------------------------------------------------------------------------ ms.tex ApJ, 2008, in press Message-ID: MIME-Version: 1.0 Content-Type: MULTIPART/MIXED; BOUNDARY="-478107720-106629879-1215638649=:3940" X-Virus-Scanned: by UF Astronomy Mail Virus Scanner ks/14/4/2005 X-MailScanner-Information: Please contact the postmaster@aoc.nrao.edu for more information X-MailScanner-ID: m69LOF1b016645 X-MailScanner: Found to be clean X-MailScanner-SpamCheck: not spam, SpamAssassin (not cached, score=-4, required 5, autolearn=disabled, RCVD_IN_DNSWL_MED -4.00) X-MailScanner-From: reba@astro.ufl.edu X-Spam-Status: No This message is in MIME format. The first part should be readable text, while the remaining parts are likely unreadable without MIME-aware tools. ---478107720-106629879-1215638649=:3940 Content-Type: TEXT/PLAIN; charset=en_US Content-Transfer-Encoding: QUOTED-PRINTABLE %astro-ph/0807.0657 \documentclass[12pt,preprint]{aastex} \usepackage{epsfig} \shorttitle{CXOGC~J174536.1-285638: X-ray and Infrared Variability} \shortauthors{Mikles et al.} \begin{document} \title{Discovery and Interpretation of an X-ray Period in the Galactic Cent= er Source CXOGC~J174536.1-285638} \author{Valerie J. Mikles} \affil{Department of Astronomy, University of Florida, Gainesville, FL 3261= 1} \email{mikles@astro.ufl.edu} \author{Stephen S. Eikenberry} \affil{Department of Astronomy, University of Florida, Gainesville, FL 3261= 1} \email{eikenberry@astro.ufl.edu} \author{Reba M. Bandyopadhyay} \affil{Department of Astronomy, University of Florida, Gainesville, FL 3261= 1} \email{reba@astro.ufl.edu} \and=20 \author{Michael P. Muno} \affil{Space Radiation Laboratory, Caltech, Pasadena, CA 91125} \email{mmuno@srl.caltech.edu} \begin{abstract} We present X-ray and infrared observations of the X-ray source CXOGC~J17453= 6.1-285638. Previous observations suggest that this source may be an accret= ing binary with a high-mass donor (HMXB) or a colliding wind binary (CWB). = Based on the {\it Chandra} and {\it XMM-Newton} light curve, we have found = an apparent $189 \pm 6$ day periodicity with better than 99.997\% confidenc= e. We discuss several possible causes of this periodicity, including both o= rbital and superorbital interpretations. We explore in detail the possibili= ty that the X-ray modulation is related to an orbital period and discuss th= e implications for two scenarios; one in which the variability is caused by= obscuration of the X-ray source by a stellar wind, and the other in which = it is caused by an eclipse of the X-ray source. We find that in the first c= ase, CXOGC~J174536.1-285638 is consistent with both CWB and HMXB interpreta= tions, but in the second, CXOGC~J174536.1-285638 is more likely a HMXB. \end{abstract} \keywords{accretion, accretion disks, binaries: X-ray, infrared: stars, X-r= ay: stars, stars:individual: CXOGC~J174536.1-2856} \section{Introduction} {\it Chandra} observations of the Galactic Center (GC) have revealed a larg= e new population of low luminosity X-ray sources with $L_X[D/8kpc]^2 \sim 1= 0^{31} - 10^{35} erg~s^{-1}$ \citep{mun03}. In addition, the Swift and INTE= GRAL missions have recently revealed a new population of highly-absorbed X-= ray sources, believed to be high-mass X-ray binaries \citep[HMXBs; e.g.][]{= beck05,bod07,neg07}. In 2005, we identified an infrared (IR) star as the fi= rst spectroscopically confirmed IR counterpart to the low luminosity {\it C= handra} source CXOGC~J174536.1-285638 \citep[hereafter Paper I]{mikles06}. = Based on the X-ray and IR spectra and the X-ray to IR luminosity ratio, we = showed that the source is most likely a massive star in a binary system. Th= e source shows strong HeI (2.114-$\mu$m), Brackett-$\delta$ (1.945-$\mu$m),= and Brackett-$\gamma$ (2.166-$\mu$m) emission lines, typical of both accre= tion-powered binaries and CWBs. Additionally, we observe Brackett series, H= eI, HeII, CIII, and NIII line emission. P~Cygni profiles are visible in sev= eral HeII lines, suggesting wind activity around a massive star. The X-ray = spectrum of this source is particularly intriguing, having prominent Fe-XXV= emission centered at 6.7~keV with an equivalent width of 2.2~keV. This is = one of the highest equivalent width Fe-XXV lines ever seen (Paper I).=20 Our initial IR observations were aimed toward the discovery of a short ($<1= $~d) period in the CXOGC~J174536.1-2856 binary. We use IR spectroscopy to s= earch for variations in CXOGC~J174536.1-285638's IR emission features. We a= nalyze {\it Chandra} and {\it XMM-Newton} archival data to search for X-ray= variability over short and long baselines. From the combined X-ray light c= urve, we find a period of $189 \pm 6$ days. We discuss CXOGC~J174536.1-2856= 's variability in the IR and X-ray, and examine the implications of a 189~d= period for the nature of the source. In \S 2, we summarize our IR and X-ra= y observations and analysis, detailing both our IR radial velocity study an= d X-ray period analysis. In \S 3, we discuss the IR and X-ray variability i= n CXOGC~J174536.1-2856, specifically exploring an orbital period interpreta= tion of the identified X-ray period.=20 \section{Observations and Analysis} \subsection{Infrared Counterpart to CXOGC~J 174536.1-285638} CXOGC~J174536.1-285638 was discovered as part of a {\it Chandra} survey of = the GC region. The survey, conducted by \citet{mun03}, identified 2357 sere= ndipitous X-ray sources with $L_X[D/8kpc]^2 \sim 10^{31} - 10^{35} erg s^{-= 1}$ within $\sim 10$-arcmin of Sgr A*. The coordinates of the source are 26= 6.40060, -28.94407 with positional uncertainty of 0.4-arcsec \citep{mun04}. We searched for potential IR counterparts using the 2MASS catalog and ident= ified the 2MASS source 17453612-2856386 as the likely counterpart. The blen= ded 2MASS source is clearly resolved into two stars in our IRTF observation= s. In Figure \ref{fig1}, we show a 15'' x 15'' 2MASS image and IRTF SpeX = slit image with a 1.5'' circle at the {\it Chandra} coordinate center. The = two stars, blended in 2MASS, are well separated in the IRTF finder image. T= he IR astrometric solution is derived from 2MASS which has a stated astrome= tric accuracy of 15mas. Due to the proximity of the two potential IR counte= rparts, we were able to obtain simultaneous spectra of both stars; we plot = both spectra in Figure \ref{fig2}. In Paper I, we identify ``star 1,'' the = emission line source, as ``Edd-1,'' the counterpart to the {\it Chandra} ob= ject. The second source is an evolved star of type K or cooler, with no evi= dence for emission lines which are signatures of high energy processes, suc= h as accretion or wind collison. It thus seems unlikely that this second so= urce is the IR counterpart to the X-ray source. =20 \subsection{Search for short-period IR variability} On 2006 Aug 02-04 UT, we obtained J, H, and K band (1.1-2.4 $\mu$m) spectra= of the IR counterpart to CXOGC~J174536.1-285638 using SpeX on IRTF \citep{= rayner03}, in hopes of finding short period ($<1$~d) variability in the sou= rce. Dithering along the 0.5 arcsec slit, we obtained 184 exposures of 120~= s each over the course of three half-nights, giving us a time baseline of 3= -4 hours per night. The procedure for our IR analysis of the SpeX data is d= escribed in Paper I. We extract spectra using the standard SpexTool procedu= re for AB nodded data, resulting in a series of sky-subtracted, wavelength-= calibrated spectra \citep{vacca03,cushing04}. We interpolate over the intri= nsic Brackett absorption features in the G0V-star spectrum, then divide the= target spectrum by the G0V-star in order to remove atmospheric absorption = bands. We multiply the resultant spectrum by a 5900~K blackbody spectrum, c= orresponding to the temperature of the G0V-star. =20 Using our previous observations taken on 2005 July 1 UT (Paper I), we adopt= a reddening value of $A_V=3D29$ mag and apply this correction to all data.= =20 We used spectra from each night to test variability on multiple time scales= =2E Figure \ref{fig3} shows the series of twenty-one K-band spectra taken o= ver the course of our observations, with integration times between 8 and 20= minutes per spectrum. We list the specific time stamps and exposure times = of these spectra in Table \ref{tbl-1}. To search for radial velocity variat= ions in the emission lines, we track the line centers with two methods: fir= st by taking a statistical mean of the wavelength around the line center, w= eighted by flux, and second by fitting a Gaussian to the line. We find no r= adial velocity variations, nor do we find significant flux variations in th= e lines. We checked for IR variability on 1 year, 3 day, 3 hour, 1 hour, an= d 30 minute baselines and found no evidence of periodic variability or flar= es in this sample. The only apparent variation is in the structure of the = Br-$\gamma$ line complex (see Figure \ref{fig4}), but this does not often v= ary more than $\sim 5$ times the RMS spectral difference in the vicinity of= the $\lambda 2.164$ Helium component. Further we note that this region is = affected by our data reduction process (i.e., the removal of the intrinsic = Brackett absorption in the G0V). \subsection{X-ray Variability} {\it Chandra} observations of CXOGC~J174536.1-285638 revealed long baseline= intensity variations by a factor of $\sim$3 in X-ray in the 2-8~keV range.= The variation, which was observed initially with {\it Chandra} in 2002 as = a drop in flux, repeated in 2006 with similar morphology, prompting us to s= earch archival X-ray data for additional information about this source's lo= ng term (month to year) variability. We list the {\it Chandra} data used in= our analysis in Table \ref{tbl-2}. We supplement the {\it Chandra} observa= tions with {\it XMM-Newton} archival data, listed in Table \ref{tbl-3}. CXO= GC~J174536.1-285638 is easily identified as an isolated source $\sim$10~pc = from the GC and is not confused with any other source detection in either t= he {\it Chandra} or {\it XMM} images. The positions of the {\it XMM} and {\= it Chandra} sources are consistent within the respective astrometric accura= cy of the two instruments \citep[0.4'' and 1'' respectively;][]{mun04b, kir= sch05}. We identify the {\it XMM} counterpart to the {\it Chandra} source a= nd show that in Figure \ref{fig5}. In addition to the astrometric accuracy,= strong Fe-emission is detected in both the {\it Chandra} and {\it XMM} dat= a (see Fig. \ref{fig6}), confirming that the {\it XMM} source is the same a= s the {\it Chandra} source. The Fe-XXV emission in this source is unusually= strong and it would be extremely unlikely to detect emission in both the {= \it Chandra} (see Paper 1) and {\it XMM} (this paper) spectra were they not= the same source. =20 Due to the relative faintness of CXOGC~J174536.1-285638 in the X-ray (usual= ly $<$20 cts/hr), many of the {\it XMM} observations suffer from low signal= -to-noise. While the 2001-2002 data consist of fairly short observations (e= xposure time $< 7$~h), in 2004 there are four observations of 40 consecutiv= e hours each. Following standard {\it XMM} data reduction techniques, we ge= nerate an astrometrically calibrated event list. From this, we located CXOG= C~J174536.1-285638 and extracted light curves and spectra from a circle wit= h a 200-pixel radius. The background was calculated from a ring extending 3= 00-500 pixels from the source center. We show the extraction region around = the source in Figure \ref{fig5}. We set the spectral bin size at 200~eV and= plot two representative spectra in Figure \ref{fig6}. The flux varies by a= factor of three between these observations. Because of the extremely low c= ount rate, we cannot meaningfully constrain the fainter spectrum with XSpec= models.=20 We extract light curves at five hour intervals over the full 2-8~keV band, = as well as from the ``soft'' 2-4~keV band and the ``hard'' 4-8~keV band sep= arately for the 2004 {\it XMM} observations. The time resolution is chosen = to ensure sufficient counts in each bin to test for variability. The X-ray = flux varies aperiodically by less than a factor of 2 over the course of eac= h individual observation and the hardness ratio is consistent with zero. Ap= eriodic variability is not uncommon in stellar X-ray sources on these times= cales. We observe no periodic variability on timescales less than 40 hours.= =20 Because the X-ray flux is relatively constant over the course of each {\it = XMM} observation, we calculate a single flux value for each observation epo= ch and combine these measurements with the {\it Chandra} light curve in Fig= ure \ref{fig7}. Using the combined light curve, we can test for the presenc= e or absence of periodic flux variations on timescales longer than 40 hours= =2E The most notable flux variation in the {\it XMM} is a 4$\sigma$ variati= on in consecutive observations separated by five months (see Fig. \ref{fig6= }). If periodic, this suggests a longer timescale variability. Using the me= thod of \citet{horne86}, we perform a periodogram analysis of the combined = {\it Chandra} and {\it XMM} light curve and find a period of $189 \pm 6$ da= ys. In Figure \ref{fig8}, we show the resultant periodogram which tests for= periodicity on scales of 1-1500 days. The peak at 189~d is clearly distinc= t and additional peaks are visible at integer multiples of the period. In F= igure \ref{fig7}, we plot the X-ray light curve folded on the 189 d period.= Analytically estimating the significance of a signal in non-uniformly samp= led data is non-trivial. Thus, in order to estimate the confidence of this = detection, we perform a Monte Carlo simulation as follows. We take the exis= ting data set and maintain the same sampling intervals throughout. For each= Monte Carlo realization, we randomly reassign the observed flux values to= the time samples, effectively scrambling the light curve. We plot the resu= lts of these simulations in Figure \ref{fig9}. In 30,000 trials, we do not = achieve a peak power approaching the power of our original periodogram, imp= lying that the 189-day period is not due to random noise with a confidence = level greater than 99.997\%. The previous test accounts for white noise variability; however, red noise = is a significant source of false peaks in X-ray power spectra of X-ray bina= ries \citep{tit07}. Red noise is a flux variation in the power spectrum tha= t can be parameterized with a frequency dependence $f^{-\beta}$. A white no= ise process will generate a flat power spectrum such that $\beta \sim 0$; a= value of $\beta \sim 2$ describes random walk noise \citep{tk95}. A $\beta= \sim 1$ dependence has been identified in stellar-mass black hole candidat= es and may be strongly related to accretion physics in the system \citep{mi= n94,tk95,tit07}. Following the method of \citet{tk95}, we test the possibi= lity of red noise creating a false signal matching the strength of our peri= odogram. Simulating a number of red noise dominated light curves of varying= power law slope, $\beta$, we find that as $\beta$ increases, more noise ge= ts shunted near the period frequency, and the significance of our detection= decreases. We show the results of our tests in Figure \ref{fig10}. We find= the significance of our period detection remains above $3\sigma$ for value= s of $\beta \leq 1.0$ and above $2.5\sigma$ for $\beta \leq 1.5$, showing t= hat the significance decreases slowly as red noise is increased. \section{Discussion} \subsection{Infrared Variability} We test the IR spectra for variations on short timescales (hours to days). = Due to limits of our spectral resolution, we cannot observe radial velocit= y variations if the orbital velocity is less than 70km/s. In Figure \ref{fi= g4}, we show several close-ups of the Brackett-$\gamma$ region of CXOGC~J17= 4536.1-285638's spectrum over the course of our three night IRTF run in 200= 6. To the left of the $\lambda$2.164$\mu$m marker, we see minor variances i= n the Helium contribution to the line. Because this line cannot be resolved= from the larger Br-$\gamma$ contribution, it is difficult to determine the= significance of this change. The RMS spectral difference rarely reaches 5$= \sigma$ between any two events which are separated by $\sim 1$ hour. The ob= served differences are primarily in the wings of the line ($\sim 2.164 \mu$= m or $\sim 2.168 \mu$m). Higher resolution spectroscopy is required to dete= rmine whether the changes in the Helium contribution are intrinsic to CXOGC= ~J174536.1-285638 rather than an artifact of the data analysis. The observe= d variations do not have any detectable periodicity. It should also be note= d that this region is affected by the data reduction process, as described = in \S 2. Our 2006 IR spectra were obtained about two days after the {\it Ch= andra} observations on Day 2402 in the X-ray light curve (see Fig. \ref{fig= 7}), where the object is transitioning from an apparent low-flux state to a= high-flux state. Since we have no IR data consistent with the lowest X-ray= flux events, it is impossible to determine from these IR data if the appar= ent Helium variability at $\lambda 2.164\mu$m we observe is associated with= this X-ray flux transition. We also search for wind variations in the P Cygni profiles. In our initial = discovery spectrum, we identified three HeII lines with P~Cygni profiles: 2= =2E0379 $\mu$m, 2.1891 $\mu$m, and 2.3464 $\mu$m (Paper I). In our 2005 ana= lysis, we estimated the P~Cygni velocity at $170 \pm 70$ km/s. We repeat ou= r analysis on the 2006 data to search for variations and find the approxima= te velocity of the wind is $200 \pm 70$ km/s. The error is dominated by the= spectral resolution. We find no evidence of changes in the P~Cygni profile= or velocity over our three day observations. Also, the 2005 and 2006 spect= ra have consistent P~Cygni profiles and velocities. Unfortunately, it was not until after completion of our IR observation camp= aign that we discovered the 189~d X-ray periodicity in the source. Thus we = were not able to schedule our IR observations to sample different X-ray pha= ses; as a result, both our 2005 and 2006 observations sample the same phase= (indicated in Figure \ref{fig7}). The lack of IR radial velocity variation= s is consistent with the observations being at the same phase of a long per= iod system.=20 \subsection{X-ray Variability} Long term {\it Chandra} observations of this source revealed repeated X-ray= flux variations, prompting us to search for periodicity by combining {\it = XMM} and {\it Chandra} data, and revealing a 189-d period. In Paper I, we a= rgue that CXOGC~J174536.1-285638 contains at least one massive star based o= n the presence of P~Cygni profiles in the IR spectrum. Although we consider= the possibility of both an isolated massive star or a massive star in a bi= nary system in Paper I, here we favor a binary interpretation because X-ray= variability similar to that seen in CXOGC~J174536.1-285638 is not observed= in isolated massive stars \citep{cohen00}. In comparing CXOGC~J174536.1-28= 5638 to other systems containing massive stars, we showed that the X-ray to= IR luminosity ratio, $L_X/L_K \sim 10^{-4}$, is consistent with both colli= ding wind binary (CWB) and high-mass X-ray binary (HMXB) systems (Paper I). In the standard models for CWBs, X-ray emission arises from the shock front= of colliding winds in two massive stars \citep[see, e.g.][]{luo90,sana04,d= ebeck06}. Observed variability is often attributed to phase-locked flux mod= ulations due to the effect of variations in absorption along the line of si= ght and variations in X-ray emission as a function of orbital phase. In thi= s situation, the X-ray periodicity reflects an orbital period. Alternativel= y, it is possible that stellar rotation or photospheric pulsation may also = produce periodic X-ray modulations. Models of such behavior are often emplo= yed to explain the 84~d quasi-periodicity in $\eta$ Carinae \citep{dav98}. = In these situations, the modulation of the X-ray flux is correlated to recu= rrent behavior affecting the wind emission, but not related to the orbital = period. =20 However, in HMXBs, periodic X-ray flux changes can be the result of either = orbital or superorbital motion. A superorbital periodicity is defined as an= y periodicity apparent in the periodogram that is greater than the orbital = period. The predominant model for superorbital periodicity is that of a pre= cessing warped accretion disk; however, long period variations may also be = due to the precession of a compact object (not applicable to black hole sys= tems), periodic modulation of the mass accretion rate, or the influence of = a third body \citep{paul00,og01,clarkson03}. Superorbital variations divide= into two broad observational classes. The first class is characterized by = clear, stable X-ray variations of about $\sim$30 days, while the second cla= ss has longer, quasi-periodic variations ranging from $\sim 50-200$ days \c= itep{clarkson03}. The second class is considered quasi-periodic, because lo= ng term monitoring shows a broad power peak in the periodogram, often super= posed on a red noise spectrum \citep[e.g., Cyg X-2;][]{paul00}. Cen~X-3, Cy= g~X-1, and Vela X-1 are all high-mass binary systems showing both orbital a= nd superorbital periods. They range in X-ray luminosity from $L_X (2-8keV) = \sim 10^{33.3 - 37.7} erg s^{-1}$ \citep[][and references therein]{mikles06= }. The superorbital periods of these systems are 140 d, 142 d, and 93 d res= pectively and their orbital periods are 2.1 d, 5.6 d, and 8.9 d \citep{og01= }. \citet{sood07} interprets these superorbital periods as unstable. Of the= $\sim$20 sources for which both the orbital and superorbital period are kn= own, no definitive empirical trend defines the relationship \citep[see Fig.= 1 of][]{sood07}.=20 The morphology of CXOGC~J174536.1-285638's X-ray light-curve is not incons= istent with that caused by a precessing accretion disk, in that the flux ap= pears to vary uniformly in the hard and soft X-rays. However, there is pres= ently no direct observational test to confirm that a period is superorbital= rather than orbital. In order to verify the presence and physical cause of= a superorbital period, additional physical parameters of the system are re= quired, including the mass ratio of the system, the inclination of the disk= with respect to the orbital plane, the orbital period, and the orbital sep= aration \citep{clarkson03}. Thus, while we cannot rule out the possibility = that this periodicity is superorbital, as yet, we do not have sufficient in= formation to place meaningful constraints on the superorbital hypothesis. T= hus for the remainder of this discussion, we restrict ourselves to explorin= g the possibility that the 189~d period is orbital rather than superorbital= =2E \subsection{The Orbital Period Assumption} For both the CWB and HMXB cases, the X-ray periodicity can trace the orbit= al period. CWBs have periods of days to years while HMXBs have shorter peri= ods ranging from hours to days \citep{van98,mc03}. %Several HMXBs in the \c= itet{liu06} catalog have orbital periods of $>100~d$. These long-period sys= tems are often X-ray transients with Be~star counterparts in eccentric orbi= ts \citep{lewin97}. In Paper I, we determine an absolute IR magnitude $M_K =3D -7.6 \pm 0.3$~m= ag for CXOGC~J174536.1-285638 using a distance of 8~kpc, reddening of $A_K = =3D3.4$, and a 2MASS magnitude of $K_S =3D10.33$~mag. Given that the source= appears blended in 2MASS, we verify the magnitude using the UKIDSS Galacti= c Plane Survey where the source is clearly resolved \citep{lawrence07,lucas= 08}. The UKIDSS survey lists the magnitude as $K =3D 10.390 \pm 0.001$~mag,= which is consistent with 2MASS, given the photometric transform between th= e relevant filters in these two surveys is $<0.1$ mag.=20 We can use CXOGC~J174536.1-285638's exceptional brightness and the X-ray p= eriod to place constraints on the nature of the system. For our purposes, t= he ``primary'' star (mass, $M_{OB}$) will refer to the massive OB-star and = the ``secondary'' star (mass, $M_2$) will refer to the companion whose natu= re has yet to be identified.=20 Using the mass function \begin{displaymath} f(q,i)=3D\frac{(q \sin{i})^3}{(1+q)^2} =3D \frac{P v_{orb}^3}{2 \pi G M_{OB= }} \end{displaymath} where $q=3DM_2/M_{OB}$, we can generate a parameter space of orbital veloci= ties and mass ratios for the system. Massive OB stars can range from $20 - = 100 M_\odot$ and still emit strongly in the IR \citep[see, e.g.][]{cox00,gi= rardi02}.=20 In Figure \ref{fig11}, we plot the mass ratio as a function of the inferred= orbital velocity for a range of primary masses and note that the orbital v= elocity is less than our IR spectral resolution of $70$~$km/s$ for cases of= mass ratio $q<0.5$. Even for higher mass ratios, a radial velocity variati= on would have low signal-to-noise with our current observations. Thus, we r= equire higher resolution spectroscopy in order to observe radial velocity v= ariations in the IR associated with this periodicity. In the next two sections we discuss the possibility that the modulations in= X-ray flux are caused by (1) obscuration of the X-ray source by stellar wi= nd; and (2) eclipse of the X-ray source.=20 \subsubsection{Wind Obscuration Scenario} Wind obscuration resulting in variable column absorption may be responsible= , in part, for the X-ray flux variations observed in CXOGC~J174536.1-285638= =2E This assumption would be most practically tested by analyzing the chang= e in hardness, as softer X-ray photons are absorbed preferentially. Such an= alysis is hindered by the relative faintness of the X-ray source, i.e., the= low count rate. For the spectra shown in Figure \ref{fig6}, the total inte= gration time for each observation is 40 hours. For the higher flux observat= ion on August 31, 2004, we observe a hardness ratio of $0.11 \pm 0.06$, whe= re the soft counts are summed from $2-4$~keV, the hard counts $4-8$~keV. Th= e hardness ratio is $(S-H)/(S+H)$ and the error is estimated from Poisson n= oise. For the second spectrum at the lower flux stage, taken on March 30, 2= 004, the hardness ratio is $0.01 \pm0.09$. The errors of these two measurem= ents make them consistent with no change in hardness. However, the low coun= t rate makes it difficult to estimate the robustness of this result.=20 Energy-independent X-ray variations in the spectrum could result if electro= n scattering is an important source of absorption. By testing the possibili= ty that an obscuring wind is solely responsible for the flux variations, we= can find the upper limit of the mass loss rate of the massive star compone= nt of the system. If wind obscuration is only partially responsible for the= flux variation, a lower mass-loss rate results. Thus, here, we are determi= ning the most extreme wind-producing source required to produce the flux va= riations we observe. =20 CXOGC~J174536.1-285638's X-ray light-curve shows a maximum flux variation = by a factor of 4 over the course of the 189~d period. Using this informatio= n, if we assume that the X-ray emitting source is being obscured by a windy= counterpart, we can calculate the column density of the wind required to c= ause such absorption. Because there are insufficient counts in the low-flux= state to fit the X-ray spectrum, we use the model fit from the high-flux s= tate and create a dummy response with XSPEC to measure the amount of absorp= tion required to decrease the flux by a factor of four. Given our initial $= N_H =3D 5.2 \times 10^{22} cm^{-2}$ (see Paper I), we find the column densi= ty from the obscuring wind must reach $N_H \approx 2.5 \times 10^{23} cm^{-= 2}$ to cause the flux variation observed in CXOGC~J174536.1-285638.=20 To estimate the absorption column caused by a dense stellar wind, we use th= e equation: \begin{equation} N_H =3D \int_R^\infty \rho(r) dl. \end{equation} For a spherically symmetric shell, and a star with mass-loss rate $\dot{M}$= and escape velocity $V_\infty$, \begin{equation} \rho(r) =3D \frac{\dot{M}}{4\pi r^2 v_\infty}. \end{equation} For an edge-on view of the system, $dl =3D dr$, thus \begin{equation} N_H =3D \int_R^\infty \frac{\dot{M}}{4\pi v_\infty}\frac{dr}{r^2} =3D \frac= {\dot{M}}{4\pi v_\infty R_{OB}}. \end{equation} Normalizing for typical values of $v_\infty =3D1000km/s$ and $\dot{M} =3D10= ^{-6}M_\odot yr^{-1}$ \citep[see, e.g.][]{mok07}, this becomes \begin{equation} \label{eq:nhmdot} \frac{N_H}{10^{23} cm^{-2}} =3D 4.3 \times \left(\frac{\dot{M}}{10^{-6}M_\o= dot yr^{-1}}\right) \left(\frac{v_\infty}{1000km s^{-1}}\right)^{-1} \left(= \frac{R_{OB}}{R_\odot}\right)^{-1}. \end{equation} If we are not viewing the system edge-on, we must take into account the ang= le through which we are viewing the wind as an effect on the observed absor= ption column. We can parameterize this in terms of an impact factor $b$ suc= h that $b=3Dr\cos{\theta}$. In this case, $dl =3D bd\theta$ and=20 \begin{equation} N_H =3D \frac{\dot{M}}{4 \pi v_\infty b} \int_{\theta_0}^{\pi/2} \cos^2{\th= eta}d\theta =3D \frac{\dot{M}}{4 \pi v_\infty b} \left[\frac{\pi}{2} - arcc= os\frac{b}{R} -\frac{b\sqrt{R^2-b^2}}{R^2}\right], \end{equation} where $\cos{\theta_0} =3D b/R$. Larger impact values require windier stars = to create the same absorption column, thus the values of $\dot{M}$ estimate= d with Equation \ref{eq:nhmdot} should be considered a lower limit of the $= \dot{M}$ required to produce the absorption column that causes the flux cha= nge in CXOGC~J174536.1-285638. We estimate the mass loss for two special cases. In the first case, we post= ulate the IR light is dominated by a single bright source. In HMXBs, the st= ar is expected to contribute more heavily to optical and IR emission than t= he accretion disk \citep{lewin97}. In certain CWB cases, especially of lowe= r mass ratios, it is possible that a single source dominates emission \cite= p{lepine}. Thus for CWB and HMXB scenarios in which a single star dominates= the IR emission, we use CXOGC~J174536.1-285638's IR luminosity and estimat= e stellar characteristics based on the isochrones of \citet{girardi02} and = find that a star with $M_{K} \sim -7.6$ will likely have a radius $R_{OB} \= sim 80 R_\odot$ valid for a range of masses $20-100 M_\odot$. Using equatio= n \ref{eq:nhmdot}, we get a mass-loss rate of $\dot{M} \sim 4 \times 10^{-5= } M_\odot yr^{-1}$. In the second case, we consider a system that contains = two massive stars, each contributing half of the IR luminosity which is onl= y consistent for CWBs containing two stars of similar bolometric luminosity= =2E These stars would have $R_{OB} \sim 20R_\odot$ and $\dot{M} \sim 1 \tim= es 10^{-5} M_\odot yr^{-1}$. Typical massive O-stars are reported to have m= ass-loss rates of $10^{-6} - 10^{-5} M_\odot yr^{-1}$ \citep{mok07}. Thus, = even in the most extreme case, where the flux variation is caused entirely = by absorption, a relatively windy star is necessary to produce the flux var= iations that we observe, but the mass-loss rate is not unreasonable. \subsubsection{The Eclipsing Binary Scenario} Assuming that the X-ray variability is caused by an eclipse has the greates= t potential for constraining the nature of the system components, and also = involves the most stringent physical constraints. We note that the X-ray li= ght curve (Fig. \ref{fig7}) is atypical for a standard eclipsing source, bo= th in the morphology of the dip and the phase duration of the low flux stat= e. In a HMXB or CWB, the X-ray emitting region is small compared to the mas= sive star. For a binary system in circular orbit, the eclipse of the X-ray = region causes a decrease in X-ray emission that is relatively brief compare= d to the orbital period. For a binary system in an elliptical orbit, it is = likely that the X-ray emitting region would experience periodic enhancement= while the sources are in close approach. Our source spends approximately e= qual time at the high flux and low flux stage and transitions smoothly betw= een the two. Despite this, we find it useful to explore the eclipsing assum= ption, as it allows us to define the limits of system in which the variatio= n is caused by a combination of multiple effects (e.g., an eclipse plus win= d obscurration). By assuming that the low-flux portion of the dip is caused by an eclipse of= the X-ray region, we estimate a transit time of $\tau \sim 50$ d for the p= utative eclipse, limited by adjacent observations of the high-flux stage. W= e convert the transit time to a velocity by estimating $v_{orb} =3D 2 R_{OB= }/ \tau$. Combining this with the mass function, we get \begin{equation} \label{eq:massfunction} \frac{(q \sin{i})^3}{(1+q)^2} =3D \frac{4 P R_{OB}^3}{\pi G M_{OB} \tau^{3}= } =3D 6.5\times 10^{-7} \frac{(P/189d)}{(\tau/50d)^3}\frac{r_{OB}^3}{m_{OB}= }, \end{equation} where $m_{OB}$ and $r_{OB}$ are in units of solar masses and solar radii re= spectively. Assuming $\sin{i} =3D1$, we then solve the cubic equation for d= ifferent scenarios. In Table \ref{tbl-4}, we list a series of mass ratios, = $q$, associated with varying fractions $r_{OB}^3/m_{OB}$. As an example, we= can examine the two cases as we did above. To complete this numerical exer= cise, we choose a median primary mass $M_{OB} =3D40M_\odot$ (while acknowle= dging that a wide range of masses is possible).=20 If two massive stars each contribute half of the IR luminosity, then $R_{OB= } \sim 20R_\odot$, $r_{OB}^3/m_{OB} =3D 200$, and the mass ratio is $q \sim= 0.05$. This resulting mass ratio is inconsistent with our initial assumpti= on of two massive stars contributing equally to the emission. If a single m= assive star dominates the IR emission, then $R_{OB} \sim 80 R_\odot$, $r_{O= B}^3/m_{OB} =3D 12800$, and the mass ratio is $q \sim 0.2$. We find that ad= justing the inclination does not significantly alter this result because ``= eclipsing'' scenarios do not exist at low inclinations \citep[$i<82^o$;][]{= terrell05}. In Figure \ref{fig12}, we plot the mass ratio as a function of transit time= , to explore the possibility that only a fraction of the flux variation is = caused by an eclipse of the X-ray source. We convert the transit time into = an orbital velocity using the radii $20 R_\odot$ and $80 R_\odot$ as we did= above to represent systems where two massive stars contribute to the IR lu= minosity and systems where a single source dominates the IR emission. We fi= nd that for transit times above $\sim$10 days ($v_{orb} < 130 km/s$), the s= ystem is consistent with low mass ratios ($q < 0.4$). In systems where two = stars are contributing equally to the IR luminosity (valid only for CWBs), = the transit time would be $< 2$ days, corresponding to an orbital velocity = $v_{orb} > 160$ days. Variability of this nature and on this timescale shou= ld have been apparent in our IR observations. Since we do not see those var= iations, we find eclipsing scenarios more likely for systems with lower mas= s ratios. Thus if the system is a CWB, it would have to have a relatively low mass ra= tio with the IR emission dominated by a single source. This implies that th= e wind emission of one source overwhelms that of its companion \citep{luo90= }. It is possible for CWBs to have lower mass ratios if the secondary is a = Wolf-Rayet (WR) star. By the time a massive star reaches the WR stage, it m= ay have a relatively small mass, but still have enormously powerful winds \= citep{crowther07}. For example, $\gamma ^2 Velorum$ is a WR+O star with a m= ass ratio $q \sim 0.35$ \citep{vanderhutch01}. In the case of $\gamma ^2 Ve= lorum$, the WR star dominates the IR emission, so the source appears Helium= rich \citep{crowther07}. It is possible that the Helium emission we observ= e in CXOGC~J174536.1-285638 is evidence of an obscured WR companion. Howeve= r, because Brackett series emission rather than Helium emission dominates t= he IR spectrum, we find this scenario less likely. In Table \ref{tbl-5}, we= list line ratios of Br-$\gamma$ to HeI 2.114$\mu$m and Br-$\gamma$ to HeII= 2.189$\mu$m for known CWBs and XRBs. In known WR+O binaries, the HeII 2.18= 9$\mu$m is notably stronger than Br-$\gamma$. Comparatively, CXOGC~J174536.= 1-285638 has much stronger Br-$\gamma$ emission, and hence a quite differen= t Br-$\gamma$/ HeII line ratio from what is observed in WR+O systems. In fa= ct, we note the Br-$\gamma$/HeI and Br-$\gamma$/HeII line ratios in CXOGC~J= 174536.1-285638 are more consistent with HMXBs than either O+O or O+WR CWBs= =2E Thus if CXOGC~J174536.1-285638 is a WR+O CWB, it is very unusual. In th= e eclipsing binary scenario, CXOGC~J174536.1-285638 would more likely be an= HMXB. \subsubsection{CXOGC~J174536.1-285638 as a Wind-Accreting HMXB} In Paper I, we showed that the X-ray luminosity of CXOGC~J174536.1-285638 (= $1.1 \times 10^{35}$ erg~s$^{-1}$) is consistent with HMXBs, within the obs= erved range of X-ray luminosities between INTEGRAL sources identified as HM= XBs \citep[$\sim 10^{34}$ erg~s$^{-1}$;][]{tomsick06, sidoli06} and the can= onically bright sources such as Cyg~X-1 and Cen~X-3 \citep[$\sim 10^{37}$ e= rg~s$^{-1}$;][]{nag92,schulz02}. We explore the implications of the observe= d period for the case where CXOGC~J174536.1-285638 is an accreting binary s= ystem with a compact object. Since the IR data suggest that CXOGC~J174536.= 1-285638 contains a high-mass star, we focus on the case of wind-fed accret= ion.=20 Taking the standard accretion luminosity as=20 \begin{equation} L_X =3D \epsilon \dot{M} c^2, \end{equation} where $\epsilon$ is the efficiency of converting energy into X-ray light an= d $\dot{M}$ is the accretion rate, we can rewrite this in terms of the mass= loss rate of the donor star due to wind such that \begin{equation} L_X \approx 5.7\times 10^{37} \epsilon \left(\frac{\dot{M}}{-10^{-4}\dot{M}= _{wind}}\right)\left(\frac{-\dot{M}_{wind}}{10^{-5}M_\odot yr^{-1}}\right)e= rg~s^{-1}. \end{equation} We have normalized the mass loss rate of the primary due to wind and the ac= cretion efficiency of the system with typical values found in \citet{fkr02}= =2E \citet{fkr02} estimate the accretion efficiency, $\dot{M}/ \dot{M}_{win= d}$, by comparing the mass flux within an accretion cylinder to the total m= ass loss of the donor star. The accretion cylinder is estimated from the gr= avitational potential of the compact object, giving \begin{displaymath} \frac{\dot{M}}{-\dot{M}_{wind}} =3D \frac{\pi r_{acc}^2 v_{wind}(a)}{4 \pi = a^2 v_{wind}(a)} \end{displaymath} where $r_{acc} \sim 2 G M_2 / v_{wind}^2$, $v_{wind} \sim (2 G M_1 /R_1)^{1= /2}$, and $a$ is the orbital separation. This gives us \begin{equation} \frac{\dot{M}}{-\dot{M}_{wind}} \simeq \frac{1}{4}\left(\frac{M_{2}}{M_{OB}= }\right)^2\left(\frac{R_{OB}}{a}\right)^2. \end{equation} Normalizing to standard values, and using our known values, we get \begin{equation} \label{eq:qmdotlx} \frac{L_X}{10^{35} erg~s^{-1}} \approx 35 \epsilon \left[\frac{r_{OB}^3}{m_= {OB}} \frac{q^3}{(1+q)}\right] \left[\frac{-\dot{M}_{wind}}{10^{-5}M_\odot = yr^{-1}}\right]\left[\frac{P}{189~d}\right], \end{equation} where $r_{OB}$ and $m_{OB}$ are normalized to solar radii and solar masses = respectively. This form is useful for exploring the scenarios put forth in = the previous sections. Because we are considering a wind-accreting HMXB, we= use our previous estimate where a single massive star dominates the system= , for mass between $20 - 100 M_\odot$ and radius $R \sim 80 R_\odot$. The wind obscuration scenario gave an estimate of $\dot{M}_{wind} \approx 4= \times 10^{-5} M_\odot/yr$. We can then use Equation \ref{eq:qmdotlx} and = find that the mass ratio of the system is $q \sim 0.01$. This suggests a ma= ssive $M > 80 M_\odot$ donor for a typical neutron star companion. By relax= ing the estimate of the massive star radius, $R_{OB}$, we find that $q$ wil= l increase and more compact object solutions exist over a wider range of pr= imary masses. The estimate of $R_{OB} =3D 80 R_\odot$ is derived from the g= ravitational potential as estimated by \citet{girardi02}. In Table \ref{tbl= -6}, we list a series of solutions for Equation \ref{eq:qmdotlx}. Because the eclipsing scenario case places firm constraints on the mass rat= io of the system, we use Equation \ref{eq:qmdotlx} to calculate the mass lo= ss rates associated with various scenarios. We list those values in Table \= ref{tbl-4}. For the case where the mass ratio is $q \sim 0.2$, the associat= ed mass loss rate is low ($\dot{M}_{wind} \sim 2 \times 10^{-7} M_\odot /yr= $), for an efficiency $\epsilon \sim 0.1$. This is not unreasonable for mas= sive stars \citep{mok07}. Interestingly, in both the wind obscuration and t= he eclipsing binary scenario, the X-ray luminosity is consistent with a low= mass ratio for the system. \section{Conclusions} We have searched for evidence of periodic variability in the IR spectra and= long-term X-ray light-curve of the GC X-ray source CXOGC~J174536.1-285638.= We find no evidence of IR variability on short ($<3~d$) timescales or betw= een the 2005 and 2006 spectra. We compare the IR line ratios Br-$\gamma$/He= I and Br-$\gamma$/HeII in CXOGC~J174536.1-285638 to known HMXBs and CWBs an= d find the relative emission line strengths to be more consistent with an H= MXB. We have identified an apparent $189 \pm 6$ d period in the CXOGC~J1745= 36.1-285638 X-ray light curve. We find no evidence of periodic X-ray variab= ility at timescales less than 189~d. Using a Monte Carlo simulation, we tes= t the significance of the 189~d period detection; despite our fairly sparse= time sampling, we find this period is significant with a confidence level = greater than 99.997\%. We explore several interpretations of the X-ray modu= lation. It is plausible, if the source is a HMXB, that the periodic modulation is s= uperorbital in nature and related to a precessing accretion disk, in which = case, further observations are required to determine the orbital period of = the system and thus the nature of the system components. If the source is a= HMXB and the 189~d period is superorbital, then we expect to find a shorte= r orbital period. This putative orbital periodicity is not necessarily obse= rvable in the IR as in this scenario, the IR emission is dominated by a sin= gle bright source. If the orbital period is detectable in the X-ray, target= ed observations with a sensitive detector over a time interval of $\sim 1-2= $ weeks during the high flux stage are required to ensure sufficient counts= to test for variability. We also explore an orbital period interpretation and summarize scenarios fo= r this in Table \ref{tbl-7}. If the observed period is orbital in nature, a= nd the X-ray modulation is caused by obscuration of the X-ray source due to= a dense wind, then CXOGC~J174536.1-285638 is consistent with both CWB and = HMXB interpretations. The further constraint of the X-ray luminosity is con= sistent with a massive ($M_{OB} > 80 M_\odot$) donor with a neutron star co= mpanion. If X-ray modulation is caused by an eclipse, the mass ratio is low= and CXOGC~J174536.1-285638 is more consistent with an HMXB interpretation.= If the 189~d period is orbital, we may be able to identify the source natu= re by obtaining long term photometric observations in the IR. Also, targete= d IR follow-up spectroscopy to cover multiple phases of the source period w= ill allow us to search for a relationship between the X-ray and IR variabil= ity in this system. In the low flux phase, additional IR spectroscopic line= features (e.g., absorption, P~Cygni variation) may become apparent that ca= n help us discern the nature of the stellar components. Recently, \citet{hyodo08} reported the discovery of an early-type, Galactic Center source which appears to have many characteristics in common with CXOGC~J174536.1-285638. The source, CXOGC J174645.3-281546, has an unusually strong FeXXV line ($\sim$1~keV), shows X-ray variability of a factor of $\sim 2$ on a $\sim$year timescale, and appears to have a high-mass star as its likely IR counterpart. As in CXOGC~J174536.1-285638, its X-ray to IR luminosity ratio is $\sim 10^{-4}$. These intriguing similarities in X-ray spectral appearance, variability timescale, and luminosity lead us to suggest that it would be interesting in future observations to study this source in concert with CXOGC~J174536.1-285638. Although there are only two sources with these properties known at present, it is possible that they could ultimately define a new (sub)class of early-type Galactic sources with strong FeXXV emission. \acknowledgements The authors make use of observations obtained with XMM-Newton, an ESA scien= ce mission with instruments and contributions directly funded by ESA Member= States and NASA. Authors are Visiting Astronomers at the Infrared Telescope Facility, which = is operated by the University of Hawaii under Cooperative Agreement no. NCC= 5-538 with the National Aeronautics and Space Administration, Office of Sp= ace Science, Planetary Astronomy Program. Many thanks to the IRTF support s= taff who assisted us with remote observing on this run.=20 VJM, SSE, and RMB are supported in part by an NSF Grant (AST-0507547). MPM was supported by the National Aeronautics and Space Administration thro= ugh Chandra Award Number=A0GO6-7135 issued by the Chandra X-ray Observatory= Center, which is operated by the Smithsonian Astrophysical Observatory for= and on behalf of the National Aeronautics Space Administration under contr= act NAS8-03060. \begin{thebibliography}{} \bibitem[Beckmann et al.(2005)]{beck05} Beckmann, V., et al. \ 2005, \apj, 631, 506=20 \bibitem[Benaglia et al.(2001)]{ben01} Benaglia, P., Cappa,=20 C.~E., \& Koribalski, B.~S.\ 2001, \aap, 372, 952=20 \bibitem[Bodaghee et al.(2007)]{bod07} Bodaghee, A., et al. \ 2007, \aap, 467, 585=20 \bibitem[Clark \& Dolan(1999)]{clark99} Clark, L.~L., \& Dolan,=20 J.~F.\ 1999, \aap, 350, 1085=20 \bibitem[Clark et al.(2003)]{clark03} Clark, J.~S., Charles,=20 P.~A., Clarkson, W.~I., \& Coe, M.~J.\ 2003, \aap, 400, 655 \bibitem[Clarkson et al.(2003)]{clarkson03} Clarkson, W.~I.,=20 Charles, P.~A., Coe, M.~J., Laycock, S., Tout, M.~D.,=20 \& Wilson, C.~A.\ 2003, \mnras, 339, 447=20 \bibitem[Cohen(2000)]{cohen00} Cohen, D.~H.\ 2000, IAU=20 Colloq.~175: The Be Phenomenon in Early-Type Stars, 214, 156=20 \bibitem[Corbet(1986)]{corbet86} Corbet, R.~H.~D.\ 1986, \mnras,=20 220, 1047 \bibitem[Cox(2000)]{cox00} Cox, A.~N.\ 2000, Allen's=20 Astrophysical Quantities, 4th ed.~Publisher: New York: AIP Press; Springer,= =20 2000.~Edited by Arthur N.~Cox.~ ISBN: 0387987460 \bibitem[Crowther(2007)]{crowther07} Crowther, P.~A.\ 2007, \araa,=20 45, 177=20 \bibitem[Cushing et al.(2004)]{cushing04} Cushing, M.~C., Vacca,=20 W.~D., \& Rayner, J.~T.\ 2004, \pasp, 116, 362=20 \bibitem[Davidson et al.(1998)]{dav98} Davidson, K.,=20 Ishibashi, K., \& Corcoran, M.~F.\ 1998, New Astronomy, 3, 241=20 \bibitem[De Becker et al.(2006)]{debeck06} De Becker, M., Rauw,=20 G., Sana, H., Pollock, A.~M.~T., Pittard, J.~M., Blomme, R., Stevens,=20 I.~R., \& van Loo, S.\ 2006, \mnras, 371, 1280=20 \bibitem[Eikenberry et al.(2001)]{eiken01} Eikenberry, S.~S.,=20 Cameron, P.~B., Fierce, B.~W., Kull, D.~M., Dror, D.~H., Houck, J.~R.,=20 \& Margon, B.\ 2001, \apj, 561, 1027=20 \bibitem[Fender et al.(2003)]{fender03} Fender, R., Migliari,=20 S., \& M{\'e}ndez, M.\ 2003, New Astronomy Review, 47, 481=20 \bibitem[Figer et al.(1997)]{figer97} Figer, D.~F., McLean, I.~S., \& Najar= ro, F.\ 1997, \apj, 486, 420=20 \bibitem[Filliatre \& Chaty(2004)]{fill04} Filliatre, P., \&=20 Chaty, S.\ 2004, \apj, 616, 469=20 \bibitem[Frank et al.(2002)]{fkr02} Frank, J., King, A., \&=20 Raine, D.~J.\ 2002, Accretion Power in Astrophysics, by Juhan Frank and=20 Andrew King and Derek Raine, pp.~398.~ISBN 0521620538.~Cambridge, UK:=20 Cambridge University Press, February 2002. \bibitem[Girardi et al.(2002)]{girardi02} Girardi, L., Bertelli,=20 G., Bressan, A., Chiosi, C., Groenewegen, M.~A.~T., Marigo, P., Salasnich,= =20 B., \& Weiss, A.\ 2002, \aap, 391, 195=20 \bibitem[Hanson et al.(1996)]{hanson96} Hanson, M.~M., Conti,=20 P.~S., \& Rieke, M.~J.\ 1996, \apjs, 107, 281=20 \bibitem[Horne \& Baliunas(1986)]{horne86} Horne, J.~H., \&=20 Baliunas, S.~L.\ 1986, \apj, 302, 757=20 \bibitem[Hyodo et al.(2008)]{hyodo08} Hyodo, Y., Tsujimoto, M.,=20 Koyama, K., Nishiyama, S., Nagata, T., Sakon, I., Murakami, H.,=20 \& Matsumoto, H.\ 2008, \pasj, 60, 173=20 \bibitem[Kirsch et al.(2005)]{kirsch05} Kirsch, M.~G., et al.\=20 2005, \procspie, 5898, 212=20 \bibitem[Lawrence et al.(2007)]{lawrence07} Lawrence, A., et al.\=20 2007, \mnras, 379, 1599=20 \bibitem[L{\'e}pine et al.(2001)]{lepine} L{\'e}pine, S.,=20 Wallace, D., Shara, M.~M., Moffat, A.~F.~J., \& Niemela, V.~S.\ 2001, \aj,= =20 122, 3407 \bibitem[Lewin et al.(1997)]{lewin97} Lewin, W.~H.~G., van=20 Paradijs, J., \& van den Heuvel, E.~P.~J.\ 1997, X-ray Binaries, Edited by= =20 Walter H.~G.~Lewin and Jan van Paradijs and Edward P.~J.~van den Heuvel,=20 pp.~674.~ISBN 0521599342.~Cambridge, UK: Cambridge University Press,=20 January 1997. \bibitem[Lewin \& van der Klis(2006)]{mc03} Lewin, W.~H.~G.,=20 \& van der Klis, M.\ 2006, Compact stellar X-ray sources. \bibitem[Liu et al.(2006)]{liu06} Liu, Q.~Z., van Paradijs,=20 J., \& van den Heuvel, E.~P.~J.\ 2006, \aap, 455, 1165=20 \bibitem[Lucas et al.(2007)]{lucas08} Lucas, P.~W., et al.\=20 2007, ArXiv e-prints, 712, arXiv:0712.0100 \bibitem[Luo et al.(1990)]{luo90} Luo, D., McCray, R., \& Mac=20 Low, M.-M.\ 1990, \apj, 362, 267 \bibitem[Mikles et al.(2006)]{mikles06} Mikles, V.~J.,=20 Eikenberry, S.~S., Muno, M.~P., Bandyopadhyay, R.~M., \& Patel, S \ 2006,= =20 \apj, 651, 408=20 \bibitem[Mineshige et al.(1994)]{min94} Mineshige, S., Ouchi,=20 N.~B., \& Nishimori, H.\ 1994, \pasj, 46, 97=20 \bibitem[Mokiem et al.(2007)]{mok07} Mokiem, M.~R., et al.\=20 2007, \aap, 465, 1003=20 \bibitem[Muno et al.(2003)]{mun03} Muno, M.~P., et al.\ 2003,=20 \apj, 589, 225=20 \bibitem[Muno et al.(2004a)]{mun04b} Muno, M.~P., et al.\ 2004a,=20 \apj, 613, 326=20 \bibitem[Muno et al.(2004b)]{mun04} Muno, M.~P., et al.\ 2004b,=20 \apj, 613, 1179=20 \bibitem[Muno et al.(2006)]{mun06} Muno, M.~P., Bauer, F.~E.,=20 Bandyopadhyay, R.~M., \& Wang, Q.~D.\ 2006, \apjs, 165, 173=20 \bibitem[Muno et al.(2006)]{mun05} Muno, M.~P., Bower, G.~C.,=20 Burgasser, A.~J., Baganoff, F.~K., Morris, M.~R.,=20 \& Brandt, W.~N.\ 2006, \apj, 638, 183=20 \bibitem[Nagase et al.(1992)]{nag92} Nagase, F., Corbet,=20 R.~H.~D., Day, C.~S.~R., Inoue, H., Takeshima, T., Yoshida, K., \& Mihara,= =20 T.\ 1992, \apj, 396, 147=20 \bibitem[Negueruela et al.(2005)]{neg05} Negueruela, I.,=20 Smith, D.~M., Reig, P., Chaty, S., \& Torrej{\'o}n, J.~M.\ 2006, The X-ray= =20 Universe 2005, 604, 165=20 \bibitem[Negueruela \& Schurch(2007)]{neg07} Negueruela, I.,=20 \& Schurch, M.~P.~E.\ 2007, \aap, 461, 631=20 \bibitem[Ogilvie \& Dubus(2001)]{og01} Ogilvie, G.~I., \&=20 Dubus, G.\ 2001, \mnras, 320, 485 \bibitem[Paul et al.(2000)]{paul00} Paul, B., Kitamoto, S.,=20 \& Makino, F.\ 2000, \apj, 528, 410=20 \bibitem[Predehl \& Schmitt(1995)]{ps95} Predehl, P., \&=20 Schmitt, J.~H.~M.~M.\ 1995, \aap, 293, 889 \bibitem[Rayner et al.(2003)]{rayner03} Rayner, J.~T., Toomey,=20 D.~W., Onaka, P.~M., Denault, A.~J., Stahlberger, W.~E., Vacca, W.~D.,=20 Cushing, M.~C., \& Wang, S.\ 2003, \pasp, 115, 362=20 \bibitem[Sana et al.(2004)]{sana04} Sana, H., Stevens, I.~R.,=20 Gosset, E., Rauw, G., \& Vreux, J.-M.\ 2004, \mnras, 350, 809=20 \bibitem[Schulz et al.(2002)]{schulz02} Schulz, N.~S., Cui, W.,=20 Canizares, C.~R., Marshall, H.~L., Lee, J.~C., Miller, J.~M., \& Lewin,=20 W.~H.~G.\ 2002, \apj, 565, 1141=20 \bibitem[Sidoli et al.(2006)]{sidoli06} Sidoli, L., Mereghetti,=20 S., Favata, F., Oosterbroek, T., \& Parmar, A.~N.\ 2006, \aap, 456, 287=20 \bibitem[Sood et al.(2007)]{sood07} Sood, R., Farrell, S.,=20 O'Neill, P., \& Dieters, S.\ 2007, Advances in Space Research, 40, 1528=20 \bibitem[Terrell \& Wilson(2005)]{terrell05} Terrell, D., \&=20 Wilson, R.~E.\ 2005, \apss, 296, 221=20 \bibitem[Timmer \& Koenig(1995)]{tk95} Timmer, J., \& Koenig, M.\ 1995, \aa= p, 300, 707=20 \bibitem[Titarchuk et al.(2007)]{tit07} Titarchuk, L.,=20 Shaposhnikov, N., \& Arefiev, V.\ 2007, \apj, 660, 556=20 \bibitem[Thaller(1997)]{thaller97} Thaller, M.~L.\ 1997, \apj,=20 487, 380=20 \bibitem[Tomsick et al.(2006)]{tomsick06} Tomsick, J.~A., Chaty,=20 S., Rodriguez, J., Foschini, L., Walter, R., \& Kaaret, P.\ 2006, \apj,=20 647, 1309=20 \bibitem[Vacca et al.(2003)]{vacca03} Vacca, W.~D., Cushing,=20 M.~C., \& Rayner, J.~T.\ 2003, \pasp, 115, 389=20 \bibitem[Vanbeveren et al.(1998)]{van98} Vanbeveren, D., de=20 Donder, E., van Bever, J., van Rensbergen, W., \& de Loore, C.\ 1998, New= =20 Astronomy, 3, 443=20 \bibitem[van der Hucht(2001)]{vanderhutch01} van der Hucht, K.~A.\=20 2001, New Astronomy Review, 45, 135=20 \end{thebibliography} %figures \clearpage \begin{figure*} \epsscale{0.8} \plottwo{f1a.eps}{f1b.eps} \caption{Left: A 15'' x 15'' image of the 2MASS region near the {\it Chandr= a} X-ray coordinate center. A 1.5 arcsec circle is drawn around the {\it C= handra} source coordinates. A second circle is drawn around the blended sou= rce. Right: A 15'' x 15'' IRTF slit image of the same region. The stars ble= nded in the 2MASS region are clearly resolved on the slit.} \label{fig1} \end{figure*} \begin{figure*} %\epsscale{0.8} \plotone{f2.eps} \caption{K-band spectra of CXOGC~J174536.1-285638 and the neighbor star. Th= e two objects are blended in 2MASS, but clearly resolved by IRTF. Source 1 = is the likely X-ray counterpart. Source 2 is a type K or cooler evolved sou= rce, lacking emission lines which would be indicative of energetic processe= s.} \label{fig2} \end{figure*} \begin{figure*} \plotone{f3.eps} \caption{The K-band spectra of CXOGC~J174536.1-285638. We show the original= 2005 spectrum at the bottom and the twenty-minute combinations of the 2006= spectra over the three nights. These are offset by time of observation, su= ch that the earliest spectra are lower and later are higher. The relative t= imes of these spectra are listed in Table \ref{tbl-1}.} \label{fig3} \end{figure*} \begin{figure*} \epsscale{0.8} \plotone{f4.eps} \caption{The Br-$\gamma$ region of select CXOGC~J174536.1-285638 spectra ta= ken from 2006 Aug 02-04. The region shows apparent non-periodic variation, = mostly around the $2.164 \mu$m Helium contribution. These variations are on= ly occasionally greater than 5-times the RMS spectral difference. Higher re= solution spectroscopy is needed to show whether this is intrinsic to CXOGC~= J174536.1-285638 or an artifact of the data reduction. The relative times o= f these spectra are listed in Table \ref{tbl-1}.} \label{fig4} \end{figure*} \begin{figure*} \epsscale{0.8} \plotone{f5.eps} \caption{A 60'' x 60'' {\it XMM} image centered around the {\it Chandra} so= urce coordinates (denoted by the inner circle). The concentric circles deno= te the region of source counts and background counts used in analysis of th= e {\it XMM} data.} \label{fig5} \end{figure*} \begin{figure*} \includegraphics[angle=3D90,width=3D\textwidth]{f6.eps} \caption{Two representative {\it XMM} spectra separated by 0.7 in phase. Wh= ile the strength of the FeXXV line is consistent between the two observatio= ns, the continuum level drops significantly. If such variation were caused = entirely by column absorption due to a stellar wind, then $N_H$ would incre= ase by $2.5 \times 10^{23} cm^{-2}$.} \label{fig6} \end{figure*} \begin{figure*} \epsscale{0.8} \plotone{f7.eps} \caption{The X-ray light curve (top) and folded light curve (bottom) of CXO= GC~J174536.1-285638. The light curve is folded on a 189 d period. The squar= es are {\it XMM} data; the diamonds are {\it Chandra} data. The arrow indic= ates the data of the IR spectra.} \label{fig7} \end{figure*} \begin{figure*} \includegraphics[width=3D\textwidth]{f8.eps} \caption{A periodogram analysis of the X-ray light curve. The most signific= ant period is $189 \pm 6$ days. Subsequent peaks appear at integer multiple= s of this period.} \label{fig8} \end{figure*} \begin{figure*} %\epsscale{0.8} \plotone{f9.eps} \caption{Monte Carlo simulation testing the possibility of a random periodo= gram peak of the observed power (see Fig. \ref{fig8}) at this sampling. The= vertical line indicates the power of the original signal. We find that our= period is significant with a confidence level of 99.997\%.} \label{fig9} \end{figure*} \begin{figure*} %\epsscale{0.8} \plotone{f10.eps} \caption{Monte Carlo simulations testing for power peaks, as in Figure \ref= {fig9}, but assuming different levels of red noise in the system (see text)= =2E As more red noise is assumed in the observation, the strength of the si= gnal decreases.} \label{fig10} \end{figure*} \begin{figure*} \includegraphics[angle=3D90, width=3D\textwidth]{f11.eps} \caption{Using the mass function and the putative period of 189 days, we ca= lculate the expected mass ratio, $q=3DM_2/M_{OB}$, for primary masses $M_{O= B} =3D 20 - 100 M_\odot$. The primary mass is indicated to the left of each= line. The vertical dashed line represents the limiting IR spectral resolut= ion.} \label{fig11} \end{figure*} \begin{figure*} \includegraphics[angle=3D90, width=3D\textwidth]{f12.eps} \caption{In a similar manner to Figure \ref{fig11}, we compute the mass rat= io of the system for a variety of transit times related to the orbital velo= city of the system for primary sources ranging from $M_{OB} =3D 20 - 100 M_= \odot$. The solid lines indicate systems in which a single massive source d= ominates the IR emission ($R =3D 80 R_\sun$) and the dashed line is for two= massive sources contributing approximately equally to the emission ($R =3D= 20 R_\sun$). See details in text. } \label{fig12} \end{figure*} %tables \begin{deluxetable}{llll} \tablecaption{{\bf Observing Log: IR Spectra}} \tablewidth{0pt} \tablehead{\colhead{Obs ID} & \colhead{Date} & \colhead{Time (UT)} & \colhe= ad{Exposure Time (min)}} \startdata a & 2006-08-02 & 6:38 & 20 \\ b & 2006-08-02 & 7:05 & 20 \\ c & 2006-08-02 & 8:25 & 20 \\ d & 2006-08-02 & 8:55 & 16 \\ e & 2006-08-02 & 9:46 & 8 \\ f & 2006-08-03 & 5:27 & 16 \\ g & 2006-08-03 & 5:52 & 16 \\ h & 2006-08-03 & 6:37 & 20 \\ i & 2006-08-03 & 7:07 & 20 \\ j & 2006-08-03 & 7:49 & 20 \\ k & 2006-08-03 & 8:20 & 20 \\ l & 2006-08-03 & 9:02 & 20 \\ m & 2006-08-03 & 9:32 & 12 \\ n & 2006-08-04 & 5:51 & 20 \\ o & 2006-08-04 & 6:27 & 16 \\ p & 2006-08-04 & 6:46 & 16 \\ q & 2006-08-04 & 7:23 & 20 \\ r & 2006-08-04 & 7:58 & 16 \\ s & 2006-08-04 & 8:11 & 18 \\ t & 2006-08-04 & 8:47 & 20 \\ u & 2006-08-04 & 9:10 & 20 \\ =09=09=09=09 \tableline=09=09=09 \enddata=09=09=09 \tablecomments{These observation IDs are associated with Figures 1 and 2. T= he days align with days 2404-2406 on our X-ray light curves.} \label{tbl-1} \end{deluxetable} \begin{deluxetable}{ccccccc} \tablecaption{{\bf Observing Log: Chandra }} \tablewidth{0pt} \tablecolumns{7} \tablehead{Date & Time & Obs. ID & Exp. Time & R.A. & Declination & Roll \\ \multicolumn{2}{c}{(UT)} & & (ks) & \multicolumn{2}{c}{(J2000)} & (deg)} \startdata 2000-10-26 & 18:15:11 & 1561a & 35.7 & 266.41344 & -29.01281 & 264.7\\ 2001-07-14 & 01:51:10 & 1561b & 13.5 & 266.41344 & -29.01281 & 264.7\\ 2001-07-18 & 14:25:48 & 2284 & 10.6 & 266.40415 & -28.94090 & 283.8\\ 2002-05-22 & 22:59:15 & 2943 & 34.7 & 266.41991 & -29.00407 & 75.5\\ 2002-02-19 & 14:27:32 & 2951 & 12.4 & 266.41867 & -29.00335 & 91.5\\ 2002-03-23 & 12:25:04 & 2952 & 11.9 & 266.41897 & -29.00343 & 88.2\\ 2002-04-19 & 10:39:01 & 2953 & 11.7 & 266.41923 & -29.00349 & 85.2\\ 2002-05-07 & 09:25:07 & 2954 & 12.5 & 266.41938 & -29.00374 & 82.1\\ 2002-05-25 & 15:16:03 & 3392 & 165.8 & 266.41992 & -29.00408 & 75.5\\ 2002-05-28 & 05:34:44 & 3393 & 157.1 & 266.41992 & -29.00407 & 75.5\\ 2003-06-19 & 18:28:55 & 3549 & 24.8 & 266.42092 & -29.01052 & 346.8\\ 2002-05-24 & 11:50:13 & 3663 & 38.0 & 266.41993 & -29.00407 & 75.5\\ 2002-06-03 & 01:24:37 & 3665 & 89.9 & 266.41992 & -29.00407 & 75.5\\ 2004-07-05 & 22:33:11 & 4683 & 49.5 & 266.41605 & -29.01238 & 286.2\\ 2004-07-06 & 22:29:57 & 4684 & 49.5 & 266.41597 & -29.01236 & 285.4\\ 2004-08-28 & 12:03:59 & 5360 & 5.1 & 266.41477 & -29.01211 & 271.0\\ 2005-07-24 & 19:58:27 & 5950 & 48.5 & 266.41519 & -29.01222 & 276.7\\ 2005-07-27 & 19:08:16 & 5951 & 44.6 & 266.41512 & -29.01219 & 276.0\\ 2005-07-29 & 19:51:11 & 5952 & 43.1 & 266.41508 & -29.01219 & 275.5\\ 2005-07-30 & 19:38:31 & 5953 & 45.4 & 266.41506 & -29.01218 & 275.3\\ 2005-08-01 & 19:54:13 & 5954 & 18.1 & 266.41502 & -29.01215 & 274.9\\ 2005-02-27 & 06:26:04 & 6113 & 4.9 & 266.41870 & -29.00353 & 90.6\\ 2006-07-17 & 03:58:28 & 6363 & 29.8 & 266.41541 & -29.01228 & 279.5\\ 2006-04-11 & 05:33:20 & 6639 & 4.5 & 266.41890 & -29.00369 & 86.2\\ 2006-05-03 & 22:26:26 & 6640 & 5.1 & 266.41935 & -29.00383 & 82.8\\ 2006-06-01 & 16:07:52 & 6641 & 5.1 & 266.42018 & -29.00440 & 69.7\\ 2006-07-04 & 11:01:35 & 6642 & 5.1 & 266.41633 & -29.01237 & 288.4\\ 2006-07-30 & 14:30:26 & 6643 & 5.0 & 266.41510 & -29.01218 & 275.4\\ 2006-08-22 & 05:54:34 & 6644 & 5.0 & 266.41484 & -29.01202 & 271.7\\ 2006-09-25 & 13:50:35 & 6645 & 5.1 & 266.41448 & -29.01195 & 268.3\\ 2006-10-29 & 03:28:20 & 6646 & 5.1 & 266.41425 & -29.01178 & 264.4\\ \tableline \enddata \label{tbl-2} \end{deluxetable} \begin{deluxetable}{llll} \tablecaption{{\bf Observing Log: XMM-Newton }} \tablewidth{0pt} \tablecolumns{11} \tablehead{\colhead{Observation ID} & \colhead{Date} & \colhead{Time (h)} = & \colhead{Exposure Time (h)}} \startdata 0112972101 & 2001-09-04 & 01:19:34 & 7.5 \\ 0111350101 & 2002-02-26 & 03:11:27 & 14 \\ 0111350301 & 2002-10-03 & 06:36:49 & 5 \\ 0202670501 & 2004-03-28 & 14:37:16 & 40 \\ 0202670601 & 2004-03-30 & 14:29:07 & 40 \\ 0202670701 & 2004-08-31 & 02:54:31 & 40 \\ 0202670801 & 2004-09-02 & 02:44:08 & 40 \\ \tableline \enddata \label{tbl-3} \end{deluxetable} \begin{deluxetable}{lll} \tablecaption{{\bf Mass ratio estimations for the eclipsing scenario}} \tablewidth{4in} \tablehead{\colhead{$r_{OB}^3 / m_{OB}$} & \colhead{$M_2/M_{OB}$} & $M_{win= d}/ M_\odot~yr^{-1}$ \\} \startdata $10^5$ & 0.5 & $8 \times 10^{-9}$ \\ $10^4$ & 0.2 & $2 \times 10^{-7}$ \\ $10^3$ & 0.09 & $4 \times 10^{-6}$ \\ $10^2$ & 0.04 & $1 \times 10^{-4}$ \\ $10^1$ & 0.02 & $3 \times 10^{-2}$ \\ \tableline \enddata \tablecomments{The mass ratio expected for a primary of the given mass to r= adius ratio in the eclipsing binary scenario. In Column 1, the ratios are i= n units of $R_\odot^3/M_\odot$. Values of $r_{OB}^3 / m_{OB} > 10^{4}$ are = more typical of brighter stars ($M_K \sim -7.6$) and thus consistent with c= ases where a single massive star is dominating CXOGC~J174536.1-285638's IR = emission. Values of $r_{OB}^3 / m_{OB} < 10^{4}$ are more consistent with $= M_K \sim -4$ stars such that CXOGC~J174536.1-285638's IR emission is compos= ed of the flux from two bright stars. The estimation of $M_{wind}$ is based= on Equation \ref{eq:qmdotlx}, which is only valid for the HMXB case.} \label{tbl-4} \end{deluxetable} \begin{deluxetable}{lllllll} \tablecaption{{\bf Infrared Line Ratios}} \tablewidth{\textwidth} \scriptsize \tablehead{Source & \multicolumn{3}{l}{Equivalent Width ($\AA$)} & Ref. & B= r$\gamma$/HeI & Br$\gamma$/HeII\\ & HeI & Br$\gamma$ & HeII & & & \\ & 2.114$\mu$m & 2.166$\mu$m & 2.189$\mu$m & & & } \startdata CXOGC~J174536.1-285638 & 13.8 & 36.6 & $<$2 & 1 & 2.65 & $>$18.3 \\ \tableline {\bf HMXB} & & & & & & \\ Cir X-1 & & 24.2 & 1.3 & 2 & & 18.62 \\ IGR J16318-4848 (sgB[e]) & 5 & 45 & & 5 & 9 & \\ HD 34921 (B0I) & 1 & 6 & & 6 & 6 & \\ HD 24534 (O9III-Ve) & 2.7 & 14.5 & & 2 & 5.37 & \\ EXO2030+375 & 1.7 & 4 & & 2 & 2.35 & \\ V725Tau (O9.7IIe) & & 13 & $<1$ & 6 & & $>13$ \\ \tableline {\bf O+O} & & & & & & \\ HD 93205 (O3V)=09=09& &=092 &=091.1 & =096 & & 1.82 \\ HD 206267 (O6.5V)=09& &=091.2 &=090.4 &=096 & & 3 \\ HD 152248 (O7Ib)=09& &=094 &=091.8 &=096 & & 2.22 \\ HD 57060 (O7Ia)=09=09& &=095 &=091.1 &=096 & & 4.55 \\ HD 47129 (O8)=09=09& &=097 &=09$<0.5$ &=096 & & $>14$\\ HD 37043 (O9III)=09& &=091.6 &=090.2 &=096 & & 8\\ HD 47129(O7.5I+O6I)=09& &=097 &=09$<0.5$ &=096 & & $>14$\\ HD 15558 (O5III)=09& &=091.4 &=090.4 &=096 & & 3.5 \\ HD 199579 (O6V)=09=09& &=091.4 &=090.6 &=096 & & 2.33 \\ \tableline {\bf O+WR} & & & & & & \\ WR138 (WN5+O9)=09 & 12 &=0934 &=0952 &=094 &=092.83 &=090.65 \= \ WR139 (WN5+O6)=09 & 15 &=0928 &=0966 &=094 &=091.87 &=090.42\\ WR133 (WN4.5+O9.5) &=09 &=0930 &=0920 &=094 &=09 &=090.63\\ WR127 (WN4+O9.5) & 16 &=0941 &=0977 &=094 &=092.56 &=090.53\\ WR151 (WN4+O8)=09 & 16 &=0936 &=0981 &=094 &=092.25 &=090.44\\ \tableline=09=09=09 \enddata=09=09=09 \tablecomments{IR line ratios. We compare the relative strength of HeI and = HeII lines to Br-$\gamma$ in CXOGC~J174536.1-285638 and a selection of HMXB= s and CWBs. Note that the HeII 2.189$\mu$m line in CXOGC~J174536.1-285638 h= as a P Cygni profile. We group O+O and O+WR binaries separately, as the for= mer systems are less likely to produce low mass ratios. In known WR+O syste= ms, the Br-$\gamma$/HeII line ratio is significantly different than that ob= served in CXOGC~J174536.1-285638. REFERENCES - (1) Paper 1; (2) \citet{clar= k99}; (3) \citet{clark03}; (4) \citet{figer97}; (5) \citet{fill04}; (6) \ci= tet{hanson96}.} \label{tbl-5} \end{deluxetable} \begin{deluxetable}{llll} \tablecaption{{\bf Mass ratio estimations for the wind obscuration scenario= in the case of a HMXB}} \tablewidth{4in} \tablehead{\colhead{$R_{OB}/ R_\odot$} & \colhead{$M_{OB}/M_\odot$} & \col= head{$M_2/M_{OB}$} & \colhead{$M_2/M_\odot$} } \startdata 80 & 20 & 0.010 & 0.2 \\ \tableline 80 & 60 & 0.011 & 0.6 \\ \tableline 80 & 100 & 0.016 & 1.6 \\ \tableline\tableline 50 & 20 & 0.015 & 0.3 \\ \tableline 50 & 60 & 0.022 & 1.3 \\ \tableline 50 & 100 & 0.026 & 2.6 \\ \tableline\tableline 20 & 20 & 0.03 & 0.6 \\ \tableline 20 & 60 & 0.06 & 3.6 \\ \tableline 20 & 100 & 0.07 & 7.0 \\ \tableline \enddata \tablecomments{The mass ratio and compact object mass expected for a primar= y of the given mass to radius ratio in the wind obscuration scenario, valid= for the HMXB case. The estimation of $q$ is based on Equation \ref{eq:qmdo= tlx}. We use $L_X =3D 1.1 \times 10^{35} erg~s^{-1}$ and assume an efficien= cy $\epsilon =3D 0.1$, and a mass loss rate $\dot{M} =3D 4 \times 10^{-5} M= _\odot~yr^{-1}$. The value $R_{OB} =3D 80 R_\odot$ is most consistent with = our observed IR luminosity \citep{girardi02}.} \label{tbl-6} \end{deluxetable} \begin{deluxetable}{p{3.2in}|p{3.2in}} \tablecaption{{\bf Summary of scenarios under the orbital period assumption= }} \tablewidth{\textwidth} \tablehead{\colhead{\bf WIND OBSCURATION SCENARIO} & \colhead{\bf ECLIPSING= BINARY SCENARIO}} \startdata \multicolumn{2}{c}{Two stars contributing equally to the IR luminosity (CWB= )}\\ \multicolumn{2}{c}{$R_{OB} \sim 20 R_\odot$}\\ $\dot{M} =3D 10^{-5} M_\odot /yr$ (Eq. \ref{eq:nhmdot}) & $q \approx 0.05$ = (Eq. \ref{eq:massfunction})\\ {\it consistent} & {\it inconsistent with initial assumptions} \\ \tableline \multicolumn{2}{c}{One star dominating the IR luminosity (CWB)}\\ \multicolumn{2}{c}{$R_{OB} \sim 80 R_\odot$}\\ $\dot{M} =3D 4\times 10^{-5} M_\odot /yr$ (Eq. \ref{eq:nhmdot}) & $q \appro= x 0.2$ (Eq. \ref{eq:massfunction}) \\ {\it consistent} & {\it IR line ratios inconsistent with known WR+O system= s} \\ \tableline \multicolumn{2}{c}{One star dominating the IR luminosity (HMXB)}\\ \multicolumn{2}{c}{$R_{OB} \sim 80 R_\odot$}\\ \multicolumn{2}{c}{$L_X =3D 1.1 \times 10^{35} erg~ s^{-1}$}\\ $\dot{M} =3D 4\times 10^{-5} M_\odot /yr$ (Eq. \ref{eq:nhmdot}) & $q \appro= x 0.2$ (Eq. \ref{eq:massfunction})\\ $q \sim 0.01$ (Eq. \ref{eq:qmdotlx}) & $\dot{M} =3D 2 \times 10^{-7} M= _\odot /yr$ (Eq. \ref{eq:qmdotlx})\\ {\it radius constraint suggests $M_{OB} > 80 M_\odot$} & {\it consistent} = \\ \enddata \label{tbl-7} \tablecomments{See details of more general cases and caveats in Section 3.3= =2E3.} \end{deluxetable} \end{document} =20 -- Dr. Reba M. Bandyopadhyay EMAIL: reba@astro.ufl.edu Assistant Scientist reba@alum.mit.edu Department of Astronomy rmb@astro.ox.ac.uk University of Florida reba@xeus.nrl.navy.mil 211 Bryant Space Science Centre PHONE: +1-352-392-2052, ext. 232 Gainesville, FL 32611-2055 USA WWW: http://www.astro.ufl.edu/~reba/ ---478107720-106629879-1215638649=:3940--