288:  52941.345 & -50.8 & 1.0  & 141.1 & 1.0 & 51.1 & 1.0 & -6.8 & 0.4 \\
289:  53273.479 & -53.1 & 1.0  & 139.3 & 1.0 & 41.0 & 1.0 & 7.9 & 1.0 \\
290:  53278.490 & 93.6 & 1.0  & -84.9 & 1.0 & - & - & - & - \\
291:  53637.454 & -55.3 & 1.0  & 134.5 & 1.0 & 12.9 & 1.0 & 42.4 & 1.0 \\
292:  54001.454 & -42.5 & 1.0  & 113.7 & 1.0 & 3.6 & 0.4 & 66.6 & 1.0 \\
293:  54003.423 & 89.9 & 1.0  & -90.3 & 1.0 & - & - & - & - \\
294:  54005.351 & -58.5 & 1.0  & 137.5 & 1.0 & 62.1 & 1.0 & - & - \\
295:  54005.506 & -57.7 & 1.0  & 134.6 & 1.0 & 60.4 & 1.0 & 0.9 & 0.4 \\
296: \enddata
297: \tablenotetext{a}{An observation of Aa1 of unit weight has a standard error of 2.0 ${\rm km\,s^{-1}}.$}
298: \tablenotetext{b}{An observation of Aa2 of unit weight has a standard error of 2.3 ${\rm km\,s^{-1}}.$}
299: \tablenotetext{c}{An observation of Ab1 of unit weight has a standard error of 2.6 ${\rm km\,s^{-1}}.$}
300: \tablenotetext{d}{An observation of Ab2 of unit weight has a standard error of 2.6 ${\rm km\,s^{-1}}.$}
301: \end{deluxetable*} 
302: %\clearpage
303: 
304: From 1984 January through 2006 September we obtained 82 spectrograms
305: of 88~Tau with the Kitt Peak National Observatory (KPNO) 0.9 m coud\'e 
306: feed telescope, coud\'e spectrograph, and a TI~CCD detector.  Sixty-eight 
307: spectrograms are centered in the red at 6430~\AA, cover a wavelength
308: range of about 80~\AA, and have a two pixel resolution of 0.21~\AA.  
309: Those spectra have signal-to-noise ratios of $\sim$250.  The remaining
310: 14 spectrograms are centered in the blue at 4500~\AA, cover a 
311: wavelength range of 85~\AA, and have a resolution of 0.22~\AA. 
312: Signal-to-noise ratios of $\sim$300 are typical.  
313: 
314: %\clearpage
315: \begin{deluxetable*}{lcccccccc}
316: %%\label{tab:datafairborn}
317: \tablecaption{Fairborn Observatory Radial Velocity data for 88 Tau A\label{tab:datafairborn}}
318: \tablewidth{0pt}
319: \tablehead{
320: HJD-2400000.5 &  $V_{Aa1}$& Weight\tablenotemark{a} & $V_{Aa2}$ & Weight\tablenotemark{b} & $V_{Ab1}$ & Weight\tablenotemark{c} & $V_{Ab2}$ & Weight\tablenotemark{d} \\
321:                               &   (${\rm km\,s^{-1}}$) & & (${\rm km\,s^{-1}}$) & & (${\rm km\,s^{-1}}$) & & (${\rm km \,s^{-1}}$) & 
322: }
323: \startdata 
324: 53020.204 & -38.5 & 1.0  & 118.0 & 1.0 & 50.8 & 1.0 & - & - \\
325:  53032.197 & 101.1 & 1.0  & -90.3 & 1.0 & -6.3 & 1.0 & 53.7 & 1.0 \\
326:  53052.273 & -43.0 & 1.0  & 126.0 & 1.0 & 50.9 & 1.0 & - & - \\
327:  53276.489 & -37.3 & 1.0  & 110.0 & 1.0 & - & - & 53.0 & 1.0 \\
328:  53285.468 & 85.9 & 1.0  & -69.9 & 1.0 & 10.8 & 1.0 & 41.8 & 1.0 \\
329:  53314.392 & 102.0 & 1.0  & -95.0 & 1.0 & 8.8 & 1.0 & 40.6 & 1.0 \\
330:  53350.390 & 93.6 & 1.0  & -85.3 & 1.0 & 46.3 & 1.0 & 7.7 & 1.0 \\
331:  53395.275 & -28.2 & 1.0  & 97.3 & 1.0 & -2.4 & 1.0 & 58.0 & 1.0 \\
332:  53405.184 & -44.4 & 1.0  & 127.0 & 1.0 & 39.9 & 1.0 & 11.9 & 1.0 \\
333:  53630.518 & -55.6 & 1.0  & 139.5 & 1.0 & 2.7 & 1.0 & 57.9 & 1.0 \\
334:  53644.477 & -48.0 & 1.0  & 127.2 & 1.0 & 38.3 & 1.0 & 20.1 & 1.0 \\
335:  53659.486 & -33.5 & 1.0  & 102.9 & 1.0 & 58.4 & 1.0 & - & - \\
336:  53700.387 & 90.5 & 1.0  & -82.3 & 1.0 & - & - & - & - \\
337:  53741.243 & -56.5 & 1.0  & 139.1 & 1.0 & -1.2 & 1.0 & 56.7 & 1.0 \\
338:  54191.132 & -61.0 & 1.0  & 136.8 & 1.0 & 7.3 & 1.0 & 63.7 & 1.0 \\
339:  54194.134 & -28.7 & 1.0  & 85.1 & 1.0 & - & - & - & - \\
340:  54194.156 & -30.2 & 1.0  & 89.5 & 1.0 & - & - & - & - \\
341:  54198.110 & -56.9 & 1.0  & 132.4 & 1.0 & - & - & - & - \\
342: \enddata
343: \tablenotetext{a}{An observation of Aa1 of unit weight has a standard error of 2.0 ${\rm km\,s^{-1}}.$}
344: \tablenotetext{b}{An observation of Aa2 of unit weight has a standard error of 2.3 ${\rm km\,s^{-1}}.$}
345: \tablenotetext{c}{An observation of Ab1 of unit weight has a standard error of 2.6 ${\rm km\,s^{-1}}.$}
346: \tablenotetext{d}{An observation of Ab2 of unit weight has a standard error of 2.6 ${\rm km\,s^{-1}}.$}
347: \end{deluxetable*} 
348: %\clearpage
349: 
350: \begin{figure}
351: \figurenum{2}
352: %\epsscale{0.8}
353: %\plotfiddle{1458f1.eps}{1cm}{-90}{0.8}{0.8}{0}{0}
354: %%\includegraphics[scale=0.45,angle=90 ]{f2.eps}
355: \plotone{f2.eps}
356: \figcaption{Two spectra of 88 Tau, (a) JD 2,452,539.97 and 
357: (b) JD 2,454,006.01, compared with (c) the spectrum of the IAU radial 
358: velocity standard 10 Tau.  The four components of 88 Tau A are identified 
359: for two different lines. Solid tick marks indicate lines of the Aa (3.57 day 
360: period) binary with the longer tick mark identifying the primary.  Dotted tick 
361: marks indicate lines of the Ab (7.89 day period) binary with the longer tick mark 
362: identifying the primary. 
363: \label{fig2}}
364: \end{figure}
365: %\clearpage
366: 
367: From 2004 January through 2007 April we acquired 29 spectrograms  
368: with the Tennessee State University 2 m automatic spectroscopic
369: telescope (AST), fiber-fed echelle spectrograph, and a 2048 x 4096 SITe
370: ST-002A CCD.  The echelle spectrograms have 21 orders, covering the
371: wavelength range 4920--7100~\AA\ with an average resolution of 0.17~\AA.
372: The typical signal-to-noise ratio is $\sim$50.
373: \citet{ew04} have given a more extensive description of the telescope,
374: situated at Fairborn Observatory near Washington Camp in the Patagonia
375: Mountains of southeastern Arizona, and its operation.
376: 
377: For the KPNO spectrograms we determined radial velocities with the
378: IRAF cross-correlation program FXCOR \citep{f93}, fitting  Gaussian
379: functions to the individual cross-correlation peaks.  
380: Double Gaussian fits were required to obtain individual velocities from blended cross-correlation 
381: peaks.  The IAU radial velocity standard
382: star 10~Tau was used as the cross-correlation reference star for the
383: red-wavelength spectrograms.  Its velocity of 27.9 km~s$^{-1}$ was adopted
384: from \citet{setal90}.  Lines in the wavelength region redward of 6445~\AA\
385: are not particularly suitable for measurement because most features are
386: blends, and there are a number of modest strength water vapor lines.
387: Thus, the radial velocities were determined from lines in the region
388: 6385--6445~\AA.  However, this 60~\AA\ portion of the spectrum is
389: so small that a spectrum mismatch, caused by the varying strength of line
390: blends with temperature, between the A and F spectral type components of 
391: 88~Tau and the F9~IV-V \citep{km89} spectral type of the reference star
392: 10~Tau, can significantly alter the measured velocity.  Thus,
393: instead of cross-correlating this entire 60~\AA\ wavelength region, only
394: the wavelength regions around two or three of the strongest and 
395: least-blended lines, usually the Fe~I lines at 6394 and 6412~\AA\ 
396: plus the Ca~I line at 6439~\AA, were cross-correlated.
397: 
398: At blue wavelengths the Am star dominates the spectrum.  To compute 
399: velocities from those spectrograms, 68~Tau, spectral type A2~IV 
400: \citep{am95}, which has a velocity of 39.0 km~s$^{-1}$ \citep{f99}, 
401: was used as the reference star.  The region between 4485 and 4525~\AA\ 
402: was cross correlated. Velocities for our KPNO spectra
403: are given in Table \ref{tab:dataknpo}.
404: 
405: For the Fairborn Observatory AST spectra, lines in approximately 100 
406: regions, centered on the rest wavelengths \citep{metal66} of relatively 
407: strong lines (mostly of Fe~I and Fe~II) that were not strong blends,
408: were measured.  Lines at the ends of each echelle
409: order were excluded because of their lower signal-to-noise ratios.
410: A Gaussian function was fitted to the profile of each component.  
411: Double Gaussian fits were required to represent blended
412: components.  The difference between the observed wavelength and that 
413: given in the solar line list of \citet{metal66} was used to compute 
414: the radial velocity, and a heliocentric correction was applied.  The 
415: final mean velocity for each observation is given in Table~\ref{tab:datafairborn}.  
416: Unpublished velocities of several IAU standard stars with F dwarf 
417: spectral types indicate that the Fairborn Observatory velocities have 
418: a small zero-point offset of $-$0.3 km~s$^{-1}$ relative to the 
419: velocities of \citet{setal90}. 
420: 
421: \subsection{Preliminary spectroscopic analysis}
422: In a study of lithium in Am stars \citet{bc88} acquired two high-resolution
423: spectrograms of 88 Tau in the 6710~\AA\ region.  Comparing two
424: sets of lines in the two spectra, they reported detecting the lines of 
425: 5 different components.  Figure~2 presents two spectra of 88~Tau~A in the 
426: 6430~\AA\ region that show the two components of the 3.57 day binary near 
427: opposite nodes in their spectroscopic orbit, when the components have 
428: their maximum velocity separation.  Between the two ``outside'' 
429: lines are two additional weak components.  From a careful inspection of 
430: our KPNO spectra, as well as the ones obtained at Fairborn Observatory, 
431: we find lines of only 4 components rather than the 5 reported by 
432: \citet{bc88}.  In Figure~\ref{fig2} many lines of the Am star are $\sim$5\% deep, 
433: while the lines of the other 3 components typically have line depths 
434: $\leq$2.5\%.  Correctly identifying components in such a weak-lined 
435: and complex spectrum is not easy because in many spectra two or 
436: more of the components are blended. 
437: 
438: While the identification of the components of the 3.57 day binary is 
439: straightforward, the very weak lines of the other two components are 
440: similar in strength and line width, making it difficult to tell 
441: them apart.  To determine a preliminary orbital period, we initially 
442: examined only the latter portion of our KPNO velocities, obtained 
443: from MJD 50400 to 54000.  For each observation we computed the absolute
444: value of the velocity difference between the two components and then 
445: used those results as the input data for two different period finding 
446: approaches.  First, a sine curve was fitted to the velocity differences 
447: for trial periods between 1 and 100 days with a step size of 0.0005 
448: days.  The period with the smallest sum of the squared residuals was 
449: adopted as the best period.  Next, a search over a similar period 
450: range and with the same step size was done with the least string 
451: method \citep{betal70}.  Both searches resulted in a period of 3.9435 
452: days. Doubling this period produced an orbital period of 7.887 days.  
453: Separate analyses of our earlier KPNO velocities as well as the Fairborn 
454: Observatory velocities produced a similar orbital period.  We then 
455: adopted the 7.887 day period and computed a phase diagram to identify
456: correctly the components.  Afterward we compared those results with 
457: an attempt at visual identification, based on which set of 
458: lines appeared to be stronger in each spectrum.  The visual inspection
459: correctly identified the more massive component only about half of the 
460: time.  Apparently, the lines of these two components are similar 
461: enough that weak lines from other components and noise can significantly
462: affect the apparent line strengths.  Thus, we conclude that in our spectra it is not 
463: possible to differentiate the two components based on line strength.
464: 
465: \subsection{Orbital Models}
466: 
467: In modeling the hierarchical quadruple system we make the simplifying
468: assumption that the three orbital systems do not perturb each other during 
469: the time of our observations, i.e. we use three Keplerian orbital systems, one wide (Aa-Ab) and slow
470: (18-year period), and two short period systems: Aa1-Aa2, 3.57-day period,
471: and Ab1-Ab2, 7.89-day period.  Note that one cannot simply superimpose
472: the separation vectors from the three models; this is because the
473: PHASES observable is the angle between the two Centers-of-Light (COL)
474: of the short-period systems. 
475: \begin{eqnarray}\label{couplingEquation}
476: \overrightarrow{y_{\rm{obs}}} & = & \overrightarrow{r_{\rm{Aa-Ab}}} \nonumber \\
477:                                                      &     & + \frac{ R_{Aa} - L_{Aa}}{\left(1+R_{Aa}\right)\left(1+L_{Aa}\right)}\overrightarrow{r_{\rm{Aa1-Aa2}}} \nonumber \\
478:                                                      &     & - \frac{ R_{Ab} - L_{Ab}}{\left(1+R_{Ab}\right)\left(1+L_{Ab}\right)}\overrightarrow{r_{\rm{Ab1-Ab2}}}
479: \end{eqnarray}
480: Here $R_{Aa} = M_{\rm{Aa2}}/M_{\rm{Aa1}}$ is the Aa component 
481: mass ratio and $L_{Aa} = L_{\rm{Aa2}}/L_{\rm{Aa1}}$ the luminosity ratio,
482: while $R_{Ab} = M_{\rm{Ab2}}/M_{\rm{Ab1}}$ and $L_{Ab} = L_{\rm{Ab2}}/L_{\rm{Ab1}}$
483: are the corresponding ratios for the Ba--Bb sub-system. 
484: Including this coupling term for astrometric data is 
485: important when a full analysis, including radial velocity data, is made.
486: 
487: \section{Results}
488: 
489: The best-fit orbital model was found with an iterative non-linear
490: least-squares minimization scheme. The best-fit parameters 
491: are found in Table \ref{tab:fit}. The reduced $\chi_r^2$ of the combined fit
492: to PHASES, radial velocity, and previous differential astrometry data
493: is 1.37.  This combined set has 378 data points (49 of which are 
494: two-dimensional astrometric points) and  23 free parameters. 
495: The fits to the astrometric and radial-velocity data for the various 
496: subsystems are shown in Figures 3--7.
497: We find that the two short-period systems have eccentricities 
498: consistent with zero, and we therefore held these parameters 
499: fixed at zero for the fit. The time of maximum primary apparent velocity 
500: is chosen as zero orbital phase. 
501: To investigate the consistency of two astrometric data sets we
502: also re-ran the fit, without including the previous astrometry. We find that 
503: the final results are consistent to within one sigma; including the 
504: previous astrometry does reduce certain parameter uncertainties by a small amount. 
505: In Table \ref{tab:fit}  we compare our best-fit values with those available in the 
506: literature \citep{abt,balega99} where possible; we find generally good 
507: agreement.
508: 
509: Finally, we considered the possibility of an additional massive body in the
510: system.  At this point, a full mass-period phase space search for a fifth
511: component is computationally prohibitive and beyond the scope of this
512: investigation.  However, through the course of the investigation, we discovered
513: a possible periodicity that warranted follow-up.  Before we had determined
514: the 7.89 day period of the Ab1-Ab2 system, we had searched the
515: PHASES data for this third orbit signal with the
516: period-searching program used to find the signatures of additional
517: companions\citep{limits}; because the Aa-Ab and Aa1-Aa2 periods
518: were known already, these orbits were seeded at their best fit double Kepler values. 
519: While radial velocities eventually isolated the period for Ab1-Ab2 as 7.89 days, this
520: astrometric search additionally identified a curious improvement near 56 days.  The
521: size of that orbit could have corresponded either to a brown dwarf or to a nearly equal
522: luminosity star-star subsystem.  Inspired by this identification, a 5-component, quadruple
523: Keplerian model was seeded with the three known orbits (Aa-Ab, Aa1-Aa2, and
524: Ab1-Ab2) and a fourth perturbation with a period of 56 days. Separately, the harmonic of
525: 28 days was also investigated.  This model included all astrometric and radial velocity data. 
526: The 5-component model with a candidate period near 28 days (C/Ab mass ratio $\sim$ 0.044) 
527: yields a significant improvement in the reduced $\chi^2$ of the fit over a 4-body
528: model  (491 for 30 parameters and 397 degrees of freedom vs. 547 for 23 parameters and
529: 404 degrees of freedom, corresponding to a $\sim 2.3\sigma$ improvement from the
530: $\chi^2$ ratio test).   The candidate period near 56 days yields a best-fit $\chi^2 \sim 474$
531: for a $3.7\sigma$ improvement over the 4-component model, yet with a very large best-fit eccentricity (0.93). 
532: However, we note that the 28-day period is close to the product of the
533: Aa and Ab orbital periods ($\sim28.1$ days) and the 56 day period is a harmonic of this,
534: raising suspicion regarding the reliability of the solution.  Hence the astrophysical
535: significance of this additional component remains unclear.  We do not claim this as a
536: detection, but it is an intriguing result worthy of observational follow-up.
537: 
538: %\clearpage
539: \begin{figure*}[t]
540: \figurenum{3}
541: %\epsscale{1.0}
542: \plottwo{f3a.eps}{f3b.eps}
543: \caption[]{\label{fig:astro_wide} (left)The best-fit visual orbit of the 88 Tau Aa-Ab system, together
544: with previously available astrometric data, and our PHASES astrometry. Note that the 
545: error ellipses of the PHASES data appear smaller than the points used to indicate the 
546: data. (right) A close-in view 
547: of a subsection of the PHASES astrometry, together with the best-fit orbital model.}
548: \end{figure*}
549: 
550: \begin{figure*}[t]
551: \figurenum{4}
552: %\epsscale{1.0}
553: \plottwo{f4a.eps}{f4b.eps}
554: \caption[]{\label{fig:astro_narrow_01} The astrometric orbit of the 88
555:   Tau Aa1-Aa2 subsystem, projected along two different 
556:   axes (145 degrees East of North deg on the left, 38 degrees on the right.) In each case the
557:   motion of the Aa--Ab and other sub-system has been removed. The
558:   axes correspond to the two most common orientations of the minor axis
559:   of the positional error ellipses (which vary slightly from night to
560:   night, and between baselines).  For clarity, only those observations where the projected
561:   uncertainty is less than 300 $\mu$as have been included in the plot
562:   (all observations are included in the fit.)  }
563: \end{figure*}
564: 
565: \begin{figure}[t]
566: \figurenum{5}
567: %\epsscale{0.5}
568: \plotone{f5.eps}
569: \caption[]{\label{fig:astro_narrow_02} The astrometric orbit of the 88
570:   Tau Ab1-Ab2 sub-system.  The
571:   motion of the Aa--Ab and other sub-system has been removed. The
572:   separations shown are projected along an axis oriented 145 degrees
573:   East of North, corresponding to mean orientation of the minor axis
574:   of the positional error ellipses (which vary slight from night to
575:   night).  For clarity, only those observations where the projected
576:   uncertainty is less than 300 $\mu$as have been included in the plot
577:   (all observations are included in the fit).  }
578: \end{figure}
579: 
580: 
581: 
582: \begin{figure*}[t]
583: \figurenum{6}
584: %\epsscale{1.0}
585: \plottwo{f6a.eps}{f6b.eps}
586: \caption[]{\label{fig:rv_narrow} (left) The measured and model radial velocities of the 
587: Aa1--Aa2 subsystem, phased about the best-fit orbital model, and with the 
588: motions due to the Aa--Ab orbit removed. (right) Measured and model 
589: radial velocities of the Ab1--Ab2 subsystem, with the Aa- -Ab motion removed. }
590: \end{figure*}
591: 
592: \begin{figure}[t]
593: \figurenum{7}
594: %\epsscale{0.5}
595: \plotone{f7.eps}
596: \caption[]{\label{fig:rv_wide} The measured and modeled radial velocities 
597:  of the 88 Tau Aa-Ab system, with the motion due to the Aa1--Aa2 and Ab1--Ab2
598:  systems subtracted. }
599: \end{figure}
600: %\clearpage
601: 
602: \begin{deluxetable}{l ll}
603: \tablecaption{Best-fit Orbital Parameters for 88 Tau A\label{tab:fit}}
604: \tablewidth{0pt}
605: \tablehead{
606:  Parameter & Value \& Uncertainty & Previous Value 
607: }
608: \startdata 
609: $ \chi^2     $ & $   547   $ &  \\
610: $ \chi_{r}^2   $ & $     1.37 $ & $  $ \\
611: $ {\rm No. Param.} $ & $    23 $ & $  $ \\
612: $ {\rm P_{AaAb}(days)} $ & $  6585  \pm 12 $ &  $6593\pm44\tablenotemark{c}$ \\
613: $ {\rm e_{AaAb}} $ & $     0.0715\pm 0.0026 $ & $0.083\pm0.008\tablenotemark{c}$ \\
614: $ {\rm i_{AaAb}(deg.)} $ & $    69.923 \pm 0.048 $ & $70.4\pm0.4\tablenotemark{c}$  \\
615: $ {\rm \omega_{AaAb}(deg.)} $ & $   205.7  \pm 1.2 $ & $222.0\pm3.0\tablenotemark{c}$\\
616: $ {\rm T_{AaAb} (HMJD)} $ & $ 55261  \pm 22 $ & \\
617: $ {\rm \Omega_{AaAb}(deg.)} $ & $   146.734 \pm 0.067 $ & $146.6\pm0.3\tablenotemark{c}$\\
618: $ {\rm M_{Aa}(M_{\odot})} $ & $     3.42 \pm 0.18 $ & \\
619: $ {\rm M_{Ab}(M_{\odot})} $ & $     2.13  \pm 0.13 $ & \\
620: $ {\rm d (pc)} $ & $    50.70  \pm 0.88 $ & \\
621: \tableline
622: 
623: $ {\rm P_{Aa1Aa2}(days)} $ & $     3.571096 \pm 0.000003 $ &  $3.571391\pm0.000008\tablenotemark{d}$\\
624: $ {\rm e_{Aa1Aa2}} $ & $     0.0   $ & $0\tablenotemark{d}$\\
625: $ {\rm i_{Aa1Aa2}(deg.)} $ & $   110.6  \pm 2.7 $ & \\
626: $ {\rm \omega_{Aa1Aa2}(deg.)} $ & $     0.0  $ & \\
627: $ {\rm T_{0,Aa1Aa2}(MHJD)} $ & $ 53389.3824  \pm 0.0030 $ & $43108.103\pm0.013\tablenotemark{d}$\\
628: $ {\rm \Omega_{Aa1Aa2}(deg.)} $\tablenotemark{a} &$  287.5 \pm 1.8 $ & \\
629: $ {\rm M_{Aa2}/M_{Aa1}} $ & $     0.6602  \pm 0.0028 $ & \\
630: $ {\rm L_{Aa2}/L_{Aa1}}$(K-band)\tablenotemark{a}  & $     0.249 \pm 0.035 $ & \\
631: \tableline
632: 
633: $ {\rm P_{Ab1Ab2}(days)} $ & $     7.886969 \pm 0.000066 $ & \\
634: $ {\rm e_{Ab1Ab2}} $ & $     0.0 $ &  \\
635: $ {\rm i_{Ab1Ab2}(deg.)} $ & $    27.23  \pm 0.72 $ & \\
636: $ {\rm \omega_{Ab1Ab2}(deg.)} $ & $     0.0  $ &  \\
637: $ {\rm T_{0,Ab1Ab2}(MHJD)} $ & $ 52507.31  \pm 0.02 $ & \\
638: $ {\rm \Omega_{Ab1Ab2}(deg.)} $\tablenotemark{b} & $    34.0  \pm 8.2 $ & \\
639: $ {\rm M_{Ab2}/M_{Ab1}} $ & $     0.988 \pm 0.024 $ & \\
640: $ {\rm L_{Ab2}/L_{Ab1}}$(K-band)\tablenotemark{b}  & $     0.865 \pm 0.028 $ & \\
641: $ {\rm V_0 (KPNO, km\,s^{-1})}  $ & $ 23.70  \pm 0.17 $ & \\
642: $ {\rm V_0 (Fairborn,km\,s^{-1})} $ & $ 23.91 \pm 0.31 $ & \\
643: \enddata
644: \tablenotetext{a}{An alternate, but disfavored, solution has $L_{Aa2}/L_{Aa1} = 1.48 $ and $ \Omega_{Aa1Aa2} =326$.}
645: \tablenotetext{b}{An alternate solution has $L_{Ab2}/L_{Ab1} = 1.10$ and $\Omega_{Ab1Ab2} = 205$.}
646: \tablenotetext{c}{From \citet{balega99}}
647: \tablenotetext{d}{From \citet{abt}}
648: \end{deluxetable} 
649: 
650: \begin{deluxetable}{l ll}
651: 
652: \tablecaption{Derived System Parameters for 88 Tau A\label{tab:params}}
653: \tablewidth{0pt}
654: \tablehead{
655:  Parameter & Value & Uncertainty
656: }
657: \startdata 
658: $ {\rm \Phi_{AaAb-Aa1Aa2}(deg.)} $ &  143.3 & $ \pm 2.5 $ \\
659: $ {\rm \Phi_{AaAb-Ab1Ab2}(deg.)} $ &  82.0\tablenotemark{a} & $ \pm 3.3 $ \\
660: $ {\rm \Phi_{Aa1Aa2-Ab1Ab2}(deg.)} $ &  115.8\tablenotemark{b} & $ \pm 4.6 $ \\
661: $ \pi {\rm (asec)} $ &  0.01973 & $ \pm 0.00034 $ \\
662: $ {\rm a_{AaAb}(milliarcsec)} $ &  240.1 & $ \pm  5.3 $ \\
663: $ {\rm a_{Aa1Aa2}(milliarcsec)} $ &  1.359 & $ \pm  0.034 $ \\
664: $ {\rm a_{Aa1Aa2,col} (milliarcsec)} $ &  0.270 & $ \pm  0.032  $ \\
665: $ {\rm a_{Ab1Ab2}(milliarcsec)} $ & 1.967 & $ \pm  0.054 $ \\
666: $ {\rm a_{Ab1Ab2,col}(milliarcsec)} $ & $0.065 $ & $ \pm  0.020 $ \\
667: $ {\rm a_{AB}(AU)} $ &  12.17 & $ \pm  0.17$ \\
668: $ {\rm a_{Aa1Aa2}(AU)} $ & 0.0689 & $ \pm  0.0012 $ \\
669: $ {\rm a_{Ab1Ab2}(AU)} $ & 0.0997 & $ \pm  0.0021 $ \\
670: %$ {\rm K_{Aa} (km\,s^{-1})} $ & 7.26 & $ \pm  0.38$ \\
671: %$ {\rm K_{Ab} (km\,s^{-1})} $ & 11.68 & $ \pm  0.42$ \\
672: %$ {\rm K_{Aa1} (km\,s^{-1})} $ & 78.118 & $ \pm  2.6$ \\
673: %$ {\rm K_{Aa2} (km\,s^{-1})} $ & 118.326 & $ \pm  3.3$ \\
674: %$ {\rm K_{Ab1} (km\,s^{-1})} $ & 31.280 & $ \pm  1.4$ \\
675: %$ {\rm K_{Ab2} (km\,s^{-1})} $ & 31.655 & $ \pm  1.4$ \\
676: $ {\rm M_{Aa1} (M_{\odot})} $ & 2.06& $ \pm  0.11$ \\
677: $ {\rm M_{Aa2} (M_{\odot})} $ & 1.361 & $ \pm  0.073$ \\
678: $ {\rm M_{Ab1} (M_{\odot})} $ & 1.069 & $ \pm  0.069$ \\
679: $ {\rm M_{Ab2} (M_{\odot})} $ & 1.057 & $ \pm  0.068$ \\
680: $ M_{K,Aa1} $ & 0.69    & $ \pm  0.26$ \\
681: $ M_{K,Aa2}  $ & 2.20    & $ \pm  0.28$ \\
682: $ M_{K,Ab1}  $ & 2.31 & $ \pm  0.27$ \\
683: $ M_{K,Ab2}  $ & 2.00 & $ \pm  0.27$ \\
684: \enddata
685: \tablenotetext{a}{An alternate solution has $\Phi_{AaAb-Ab1Ab2}=58$ deg. if $L_{Ab2}/L_{Ab1} >1$.}
686: \tablenotetext{b}{An alternate solution has $\Phi_{Aa1Aa2-Ab1Ab2}=107$ deg. if $L_{Ab2}/L_{Ab1} > 1$.}
687: \tablecomments{The parameters derived from the best-fit model values in Table \ref{tab:fit}
688: and their uncertainties. ``COL" refers to the amplitude of the motion of the Center of Light of the 
689: subsystem in question. Note that our combined astrometry and radial velocity 
690: model fits for the system masses directly.}
691: \end{deluxetable} 
692: %\clearpage
693: 
694: \subsection{Relative Orbital Inclinations}
695: 
696: The mutual inclination $\Phi$ of two orbits is given by
697: \begin{equation}\label{MutualInclination}
698: \cos \Phi = \cos i_1 \cos i_2  + \sin i_1 \sin i_2 \cos\left(\Omega_1 - \Omega_2\right)
699: \end{equation}
700: \noindent where $i_1$ and $i_2$ are the orbital inclinations and $\Omega_1$ and $\Omega_2$ are the 
701: longitudes of the ascending nodes. For this quadruple system we derive 
702: three separate mutual inclinations, corresponding to the three possible
703: pairwise comparisons of the three orbits in this system (i.e. $Aab-Aa1Aa2$, $Aab-Ab1Ab2$ and 
704: $Aa1Aa2-Ab1Ab2$). The resulting values found from our combined 
705: orbital solutions are given in Table \ref{tab:params}.  The large 
706: mutual inclinations indicate that the system orbits are not co-planar.
707: 
708: It should be noted that even with both COL-astrometry and radial-velocity data, 
709: there exists a parameter degeneracy corresponding to an exchange of the 
710: ascending and descending nodes together with a change in the luminosity 
711: ratio (interchanging which is the brighter star). Given one solution for the 
712: mass and luminosty ratios ($R$ and $L_1$), the other possible luminosity 
713: ratio can be found from
714: \begin{equation}
715: L_2 = \frac{2R+RL_1 - L_1}{1+2L_1-R}
716: \end{equation}
717: In a quadruple stellar system such as 88 Tau A there are 4 possible 
718: model solutions. However, as can be seen in Fig. \ref{fig2}
719: the luminosity of the $Aa1$ component is clearly greater than the $Aa2$
720: component, hence we                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                              choose the solution where $L_{Aa1Aa2} = 0.24$.
721: However, given the nearly-equal masses of the $Ab1$ and $Ab2$
722: components, it is not entirely clear which is the preferred solution
723: ($L_{Ab1Ab2} = 0.87$ or $1.10$), and we calculate the two 
724: possible values for the corresponding mutual inclinations. 
725: 
726: The two possible mutual inclination values of the Ab system 
727: are both in the range for inclination-eccentricity 
728: oscillations of 39.2 -- 140.8 degrees \citep{kozai}, while the Aa 
729: system is not (though only 1 standard deviation away from the limit).  
730: During these``Kozai-cycles" the orbital eccentricity varies
731: on a timescale $\propto P_{out}^2/P_{in} (\sim 10^4$ years for
732: the $Ab$ system;  Kiseleva-Eggleton \& Eggleton 2001\nocite{kis01}),
733: and in the absence of damping factors the eccentricity 
734: values would range from $\sim 0$ to $\sim 0.98$ (Kiseleva et al., 1998\nocite{kis98};
735: however, see Kiseleva-Eggleton \& Eggleton 2001 for a discussion 
736: of various factors that may limit these excursions).
737: A full treatment of the dynamics of this system, including the 
738: effects of tidal friction, quadrupolar distortion and general relativity 
739: is beyond the scope of this paper.
740: 
741: Recently, \citet{ft07} studied the effect that Kozai oscillations 
742: would have on the distributions of orbital properties 
743: of triple systems. Systems with mutual inclinations in the 
744: range where Kozai cycles occur evolve rapidly as tidal 
745: dissipation during the high-eccentricity (and hence close approach)
746: phase of the oscillation causes the orbit of the inner binary to shrink. 
747: The resulting mutual  inclination distribution is strongly 
748: peaked near the critical values of 39 and 141 degrees. 
749: It is therefore interesting to note that the Aa system is 
750: so close to one of these predicted values.  The Ab system
751: is more ambiguous: the preferred (albeit only weakly) solution 
752: for the Ab system yields a mutual inclination that would
753: be expected to result in rapid orbital evolution. Interestingly, the
754: second possible solution is also relatively close to the 
755: critical limit. Clearly, further observations will be needed 
756: in order to remove the ambiguity in the node and hence mutual 
757: inclination.  
758: 
759: \subsection{Component Masses and Distance}
760: 
761: This study represents the first determination of the orbital inclinations
762: in this system, and hence the first time the masses have been determined;
763: the precision achieved is $\sim$ 5\% for the Aa components 
764: and $\sim 6$ \% for the Ab components. The parallax is found to
765: be $ 19.73  \pm 0.34$ mas (1.7\% uncertainty), placing the 
766: system at a greater distance than that estimated by {\em Hipparcos} ($21.68 \pm 0.82$ mas; ESA 1997\nocite{hipparcos}).
767: Our greater distance resolves the mass/luminosity 
768: discrepancy pointed out by \citet{balega99}, which arises if one 
769: assumes the {\em Hipparcos} distance to this system.
770: 
771: \citet{stal98} and \citet{balega99} stated
772: that the 3.57 day binary is eclipsing but gave no reference for
773: this claim.  Our inclination for that binary is $110.6\arcdeg$, a value
774: that is about $10\arcdeg$ above the maximum value required for 
775: eclipses to occur if canonical values are assumed for the radii of 
776: the two stars \citep{g92}.  Thus, the system is not eclipsing.  
777: We also note that because of our lower inclination, our masses are 
778: about $0.2 M_{\odot}$ larger than those adopted by \citet{balega99}.
779: 
780: \subsection{Component Luminosities}
781: 
782: As part of the combined astrometric and radial velocity fit we can 
783: solve for the K-band luminosity ratios of the components; this 
784: is because the distance and subsystem total masses are
785: essentially determined by the observations of the wide Aa--Ab system,
786: while the subsystem mass ratios are found from the subsystem
787: radial velocities; this leaves only the component luminosity ratios 
788: dependent on the size of the observed astrometric perturbation. 
789: 
790: While PTI cannot provide precise determinations of the total
791: system magnitude $m_K$ or the Aa--Ab system differential 
792: magnitude $\Delta m_K$, these can be found in the literature.
793: \cite{balega01} gave $\Delta m_{K'} = 1.29 \pm 0.12$ for the 
794: Aa-Ab system.  We derive the resulting absolute K magnitudes 
795: and list them in Table \ref{tab:params}; the results are dominated by
796: the uncertainty in the total magnitude. Note that we assume the 
797: solution where $L_{Ab2}/L_{Ab1} <1$.
798: We compare the determined masses and K-band absolute 
799: magnitudes to published theoretical 
800: isochrones \citep{g02} in the range 0.7--2 Gyr (Z=0.019). We find that the 
801: Aa components are consistent with the isochrones, but the
802: Ab system components appear over-luminous by $\sim$0.5 mag
803: (Figure \ref{fig:massk}).  However, this must be considered a preliminary finding
804: since it is dependent on a single measurement of the Aa--Ab 
805: magnitude difference, and we have not taken into account the 
806: (presumably minor) effects of the slightly different bandpasses ($K$ vs. $K'$).
807: %\clearpage
808: \begin{figure}[t]
809: \figurenum{8}
810: %\epsscale{0.5}
811: \plotone{f8.eps}
812: \caption[]{\label{fig:massk} Mass vs. K magnitude for the components 
813: of the 88 Tau A system, together with isochrones from \citet{g02}.}
814: \end{figure}
815: %\clearpage
816: 
817: \subsection{Spectral Classes and $v$ sin$i$ }
818: 
819: \citet{sf90} identified several luminosity-sensitive and temperature-sensitive
820: line ratios in the 6430-6465~\AA\ region.  Those critical line ratios and
821: the general appearance of the spectrum were employed as spectral-type
822: criteria.  However, for stars that are hotter than about early-G, the line
823: ratios in the 6430~\AA\ region have little sensitivity to luminosity, so
824: only the spectral class of an A or F star can be determined.  The luminosity
825: class is found by computing the absolute visual magnitude with the
826: {\it Hipparcos} or our orbital parallax and comparing that magnitude to
827: evolutionary tracks or a table of canonical values for giants and dwarfs.
828: 
829: The red wavelength spectrum of 88 Tau was compared with those of 
830: slowly rotating Am, F and G dwarfs. These reference stars, identified 
831: mostly from the lists of \citet{am95}, \citet{km89}, and \citet{f97}, 
832: were observed at KPNO with the same telescope, spectrograph, and 
833: detector as our spectra of 88 Tau.  With a computer program developed 
834: by \citet{hb84} and \citet{b85}, various combinations of reference-star 
835: spectra were rotationally broadened, shifted in radial velocity, 
836: appropriately weighted, and added together in an attempt to reproduce 
837: the spectrum of 88~Tau in the 6430~\AA\ region.  
838: \citet{am95} classified 88 Tau as an Am star with spectral classes of
839: A4, A6, and A7 for the calcium, hydrogen, and metal lines, respectively.
840: Their classification of HR~3526 was identical to that of 88~Tau, so
841: we adopted the spectrum of HR~3526 as the proxy for the primary of
842: the 3.57 day binary, which dominates the spectrum at blue wavelengths
843: and is still the strongest component in our red-wavelength region (Fig.~2).  
844: The 3 subclass difference between the calcium K line type and metal line 
845: type indicates that this star is a marginal or mild Am star \citep{ab69}.
846: A good fit to the lines of the 3.57 day secondary was produced by Procyon, 
847: spectral type F5~IV-V \citep{jm53}.  Lines of the two components in the 
848: 7.89 day binary are similar in strength and rotation and were well 
849: represented by a spectrum of 70 Vir, spectral type G4~V \citep{km89}.
850: A fit with HR 483, spectral type G1.5 V \citep{km89}, used 
851: as a proxy for the 7.89 day binary pair, was nearly as good.
852: Thus, the spectral classes of the four stars are A6m, F5, G2-3:, and G2-3:,
853: where the colon indicates that the spectral class is more uncertain than
854: usual because of the weakness of the lines. 
855: As shown by their positions in Figure~8, all four stars are dwarfs.
856: The abundances of Procyon, 70 Vir, and HR 483 are essentially 
857: solar, indicating that the abundances of the components of 88~Tau, except for the Am star, 
858: are also close to solar.   
859: 
860: The continuum intensity ratio of our best reference star combination, 
861: fitted to the spectrum of 88~Tau, is 0.79:0.11:0.05:0.05.
862: If we adopt the continuum intensity ratios as the luminosity
863: ratios at 6430~\AA, we obtain a magnitude difference of 2.1 $\pm$ 0.3 
864: between the 3.57 day pair, and 2.4 $\pm$ 0.3 between the astrometric 
865: components, where the uncertainties are estimated. The 6430~\AA\ 
866: wavelength is about 0.6 of the way between the center of the Johnson 
867: $V$ and $R$ bandpasses.
868: 
869: With the procedure of \citet{f97}, we determined projected rotational
870: velocities for the 4 components of 88~Tau from 10 KPNO red-wavelength
871: spectra.  For each spectrum the full-widths at half-maximum of 2 or 3 
872: unblended lines in the 6430~\AA\ region were measured and the results 
873: averaged for each component.  The instrumental broadening
874: was removed, and the calibration polynomial of \citet{f97} was used to
875: convert the resulting broadening in angstroms into a total line broadening
876: in km~s$^{-1}$.  From \citet{f97,f03} we assumed a 
877: macroturbulence of 0.0 for the Am star, 4~km~s$^{-1}$ for the mid-F star, 
878: and 3 km~s$^{-1}$ for the G stars.  The resulting {\it v}~sin~{\it i} 
879: values are 37 $\pm$ 2 and 17 $\pm$ 2 km~s$^{-1}$ for the primary and 
880: secondary of the 3.57 day binary, respectively, and 5 $\pm$ 3 km~s$^{-1}$ 
881: for both components of the 7.89 day binary.  Our value for the Aa1 
882: component is consistent with the determination of {\it v}~sin~{\it i} =
883: 36 km~$s^{-1}$ by \citet{r02}.
884: 
885: To determine whether the rotational velocities of the binary components
886: are synchronized, we assumed that the rotational axis of each component 
887: is parallel to its orbital axis.  We then computed the equatorial 
888: rotational velocity for each component and compared it with its expected 
889: synchronous velocity, computed with its canonical radius \citep{g92}. 
890: The resulting equatorial rotational velocities are 40, 18, 11, and 11 
891: km~s$^{-1}$, for components Aa1, Aa2, Ab1, and Ab2, respectively.  The 
892: computed synchronous rotational velocities are 24, 19, 6.5, and 6.5 
893: km~s$^{-1}$, respectively.
894:  
895: Component Aa1, the Am star and primary of the 3.57 day binary, is 
896: the only one of the four components that does not have a convective
897: atmosphere.  It is also the only component that is clearly not
898: rotating synchronously: its equatorial rotational velocity is
899: 1.67 times faster than synchronous.  On the other hand its F5 
900: companion, component Aa2 is rotating synchronously.  Because of 
901: the relatively large uncertainties of our v sin i measurements for 
902: Ab1 and Ab2, it is possible that the two components of the 7.89 day 
903: are also synchronously rotating.
904: 
905: 
906: \section{Conclusion}
907: 
908: PHASES interferometric astrometry has been used together with 
909: radial velocity data to measure the orbital
910: parameters of the quadruple star system 88 Tau A, and in particular
911: to resolve the apparent orbital motion of the close Aa1-Aa2 and Ab1-Ab2 pairs.
912: We have made the first determination of the period of the 
913: Ab binary system and found it to consist of a pair of nearly equal-mass
914: G stars. The amplitude of the Ab1-Ab2 Center-of-Light motion is only $\sim 65 \mu$as, 
915: indicating the level of astrometric precision attainable
916: with interferometric astrometry. We are able to resolve the orbital
917: motion of all of the components, and hence determine the 
918: orbital inclinations and component masses with a precision of 
919: a few percent. Finally, we are able to determine the mutual 
920: inclinations of the various orbits. 
921: 
922: \acknowledgements We wish to acknowledge the extraordinary
923: observational efforts of K. Rykoski. Observations with PTI are made
924: possible thanks to the efforts of the PTI Collaboration, which we
925: acknowledge. This research has made use of services from
926: the Michelson Science Center, California Institute of Technology,
927: http://msc.caltech.edu.  Part of the work described in this paper was
928: performed at the Jet Propulsion Laboratory under contract with the
929: National Aeronautics and Space Administration. This research has made
930: use of the Simbad database, operated at CDS, Strasbourg, France, and
931: of data products from the Two Micron All Sky Survey, which is a joint
932: project of the University of Massachusetts and the Infrared Processing
933: and Analysis Center/California Institute of Technology, funded by the
934: NASA and the NSF.  We thank T. Willmitch for measuring some of the early KPNO spectra.
935: The work of FCF and MW has been supported in part by NASA 
936: grant NCC5-511 and NSF grant HRD-9706268. 
937: PHASES is funded in part by the California Institute of Technology Astronomy Department, 
938: and by the National Aeronautics and Space Administration under grant 
939: NNG05GJ58G issued through the Terrestrial Planet Finder Foundation Science Program. 
940: This work was supported in part by the National Science Foundation
941:  through grants AST-0300096, AST-0507590 and AST-005366. MWM is grateful for the 
942:  support of a Townes fellowship. MK is supported by NASA through grant NNG04GM62G and
943: the Polish Ministry of Education and Science through 
944: grant 1P03D 021 29.
945: 
946:  
947: \begin{thebibliography}{}
948: 
949: 
950: \bibitem[Abt \& Bidelman(1969)]{ab69}
951: Abt, H. A., \& Bidelman, W. P. 1969, 158, 1091
952: 
953: \bibitem[Abt \& Levy(1985)]{abt} 
954: Abt, H.~A., \& Levy,  S.~G.\ 1985, \apjs, 59, 229 
955: 
956: \bibitem[Abt \& Morrell(1995)]{am95}
957: Abt, H. A., \& Morrell, N. I. 1995, \apjs, 99, 135
958: 
959: \bibitem[Barden(1985)]{b85}
960: Barden, S. C. 1985, \apj, 295, 162
961: 
962: \bibitem[Balega et al.(1999)]{balega99} Balega, I.~I., Balega, 
963: Y.~Y., Hofmann, K.-H., Tokovinin, A.~A., \& Weigelt, G.~P.\ 1999, Astronomy 
964: Letters, 25, 797
965: 
966: % IR speckle delta-m
967: \bibitem[Balega et al.(2001)]{balega01} 
968: Balega, I.~I., Balega, Y.~Y., Hofmann, K.-H., \& Weigelt, G.\ 2001, Astronomy Letters, 27, 95 
969: 
970: \bibitem[Borkovits et al.(2004)]{bork04} Borkovits, T., 
971: Forg{\'a}cs-Dajka, E., \& Reg{\'a}ly, Z.\ 2004, \aap, 426, 951 
972: 
973: \bibitem[Boden et al.(2000)]{boden00}
974: Boden, A., Creech-Eakman, M., Queloz, D., 2000, \apj, 536, 880-890.
975: 
976: 
977: 
978: \bibitem[Bopp et al.(1970)]{betal70} 
979: Bopp, B. W., Evans, D. S., \& Laing, J. D. 1970, \mnras, 147, 355
980: 
981: \bibitem[Burkhart \& Coupry(1988)]{bc88}
982: Burkhart, C., \& Coupry, M. F. 1988, \aap, 200, 177
983: 
984: 
985: \bibitem[Colavita et al.(1999)]{colavita99}  
986: Colavita, M. M.,et al.~ 1999, \apj, 510 , 505.
987: 
988: \bibitem[Colavita(1999b)]{colavita99b}  
989: Colavita, M. M., 1999, \pasp, 111, 111. 
990: 
991: \bibitem[Colavita et al.(2003)]{colavita03} 
992: Colavita, M., et al. 2003, \apjl, 592, L83 
993: 
994: \bibitem[Cowley et al.(1969)]{c69} Cowley, A., Cowley, C., 
995: Jaschek, M., \& Jaschek, C.\ 1969, \aj, 74, 375 
996: 
997: \bibitem[Cox(1999)]{allen}
998: Cox, A. N., 1999, Allens Astrophysical Quantities, Springer-Verlag, New York.
999: 
1000: \bibitem[Eaton \& Williamson(2004)]{ew04}
1001: Eaton, J. A., \& Williamson, M. H. 2004, SPIE, 5496, 710
1002: 
1003: \bibitem[ESA(1997)]{hipparcos}
1004: ESA 1997, The Hipparcos and Tycho Catalogues, ESA SP-1200.
1005: 
1006: \bibitem[Fabrycky \& Tremaine(2007)]{ft07} 
1007: Fabrycky, D., \& Tremaine, S.\ 2007, ArXiv e-prints, 705, arXiv:0705.4285 
1008: 
1009: \bibitem[Fekel(1997)]{f97}
1010: Fekel, F. C. 1997, \pasp, 109, 514
1011: 
1012: \bibitem[Fekel(1999)]{f99}
1013: Fekel, F. C. 1999, IAU Colloquium 170, Precise Stellar Radial 
1014: Velocities, eds. J. B. Hearnshaw and C. D. Scarfe (San Francisco: ASP),
1015: 378
1016: 
1017: \bibitem[Fekel et al.(2000)]{fekel00} 
1018: Fekel, F.~C., Joyce, R.~R., Hinkle, K.~H., \& Skrutskie, M.~F.\ 2000, \aj, 119, 1375 
1019: 
1020: \bibitem[Fekel et al.(2002)]{f02} Fekel, F.~C., Scarfe, 
1021: C.~D., Barlow, D.~J., Hartkopf, W.~I., Mason, B.~D., \& McAlister, H.~A.\ 
1022: 2002, \aj, 123, 1723 
1023: 
1024: \bibitem[Fekel(2003)]{f03}
1025: Fekel, F. C. 2003, \pasp, 115, 807
1026: 
1027: 
1028: \bibitem[Fitzpatrick(1993)]{f93}
1029: Fitzpatrick, M. J. 1993, in ASP Conf. Ser. 52, Astronomical
1030: Data Analysis Software and Systems II, ed. R. J. Hanisch, R. V. J.
1031: Brissenden, \& J. Barnes (San Francisco: ASP), 472
1032: 
1033: \bibitem[Girardi et al.(2002)]{g02} Girardi, L., Bertelli, 
1034: G., Bressan, A., Chiosi, C., Groenewegen, M.~A.~T., Marigo, P., Salasnich, 
1035: B., \& Weiss, A.\ 2002, \aap, 391, 195 
1036: 
1037: \bibitem[Gray(1992)]{g92}
1038: Gray, D. F. 1992, The Observation and Analysis of Stellar Photospheres
1039: (Cambridge: Cambridge University Press) 
1040: 
1041: \bibitem[Hartkopf et al.(2001)]{int4} 
1042: Hartkopf, W.~I., McAlister, H.~A., \& Mason, B.~D.\ 2001, \aj, 122, 3480 
1043: 
1044: 
1045: \bibitem[Huenemoerder \& Barden(1984)]{hb84}
1046: Huenemoerder, D. P., \& Barden, S. C. 1984, \baas, 16, 510
1047: 
1048: \bibitem[Johnson \& Morgan(1953)]{jm53}
1049: Johnson, H. L., \& Morgan, W. W. 1953, \apj, 117, 313
1050: 
1051: \bibitem[Keenan \& McNeil(1989)]{km89}
1052: Keenan, P. C., \& McNeil, R. C. 1989, \apjs, 71, 245
1053: 
1054: \bibitem[Kozai(1962)]{kozai} Kozai, Y.\ 1962, \aj, 67, 579 
1055: 
1056: \bibitem[Kiseleva et al.(1998)]{kis98} 
1057: Kiseleva, L.~G., Eggleton, P.~P., \& Mikkola, S.\ 1998, \mnras, 300, 292 
1058: 
1059: \bibitem[Kiseleva-Eggleton \& Eggleton(2001)]{kis01} 
1060: Kiseleva-Eggleton, L., \& Eggleton, P.~P.\ 2001, Evolution of Binary and 
1061: Multiple Star Systems, 229, 91 
1062: 
1063: %
1064: \bibitem[Lane \& Muterspaugh(2004)]{lm04} Lane, B.~F., \& 
1065: Muterspaugh, M.~W.\ 2004, \apj, 601, 1129 
1066: 
1067: \bibitem[Lestrade et al.(1993)]{l93} Lestrade, J.-F., 
1068: Phillips, R.~B., Hodges, M.~W., \& Preston, R.~A.\ 1993, \apj, 410, 808 
1069: 
1070: % discovery of the speckle system
1071: \bibitem[McAlister et al.(1987)]{mac87} 
1072: McAlister, H.~A., Hartkopf, W.~I., Hutter, D.~J., \& Franz, O.~G.\ 1987, \aj, 93, 688 
1073: 
1074: % delEq
1075: \bibitem[Muterspaugh et al.(2005)]{delEqu} Muterspaugh, M.~W., 
1076: Lane, B.~F., Konacki, M., Burke, B.~F., Colavita, M.~M., Kulkarni, S.~R., 
1077: \& Shao, M.\ 2005, \aj, 130, 2866 
1078: 
1079: %
1080: \bibitem[Muterspaugh et al.(2006a)]{v819Her} Muterspaugh, M.~W., 
1081: Lane, B.~F., Konacki, M., Burke, B.~F., Colavita, M.~M., Kulkarni, S.~R., 
1082: \& Shao, M.\ 2006, \aap, 446, 723 
1083: 
1084: %
1085: \bibitem[Muterspaugh et al.(2006b)]{kapPeg} Muterspaugh, M.~W., 
1086: Lane, B.~F., Konacki, M., Wiktorowicz, S., Burke, B.~F., Colavita, M.~M., 
1087: Kulkarni, S.~R., \& Shao, M.\ 2006, \apj, 636, 1020 
1088: 
1089: %
1090: \bibitem[Muterspaugh et al.(2006c)]{limits} Muterspaugh, M.~W., 
1091: Lane, B.~F., Kulkarni, S.~R., Burke, B.~F., Colavita, M.~M., \& Shao, M.\ 
1092: 2006, \apj, 653, 1469 
1093: 
1094: \bibitem[Muterspaugh et al.(2007)]{muOri}
1095: Muterspaugh et al. in prep.
1096: 
1097: \bibitem[Moore et al.(1966)]{metal66}
1098: Moore, C. E., Minnaert, M. G. J., \& Houtgast, J. 1966, The Solar
1099: Spectrum 2935 \AA\ to 8770 \AA, National Bureau of Standards Monograph
1100: 61 (Washington, D. C.: U. S. Government Printing Office)
1101:  
1102: \bibitem[Royer et al.(2002)]{r02}  
1103:  Royer, F., Grenier, S., Baylac, M.-O., G{\'o}mez, A.~E., \& Zorec, J.\ 2002, \aap, 393, 897 
1104:  
1105: \bibitem[Scarfe et al.(1990)]{setal90}
1106: Scarfe, C. D., Batten, A. H., \& Fletcher, J. M. 1990, Publ. Dominion
1107: Astrophy. Obs., 18, 21
1108: 
1109: \bibitem[Schoeller et al.(1998)]{stal98}
1110: Schoeller, M., Balega, I. I., Balega, Y. Y., Hofmann, K.-H., 
1111: Reinheimer, T., \& Weigelt, G. 1998, Astronomy Letters, 24, 337
1112: 
1113: 
1114: %2MASS
1115: \bibitem[Skrutskie et al.(2006)]{2mass}
1116: Skrutskie, M.~F. et al, 2006, \aj, 131, 1163.
1117: 
1118: \bibitem[Strassmeier \& Fekel(1990)]{sf90}
1119: Strassmeier, K. G., \& Fekel, F. C. 1990, \aap, 230, 389
1120: 
1121: % importance of mutual inclinations
1122: \bibitem[Sterzik \& Tokovinin(2002)]{st02} 
1123: Sterzik, M.~F., \& Tokovinin, A.~A.\ 2002, \aap, 384, 1030 
1124: 
1125: %the CPM 
1126: \bibitem[Tokovinin \& Gorynya(2001)]{tg01} 
1127: Tokovinin, A.~A., \& Gorynya, N.~A.\ 2001, \aap, 374, 227 
1128: 
1129: % the 4th component and CPM
1130: \bibitem[Tokovinin et al.(2006)]{toko06} Tokovinin, A., 
1131: Thomas, S., Sterzik, M., \& Udry, S.\ 2006, \aap, 450, 681 
1132: 
1133: %MSC
1134: \bibitem[Tokovinin(1997)]{msc} Tokovinin, A.~A.\ 1997, 
1135: \aaps, 124, 75 
1136: 
1137: %castor
1138: \bibitem[Torres \& Ribas(2002)]{tr02} 
1139: Torres, G., \& Ribas, I.\ 2002, \apj, 567, 1140 
1140: 
1141: \end{thebibliography}
1142: 
1143: 
1144: 
1145: 
1146: \end{document}
1147: 
1148: 
1149: 
1150: 
1151: 
1152: 
1153: