1: %\documentclass[manuscript]{rmaa}
2: \documentclass[proof]{rmaa}
3: \usepackage[latin1]{inputenc}
4: \usepackage{paralist}
5:
6: \title{New insights into the nature of the SMC WR/LBV binary HD 5980\footnote{Based on observations made with the NTT, MPI 2.2m and Danish 1.5m telescopes at ESO, La Silla, Chile}
7: }
8: \author{C. Foellmi\altaffilmark{1}, %cedric.foellmi@obs.ujf-grenoble.fr
9: G. Koenigsberger\altaffilmark{2}, %gloria@fis.unam.mx
10: L. Georgiev\altaffilmark{3}, %georgiev@astroscu.unam.mx
11: O. Toledano\altaffilmark{2} %oswaldotoledano@hotmail.com
12: S.V. Marchenko \altaffilmark{4}, %sergey@astro.wku.edu
13: P. Massey\altaffilmark{5}, %Phil.Massey@lowell.edu
14: T. H. Dall \altaffilmark{6}, %tdall@gemini.edu
15: A.F.J. Moffat\altaffilmark{7}, %moffat@astro.umontreal.ca
16: N. Morrell\altaffilmark{8}, %nidia@lcoeps1.lco.cl; nidia@fcaglp.unlp.edu.ar
17: M. Corcoran\altaffilmark{9,10}, %Michael.F.Corcoran.1@gsfc.nasa.gov
18: A. Kaufer\altaffilmark{11}, %akaufer@eso.org
19: Y. Naz\'e\altaffilmark{12}, %naze@astro.ulg.ac.be
20: J. Pittard\altaffilmark{13}, %jmp@ast.leeds.ac.uk
21: N. St.-Louis\altaffilmark{7}, % @astro.umontreal.ca
22: A. Fullerton\altaffilmark{14}, %fullerton@stsci.edu
23: D. Massa\altaffilmark{15} %massa@taotaomona.gsfc.nasa.gov
24: A. M. T. Pollock\altaffilmark{17} %Andy.Pollock@sciops.esa.int
25: }
26: \altaffiltext{1}{Laboratoire d'AstrOphysique de Grenoble, 414 Rue de la Piscine, Domaine Universitaire, 38400 Saint-Martin d'H\`eres, France} %Cedric
27: \altaffiltext{2}{Instituto de Ciencias F\'{\i}sicas, Univer\-sidad Nacio\-nal Aut\'o\-noma de M\'exico, Cuernavaca, Mor., Mexico}
28: \altaffiltext{3}{Instituto de Astronom\'{\i}a, Univer\-sidad Nacio\-nal Aut\'o\-noma de M\'exico, M\'exico D.~F. 04510}
29: \altaffiltext{4}{Department of Physics and Astronomy, Western Kentucky University, 1906 College Heights Blvd \#11001, Bowling Green, KY 42101-1101, USA}
30: \altaffiltext{5}{Lowell Observatory, 1400 W Mars Hill Rd., Flagstaff, AZ, 86001,USA} %Phil Massey
31: \altaffiltext{6}{Gemini Observatory, 670 N. A'ohoku Pl., Hilo, HI 96720, USA}
32: \altaffiltext{7}{ D\'epartement de Physique, Universit\'e de Montr\'eal, C.P. 6128, Succ. C-V, Montr\'eal, QC, H3Y 1C5, and Observatoire du mont M\'egamtic, Canada}
33: \altaffiltext{8}{Las Campanas Observatory, Carnegie Observatories, Casilla 601, La Serena, Chile}
34: \altaffiltext{9}{CRESST and X-ray Astrophysics Laboratory NASA/GSFC, Greenbelt, MD 20771, USA.}
35: \altaffiltext{10}{Universities Space Research Association, 10211 Wincopin Circle, Suite 500 Columbia, MD 21044, USA.}
36: \altaffiltext{11}{European Southern Observatory, Alonso de Cordova 3107, Casilla 19001, Santiago 19, Chile} %Andreas
37: \altaffiltext{12}{FNRS and Institut d'Astrophysique et de G\'eophysique, Universit\'e de Li\`ege, All\'ee du 6 Ao\^ut 17, Bat B5C, B4000-Li\`ege, Belgium}
38: \altaffiltext{13}{School of Physics and Astronomy, University of Leeds, Leeds, UK}
39: \altaffiltext{14}{Space Telescope Science Institute, 3700 San Martin Drive, Baltimore, MD 21218} %Alex
40: \altaffiltext{15}{NASA/GSFC, Mailstop 665.0, Greenbelt, MD 20771} %Derck
41: \altaffiltext{17}{European Space Agency, XMM-Newton SOC, European Space Astronomy Centre, Apartado 78, 28691 Villanueva de la Caqada, Madrid, Spain.} %Andy Pollock
42:
43: \shortauthor{Foellmi et~al.}
44: \Note{Version: 12 September 2007}
45: \shorttitle{New insights into HD 5980}
46: \fulladdresses{
47: %\item Leonid Georgiev:
48: %Ins\-titu\-to de Astro\-nom\'{\i}a, UNAM, Apartado Pos\-tal 70--264,
49: %04510 M\'exico D.~F., M\'exico
50: %
51: %\item Gloria Koenigsberger, Oswaldo Toledano: Instituto de Ciencias F\'{\i}sicas,
52: % UNAM, Apartado Postal 48--3, 62251 Cuernavaca, Mor., M\'exico.(gloria@fis.unam.mx;oswaldo@fis.unam.mx)
53: }
54:
55: \listofauthors{M. Corcoran, T. H. Dall, C. Foellmi, A. Fullerton, L. Georgiev, A. Kaufer,
56: G. Koenigsberger, S. Marchenko, D. Massa, P. Massey, A. Moffat, N. Morrell, Y. Naz\'e, J. Pittard,
57: A. Pollock, N. St-Louis, O. Toledano }
58: \indexauthor{Corcoran, M.}
59: \indexauthor{Dall, T.H.}
60: \indexauthor{Foellmi, C.}
61: \indexauthor{Fullerton, A.}
62: \indexauthor{Georgiev, L.}
63: \indexauthor{Kaufer, A.}
64: \indexauthor{Koenigsberger, G.}
65: \indexauthor{Marchenko, S.}
66: \indexauthor{Massa, D.}
67: \indexauthor{Massey, P.}
68: \indexauthor{Moffat, A.}
69: \indexauthor{Morrell, N.}
70: \indexauthor{Naz\'e, Y.}
71: \indexauthor{Pittard, J.}
72: \indexauthor{Pollock, A.}
73: \indexauthor{St-Louis, N.}
74: \indexauthor{Toledano, O.}
75:
76: \SetVolume{00}
77: \SetYear{2008}
78:
79: \ReceivedDate{2007 July 6 }
80: \AcceptedDate{}
81: \resumen{Presentamos los resultados de una campa\~na de observaci\'ones \'opticas del
82: sistema m\'ultiple HD 5980 cuya estrella primaria sufri\'o una erupci\'on importante
83: en 1994. Atribuyendo la variabilidad de las l\'{\i}neas en emisi\'on al movimiento
84: orbital de las dos estrellas con \'orbita de 19.3 dias, deducimos sus masas. El
85: comportamiento de las l\'{\i}neas fotosf\'ericas indica que la fuente de ``tercera luz"
86: del sistema es probablemente tambi\'en un sistema binario como fu\'e propuesto por
87: Scheickhardt (2002). Los datos presentados en este art\'{\i}culo estar\'an publicamente
88: disponibles.}
89:
90: \abstract{We present the results of optical wavelength observations of the unusual SMC
91: eclipsing binary system HD 5980 obtained in 1999 and 2004--2005. Radial velocity curves for
92: the erupting LBV/WR object ({\it star A}) and its close WR-like companion ({\it star B}) are
93: obtained by deblending the variable emission-line profiles of NIV and NV lines. The derived
94: masses $M_A=$58--79 M$_\odot$ and $M_B=$51--67 M$_\odot$, are more consistent with the the
95: stars' location near the top of the HRD than previous estimates. The presence of a wind-wind
96: interaction region is inferred from the orbital phase-dependent behavior of He I P Cygni
97: absorption components.
98: The emission-line intensities continued with the declining trend previously seen in UV spectra.
99: The behavior of the photospheric absorption lines is consistent with the results of Schweickhardt
100: (2002) who concludes that the third object in the combined spectrum, {\it star C}, is also a binary
101: system with $P_{starC}\sim$96.5 days, $e=$0.83.
102: % it is also a curious fact that its apparent orbital period is nearly a factor of 5 larger that
103: %the orbital period in the{\it star A+B} pair.
104: }
105: \addkeyword{Binaries: close}
106: \addkeyword{Stars: mass loss}
107: \addkeyword{Stars: individual (HD5980)}
108:
109: \begin{document}
110: \RescaleTitleLengths{0.95}
111: \maketitle
112:
113: \section{Introduction}
114:
115: The Small Magellanic Cloud system HD 5980 consists of two very luminous components
116: referred to as {\it star A} and {\it star B} in a relatively close, eclipsing and eccentric
117: orbit and a third equally luminous source, referred to as {\it star C}, that may
118: merely be a line-of-sight coincidence. The eclipsing nature of HD 5980 was discovered by
119: Hoffmann, Stift \& Moffat (1978), although the actual 19.3-day orbital period was found by
120: Breysacher \& Perrier (1980; henceforth BP80), and more recently refined by Sterken \& Breysacher
121: (1997). The two eclipses were not separated equally in time, indicating an eccentric orbit
122: ($e=0.32$, Breysacher \& Perrier 1991; $e=0.27$ Moffat et al. 1998; $e=0.30$ Kaufer et al. 2002). The
123: presence of strong and broad emission lines as well as photospheric absorption lines led to its
124: early classification as a WN+OB binary system (Azzopardi \& Vigneau 1975; Walborn 1977;
125: Breysacher \& Westerlund 1978). However, Niemela (1988) showed that the Doppler shifts
126: in the N IV 4057 and N V 4603 emission lines moved in anti-phase, concluding that {\em both}
127: stars in the 19.3-day binary orbit were Wolf-Rayet stars of the nitrogen sequence (WN) and
128: that the absorption-line spectrum had to arise in a third object. With this result,
129: Niemela et al. (1997) estimated the masses of the two stars, M$_A\sim$50 M$_\odot$ and
130: M$_B\sim$28 M$_\odot$. From the eclipse light curve, Breysacher \& Perrier (1991, henceforth
131: BP91) derived stellar radii, R$_A\sim$21 R$_\odot$ and R$_B\sim$15 R$_\odot$. Its location in
132: the SMC implied a visual absolute magnitude M$_V^{sys}=-$7.3 mag (BP91) and the BP91 light curve analysis
133: provided good constraints on the visual absolute magnitude for each of the three stars within
134: the system, from which M$_V^A=$-6.3 mag, M$_V^B=$-5.8 mag, and M$_V^C=$-6.1 mag.
135: These and other parameters for HD 5980 are summarized in the first column of Table 1, which
136: contains information derived from data obtained prior to 1981. Results from data of other
137: epochs is listed in column 2 (1994; near the time of eruption) and in column 3 (1999-2005;
138: declining phase of the eruption).
139:
140: The spectral characteristics and visual brightness of HD 5980 gradually changed between the late 1970's
141: and 1993, when it entered an eruptive state that lasted $\sim$1 year (Bateson \& Jones 1994;
142: Barb\'a et al. 1995; Koenigsberger et al. 1995). The activity involved an increase in
143: visual brightness and mass-loss rate, and a decrease in wind velocity and effective
144: temperature, similar to the eruptive phenomena observed in LBVs. The radial velocity (RV)
145: variations in the post-eruption emission-line spectrum led Barb\'a et al. (1996, 1997) to
146: conclude that the instability producing the outbursts originated in {\it star A}. This was also
147: confirmed by independent observations of Moffat et al. (1998). The hydrogen abundance
148: in {\it star A} that was derived near the time of maximum eruption, when its emission-lines are believed
149: to have completely dominated the combined spectrum, was found to be depleted (Koenigsberger et al. 1998b).
150: Thus, {\it star A} is a very massive star that has left the Main Sequence. Another fascinating aspect of
151: the system is that in the process leading up to the eruption, {\it star A} seems to have traversed the entire
152: nitrogen-sequence (WN) of the WR classification system, starting out as a WN3 (Niemela 1988) and
153: reaching WN11 (Drissen et al. 2001) during maximum eruption. This behavior is unprecedented.
154:
155: It is important to note that the eruptive activity in HD 5980 has been recorded only once. Thus,
156: {\it star A} may still be basically in the pre-LBV evolutionary phase, resembling one of the
157: H-rich (or at least not yet H-poor), luminous WR stars of type WNh, or probably even WNha,
158: like most, if not all of the 10 other WN stars in the SMC (Foellmi et al. 2003). These are
159: not really {\it bona fide} WR stars in the classical sense, i.e. in the He-burning
160: stage.
161:
162: {\it Star B} is also inferred to be a WR star based on pre-eruption spectroscopic observations that
163: indicated emission-line RV variations consistent with this notion (Breysacher, Moffat \& Niemela 1982).
164: If both stars are, indeed, WR's, the time-variable wind of {\it star A} presents a unique
165: opportunity for studying the effects that a changing wind momentum ratio produces on
166: the wind-wind interaction (WWI). Curiously, however, the change in the WWI region during
167: declining phases of the eruptive event appears to have had little effect on the observed
168: X-ray emission (Naz\'e et al. 2007).
169:
170: %The wind momentum ratio determines the geometry of the WWI shock cone.
171: %Moffat et al. (1998) discussed the expected effects due to the different wind momentum
172: %ratios prior to, during and after the eruptions in HD 5980. Koenigsberger (2004) reviewed
173: %the UV observational data and its implications for the WWI scenario, reaching the conclusion
174: %that {\it star A}'s wind still dominated the wind momentum ratio in the early 2000's. From
175: %a series of FUSE observations, Star A's wind velocity law was deduced to be typical of
176: %hot massive stars except in the direction of its close companion
177: %(Koenigsberger et al. 2006). This raises the possibility that the WWI region
178: %may reach far ``down" into the wind of {\it star A} (i.e., near its photosphere), even though
179: %this star's wind dominates the overall momentum ratio. Such a phenomenon could result from
180: %``sudden radiative braking" (Gayley et al. 1997) due to the strong UV radiation field of {\it star B}.
181: %Alternatively, this result could be indicating that {\it star A}'s wind structure on the
182: %hemisphere facing its companion is {\em intrinsically} different from the wind structure
183: %emerging in other directions. Koenigsberger \& Moreno (2007) have recently suggested
184: %that the tidal flows on the surface of {\it star A} could lead to a shearing action. The
185: %associated energy dissipation rates, largest on the side facing the companion, could
186: %alter the wind acceleration process.
187: %
188: %Central to resolving these issues is the line profile variability and the degree to which it
189: %is produced by the WWI. The time-dependent shape of the emission lines is determined by a combination of
190: %mechanisms that include wind and physical eclipses, wind-wind interactions, sudden radiative braking
191: %and X-ray heating. Of these, the effects on the line profiles due to eclipses can be easily
192: %modeled (Auer \& Koenigsberger 1994; Stevens et al. 1992). However, the effects due to wind-wind
193: %collisions are more difficult to assess.
194: %If we assume that the WWI zone merely represents a boundary between the stellar winds of
195: %the two stars, the computed line profile variability (Stevens 1993; Georgiev \&
196: %Koenigsberger 2004a) consists of {\em the absence} of emission at particular wavelengths,
197: %or the weakening of P Cygni absorption components at particular orbital phases. On the
198: %other hand, if the WWI zone is assumed to be a thin conical layer of enhanced density
199: %where additional emission (with respect to the stellar winds) is produced, the line
200: %profile variability can be modeled in terms of variable emission peaks that are superposed
201: %upon the underlying broad wind emission (L\"uhrs et al. 1997; St.-Louis et al. 2006).
202: %Very little guidance is provided from theoretical models regarding the emissivity of
203: %the WWI region at optical and UV wavelengths.
204: %
205: The masses of {\it star A} and {\it star B} are the central issue for establishing their
206: evolutionary state. But a vexing problem arises when radial velocity curves of the emission lines
207: are used to derive the stellar masses. The centroid of an emission line is ill-defined
208: when the line is asymmetrical and undergoes strong variability over the orbital cycle, as
209: is the case for HD 5980. Without a clear understanding of the line profile variations (lpvs),
210: the orbital motion cannot be reliably ascertained and, hence, one cannot establish the primary
211: cause of the variability.
212:
213: The objectives of this paper are to use new and archival observations to establish constraints
214: on the sources of the line-profile variability in HD 5980 and thus obtain RV curves that
215: may be more representative of the orbital motion. Section 2 summarizes the data; in Section 3
216: we describe the line profile variability, leading to a new set of RV curves (Section 4) from which the
217: masses of {\it star A} and {\it star B} are estimated; Section 5 summarizes the evidence suggesting
218: that the wind density of {\it star A} is diminishing over long timescales; Section 6 describes the
219: stationary photospheric absorptions; Section 7 presents the results of the photometric monitoring;
220: in Section 8 we discuss the wind-wind interaction region; and in Section 9 we present the conclusions.
221:
222: \section{Observational data}
223:
224: The 2005/2006 observation campaign at ESO was coordinated by CF and GK.
225: Spectroscopy was obtained under service observing mode with the Fiber-Fed
226: Extended Range Optical Spectrograph (FEROS) mounted on the MPI 2.2m telescope at
227: La Silla, and the ESO Multi Mode Instrument (EMMI) mounted on the NTT.
228: In addition to these data, a set of archival FEROS observations obtained in 1999
229: on the 1.5m telescopes at La Silla (Schweickhardt, 2000; Kaufer et al. 2002) were
230: retrieved. Photometric observations were obtained at Las Campanas in 2003, at
231: ESO in 2005, and at CTIO in 2005-2006.
232:
233: \subsection{Spectroscopy}
234:
235: The 33 FEROS echelle spectra reported in this paper were obtained with a
236: 2 k $\times$ 4 k EEV CCD, and they cover the whole optical range ($\sim$3750-9200 \AA).
237: The spectral resolving power of FEROS is 48000. Typical exposure times for the 1999 data were
238: between 40 and 50 minutes while those of the 2005 data were typically 27 minutes. The
239: S/N$\sim$30-70 per resolution element in the collapsed 1D spectrum, lower values corresponding
240: to the blue spectral region. Tables 2 and 3 contain the ephemeris of these observations.
241:
242: The ``pipeline" FEROS data reduction was checked by LG against sample detailed processing using
243: MIDAS and IDL routines and was found to provide similar results, and thus we chose to use
244: the ``pipeline" reduced spectra. Data analysis was performed by GK. The merged spectrum
245: in the wavelength range 3900--8500 \AA\ was
246: normalized in a three-step process. In the first step, all the spectra were combined
247: for each epoch (1999 and 2005, respectively) so as to automatically remove cosmic rays.
248: In the second step, the continuum was traced using a linear interpolation between
249: small wavelength intervals over which the location of the continuum could be relatively
250: well established. This yielded a curve that approximately describes the shape of the
251: continuum on the average spectrum and that was then used to rectify the continuum on each
252: of the individual spectra. After the spectra were normalized in this manner, the third
253: step consisted in fitting a spline function (``spline3" in IRAF) of order 3, 6 or 15 to
254: arrive at the normalized spectrum for the whole wavelength range stated above. However,
255: for the detailed comparison of the line profiles and RV measurements, selected
256: spectral windows were extracted from the whole normalized spectrum, and renormalized
257: individually by fitting a second order Legendre function to the neighboring continuum levels.
258: Cosmic rays in spectral windows to be measured were removed individually, and a Gaussian smoothing
259: procedure was applied to enhance the S/N ratio per wavelength interval.
260:
261: Spectra obtained near mid-continuum eclipses need to be renormalized
262: so that the comparison of line profiles is meaningful. Because the orbit is eccentric,
263: eclipses are not equidistant in phase; they occur at $\phi=$0 ({\it star A} ``in front") and
264: at $\phi=$0.36 ({\it star B} ``in front"). From the Breysacher \& Perrier (1991; henceforth, BP91) light curve
265: solution, $\phi=$0 eclipse is total while at $\phi=$0.36, only $\sim$50\% of {\it star A}'s continuum
266: is occulted. The relative continuum levels for stars A, B and C are, respectively, I$_A=$0.41, I$_B=$0.26
267: and I$_C=$0.33. Hence, to correct for the continuum eclipses, we added I$_B$ at $\phi=$0.00 and
268: 0.7I$_A$ at $\phi=$0.36--0.37 (see below), and then renormalized the continuum to unity.
269: It must be emphasized that this renormalization is only a gross approximation, since no recent (i.e., for data
270: taken after 1980) light curve solution is available to provide current values of the stellar radii. However,
271: under the assumption that only {\it star A} changed its size, it is pretty certain that the eclipse at $\phi=$0 is
272: a total eclipse. In addition, FUSE observations lead to the conclusion that the ratio of {\it star A} and {\it star B}
273: radii in 2002 is similar to that obtained by BP91. This would suggest that the correction at this
274: phase be only 0.5I$_A$. However, this correction leads to an increase in the line-to-continuum ratio at
275: eclipse, contrary to what is expected, especially since a fraction of {\it star A}'s line emission must be
276: eclipsed, leading to the conclusion that a larger correction is required. This leads to the correction value
277: of 0.7I$_A$ given above. Such a larger correction is justified when limb-darkening is taken into account.
278:
279: %during eclipses, part of the continuum emission is missing, making the line/continuum
280: %ratio rise. To correct for this, we need to "replace" the missing continuum-->add.
281: %April 3, 2007: I checked several of the figures "montage_profiles_*" and found that
282: %using the new renormalized spectra yields indistinguishable results from the previous
283: %renormalization, where {\it star A} and {\it star B} were assumed to contribute 30\% of the continuum.
284: %Thus, all comments apply to the new normalization figures.
285:
286: The ESO Multi Mode Instrument (EMMI) was used in the red mode (RILD) with the low-resolution grism 2,
287: leading to a resolution of 300-1700 in the wavelength range 4000-10000 \AA. Typical exposures
288: were 20s for HD5980.
289: %RILD=Red Imaging and Low Dispersion spectroscopy.
290: %R=280-1670, in the wavelength range 4000-10000 \AA. Grism dispersion is 1.74 \AA/pixel.
291: %60 seconds for calibration stars
292: EMMI data were processed by OT using standard IRAF routines. This processing consists of cosmic
293: ray excision, bias subtraction, flat-field correction, wavelength calibration, background/sky
294: subtraction and extraction to produce a 1-dimensional spectrum. Background includes the strong
295: ISM lines that are produced by the large H II region N66 within which HD 5980 is located. Three
296: spectra were obtained on each night, and these were averaged together and normalized using an
297: interactive polynomial fit to line-free continuum regions. Wavelength calibration was done with the
298: HeAr lamp exposures obtained on the same nights as the exposures of HD 5980.
299: %Chebychev order 6.
300: Table 4 contains a summary of the EMMI data, and Figure 1 illustrates the 4000-7500 \AA\ region
301: of the average EMMI spectrum at orbital phase 0.00. Prominent emission lines are identified.
302:
303: \subsection{Photometry}
304:
305: Observations with the Danish 1.5m telescope, using DFOSC and the V Johnson filter, were
306: coordinated by TD and obtained over the time period JD 2453561 --2453693. DFOSC is a focal
307: reducer imager and spectrograph, covering a 13 $\arcmin\times$13$\arcmin$ field of view. Data
308: processing was carried out by SM as follows. The available CCD images were reduced with standard
309: IRAF procedures. Following the general approach of Everett \& Howell (2001), a weighted
310: combination of fluxes of multiple comparison stars was introduced in order to produce differential
311: magnitude values, $<$m$_v$-C$>$.
312: %These were later converted into the V (Johnson) magnitudes based on the
313: %catalogued measurements for the following comparison stars: Cl* NGC 346 MPG 185, Sk80=AzV 232,
314: %C1*NGC 346 MPG 802, Cl* NGC 346 MPG 811, and C1*NGC 346 MPG 859.
315: % Cedric found the names of the comparison stars in SIMBAD; checked by Sergey
316: %\footnote{The comparisons used in the CCD field in are (RA2000,DEC2000):(00:58:42.7,-72:09:45), (00:58:53.3,-72:08:35),
317: %(00:59:31.8,-72:10:46), (00:59:36.4,-72:10:22), (00:59:40.1,-72:13:36), (00:59:42.2,-72:09:49),
318: %(00:59:54.9,-72:13:06.}
319: This approach provides a typical accuracy $\sigma_{instr}$=0.003-0.008 mag for each
320: individual observation, as listed in column 5 of Table 6.
321: %the $\delta m_v$ values were converted to m_v by adding 12.846 mag.
322: The additional columns in Table 6 list, in columns 1-3 the average MJD of the observations, the orbital phase, and
323: the number of differential magnitude determinations that were averaged to yield the $<m_v-C>$
324: listed in column 5. The standard deviation, s.d., for this average is listed in column 6.
325:
326: In addition to the above, we include V-band observations obtained by NM from
327: JD 2452975.56 to 2452991.57 on the 1-m Swope Telescope, at Las Campanas Observatory, and
328: % using the 2048x3150 SITe#3 CCD, which provides a scale of 0.435 arecsecs/px. Exposure times were
329: %generally 10s, and seeing typically 1 arcsec for these data.
330: %(December 2003)
331: data that were taken from JD 2453591.88 through 2454092.70 for PM by the SMARTS consortium on the
332: CTIO/SMARTS 1.0-m telescope. Both sets were processed by PM and converted to the visual magnitude scale.
333: Part of the SMARTS data overlap in time with the Danish Telescope photometry, filling in
334: the gaps in orbital phase not covered by the latter.
335: The photometry of the Swope and SMARTS data were done relative to other stars
336: in the NGC 346 cluster, with the zero-point set by comparison of the data from a single
337: (arbitrary) exposure to the photometry of Massey et al.\ (1989). Within each data
338: set, the photometry has a relative accuracy of $\sim$0.02 mag.
339: %The zero-point uncertainty $\sim$0.04.
340:
341: \section{Line profile variability}
342:
343: Line-profile variability in WR binaries is a major obstacle for determining the stellar masses.
344: This is because the amplitudes of the RV curves obtained from variable and asymmetric line
345: profiles depend on the way in which the lines are measured. In HD 5980, emission line-profile
346: variability has been a common feature since the star was first observed spectroscopically
347: (Breysacher \& Westerlund 1978). Remarkably, the nature of the variability has remained
348: qualitatively the same over many years, despite the strong changes that have occurred in the
349: stellar wind properties. For instance, emission lines are always broader near
350: elongations, while near eclipses they are visibly narrower and more sharply peaked
351: (BMN80; Moffat et al. 1998), as illustrated by the Balmer-series lines of hydrogen that are
352: plotted in Figure 2. The left panels of Figure 2 correspond to eclipse phases and those
353: on the right to orbital phases near elongations. Most of the emission lines display the same
354: pattern of variability.
355:
356: The change in full-width at half maximum (FWHM) is illustrated in Figure 3
357: obtained from the Gaussian function fits to HI (+He II) 4101, N IV 4057, He II 5411, and
358: CIV 5806/12 \AA\ as a function of orbital phase, showing that different lines behave differently.
359: However, the qualitative nature of the variations we observe in 1999 is as reported in previous
360: optical observations (Moffat et al. 1998; Breysacher et al. 1980; Breysacher 2000). This effect
361: is only one aspect of a much more complex phase-dependent line profile variability pattern that
362: involves asymmetric distortions in the line shape. Thus, the amplitude of any RV curve derived
363: from the emission lines depends on how the lines are measured; that is, it depends on where within
364: the system the emission lines are assumed to originate.
365:
366: Fortunately, optical and UV spectroscopic monitoring of HD 5980 since the late 1970's provides
367: the first of three pieces of solid information to aid in interpreting the variability: numerous
368: emission lines that are currently present in the spectrum were absent prior
369: to the early 1980's. The RV variations of the ``new" lines indicate that they arise in {\it star A}.
370: This is also true of the ``new" UV lines observed with IUE and HST (Koenigsberger 2004 and
371: references therein).
372:
373: A second fact is that lines arising primarily in the inner portions of the stellar wind are less
374: susceptible to distortions due to wind-wind interaction effects than lines that arise from
375: outer wind regions. The former tend to originate from atomic transitions between excited
376: states, and from elements of lower chemical abundance and are visible only because the
377: density of the inner, accelerating portions of the wind is sufficiently large.
378: Hence, lines from abundant ions and having large transition probabilities, such as H$\alpha +$He II 6560 \AA\
379: and He~II 4686 \AA\ should be avoided during the RV curve analyses, while lines such as N~IV 4057 \AA\
380: may better describe the orbital motion.
381:
382: The third datum that aids in the interpretation of the lpv's is that the system is eclipsing and
383: the largest RV's occur near orbital phase $\phi=$0.15. In addition, due to the geometry of the orbit,
384: during the eclipse at $\phi=$0.00 only the wind of {\it star A} contributes
385: emission or absorption at wavelengths shorter than the rest wavelength.
386:
387: Figure 4 provides a guide to the location of {\it star A} with respect to {\it star B} (located at the center
388: of this coordinate system) as seen by an observer located at the bottom of the page. The
389: red squares that are labeled with the orbital phases correspond to the location of {\it star A}
390: at that given time. Periastron passage occurs near $\phi=$0.07.
391:
392: \subsection{Emission lines}
393:
394: With the above ideas in mind, we compared the 2005 FEROS line profiles obtained at
395: orbital phases 0.13, 0.36 and 0.80 with the profiles observed at $\phi=$0.00.
396: The spectrum at $\phi=$0.0 is expected to be dominated by {\it star A} emission.
397: The individual line profiles are illustrated in Figures 5--10,
398: plotted on a velocity scale corrected for a SMC motion of $+$150 km/s. The
399: top left panel contains the $\phi=$0 spectrum. The other three orbital phases
400: are presented in the remaining panels, and compared with the $\phi=$0 profile, which has
401: been shifted in velocity scale to align major features. For example, Figure 5
402: displays the He II 4542, N V 4603 and N V 4621 \AA\ lines. The velocity shift
403: required to align the maxima of the three emission lines at $\phi=$0.13 is $+$275 km/s,
404: while for $\phi=$0.8, the He II 4542 \AA\ emissions are aligned with a shift of
405: $-$150 km/s.
406:
407: We first consider the profiles of N V 4603, 4621 (Figure 5) and the
408: weak He II lines in the $\lambda\lambda$6000--6200 \AA\ region (Figure 6). At $\phi=$0.13 excess
409: emission is apparent in the blue emission line wings, pointed out by arrows.
410: In N V 4603, 4621 \AA\ it can be seen as narrow sub-peaks that ``fill in"
411: the P Cygni absorption components. A similar excess emission at this phase appears
412: in N IV 4057, H$\delta$ (Figure 7), H$\gamma$ (Figure 8), He II 5411
413: (Figure 9) and C IV 5806-11 (Figure 10). We interpret this excess blue emission
414: as emission arising in {\it star B}, which is approaching the observer at $\phi=$0.13.
415:
416: No relative shift was applied to the spectra at $\phi=$0.37 since the N IV 4057
417: line coincides almost exactly with the profile of $\phi=$0. In most of the other
418: lines, at least the red wings match very well, but blueward of line center one can
419: easily see that excess emission is present.
420:
421: At $\phi=$0.8 the lines present two different patterns. In the first, N IV 4057,
422: the weak He II 6000-6200 \AA\ emissions and the N V 4603 P Cygni absorptions match
423: the shifted $\phi=$0 line profiles. Thus, these features are primarily formed in
424: {\it star A} and have negative RV's due to its approaching orbital motion.
425: The second pattern, however, is more prevalent and consists of the presence of
426: excess emissions on both wings, although the excess red emission is generally
427: more extended than the blue. Within the context of the interpretation given above,
428: the excess red emission at $\phi=$0.8 originates in {\it star B}, which is receding from the
429: observer at this phase. The excess blue emission, however, appears
430: to arise from an additional source. It is tempting to suggest that it
431: arises from the same source as that which produces excess emission at $\phi=$0.36, although
432: its projected velocity towards the observer is greater at $\phi=$0.8 than at $\phi=$0.36.
433:
434: Figures 11--14 are grey-scale representation of the 1999 FEROS data and provide a global view
435: of the lpv's in 4 representative lines: N IV 4057, He II 5411, H$\beta$+He II 4860 and He I 6678 \AA\.
436: They are constructed with line profile residuals, stacked from bottom to top with increasing
437: orbital phase. The residuals are the difference between each individual spectrum and the average
438: of the 15 spectra obtained over a single orbital cycle and listed in Table 2 (the first two
439: spectra of this table were not included here). The spacing between the spectra
440: is $\Delta\phi\sim$0.05 in orbital phase, and to consistently retain this spacing in the
441: dynamic plots, the gaps of missing phase coverage are filled-in with grey
442: background color. Plotting the same data twice provides a clearer picture of the orbital-phase
443: dependent trends. The white regions in these images correspond to excess emission; the black
444: to emission deficiency with respect to the mean, or absorption. The maximum dynamic range is
445: typically 10--15\% in the residuals with respect to the normalized mean profiles.
446:
447: All four dynamic plots display similar large-scale trends, specifically, the excess emission
448: traces out a curve suggestive of orbital motion associated with {\it star A}. Some of the
449: details of the plots, however, are different. For example, there is a fainter sinusoidal
450: curve that moves in anti-phase with the more dominant bright sinusoid that is clearly seen in
451: He I 6678 and H$\beta$+He II 4860, weaker in He II 5411, but not detectable in N IV 4057. In
452: He I 6678, this secondary curve has a "whisp-like" structure around $\phi=$0 in the velocity
453: range 0 -- 1000 km/s. A striking difference between N IV 4057 and H I 6678 is the very extended
454: blue emission in the phase interval $\sim$0.9--1.0 in the latter. While N IV 4057 shows a rapid
455: shift towards the red, as expected from the orbital motion of {\it star A}, He I 6678 stays very much
456: the same, extending over the velocity range $-$300 to $-$1100 km/s. This implies the presence
457: of emitting gas that is approaching the observer with a speed that is not correlated with
458: the orbital motion within this phase interval.
459:
460: %careful: maybe entire P Cyg abs is from {\it star B} wind ....? but maybe its contribution
461: %is too weak to make much of a difference...it is not the dominant source for this line.
462:
463: \subsection{ Peculiar behavior of He I P Cygni absorptions}
464:
465: A clue to the origin of the excess blue emission at $\phi=$0.80 may lie in the
466: behavior of the He I lines. As an example, consider the blue wing of He I 5875 \AA\
467: that is shown in Figure 10. At $\phi=$0.80, a sharp P Cygni absorption is present in
468: the same velocity range as the excess blue emission in the H I and He II lines shown
469: in the grey-scale plots discussed above. Such sharp and strong P Cygni absorptions are
470: usually associated with lines that are formed far from the star, in the low-density
471: regions that are expanding at a constant speed, usually the terminal wind speed, v$_\infty$.
472: The presence of the absorption would not be surprising were it not for its {\em absence} or
473: weakness at other orbital phases. This effect can be more clearly appreciated in Figures 15,
474: which displays montages of He I 4471, 5875, 6678 and 7965 \AA\ at the four orbital phases
475: $\phi=$0.0, 0.13, 0.37 and 0.80, and in Figures 16 and 17 where the fuller orbital phase
476: coverage of the 1999 data shows that the absorptions are strongest in the orbital phase
477: interval 0.7--0.9. They nearly vanish at $\phi\sim$0--0.10 (around periastron passage),
478: and re-appear briefly right before the secondary eclipse at $\phi=$0.36. They are then
479: very weak or absent (or greatly displaced) between $\phi=$0.36 and $\sim$0.7.
480:
481: There is no immediate explanation for this behavior, although similar features as described
482: here have been linked to the WWI interface in the well-documented case of V444 Cygni
483: (Marchenko et al. 1994, 1997), GP Cep (Demers et al. 2002), $\gamma^2$ Vel (De Marco, O., 2002),
484: and $\eta$ Carinae (Nielsen et al. 2007). It is also interesting to note that in these four
485: systems the observations lead to the conclusion that there are substantial differences in
486: strengths and velocities of the leading/trailing WWC arms. A similar scenario for HD 5980
487: will be discussed in Section 8.
488:
489: In summary, we find that the only lines in the optical spectrum of HD 5980 that are likely
490: to truly describe the orbital motion of {\it star A} are N IV 4057, N V 4603 and
491: some of the weak He II lines in the $\lambda\lambda$6000--6200 \AA\ range.
492:
493: \section{New constraints on the masses of the system}
494:
495: The top panel of Figure 18 illustrates the RV curves that result when the centroid
496: of the emission lines is measured using a single Gaussian fit. The corresponding data
497: are listed in columns 4, 7 and 10 of Tables 2 and 3. The RV curve for N IV 4057 \AA\ line
498: is very similar to that derived by Barb\'a et al. (1997), indicating that {\it star A} is
499: still the major contributor to this line. The H I $+$ He II 4101 and He II 5411 \AA\ display a
500: similar trend, but with different amplitudes and shapes, implying that they contain a more
501: significant contribution from {\it star B} and from the WWI region.
502:
503: Assuming that the line profile variability is in part due to the superposition
504: of emission arising in two distinct sources, we deblended the lines using a two-function fit to
505: the profiles.
506: The results of these measurements are plotted in the middle and bottom panels of Figure 18
507: (filled-in pentagons), showing two RV curves in anti-phase. Using a similar procedure, we
508: deblended the N V 4603 emission (open stars) using two Gaussians. The results for N V are more
509: uncertain for two reasons: 1) the P Cygni absorption component of the neighboring N V 4621 \AA\
510: line limits the extent of the N V 4603 \AA\ red wing; 2) at $\phi=$0.13, {\it star B}'s component
511: is shifted into {\it star A}'s P Cygni absorption component. On some spectra, this superposition
512: makes it very difficult to measure the secondary component, since it simply fills in the P Cyg
513: absorption but is not prominent enough to enter the two-function fit. Because {\it star A}'s
514: contribution is dominant, its RV curve is not severely affected by these two effects. However,
515: that of {\it star B} not only displays much larger scatter, but also a systematic deviation from
516: the RV curve derived from N IV 4057. Part of this deviation may also be due to NV emission
517: arising in the WWI region.
518:
519: The middle panel of Figure 18 also contains the RV variations of the weak He II 6074.1 and
520: He II 6118.2 \AA\ lines (small open squares), obtained by fitting a single Gaussian function.
521: These lines show a slightly smaller RV amplitude than that of N IV 4057 \AA\, consistent with
522: the notion that they arise primarily in the wind of {\it star A} but also contain a contribution
523: from {\it star B}.
524:
525: The shape of the RV curve of {\it star B} shown in the bottom panel of Figure 18 matches quite
526: well the mirror of the RV curve for {\it star A}, within a constant factor (related to the relative
527: masses of the 2 stars), thus supporting the hypothesis that the excess emission (with respect to
528: the $\phi=$0 spectrum) is in fact primarily coming from the 2 stars. The WWI excess should be
529: out of phase with the orbital motion.
530:
531: We performed a fit to the RV amplitude of NIV and NV using the genetic algorithm PIKAIA
532: (Charbonneau, 1995). The radial velocity curves of the two stars were fitted simultaneously,
533: giving the two RV semi-amplitudes (K1 and K2), the eccentricity $e$, and the longitude of the
534: periastron $\omega_{per}$.
535: The semi-amplitudes are K$_A=$198 km/s (NIV) and 180 km/s (NV), and K$_B=$222 km/s from both
536: lines. These results lead to $M_A=$58--79 M$_\odot$ and $M_B=$51--67 M$_\odot$, adopting an
537: orbital inclination angle $i=$88$^\circ$ (Moffat et al. 1998), and given the uncertainties
538: in the fit to the RV curves ($\pm$4 km/s and $\pm$16 km/s for {\it star A} and {\it star B},
539: respectively). The corresponding range in semimajor axis is $a=$143--157 R$_\odot$. The
540: orbital eccentricity derived from both N IV and NV solutions is $e=$0.30$\pm$0.16. This value
541: is compatible with $e=$0.27 (Moffat et al. 1998) and $e=$0.32 (PB91). The longitude of periastron
542: that we derive lies in the range $\omega_{per}=$283$^\circ$--294$^\circ$. This is smaller than
543: the corresponding values of $\omega_{per}=$319$^\circ \pm$6 (Moffat et al. 1998) and
544: $\omega_{per}=$313$^\circ$ (BP91). Given that the RV curve has a limited orbital phase coverage
545: and that various interaction effects are present in the emission-line spectrum, we favor the
546: values of $\omega_{per}$ derived from the photometric light curve and polarimetry (BP91 and
547: Moffat et al. 1998).
548:
549: The above masses are significantly larger than those based on single function fits to the
550: emission lines and they are more consistent with the stars' position on the H-R Diagram near the
551: 120 M$_\odot$ evolutionary track (Koenigsberger 2004).
552:
553: \section{Declining wind densities ?}
554:
555: The emission lines in 2005 are substantially weaker than observed in 1999, indicating that
556: the trend for decreasing wind densities that was observed in UV spectra (Koenigsberger 2004)
557: is continuing. In order to quantify the effect, we measured the equivalent width (EW) of
558: H $\delta +$HeII 4101 \AA\, NIV 4057 \AA\, and He II 5411 \AA\ as listed in Tables 2 and 3.
559: The orbital phase-dependent variability in 1999 is plotted in Figure 19, showing that
560: He II 5411 \AA\ displays a large modulation with a broad minimum centered around $\phi\sim$0.15,
561: while N IV 4057 \AA\ gives only a hint of an eclipse at $\phi=$0.36. No clear phase-dependent
562: changes are evident in H$\delta +$He II 4101 \AA\.
563: Figure 20 illustrates the ratio EW(He II 5411)/EW(H$\delta +$He II 4101) as a function of
564: orbital phase for the two different epochs, 1999 and 2005, leading to the conclusion
565: that the degree of ionization in HD 5980 is greater in 2005 than in 1999.
566: Consistent with this conclusion is the significanly weaker N III 4640 \AA\ blend in
567: 2005 compared to that of 1999.
568:
569: Hence, these are indications that HD 5980's wind is reverting to its state prior to the 1993/1994
570: eruptions, when the emission lines were much broader and significantly weaker (Koenigsberger
571: et al. 1998a; Moffat et al. 1998).
572:
573: \section{Stationary photospheric absorptions}
574:
575: A striking feature in the 2005 FEROS data is the strength of photospheric absorptions
576: clearly seen superposed on the emission lines (see, for example, Figures 2, 7,
577: 8, 10, and 15).
578: Similar absorptions have been observed in 1999 HST/STIS spectra (Koenigsberger et al. 2003),
579: and found to be relatively stable in wavelength; that is, the amplitude of the RV variations was
580: $\leq$40 km/s, which was within the uncertainties of the RV measurements.
581: Table 5 lists the dominant absorption lines that we identified in the 2005 FEROS data, with
582: measurements of their RV (columns 2,4,6, and 8) and their corresponding FWHM (columns 3, 5, 7, and 9),
583: measured with Gaussian fits to the profiles. It is important to note that since these features are
584: superposed on strongly variable emission lines, the extent of the absorption line wings is nearly
585: impossible to establish. Hence, the RV measurements refer to the line cores and these do not display
586: any significant RV variations. In fact, if we average the RV values for each orbital phase
587: listed in Table 5, we find that the average RV values are virtually constant, within the standard
588: deviation of the measurements at each orbital phase. The averages and standard deviations are listed
589: in the bottom rows of Table 5. The average velocity at each of the 4 orbital phases listed in
590: this table are within the range 161--172 km/s, with a corresponding range in standard deviations
591: of $\pm$16 -- $\pm$22 km/s. Hence, they are indeed stationary within the timescale of the 2005
592: observations.
593:
594: The photospheric absorptions are believed to arise in {\it star C} (Niemela, 1988;
595: Koenigsberger et al. 2001) which is a reasonable conclusion, given their apparently stationary
596: nature over the P=19.2654 day orbital timescales. Over longer timescales, however, the RV and
597: strength of the lines undergo changes, as illustrated in Figure 21, where the He I 4471 \AA\
598: lines observed in 1999 and 2005 are compared. At each of the four orbital phases plotted in the
599: figure, the 1999 profiles are significantly redshifted with respect to those of 2005. Schweickhardt (2000)
600: measured the RVs of the O III 5592.4 photospheric absorption line in 1999 FEROS data,
601: finding them to be periodic with P$_{starC} \sim$96.5 days, and suggested that
602: {\it star C} is itself a binary system. His analysis of the RV curve implies that the {\it star C}
603: system is very eccentric ($e=$0.83; $\omega=$255$^\circ$; T$_0=$2451183.6) and that the amplitude
604: of the star responsible for the O III photospheric absorption is $K=$71 km/s. Similar results
605: were derived from the analysis of the He I 4471 \AA\ photospheric absorption. Our 2005 data
606: only cover a fraction of the phase for such a period, $\phi_{starC}\sim$0.4--0.8,
607: which lies in a relatively constant portion of the {\it star C} RV curve. However, OIII 5592.4
608: RV variations do show a declining trend with total amplitude of $\sim$20 km/s which is
609: consistent with Schweickhardt's (2000) conclusion. If this binary is gravitationally bound to the
610: A$+$B pair or whether it is merely a line-of-sight projection remains to be resolved. In any
611: case, however, Schweickhardt's (2000) results suggest that {\it star C} is also an intriguing
612: object that requires further analysis.
613:
614: An explanation for why the photospheric absorption lines appear to be more prominent in
615: 2005 compared with 1999 may lie in the decreasing visual brightness of {\it star A}
616: which thereby makes {\it star C}'s photospheric absorptions appear more prominent.
617: Another possibility is that since the emission lines became weaker in 2005 compared
618: with 1999, photospheric absorptions from {\it star A} could enhance the strength of
619: absorptions in the combined spectrum.
620:
621: In Table 5 we have also listed the measurements of a photospheric-like dip observed at
622: orbital phase $\phi=$0.13 ``blue-ward" of the stationary absorptions. It is typically
623: located $-$500 km/s from the stationary absorptions, with a FWHM$\sim$500 km/s. At this
624: phase, {\it star B} is approaching the observer and it is conceivable that these broad
625: features could be associated with its photospheric absorptions. The large width and the
626: excess blue-shift (with respect to the orbital motion) could be produced by the
627: velocity gradient in its photosphere. An alternative explanation may reside in the presence
628: of a slower-moving wind region (or WWI region ?) projected upon the stars at this phase.
629: However, this issue must await further observations since, as the intensity of the emission
630: lines from {\it star A} diminishes, the characteristics of photospheric absorptions will
631: become easier to establish.
632:
633: \section{Light curve}
634:
635: Figure 22 illustrates four sets of HD 5980 photometry. The best full orbital
636: phase coverage (open squares) consists of visual magnitude observations obtained using the
637: CTIO/SMARTS telescope during 2005. The second set having full phase coverage consists
638: of the original Breysacher \& Perrier (1980) differential magnitudes (crosses), with a constant value of
639: 10.415 mag added in to make the average levels coincide with those of the SMARTS and the Swope data at orbital
640: phase 0.60. The Swope data (small plus signs) has a poorer overall phase coverage, but covers
641: portions of the ascending branch of the $\phi=$0.36 eclipse more densely. Finally, the Danish telescope
642: differential magnitudes to which the constant value of 12.57 mag has been added (triangles) are primarily
643: concentrated around $\phi=$0.99 and around the $\phi=$0.36 eclipse phase. The top
644: panel of Figure 22 illustrates the complete light curve, while the middle and lower panels focus
645: on the two eclipses.
646:
647: %Sterken \& Breysacher give HJD(min)=2443158.701+/-0.07. Delta(phi)~0.005 implies Delta(t)~0.09 day.
648: The difference between the 2003 (SWOPE) and 2005 SMARTS data sets during the ascending
649: branches of both eclipses (there is practically no 2003 data during the descending branches) is
650: noteworthy. It shows the large degree of epoch-to-epoch variability in the system. Also,
651: we should note that in order to shift the photometric minimum to the anticipated $\phi =0.36$,
652: a slightly corrected $T_{modified}=$2443158.865 is suggested, which also better accomodates
653: the photometry obtained in 1994-1995 (Moffat et al. 1998). This implies a significantly larger
654: difference with respect to the Sterken \& Breysacher (1997) T$_0=$2443158.707$\pm$0.07 than their
655: quoted uncertainty. The possibility that this shift is related to apsidal motion and a full
656: analysis of the eclipse light curve will be explored in a forthcoming investigation. It
657: is important to note that the analysis requires a methodology as that used by Breysacher \& Perrier
658: (1991) in which the extended nature of the continuum eclipsing sources is contemplated.
659:
660: The longest train of Danish data was obtained around eclipse at $\phi=$0.36--0.37, and the individual
661: data points are plotted in Figure 23, disclosing $\sim$0.$^{mag}$035 amplitude variations within a
662: timespan of half an hour. This is considerably shorter than the $\sim$6-hour periodic variations
663: (0.02 mag) reported by Sterken \& Breysacher (1997) and attributed to oscillations of {\it star A}.
664: Curiously, if we assume T$_0=$T$_{modified}=$JD 2443158.865, then the two prominent minima appear symmetrically
665: located around $\phi_{modified}=$0.36 (where $\phi_{modified}$ is the phase computed using the modified
666: value of T$_0$). This would seem to indicate the presence of a bright emitting source that is
667: eclipsed just prior to and just after mid-eclipse, but not during mid-eclipse. Or, alternatively,
668: the presence of dense, eclipsing gas located at a distance from {\it star B} that is slightly larger
669: than the radius of {\it star A}. One geometry for such a gas that comes to mind is that of a WWI
670: interacting region whose apex is transparent to optical radiation.
671:
672: Strong variability near $\phi=$0.36 on similar timescales has previously been reported in polarimetric
673: observations (Villar-Sbaffi et al. 2003), and two HST/STIS UV observations obtained one day apart
674: at these same phases suggest that the wind speeds are very variable.
675:
676: \section{The wind-wind interaction region}
677:
678: The current view of the HD 5980 system, consisting of two stars with powerful winds, implies
679: the presence of an interaction zone. Thus, at least part of the
680: variability reported in this paper may result from phase-locked wind-wind interaction effects.
681: In its simplest representation, the WWI region can be modeled in terms of a discontinuity
682: between the two winds that consists of the two thin shocks formed when the supersonic winds
683: collide and whose shape depends on the momentum ratio of the stellar winds. Figure 24
684: is a cartoon illustrating the WWI region shape for HD 5980, assuming that the wind of {\it star A}
685: dominates. This assumption is based on the fact that all emission lines in the combined spectrum
686: increased significantly in strength after the 1994 eruption and that the phase of their Doppler motions
687: corresponds to that of {\it star A} (Koenigsberger 2004). The significantly larger emission measure
688: from {\it star A} compared to {\it star B} is expected to correlate with a larger density.
689: Preliminary stellar wind models for the HD 5980 system that have been fit to FUSE 2002 data
690: imply that {\it star A}'s mass-loss rate was still large, $\dot{M}_A\sim$10$^{-4}$ M$_\odot$/yr
691: (Georgiev et al., in preparation). In addition, for {\it star B}, Moffat et al. (1998) found
692: $\dot{M}_B=$2$\times$10$^{-5}$ M$_\odot$/yr from polarimetric data obtained prior to the eruption.
693: These mass-loss rates, combined with wind terminal speeds derived from UV and FUV observations
694: lead to the WWI shock cone configuration shown in Figure 24, with a shock cone opening angle
695: $\sim$50$^\circ$.
696:
697: Four orbital phases are shown. {\it Star A} is represented with three concentric
698: circles depicting, respectively, its photosphere (R$_{A}$), the extent of its accelerating wind region in
699: the case of a ``fast" velocity law ($\sim$1.5 R$_{A}$), and the extent of the dominant emitting
700: wind region for a moderately strong, but arbitrary, emission line ($\sim$4 R$_{A}$). Note that
701: because the orbit is eccentric, the WWI discontinuity intersects the line emitting region around
702: periastron passage, but not around apastron. When the WWI region intersects the line-emitting
703: region, the intensity of the observed line-profile is truncated at the corresponding velocity range.
704: The sets of dashed lines drawn parallel to the line-of-sight in Figure 24 enclose columns of
705: wind material that are projected against the luminous stellar continua and that are expanding
706: towards the observer (located at the bottom of the figure). All P Cygni absorption components observed in
707: the spectrum arise in these columns.
708:
709: The behavior of the blue-shifted, sharp absorption components of the He I emission lines
710: suggests that they arise in the WWI region. They are strongest during the orbital
711: phase interval 0.7--0.9 (when both stars are in view and one of the WWI arms lies in the
712: foreground of {\it star B}) and they are weak at $\phi\sim$0--0.10 (when mostly {\it star A} is in view)
713: and $\phi=$0.36 -- 0.7 (when mostly {\it star B} is in view).
714: It is important to note that the sharp and deepest portion of the P Cygni absorption components
715: generally arise in the wind regions that have attained terminal speeds. In the 1999 data, at
716: orbital phase $\sim$0.8, the He I 4471 \AA\ sharp feature extends between -900 to -1340 km/s, and
717: maximum absorption at $\sim -$1130 km/s (assuming a correction for SMC motion of $-$150 km/s). This is
718: significantly slower than terminal wind speed of $-$1750 km/s inferred for {\it star A} during
719: the same epoch from HST/STIS observations of UV lines (Koenigsberger et al. 2000). This slower
720: speed is qualitatively consistent with absorption of {\it star B} radiation traversing the
721: foreground WWI region arm. Also noteworthy is the fact that excess blue-shifted emission appears
722: in many lines around orbital phase $\phi=$0.36, which is tempting to attribute to emission arising
723: in the WWI zone that is not projected against either continuum source.
724:
725: Figure 25 provides estimates of the flow velocities, V$_t$, along the shock cone walls derived
726: from the expressions given by Cant\'o et al. (1996), using the roughly estimated values for mass-loss
727: rates of the two stars. At orbital phases around 0.8, the line-of-sight from the observer to
728: {\it star B} crosses regions of the WWI having V$_t \sim$600--800 km/s. Considering that the
729: shock cone has an opening angle $\sim$50$^\circ$, the maximum projected speed of V$_t$ along the
730: line-of-sight in the stationary frame of reference is $\sim-$600 km/s. This is significantly smaller
731: than the speeds at which the He I P Cygni absorption components are being observed. Correcting for the
732: orbital motion of {\it star B} of $\sim$220 km/s brings the predicted value closer to the
733: observations at $\phi \sim$0.15, when {\it star B} is approaching the observer, but increases the
734: discrepancy round $\phi\sim$0.8. However, the idealized WWI model presented
735: in Figures 24 and 25 is a greatly simplified model. For example, if we take into account that the
736: axis of symmetry for the WWI shock cone is in reality tilted with respect to the line connecting the
737: two stars, the angle between the leading WWI arm and the line-of-sight decreases, allowing for
738: faster projected speeds. In addition, numerical simulations that include orbital motion show that
739: the shock cone arms are curved. Hence, the leading and the trailing arms could cross the
740: line-of-sight to the stars at at different orbital phases and with larger projected speeds than
741: depicted in Figure 25. Figure 26 is a cartoon that illustrates an alternative geometry for the
742: WWI region, based on an interpretation of the variable He I P Cygni absorption lines. It would
743: be interesting to analyze whether the physical mechanisms involved in the WWI can give rise to
744: such a different geometry, compared to the typical conical surfaces that are predicted by the
745: traditional calculations.
746:
747: It is important to keep in mind that the geometry and characteristics of the interaction region
748: depend on whether the shocked gas is radiative or not, and this, in turn, determines the intensity
749: and shape of line emission/absorption that the WWI region can produce. One can estimate whether the
750: gas is radiative or not using the parameter $\chi$ from Stevens et al. (1992), which is the ratio of
751: the postshock cooling timescale to the flow time out of the system. Hydrodynamical models have indeed
752: confirmed that the postshock gas cools rapidly when $\chi\sim$1.0.
753:
754: To evaluate $\chi$ we need to determine the pre-shock velocity. $\chi$ is highly
755: dependent on this value ($\propto v^4$) so small changes in $v$ make a big
756: change in $\chi$. Using the approximation to the velocity law as in St.Louis et al. (2005),
757:
758: \begin{equation}
759: v = v_0 + (v_\infty - v_0)*(1 - R_*/r)^\beta
760: \end{equation}
761: \noindent with $v_0 = 0.16 v_\infty$ and $\beta=$2.5.
762:
763: With such a velocity profile the wind accelerates so slowly that the
764: post-shock gas (from both stars) is always highly radiative, and the Canto et al. (1996)
765: prescription used in Figures 24 and 25 is appropriate. However, if the wind(s)
766: accelerate quicker (eg $\beta=$1) then there is a chance that, for example, {\it star B}'s
767: shocked wind is adiabatic near apastron (but strongly cooling at periastron).
768: %Perhaps we should give a range of values for \chi for different
769: %velocity laws? For instance using the slow wind accleration law noted above
770: %I estimate that \chi is approximately:
771: %
772: % Star A Star B
773: %%Peri 0.007 0.024
774: %Ap 0.05 0.27
775: %
776: %If beta=1.0 and v_0 = 0.0, then \chi is:
777: %
778: % Star A Star B
779: %%Peri 0.027 0.21
780: %Ap 0.2 3.0
781:
782: The fact that the X-ray emission is apparently harder near phase
783: 0.36 (Naz\'e et al. 2007) is very suggestive that the maximum pre-shock
784: velocity increases as the stars separate, although it is by no means certain
785: that the winds between the stars bear much relation to the winds from single
786: stars. Indeed, indications exist that at least {\it star A}'s wind velocity
787: structure facing the companion differs from that of its opposite hemisphere
788: (Koenigsberger et al. 2006).
789:
790: Unfortunately there have been very few hydrodynamic calculations of the wind
791: interaction in short period systems like HD 5980 where the acceleration of
792: the winds is taken into account (Iota Ori, Pittard 1998; Sand 1, St-Louis et
793: al. 2005). Antokhin et al. (2004) took a slightly different approach which
794: enabled them to calculate the X-ray emission, but the wind needed to be
795: specified by a beta velocity law and was not calculated self-consistently
796: with inhibition and braking in the way that the other hydro models noted were.
797:
798: If both winds rapidly cool after being shocked, the postshock
799: region collapses into a dense sheet and the shocks are coincident with the
800: contact discontinuity. It is worth noting that in this case one expects
801: very strong instabilities in the WWI region (specifically the non-linear
802: thin-shell instability - see Stevens et al 1992), which means that the
803: position of the contact discontinuity from the Canto et al. model should be
804: viewed only in a time-averaged sense. This is of course a possible cause of
805: the short-timescale fluctuations that are mentioned in Section 7).
806:
807: Wind clumping is an additional factor that is absent from the simplified view
808: of WWI regions. Pittard (2007) presents simulation of the effects of clumps in
809: binary systems with adiabatic colliding winds. Although the clumps are likely
810: destroyed quite rapidly in such systems, they create a broader distribution of
811: temperatures within the WWI at any given off-axis distance. This is because
812: the interclump material is heated to higher than average temperatures, while the
813: clumps are initially heated to significantly lower temperatures (the shocks
814: driven into them are slower) - the temperature within shocked clumps is lower than the
815: temperature obtained from the collision of smooth winds by a factor equal to the
816: density contrast of the clump to interclump material. If most of the wind mass is within
817: the clumps this should have observable consequences. The shocked material within the
818: clumps is then heated further by additional weaker shocks and mixing within the WWI.
819: In systems where the wind parameters are such that a WWI formed by smooth winds is
820: not far from being radiative, the presence of higher densities within clumps may be
821: enough for cooling to become important (see Walder \& Folini 2002).
822:
823: \section{Conclusions}
824:
825: In this paper we present results from spectroscopic and photometric observations of
826: HD 5980 obtained in 1999 and 2003-2005. We first analyze the line-profile variability
827: in the 2005 spectra, concluding that in lines that are formed in the dense, inner
828: wind regions, line-profile-variability may be interpreted mostly in terms of
829: the presence of two emission components, the first belonging to {\it star A} (the eruptor) and the
830: second associated with {\it star B} (the companion). Assuming that the emission-line profiles can be
831: effectively separated with a two-function fit, we find that the centroids of the two functions are
832: consistent with the orbital motion of the two stars, leading to a new estimate of the masses,
833: M$_A\sim$58--79 M$_\odot$ and M$_B\sim$51--67 M$_\odot$. These values are significantly larger than
834: previously derived (Niemela et al. 1997), but still imply that the stars must have lost considerable
835: amounts of mass before reaching their current evolutionary state, given their location on the HRD
836: near 120 M$_\odot$ ZAMS evolutionary tracks (Koenigsberger 2004). The mechanisms by which such
837: large amounts of mass may have been stripped require detailed examination. In addition, it would
838: be very interesting to get more theoretical insight into their current internal stellar structure
839: and hence, be able to establish their current evolutionary state. As pointed out by Koenigsberger
840: et al. (1998b), if {\it star B} is a {\it bona fide} WNE remnant \footnote{that is, the bare He-
841: or C-burning core of a massive progenitor that has been stripped of its outer layers, as opposed
842: to a massive star in an earlier evolutionary stage undergoing strong wind mass-loss} of the
843: originally more massive member of the binary, its interior may have already reached a very
844: advanced stage of evolution and, thus, it may be nearing the supernova stage. On the other hand,
845: the significantly larger mass that we have now derived for {\it star B} (M$_B\sim$56--67 M$_\odot$
846: {\it vs.} $\sim$28 M$_\odot$ as believed previously) suggests that it may not yet have reached the
847: core He burning evolutionary phase. Since {\it star B} is believed to be the source of the
848: highly He-enriched WR-spectrum observed in the 1970's (Breysacher, Moffat \& Niemela 1982), the
849: implications for stellar evolution theory of such a massive, He-rich object are potentially
850: very interesting.
851: %If this is the case, the presence in this star of prominent He II, N V and N IV emission lines
852: %is intriguing.
853:
854: Line strengths in the 2005 spectra are significantly weaker than those
855: observed in 1999, consistent with a similar decreasing trend observed in UV lines, and
856: suggesting that the eruptor is reverting to a more quiescent state. The question of
857: whether another activity cycle will start in the near future is an important one,
858: since the timescale for the eruptive events provides an estimate of the total amount
859: of mass that may be lost through this mechanism. In addition, the activity timescale
860: provides a clue to understanding the underlying mechanism responsible for the
861: instability. Continued observations are required to determine the long-term
862: variability pattern.
863:
864: Visual photometric observations in 2005 provide a light curve covering the entire orbital
865: cycle, as well as a detailed description of the system's brightness variations during
866: selected orbital phases. In particular, relatively large-amplitude ($\sim$0.035 mag)
867: fluctuations appear around secondary minimum, when {\it star B}, the non-erupting star,
868: is ``in front". Continued systematic observational monitoring around this orbital phase would
869: be very valuable because of the strong and rapid variations observed in visual
870: photometry, polarimetry and UV P Cygni line profiles. In addition, the X-ray appears to
871: be greater here than around the opposite eclipse.
872:
873: The temporal behavior of the P Cygni absorption components of the He I lines is qualitatively
874: consistent with the presence of a denser wind structure folding around {\it star B} which we associate
875: with the WWI region. However, the observed speeds are not consistent with the simple models.
876: In particular, the flow speeds are significantly larger than predicted in the radiative shock
877: limit and thus suggest that at least one of the shocks is adiabatic. It is interesting to note
878: that the hemisphere of {\it star A} that faces the companion is predicted to
879: be particularly active due to tidal interaction effects (Koenigsberger \& Moreno 2007),
880: raising the question of how this activity may affect the structure of the WWI region.
881:
882: The photospheric absorption lines remain stationary over the 19-day orbital cycle. However, we find a
883: systematic shift of $\sim$20 km/s in their radial velocities observed in 1999 with respect to those observed
884: in 2005. This is consistent with the conclusion of Schweickhardt (2000) that {\it star C}, the third
885: source in the combined HD 5980 spectrum, is itself a binary system, with P$_{starC} \sim$96 days.
886: It is unclear, however, whether the {\it star C} system is gravitationally bound to the
887: {\it star A+star B} pair with a much longer orbital period. It is a curious fact that the {\it star C}
888: orbital period is nearly a factor of 5 longer than the orbital period in the {\it star A+ star B} pair, which
889: may indicate that the two systems are physically bound. If this were the case, then the eruptive
890: behavior in HD 5980 could be linked to the gravitational perturbations produced during periastron
891: passages. In this case the orbital period of the {\it star C} system around the {\it star A+ star B}
892: system would be $>$25 years, since only one large eruption has been seen over the $\sim$40 years
893: that the system has been studied.
894:
895: In the Appendix we list the SWOPE and SMARTS photometry averages for each day of observation (Tables 7--10).
896: The data presented in this paper will be made publicly available (please contact CF or GK).
897:
898: \acknowledgements
899:
900: We express our gratitute to Catherine Cesarsky for having granted European Southern Observatory
901: Director Discretionary observing time that provided the 2005 FEROS and EMMI observations and
902: to the ESO observing staff for carrying out these observations; to Jens Hjorth for granting
903: time on the Danish telescope and Uffe Graae J\-orgensen, Daniel Kubas, Chlo\'e Feron and Christina
904: Th\"one for performing the observations. We thank Felix Mirabel for hosting the visit of OT to ESO in Chile.
905: Some of the photometric data were obtained via the SMARTS consortium, and we are grateful to the various
906: observers who contributed to this effort. We acknowledge support to CF from the Swiss National Science
907: Foundation; to GK from PAPIIT/DGAPA/UNAM IN 119205; to PM from NSF grant AST-056577; to AFJM and NSL from
908: NSERC (Canada) and FQRNT (Quebec); and to YN from the FNRS and the Prodex XMM-Integral contract (Belspo).
909: This research was supported in part by the Danish Natural Science Research Council through its centre
910: for Ground-Based Observational Astronomy, IJAF/IDA, and by the Gemini Observatory,
911: which is operated by the Association of Universities for Research in Astronomy, Inc., on behalf
912: of the international Gemini partnership of Argentina, Australia, Brazil, Canada, Chile, the
913: United Kingdom, and the United States of America.
914:
915: \vspace*{7pt}
916: %\begin{thebibliography}
917: \begin{itemize}
918:
919: \item Antokhin, I.I., Owocki, S.P., Brown, J.C. 2004, ApJ, 611, 434.
920: \item Auer, L.H. \& Koenigsberger, G. 1994, ApJ, 436, 859
921: \item Barb\'a R.H., Morrell, N.I., Niemela, V.S., Bosch, G.L., Gonz\'alez, J.F., Lapasset, E., Ferrer, O.E., Brandi, E.E., Cellone, S.A., Garcia, B.E., Malaroda,S.M., Levato, O.H., Donzelli, C., Feingstein, C., Rich, M. 1996, RMAACS, 5, 85.
922: \item Barb\'a R., Niemela V.S., Morrell, N. 1997, in ASP Conf. Ser. 120, Luminous Blue Variables:Massive stars in transition, ed. A. Nota \& H. Lamers (San Frnacisco:ASP), 238.
923: \item Bateson, F.M. \& Jones, A. 1994, Publ. Var. Star. Sec. R. Astron. Soc. New Zealand, 19.
924: \item Breysacher, J.,François P. 2000, A\&A, 361, 231.
925: \item Breysacher, J. \& Perrier, C. 1980, A\&A 90, 207 (BP80)
926: \item Breysacher, J. \& Perrier, C. 1991, in IAU Symp. 143, 143, Wolf Rayet Stars and Interrelations with other Massive Stars in Galaxies, eds. K. van der Hucht \& B. Hidayat (Dordrecht: Kluwer),~229 (BP91)
927: \item Breysacher, J., Moffat, A. F. J., \& Niemela, V. 1982, ApJ, 257, 116
928: \item Breysacher, J., \& Westerlund, B. E. 1978, A\&A, 67, 261
929: \item Cant\'o J., Wilkin, \& Raga, A. 1996, ApJ, 469, 729
930: \item Charbonneau, P. 1995, ApJS 101 309
931: \item De Marco, O. 2002, ASPC, 260, 517.
932: \item Demers, H., Moffat, A.F.J, \& Marchenko, S.V. 2002, ASPC 260, 563.
933: \item Drissen, L., Crowther, P.A., Smith, L.J., Robert, C., Roy, J.-R., \& Hillier, D.J. 2001, ApJ 545, 484.
934: \item Everett, M.E., \& Howell, S.B. 2001, PASP, 113, 1428
935: \item Gayley, K. et al. 1997, ApJ 475, 786
936: \item Georgiev, L. \& Koenigsberger, G. 2004a, A\&A, 423, 267.
937: \item Georgiev, L. \& Koenigsberger, G. 2004b, IAU Symp. 215, {\it Stellar Rotation}, (ed) Maeder\& Eenens, p. 19
938: \item Hoffmann, Stift \& Moffat 1978, PASP 90, 101.
939: \item L\"uhrs, S. 1997, PASP 108, 504
940: \item Kaufer, A., Schmid, H.M., Schweickhardt, J., Tubbesing, S. 2002, in ``Interacting Winds from Massive Stars", ASP Conf. Ser. Vol. 260, 489.
941: \item Koenigsberger, G., Auer, L.H., Georgiev,L. and Guinan, E. 1998a, ApJ, 496, 934
942: \item Koenigsberger, G., Pe\~na, M., Schmutz, W. and Ayala, S. 1998b, ApJ, 499, 889
943: \item Koenigsberger, G., Georgiev, L., Barb\'a, R., Tzvetanov, Z., Walborn, N.R.,Niemela, V., Morrell, N., y Schulte-Ladbeck, R. 2000, ApJ, 542, 428
944: \item Koenigsberger, G. 2004, RMAA, 40, 107
945: \item Koenigsberger, G., Fullerton, A., Massa, D., Auer, L. 2006, AJ, 132, 1527
946: \item Koenigsbeger, G. \& Moreno, E. 2007, {\it arXiv0705.1938}.
947: \item De Marco, O. 2002, ASP Conf. Ser. 260, 517
948: \item Marchenko, S.V., Moffat, A.F.J. \& Koenigsberger, G. 1994, ApJ, 422, 810
949: \item Marchenko, S.V., Moffat, A.F.J., Eenens, P.R.J., Cardona, O., Echevarria, J. Hervieux, Y. 1997, ApJ, 485, 826
950: \item Massey, P., Parker, J.W., Garmany, C.D. 1989, AJ 98, 1305.
951: \item Moffat, A.F.J., Marchenko, S.V., Bartzakos, P., Niemela, V.S., Cerruti, M.A.,Magalhaes, A.M., Balona, L., St.-Louis, N., Seggewiss, W., Lamontagne, R. 1998, ApJ 497, 896
952: \item Naz\'e, Y., Corcoran, M.F., Koenigsberger, G., Moffat, A.F.J. 2007, ApJLett, 658, 25.
953: \item Nielsen,K.E., Corcoran, M.F., Gull, T.R., Hillier, D.J., Hamaguchi, K., Ivarsson, S., Lindler, D.J. 2007, ApJ 660, 669.
954: \item Niemela, V.S. 1988, ASPCS 1, 381
955: \item Niemela, V.S., Barb\'a, R.H., Morrell, N.I. and Corti, M. 1997, in ASP Conf Ser. 120, A. Nota and H.J.G.L.M. Lamers (eds), p.222.
956: \item Pittard, J.M. 1998, MNRAS, 300, 479
957: \item Pittard, J.M. 2007, ApJ, 660, L141.
958: \item Schweickhardt, J. 2000, PhD Thesis, Ruprecht-Karls-Universit\"at, Heidelberg.
959: \item Sterken C., Breysacher J. 1997, A\&A 328, 269
960: \item Stevens, I.R., Blondin, J.M., Pollock, A.M.T. 1992, ApJ 386, 265.
961: \item Stevens, I.R. 1993, ApJ, 404, 281.
962: \item St.-Louis, N., Moffat, A.F.J., Marchenko, S.V., Pittard, J.M. 2005, ApJ, 628, 953.
963: \item St.-Louis, N., Moffat, A.F.J., Marchenko, S.V., Pittard, J.M., Boisvert, P. 2006, ASPC,
964: 348, 121.
965: \item Villar-Sbaffi, A., Moffat, A.F.J., St.-Louis, N. 2003, ApJ 590,483
966: \item Walder, R. \& Folini, D. 2002, ASP Conf. Ser. 260, 595.
967:
968: %\end{thebibliography}
969: \end{itemize}
970: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
971:
972: \begin{table*}[!t]\centering
973: \tablecols{4}
974: \setlength\tabnotewidth{0.60\textwidth}
975: \setlength{\tabcolsep}{0.9\tabcolsep}
976: \small
977: \caption{Summary of HD~5980 Estimated Parameters}
978: \begin{tabular}{llll}
979: \toprule
980: Parameter & 1979-81 & 1994 &1999-2005 \\
981: \midrule
982: $P_{orb}$ (days) & 19.2654 & $\cdots$ & $\cdots$ \\
983: Orbital inclination (deg) & 88\tabnotemark{m} & $\cdots$ & $\cdots$ \\
984: a ($R_\odot$) & 127\tabnotemark{b} & $\cdots$ & 143--157\tablenotemark{k} \\
985: e & 0.3 & $\cdots$ &0.30$\pm0.16$ \tablenotemark{k} \\
986: $M_{\rm A}$ (M$_\odot$) &50\tabnotemark{b} & $\cdots$ & 58--79\tablenotemark{k} \\
987: $M_{\rm B}$ (M$_\odot$) & 28\tabnotemark{b} & $\cdots$ & 51--67\tablenotemark{k} \\
988: M$_V^{sys}$ (mag) & $-$7.3\tabnotemark{a} &$-$8.7\tabnotemark{e} &$-$7.6\tabnotemark{h} \\
989: M$_V^A$ (mag) & $-$6.3 &$-$8.6\tabnotemark{e} & $-$6.5 \\
990: M$_V^B$ (mag) & $-$5.8 & $\cdots$ &$\cdots$ \\
991: M$_V^C$ (mag) & $-$6.1 & $\cdots$ &$\cdots$ \\
992: $R_{{\rm A}}$ (R$_\odot$) & 21\tabnotemark{a} & $> 160$\tabnotemark{d} & $\cdots$ \\
993: & $\cdots$ & 280\tabnotemark{g} & $\cdots$ \\
994: $R_{{\rm B}}$ (R$_\odot$) & 15\tabnotemark{a} & $\cdots$ &$\cdots$ \\
995: $R_{{\rm env B}}$ (R$_\odot$)& 30--40\tabnotemark{a} & $\cdots$ &$\cdots$ \\
996: $v_\infty^A$ (km/s) & $\cdots$ & 300--1700 & 1740--2200: \\
997: & $\cdots$ & 500\tabnotemark{g} & $\cdots$ \\
998: $v_\infty^B$ (km/s) & 3100: & $\cdots$ & 2600--3100: \\
999: $v_\infty^C$ (km/s) & $\cdots$ & $\cdots$ &1700:: \\
1000: N[He]/N[H]$^{aa}$$_{\rm Non-LTE}$ & $\cdots$ & 0.4 \tabnotemark{e} &$\cdots$ \\
1001: $\dot{M}_A$ (M$_\odot$/yr) & $\cdots$ & $10^{-3} $\tabnotemark{e}&$\cdots$ \\
1002: $\dot{M}_B$ (M$_\odot$/yr) & 2$\times$10$^{-5} $\tabnotemark{m} & & \\
1003: $L_A$ (L$_\odot$) & $\cdots$ & $3\times 10^6 $\tabnotemark{e} & $\cdots$ \\
1004: & $\cdots$ & $10^7 $\tabnotemark{g} &$\cdots$ \\
1005: $T_*^A$ $^\circ$K &53,000\tabnotemark{c} & 21,000\tabnotemark{d} &$\cdots$ \\
1006: & $\cdots$ & 35,500\tabnotemark{e} & $\cdots$ \\
1007: Spectrum ({\rm Star A}) & WN3\tabnotemark{b}& $B1.5Ia^+$\tabnotemark{d} & WN6\tabnotemark{d}\\
1008: Spectrum ({\rm Star B}) &WN4\tabnotemark{b} & WN4: & WN4: \\
1009: Spectrum ({\rm Star C}) &$\cdots$ & $\cdots$ & O4-6\tabnotemark{f}\\
1010: & $\cdots$ & $\cdots$ & O7I \tabnotemark{n}\\
1011: $V_{rot}^A \, sin~i$ (km/s) &$\cdots$ & $\cdots$ & 250::\tabnotemark{j} \\
1012: $V_{rot}^C \, sin~i$ (km/s) &$\cdots$ & $\cdots$ & 75\tabnotemark{i} \\
1013: P$_{orb}$(C) (days) & $\cdots$ & $\cdots$ & 96.5 \tabnotemark{s} \\
1014: \bottomrule
1015: \tabnotetext{aa}{Abundance ratio by number (0.63 by mass); \quad
1016: \textsuperscript{a}Breysacher \& Perrier 1991; \quad \textsuperscript{b}Niemela et~al. 1997;\quad
1017: \textsuperscript{c}Assuming $L_{\rm star A}$ ($L_\odot$)=const and
1018: $R_{\rm star A}$/$R_\odot$=21; \quad
1019: \textsuperscript{d}Koenigsberger et~al.\@ 1998a; \quad
1020: \textsuperscript{e}December 30,
1021: Koenigsberger et~al.\@ 1998b; \quad
1022: \textsuperscript{f}Koenigsberger et~al.\@ 2000; \quad
1023: \textsuperscript{g}Drissen et~al.\@ 2001; \quad
1024: \textsuperscript{h}S. Duffau, assuming m-M=19.1 mag;\quad
1025: \textsuperscript{i}Koenigsberger et~al.\@ 2001;\quad
1026: \textsuperscript{j}Georgiev \& Koenigsberger 2004b;\quad
1027: \textsuperscript{k} This paper;\quad
1028: \textsuperscript{m} Moffat et al. 1998; \quad
1029: \textsuperscript{n} Koenigsberger, Fullerton, Massa \& Auer 2006;\quad
1030: \textsuperscript{s} Schweickhardt, J. 2002.
1031: }
1032: \end{tabular}
1033: \vspace*{0.7\baselineskip}
1034: \end{table*}
1035:
1036:
1037: \begin{table*}[!t]\centering
1038: \tablecols{12}
1039: \setlength\tabnotewidth{0.60\textwidth}
1040: \setlength{\tabcolsep}{0.9\tabcolsep}
1041: \small
1042: \caption{FEROS 1999 data}
1043: \begin{tabular}{lrrrrrrrrrrr}
1044: \toprule
1045: spectrum& MJD & phase & NIV & 4057.76 & & HI &4101.71& & HeII & 5411.52& \\
1046: & & & RV & FW &$-$EW & RV & FW &$-$EW& RV & FW &$-$EW\\
1047: & & &km/s &km/s & \AA &km/s &km/s & \AA &km/s &km/s & \AA \\
1048: \midrule
1049: % Jan. 2, 2007: included 2005 data; for HI4101, I cut out the phot. abs. for phi=0.37, 0.98, I used
1050: % the "norm" spectrum; not the "renorm". Velocities are measured wrt lab. wavelengths. Used "k" or "d"
1051: % ?? unable to measure 2005 spectra at phi~0.13 for HI 4101 because don't know how to interpret the profile.
1052: % ?? unless the depressed blue side is due to strong photospheric abs. from {\it star B}, blueshifted by ~300 km/s ??
1053: % OR: How about start of P Cyg abs, near zero velocity, but rest of P Cyg abs. is filled-in by emission
1054: % that is NOT projected against {\it star A}'s continuum and therefore, is superposed on ``intrinsic" P Cyg profile ?
1055: % Num MJD phi RV GFWHM -EW | RV GFWHM -EW | RV GFWHM -EW |
1056: % 4057.76 \AA | 4101.71 \AA| 5411.52 \AA
1057: % &(4) &(5) & (6) & (7) & (8) &(9) & (10) & (11) & (12) \\
1058: 78781 & 51374.418 & 0.475 & 13 & 685 &3.83 & 76 & 1105 &7.94 & 203 & 907 & 14.79 \\
1059: 79411 & 51375.422 & 0.527 &-10 & 694 &4.46 & 63 & 1221 &8.99 & 221 & 993 & 16.04 \\
1060: 80871 & 51379.402 & 0.734 &-50 & 689 &2.99 & 56 & 1515 &9.03 & 201 &1217 & 15.62 \\
1061: 81451 & 51380.398 & 0.786 &-55 & 685 &3.05 & 53 & 1546 &8.80 & 211 &1306 & 15.21 \\
1062: 82121 & 51381.402 & 0.838 &-23 & 705 &2.79 & 47 & 1538 &8.58 & 218 &1249 & 14.84 \\
1063: 83181 & 51382.352 & 0.887 &-0.5 &711 &3.03 & 66 & 1505 &8.91 & 227 &1213 & 15.18 \\
1064: 83811 & 51383.355 & 0.939 & 50 & 679 &3.14 & 97 & 1527 &9.08 & 251 &1188 & 14.11 \\
1065: 84511 & 51384.309 & 0.989 &116 & 697 &3.39 &113 & 1316 &8.51 & 226 &1066 & 13.48 \\
1066: 85181 & 51385.332 & 0.042 &197 & 663 &3.72 &187 & 1072 &9.41:& 317 & 769 & 13.27 \\
1067: 85771 & 51386.391 & 0.097 &229 & 760 &3.46 &210 & 1297 &8.55 & 340 &1023 & 14.22 \\
1068: 86321 & 51388.340 & 0.198 &187 & 810 &3.98 &233 & 1553 &8.54 & 404 &1185 & 12.99 \\
1069: 86701 & 51389.383 & 0.252 &138 & 752 &3.53 &230 & 1424 &8.02 & 399 &1049 & 13.06 \\
1070: 87211 & 51390.328 & 0.301 &118 & 647 &3.46 &143 & 1373 &8.41 & 328 &1070 & 14.01 \\
1071: 87751 & 51391.332 & 0.353 & 92 & 624 &2.68 & 66 & 1215 &7.42 & 207 & 946 & 13.46 \\
1072: 88931 & 51392.328 & 0.405 & 49 & 609 &3.28 & 43 & 1170 &8.48 & 175 & 872 & 15.55 \\
1073: 89511 & 51393.371 & 0.459 & 23 & 658 &3.57 & 55 & 1118 &8.22 & 195 & 880 & 14.91 \\
1074: 90131 & 51394.391 & 0.512 &0.5 & 671 &3.76 & 54 & 1136 &8.43 & 210 & 921 & 15.82 \\
1075: \bottomrule
1076: \end{tabular}
1077: \end{table*}
1078:
1079: \begin{table*}[!t]\centering
1080: \tablecols{12}
1081: \setlength\tabnotewidth{0.60\textwidth}
1082: \setlength{\tabcolsep}{0.9\tabcolsep}
1083: \small
1084: \caption{FEROS 2005 data}
1085: \begin{tabular}{lrrrrrrrrrrr}
1086: \toprule
1087: spectrum& MJD & phase & NIV & 4057.76 & & HI &4101.71& & HeII & 5411.52& \\
1088: & & & RV & FW &$-$EW & RV & FW &$-$EW& RV & FW &$-$EW\\
1089: & & &km/s &km/s & \AA &km/s &km/s & \AA &km/s &km/s & \AA \\
1090: \midrule
1091: 0617801 & 53538.385& 0.799 &-29 & 760 &2.56 & 21 & 1860 &5.82 & 211 &1622 & 15.33 \\
1092: 0617531 & 53538.404& 0.800 &-31 & 756 &2.63 & 12 & 1903 &5.80 & 221 &1643 & 14.56 \\
1093: 0620221 & 53541.384& 0.955 & 78 & 800 &2.70 &136 & 1692 &5.31 & 278 &1454 & 12.96 \\
1094: 0620344 & 53541.403& 0.955 & 73 & 810 &2.99 &120 & 1608 &5.62 & 291 &1497 & 13.38 \\
1095: 0710740 & 53561.367& 0.992 & 98 & 745 &3.13 &141 & 1409 &6.37 & 281 &1099 & 14.07 \\
1096: 0710950 & 53561.386& 0.993 & 98 & 729 &3.11 &143 & 1395 &5.98 & 284 &1068 & 14.83 \\
1097: 0725580 & 53576.288& 0.767 &-10 & 817 &2.65 & 91 & 2003 &5.13 & 270 &1705 & 15.39 \\
1098: 0725630 & 53576.307& 0.767 &-3 & 828 &2.76 & 77 & 1921 &5.82 & 266 &1654 & 15.03 \\
1099: 0921331 & 53635.000& 0.814 &-19 & 936 &2.32:& 23 & 1952 &4.62 & 216 &1747 & 15.36 \\
1100: 0922990 & 53635.019& 0.815 &-2 & 955 &2.36 & 56 & 2026 &5.57 & 202 &1757 & 14.99 \\
1101: 0925220 & 53638.253& 0.983 &118 & 694 &2.75 &174 & 1452 &5.11 & 295 &1239 & 13.38 \\
1102: 0925481 & 53638.272& 0.984 &113 & 758 &2.54 &178 & 1427 &4.97 & 285 &1192 & 13.24 \\
1103: 0928220 & 53641.153& 0.133 &249 & 986 &2.46 & ?? & ?? &4.60 & 420 &1629 & 13.50 \\
1104: 0928161 & 53641.173& 0.134 & 260& 921 &2.40 & ?? & ?? & 4.82& 420 &1627 & 13.46 \\
1105: 1022700 & 53665.026& 0.372 & 85 & 657 &2.53 &16- & 1290 & 5.72 &159 &1088 & 15.75 \\
1106: 1022731 & 53665.045& 0.374 & 86 & 666 &2.53 &0.1 & 1288 & 5.59 &153 &1048 & 15.92 \\
1107: \bottomrule
1108: \end{tabular}
1109: \end{table*}
1110:
1111: \begin{table*}[!t]\centering
1112: \tablecols{3}
1113: \setlength\tabnotewidth{0.60\textwidth}
1114: \setlength{\tabcolsep}{0.9\tabcolsep}
1115: \small
1116: \caption{EMMI data}
1117: \begin{tabular}{lrr}
1118: \toprule
1119: spectrum & MJD & phase \\
1120: \midrule
1121: 0619 & 53541.389 & 0.955 \\
1122: 0630 & 53552.429 & 0.528 \\
1123: 0710 & 53561.400 & 0.994 \\
1124: 0921 & 53634.999 & 0.814 \\
1125: 0924 & 53638.238 & 0.982 \\
1126: 0927 & 53641.176 & 0.135 \\
1127: 1010 & 53654.110 & 0.806 \\
1128: 1012 & 53656.164 & 0.913 \\
1129: %\vspace*{0.7\baselineskip}
1130: \bottomrule
1131: \end{tabular}
1132: \end{table*}
1133:
1134:
1135: \begin{table*}[!t]\centering
1136: \tablecols{9}
1137: \setlength\tabnotewidth{0.60\textwidth}
1138: %\setlength{\tabcolsep}{0.9\tabcolsep}
1139: \small
1140: \caption{Photospheric absorptions 2005}
1141: \begin{tabular}{lrlrlrlrl}
1142: \toprule
1143: &$\phi=$&0.8&$\phi=$&0.0&$\phi=$&0.13&$\phi=$&0.37\\
1144: Atomic ID & RV & FW & RV & FW& RV & FW& RV & FW \\
1145: \midrule
1146: % FEROS 2005 observations, spectra of similar phases combined.
1147: % measured wrt laboratory wavelengths.
1148: % ID phi0.8 phi0.0 phi0.13 phi0.37
1149: % RV FWHM RV FWHM RV FWHM RV FWHM
1150: He I 3819.60 & 142&175 &190&286& --&-- &??&?? \\
1151: H 9 3835.38 & 167&260 &170&214& 186&224 &140&267 \\
1152: & --&-- &-- &-- &-120&468 & & \\
1153: He I 3888.64\tabnotemark{c} & 208&282 &198&207\tabnotemark{a}& 163&264&195&201 \\
1154: %He I 3926.54 & -38& 232 &-89&250 & -81&184& N.P.! \\
1155: He I 4026.18 & 182& 178 &153&119 &154&171&144&127 \\
1156: & -- --& &--&-- &-324&554&-- &-- \\
1157: H I 4101.76: & 194& 265 &176&213& 157&239& 156&221 \\
1158: & -- &-- &-- &---&-314\tabnotemark{b}&579& --& -- \\
1159: He II 4199.84\tabnotemark{d}&174& 278 &176&164 &171&176 &166&186 \\
1160: &-- & -- &-- &-- &-249&547 &-- &-- \\
1161: H I 4340.49 & 163& 220 &135&232& 141&247 &141&185 \\
1162: & -- & -- & & &-351&406 & -- & -- \\
1163: %He II 4341.6& --& -- &-- &-- &-390&422:&-- &-- \\
1164: He I 4387.93 & 149& 276 &175&137& 188&326 &170&189 \\
1165: He I 4471.50:& 168& 185 &167&152& 161&158 &155&132 \\
1166: He II 4541.59&190& 181 &192&126& 165&137 &168&183 \\
1167: & & & & & -249 &545 & & \\
1168: H I 4861.30 &176& 287 &113:&311&141&257 &188&326 \\
1169: He I 4921.93 &165& 207 &173&163 &156&166 &155&178 \\
1170: He I 5015.68 &181& 144 &179&148 &173&160 &167&123\tabnotemark{b} \\
1171: He II 5411.52& 181& 178 &197&128& 177&141 &189&125 \\
1172: & -- & -- &-142&145&-323&400 & -- & -- \\
1173: O III 5592.4 & 166& 150& 156&149& 168&134 &148&117 \\
1174: C IV 5801.3 & 178& 177& 203&180& 165&141 &182&170 \\
1175: C IV 5812.0 & 177& 132& 200&117& 199&136 & -- &-- \\
1176: He I 5875.8 & 173& 220& 164&190& 162&180 &162&160 \\
1177: He I 6678.15 & 436& 181& 187&113& 122&250 &-- &-- \\
1178: & --& -- &-145&324& 419&151 &-- &-- \\
1179: He I 7065.71 & 150& 250& 154&199& 147&234 &141&129 \\
1180: & --& -- & -- &--&- 549&569 & --& -- \\
1181: N IV 7111.28 & 152 & 93& 170&105& 129& 73 & 136& 74 \\
1182: & 8& 156& -- &-- & -- &-- &-- & -- \\
1183: N IV 7122.98 & 277& 110& --&-- &-- & --& --&-- \\
1184: & 16& 81& &-- &--& --& --&-- \\
1185: $<>$ & 172& -- & 173 & --& 161 & -- & 161 & -- \\
1186: s.d. &$\pm$16& -- &$\pm$22& -- &$\pm$19& -- &$\pm$18& -- \\
1187: %&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&&
1188: %DACs
1189: %He I 5875.8 &-1238&360&-1489&296& --&--& -1621&267 \\
1190: % 4471.50 &-1210&422& -- &-- & --&--& -1592&336 \\
1191: % 3888.64 &-1246&260&-1480&195& --&--& -1604&226 \\
1192: \bottomrule
1193: \tabnotetext{}{ \quad
1194: \textsuperscript{a} filled-in by emission; \quad \textsuperscript{b} very
1195: clear \quad \textsuperscript{c} blended with H I 3889.05 \quad
1196: \textsuperscript{c} blended with N III 4200. \quad
1197: }
1198: \end{tabular}
1199: \vspace*{0.7\baselineskip}
1200: \end{table*}
1201: %\RescaleFloatLengths{0.5}
1202:
1203:
1204: \begin{table*}[!t]\centering
1205: \tablecols{6}
1206: \setlength\tabnotewidth{0.60\textwidth}
1207: \setlength{\tabcolsep}{0.9\tabcolsep}
1208: \small
1209: \caption{Average differential photometry: Danish telescope}
1210: \begin{tabular}{lrrrrr}
1211: \toprule
1212: $<$MJD$>$ & $<$phase$>$ & Num. & $<$m$_v$-C$>$ & $<\sigma_{instr}>$ & s.d. \\
1213: \midrule
1214: % fotometry from Danish telescope, processed by Marchenko.
1215: % January 25, 2007: GK converted HJD to MJD's by subtracting 0.5
1216: % Mean MJD |Phi | Num.Dat | Mean (mv-C) +12.846 | Mean per point | Sigma
1217: % June 12, 2007: deleted mv because mags. are much too faint compared w/Massey
1218: % Marchenko said to use +12.865, but for phot*_all.eps I used +12.57. NEED
1219: % TO CHECK CALIBRATION TO m_V with Phil and Sergey.
1220: 53561.422 &0.995 & 20 & -1.244 & 0.008 & 0.014\\
1221: 53562.352 &0.043 & 18 & -1.397 & 0.007 & 0.008\\
1222: 53564.352 &0.147 & 12 & -1.501 & 0.008 & 0.006\\
1223: 53568.305 &0.352 & 7 & -1.234 & 0.006 & 0.003\\
1224: 53576.367 &0.771 & 43 & -1.423 & 0.008 & 0.006\\
1225: 53684.203 &0.368 &172 & -1.186 & 0.004 & 0.013\\
1226: 53685.105 &0.415 & 30 & -1.375 & 0.003 & 0.004\\
1227: 53687.207 &0.524 & 41 & -1.440 & 0.005 & 0.007\\
1228: 53689.191 &0.627 & 95 & -1.442 & 0.006 & 0.006\\
1229: 53693.184 &0.834 & 20 & -1.452 & 0.006 & 0.007\\
1230: \bottomrule
1231: \end{tabular}
1232: \end{table*}
1233:
1234: ------------------------
1235: %Original version of above table
1236: %\begin{table*}[!t]\centering
1237: %\tablecols{7}
1238: %\setlength\tabnotewidth{0.60\textwidth}
1239: %\setlength{\tabcolsep}{0.9\tabcolsep}
1240: %\small
1241: %\caption{Average photometric visual magnitudes}
1242: %\begin{tabular}{lrrrrrr}
1243: %\toprule
1244: % $<MJD>$ & $<phase>$ & Num. & $< \Delta m_v>$ & $<m_v>$& $< err>$ & sigma \\
1245: %\midrule
1246: % fotometry from Danish telescope, processed by Marchenko.
1247: % January 25, 2007: GK converted HJD to MJD's by subtracting 0.5
1248: % Mean MJD |Phi | Num.Dat | Mean (mv-C)|+12.846 | Mean per point | Sigma
1249: % 53561.422 &0.995 & 20 & -1.244 & 11.60 & 0.0082 & 0.0139\\
1250: % 53562.352 &0.043 & 18 & -1.397 & 11.45 & 0.0066 & 0.0078\\
1251: % 53564.352 &0.147 & 12 & -1.501 & 11.34 & 0.0077 & 0.0062\\
1252: % 53568.305 &0.352 & 7 & -1.234 & 11.61 & 0.0060 & 0.0026\\
1253: % 53576.367 &0.771 & 43 & -1.423 & 11.42 & 0.0081 & 0.0063\\
1254: % 53684.203 &0.368 &172 & -1.186 & 11.66 & 0.0043 & 0.0132\\
1255: % 53685.105 &0.415 & 30 & -1.375 & 11.47 & 0.0033 & 0.0040\\
1256: % 53687.207 &0.524 & 41 & -1.440 & 11.41 & 0.0047 & 0.0070\\
1257: % 53689.191 &0.627 & 95 & -1.442 & 11.40 & 0.0062 & 0.0058\\
1258: % 53693.184 &0.834 & 20 & -1.452 & 11.39 & 0.0056 & 0.0065\\
1259: % \bottomrule
1260: % \end{tabular}
1261: %\end{table*}
1262:
1263: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%Figures%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1264:
1265: \onecolumn
1266:
1267: \begin{figure}[!t]
1268: \includegraphics[width=\columnwidth]{fig1.eps} %{emmi.eps}
1269: \caption{Average $\phi\sim$0 low dispersion spectrum obtained with EMMI in 2005
1270: and rectified.
1271: The strongest line in the spectrum is He II 4686 \AA. Other lines that are
1272: mentioned in the text are labeled with their primary identifications. The same
1273: spectrum is amplified in the lower panel, to display more clearly the weaker lines.
1274: }
1275: \end{figure}
1276:
1277: \begin{figure}[!t]
1278: \includegraphics[width=\columnwidth]{fig2.eps} %{montage_profiles_hI.eps}
1279: \caption{H I (+He II) line profiles at H$\alpha$, H$\beta$, H$\gamma$, H$\delta$, H$\epsilon$
1280: and H9 3835 \AA\ from FEROS 2005 data. The spectra are averages for each of the orbital phases
1281: $\phi=$0.0, 0.13, 0.37 and 0.80, and the spectra at $\phi=$0.0 and 0.37 are renormalized to
1282: account for the corresponding eclipses. The velocity scale is corrected for the +150 km/s systemic
1283: velocity of the SMC. Sharp emission spikes, strongest at H$\alpha$, are nebular emission.
1284: }
1285: \end{figure}
1286:
1287: \begin{figure}[!t]
1288: \includegraphics[width=\columnwidth]{fig3.eps} %{FWHM.eps}
1289: \caption{Full width at half maximum intensity measured with Gaussian fits to the
1290: line profiles of NIV 4057 (open squares), He II 5411 (filled-in squares), C IV 5806
1291: (open triangles), H I 4201 (filled-in triangles) in the 1999 spectra. All lines except N IV display
1292: the sudden decrease in width when {\it star A} is ``in front" ($\phi\sim$0). A more
1293: gradual decrease is observed at the opposite eclipse ($\phi\sim$0.36). This is the same behavior
1294: seen in the late 1970's. The dotted line connects the data points. Data at
1295: eclipse phases correspond to the renormalized spectra.
1296: }
1297: \end{figure}
1298:
1299: \begin{figure}[!t]
1300: \includegraphics[width=\columnwidth]{fig4.eps} %{fig_orbit_5980.eps}
1301: \caption{Schematic representation of the orbit indicating orbital phases in
1302: a frame of references in which {\it star B} is at the origin. For this representation,
1303: the masses M$_A=$70 M$_\odot$ and M$_B=$54 M$_\odot$ (see Section 4), and
1304: the eccentricity and argument of periastron are, respectively, e=0.3 and
1305: $\omega=$133$^\circ$; and R$_A=$21 R$_\odot$, R$_A=$15 R$_\odot$ from BP91. The observer
1306: is located off the page at the bottom of the figure.
1307: }
1308: \end{figure}
1309:
1310: \begin{figure}[!t]
1311: \includegraphics[width=\columnwidth]{fig5.eps} %{montage_profiles_nv.eps}
1312: \caption{Line profiles at $\phi=$0.0 (top left) used as a template
1313: to compare with profiles at $\phi=$0.13 (top right), 0.37 (bottom left),
1314: and 0.80 (bottom right). The template is shifted by +275 km/s (top panel, right)
1315: and -150 km/s (bottom panel, right). Spectra at $\phi=$0.0 and 0.37 are renormalized
1316: to account for the eclipses. The arrows point to excess emission that appears to be
1317: associated with {\it star B}. The velocity scale is centered on the
1318: He II $\lambda_0=$4541.59 line, corrected for SMC motion ($+$150 km/s).
1319: }
1320: \end{figure}
1321:
1322: \begin{figure}[!t]
1323: \includegraphics[width=\columnwidth]{fig6.eps} %{montage_profiles_oiii.eps}
1324: \caption{Same as previous figure, but for 3 of the He II lines
1325: in the 6000-6200 \AA\ region. V=0 km/s corresponds to $\lambda_0=$6100. \AA.
1326: }
1327: \end{figure}
1328:
1329: \begin{figure}[!t]
1330: \includegraphics[width=\columnwidth]{fig7.eps} %{montage_profiles_niv.eps}
1331: \caption{Same as previous figure, but for N IV 4057 and H I 4100 \AA.
1332: Velocity scale is centered on $\lambda$ 4101.73 \AA.
1333: }
1334: \end{figure}
1335:
1336: \begin{figure}[!t]
1337: \includegraphics[width=\columnwidth]{fig8.eps} %{montage_profiles_hgamma.eps}
1338: \caption{Same as previous figure, but for H$\gamma$.Velocity scale is centered
1339: on $\lambda$ 4340.45 \AA.
1340: }
1341: \end{figure}
1342:
1343: \begin{figure}[!t]
1344: \includegraphics[width=\columnwidth]{fig9.eps} %{montage_profiles_heII5411.eps}
1345: \caption{Same as previous figure, but for He II 5411 \AA.
1346: Velocity scale is centered on $\lambda$ 5411.52 \AA.
1347: }
1348: \end{figure}
1349:
1350: \begin{figure}[!t]
1351: \includegraphics[width=\columnwidth]{fig10.eps} %{montage_profiles_5875.eps}
1352: \caption{Same as previous figure, but for He I 5875.80 \AA.
1353: }
1354: \end{figure}
1355:
1356: \begin{figure}[!t]
1357: %made in gkiraf/FEROS.data/All_norm/For_imstack_files; see All_norm/README, September 18, 2007.
1358: \includegraphics[width=\columnwidth]{fig11.eps} %{imstack_4058_axes_new.eps}
1359: \caption{Grey scale representation of residuals for N IV 4057. Phase increases from
1360: the bottom to the top, the same data being plotted twice. White corresponds to positive residuals,
1361: black to negative. Residuals are the difference between the nearly equally-spaced (in phase) 1999
1362: FEROS spectra and the average of these same spectra. Gaps in phase coverage are filled in with
1363: grey background color.}
1364: \end{figure}
1365: \clearpage
1366:
1367: \begin{figure}[!t]
1368: \includegraphics[width=\columnwidth]{fig12.eps} %{imstack_5411_axes_new.eps}
1369: \caption{Grey scale representation of 1999 FEROS data for He II 5411 as in
1370: previous figure}
1371: \end{figure}
1372: \clearpage
1373:
1374: \begin{figure}[!t]
1375: \includegraphics[width=\columnwidth]{fig13.eps} %{imstack_4860_axes_new.eps}
1376: %%made in gkiraf/FEROS.data/All_norm/
1377: \caption{Grey scale representation of 1999 FEROS data for H$\beta$+He II as in previous figure.}
1378: \end{figure}
1379: \clearpage
1380:
1381: \begin{figure}[!t]
1382: \includegraphics[width=\columnwidth]{fig14.eps} %{imstack_6678_axes_new.eps}
1383: %%made in gkiraf/FEROS.data/All_norm/
1384: \caption{Grey scale representation of 1999 FEROS data for HeI 6678 as in previous figure.}
1385: \end{figure}
1386: \clearpage
1387:
1388: \begin{figure}[!t]
1389: \includegraphics[width=\columnwidth]{fig15.eps} %{montage_profiles_heI.eps}
1390: \caption{FEROS 2005 line profiles of He I 7065, 6678, 5875 and 4471 \AA\
1391: (from top down). The spectra are averages for each of the orbital phases
1392: $\phi=$0.0 (top left), 0.13 (top right), 0.37 (bottom left) and 0.80 (bottom right).
1393: At $\phi=$0.0 and 0.37 they are renormalized to account for the eclipses. The
1394: velocity scale is corrected for the +150 km/s systemic velocity of the SMC. Note
1395: that the photospheric absorption lines are stationary. The P Cygni absorptions
1396: are particularly prominent at $\phi=$0.80.
1397: }
1398: \end{figure}
1399: \clearpage
1400:
1401: \begin{figure}[!t]
1402: \includegraphics[width=\columnwidth]{fig16.eps} %{montage_DAC_7065.eps}
1403: \caption{Montage of the 1999 individual FEROS spectra obtained over a single orbital cycle
1404: illustrating the behavior of the He I P Cygni absorption components. The template
1405: (dotted spectrum) is the profile at $\phi=$0.99, renormalized to account for the eclipse.
1406: The spectrum at $\phi=$0.35 is also renormalized. The absorption is strong during
1407: the phase interval $\phi=$0.73-0.30, but nearly vanishes at $\phi=$0.99 (eclipse of star B).
1408: The y-axis is relative intensity, each spectrum having been shifted along this axis by 0.3 units.
1409: The x-axis is velocity with respect to the laboratory wavelength, corrected for SMC systemic
1410: motion (+150 km/s).
1411: }
1412: \end{figure}
1413: \clearpage
1414:
1415: \begin{figure}[!t]
1416: \includegraphics[width=\columnwidth]{fig17.eps} %{montage_DAC_5875.eps}
1417: \caption{Same as previous figure, but for He I 5875 \AA\ (left). The same behavior is
1418: observed for the P Cygni absorptions. Note that when they are sharpest and strongest,
1419: the emission component has a very rounded shape. When the P Cygni absorption is not
1420: as prominent, the emission appears to have a sharper (possibly superposed)
1421: additional emission component. The right panel illustrates the same montage of spectra
1422: at H$\beta$. The y-axis is relative intensity, each spectrum having been shifted
1423: along this axis by 0.3 units. The x-axis is velocity with respect to the laboratory
1424: wavelength, corrected for SMC systemic motion (+150 km/s).
1425: }
1426: \end{figure}
1427: \clearpage
1428:
1429: \begin{figure}[!t]
1430: \includegraphics[width=\columnwidth]{fig18.eps} %{RV_2stars.eps}
1431: \caption{Radial velocity measurements. Top: A single Gaussian was fit to lines of
1432: N IV 4057 \AA\ (filled pentagons), H I 4101 \AA\ (squares) and He II 5411 \AA\ (triangles).
1433: Each RV curve has a different shape due to contributions from different sources.
1434: Middle: {\it Star A}'s contribution isolated using a 2-function fit to N IV 4057 \AA\
1435: (filled pentagons) and N V 4603 \AA\ (stars), and a single Gaussian to He II 6074
1436: and 6118 \AA\ (small squares). Arbitrary vertical shifts were applied to correct for
1437: different zero-point locations and align the 4 RV curves. Bottom: {\it Star B}'s deduced
1438: contribution derived from the 2-function fit to N IV 4057 \AA\ (filled pentagons)
1439: and N IV 4603 \AA\ (stars). The dashed curves are theoretical computations for a
1440: 70$+$54 M$_\odot$ binary system in an e$=$0.3, $\omega=$133$^\circ$ orbit. All RVs are
1441: corrected for the SMC systemic velocity (+150 km/s).}
1442: \end{figure}
1443:
1444: \begin{figure}[!t]
1445: \includegraphics[width=\columnwidth]{fig19.eps} %{EW.eps}
1446: \caption{The equivalent width of 1999 FEROS data as a function of orbital phase for
1447: N IV 4057 \AA\, H$\delta+$ He II 4101 \AA\ and He II 5411 \AA. }
1448: \end{figure}
1449:
1450: \begin{figure}[!t]
1451: \includegraphics[width=\columnwidth]{fig20.eps} %{EW_HeII5411_HI4101.eps}
1452: \caption{The ratio of equivalent widths of He II 5411 \AA\ and H I (+He II) 4101 \AA\
1453: as a function of orbital phase for the 1999 and 2005 FEROS observations. There is
1454: only a hint of an orbital-phase dependence in 1999. Much clearer is the overall
1455: change in the ratio from 1999 to 2005, indicating that the degree of ionization increased
1456: in 2005.}
1457: \end{figure}
1458:
1459: \begin{figure}[!t]
1460: %input_montage_profiles_phot_heI_1999
1461: \includegraphics[width=\columnwidth]{fig21.eps} %{montage_profiles_phot_heI_1999.eps}
1462: \caption{Comparison of the He I 4471 \AA\ photospheric absorption line in
1463: the 1999 (dotted) and in 2005 at four orbital phases. The velocity
1464: scale is corrected for an assumed SMC systemic velocity of $+$150 km/s, and the
1465: 1999 data were shifted upward for clarity in the figure. Note
1466: that the 2005 profiles are blue-shifted and stronger than the 1999 profiles.
1467: The 1999 spectra were obtained over the timescale MJD 51381.4-51391.4.
1468: }
1469: \end{figure}
1470: \clearpage
1471:
1472: \begin{figure}[!t]
1473: \includegraphics[width=\columnwidth]{fig22.eps} %{photometry_all.eps}
1474: % data and input file in FEROS.data/Marchenko_photometry/
1475: \caption{Visual magnitudes from SMARTS (open squares) and Swope
1476: (plus signs) compared with differential photometric observations
1477: of Breysacher \& Perrier (1980; data of July-October 1979) $+$10.415 mag (crosses);
1478: and Danish data $+$12.57 mag (filled-in triangles). Orbital phases for the
1479: Danish2005, SMARTS and Swope data sets were computed using T$_0=$JD2443158.705 and
1480: P=19.2654 days from Sterken \& Breysacher (1997). The orbital phases of BP80 data are
1481: as listed in their Table 1, which BP80 computed with T$_{BP80}=$JD2443158.771 and
1482: P$_{BP80}=$19.266 days. The difference in phase computed with P and P$_{BP80}$
1483: is $<$0.002 for the BP80 data.
1484: %August 25, 1979=JD 2444111.4792; (44111.479-43158.77)/19.266=49.4503; /19.2654=49.4518
1485: }
1486: \end{figure}
1487: \clearpage
1488:
1489: \begin{figure}[!t]
1490: \includegraphics[width=\columnwidth]{fig23.eps} %{marchenko_dates_phi36.eps}
1491: \caption{Differential photometric observations from the Danish telescope at
1492: $\phi=$0.36--0.37, after adding +12.57 mag to convert to visual magnitudes
1493: consistent with the SMARTS values. Top: plotted as a function of time (in hours).
1494: Bottom: plotted as a function of orbital phase, using P=19.2654 d and
1495: T$_{modified}=$JD2443158.865. With this T$_{modified}$, the secondary minimum is centered on $\phi=$0.36.
1496: The error bars correspond to the uncertainty per individual data point, $\sigma_{instr}$.}
1497: \end{figure}
1498: \clearpage
1499:
1500:
1501: \begin{figure}[!t]
1502: \includegraphics[width=\columnwidth]{fig24.eps} %{wwc_orbit_1999_feros.eps}
1503: \caption{Cartoon of the system geometry for different
1504: orbital phases, illustrating how the lines-of-sight to the two stars
1505: cross different portions of the WWI region and the stellar winds. Dashed
1506: lines contain the columns of gas projected onto the continuum-emitting regions,
1507: as seen by an observer situated at the bottom of the page. Discontinuous circles
1508: around {\it star A} indicate the accelerating portion of the wind (1.5R$_A$) and
1509: the assumed extent of a line-emitting region ($\sim$4R$_A$).
1510: }
1511: \end{figure}
1512: \clearpage
1513:
1514: \begin{figure}[!t]
1515: \includegraphics[width=\columnwidth]{fig25.eps} %{wwc_velocities_M70.eps}
1516: \caption{Schematic representation of the WWC shock cone at orbital phase
1517: 0.36 listing tangential flow velocities (v$_t$) computed using the relations given by
1518: Cant\'o et al. (1996). The assumed mass-loss rates are in M$_\odot$/year and
1519: velocities are in km/s. The observer is off the page to the left at $\phi=$0.36. The dashed lines
1520: contain the column of material traversed by the line-of-sight from the observer
1521: to {\it star A}. With the assumed radii (R$_A=$15 R$_\odot$ and R$_B=$21 R$_\odot$),
1522: the light from {\it star A}'s disk must pass through the WWI region {\em and} {\it star B}'s
1523: inner wind before reaching the observer. Time-dependent mass ejections
1524: from {\it star B} or instabilities along the WWI region could be responsible for the
1525: rapid variability observed at this orbital phase.
1526: }
1527: \end{figure}
1528:
1529: \begin{figure}[!t]
1530: \includegraphics[width=\columnwidth]{fig26.eps} %{ foellmi.v4.hd5980_2.eps}
1531: \caption{Cartoon illustrating an alternative geometry for the WWI
1532: region that would be consistent with the orbital phase-dependent
1533: behavior of the He I P Cygni absorption components.
1534: }
1535: \end{figure}
1536: \clearpage
1537:
1538: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1539: \begin{table*}[!t]\centering
1540: \tablecols{5}
1541: \setlength\tabnotewidth{0.60\textwidth}
1542: \setlength{\tabcolsep}{0.9\tabcolsep}
1543: \small
1544: \caption{Average photometric visual magnitudes: SWOPE and SMARTS}
1545: \begin{tabular}{lrrrr}
1546: \toprule
1547: $<MJD>$ & $<phase>$ & Num. & $<m_v>$ & s.d. \\
1548: \midrule
1549: 52838.379 & 0.464 & 16 & 11.143 & 0.017 \\
1550: 52839.362 & 0.515 & 9 & 11.154 & 0.046 \\
1551: 52840.324 & 0.565 & 12 & 11.123 & 0.019 \\
1552: 52857.303 & 0.446 & 18 & 11.145 & 0.012 \\
1553: 52858.273 & 0.497 & 3 & 11.122 & 0.010 \\
1554: 52859.257 & 0.548 & 5 & 11.086 & 0.018 \\
1555: 52860.244 & 0.599 & 1 & 11.142 & \nodata \\
1556: 52972.072 & 0.404 & 15 & 11.185 & 0.006 \\
1557: 52973.061 & 0.455 & 8 & 11.110 & 0.008 \\
1558: 52974.059 & 0.507 & 16 & 11.076 & 0.012 \\
1559: 52975.056 & 0.559 & 4 & 11.090 & 0.005 \\
1560: 52991.107 & 0.392 & 18 & 11.236 & 0.018 \\
1561: 52992.063 & 0.441 & 19 & 11.114 & 0.011 \\
1562: 52992.844 & 0.482 & 2 & 11.073 & 0.001 \\
1563: 52993.905 & 0.537 & 2 & 11.038 & 0.008 \\
1564: 53008.062 & 0.272 & 7 & 11.101 & 0.013 \\
1565: 53009.067 & 0.324 & 1 & 11.170 & \nodata \\
1566: 53010.034 & 0.374 & 4 & 11.335 & 0.013 \\
1567: 53011.039 & 0.426 & 6 & 11.180 & 0.018 \\
1568: 53159.403 & 0.127 & 4 & 11.048 & 0.010 \\
1569: 53183.377 & 0.372 & 4 & 11.357 & 0.016 \\
1570: 53210.325 & 0.771 & 7 & 11.117 & 0.005 \\
1571: 53225.305 & 0.548 & 7 & 11.109 & 0.010 \\
1572: 53226.258 & 0.598 & 11 & 11.128 & 0.011 \\
1573: 53227.387 & 0.656 & 6 & 11.154 & 0.012 \\
1574: 53228.295 & 0.703 & 6 & 11.128 & 0.010 \\
1575: 53229.240 & 0.752 & 7 & 11.099 & 0.016 \\
1576: 53230.228 & 0.804 & 8 & 11.135 & 0.011 \\
1577: 53247.271 & 0.688 & 7 & 11.153 & 0.015 \\
1578: 53254.290 & 0.053 & 20 & 11.205 & 0.013 \\
1579: 53288.193 & 0.812 & 26 & 11.137 & 0.015 \\
1580: 53591.385 & 0.550 & 2 & 11.119 & 0.003 \\
1581: 53592.405 & 0.603 & 2 & 11.114 & 0.003 \\
1582: 53593.400 & 0.655 & 1 & 11.120 & \nodata \\
1583: 53596.401 & 0.810 & 5 & 11.122 & 0.007 \\
1584: 53597.378 & 0.861 & 1 & 11.128 & \nodata \\
1585: 53599.223 & 0.957 & 2 & 11.149 & 0.041 \\
1586: 53600.401 & 0.018 & 2 & 11.325 & 0.006 \\
1587: 53602.355 & 0.119 & 2 & 11.084 & 0.005 \\
1588: 53608.313 & 0.429 & 2 & 11.189 & 0.004 \\
1589: 53615.078 & 0.780 & 1 & 11.158 & \nodata \\
1590: 53617.281 & 0.894 & 2 & 11.134 & 0.006 \\
1591: 53618.199 & 0.942 & 3 & 11.138 & 0.016 \\
1592: 53619.232 & 0.995 & 1 & 11.347 & \nodata \\
1593: 53620.238 & 0.048 & 1 & 11.174 & \nodata \\
1594: \bottomrule
1595: \end{tabular}
1596: \end{table*}
1597:
1598: \begin{table*}[!t]\centering
1599: \tablecols{5}
1600: \setlength\tabnotewidth{0.60\textwidth}
1601: \setlength{\tabcolsep}{0.9\tabcolsep}
1602: \small
1603: \caption{Average photometric visual magnitudes: SWOPE and SMARTS}
1604: \begin{tabular}{lrrrr}
1605: \toprule
1606: $<MJD>$ & $<phase>$ & Num. & $<m_v>$ & s.d. \\
1607: \midrule
1608: 53621.379 & 0.107 & 2 & 11.091 & 0.006 \\
1609: 53622.366 & 0.158 & 2 & 11.069 & 0.006 \\
1610: 53624.244 & 0.256 & 2 & 11.104 & 0.008 \\
1611: 53625.326 & 0.312 & 2 & 11.191 & 0.011 \\
1612: 53628.267 & 0.464 & 2 & 11.142 & 0.013 \\
1613: 53631.317 & 0.623 & 2 & 11.146 & 0.004 \\
1614: 53632.323 & 0.675 & 2 & 11.145 & 0.006 \\
1615: 53634.303 & 0.778 & 2 & 11.126 & 0.012 \\
1616: 53638.265 & 0.983 & 2 & 11.274 & 0.008 \\
1617: 53640.255 & 0.087 & 2 & 11.137 & 0.004 \\
1618: 53642.245 & 0.190 & 2 & 11.107 & 0.011 \\
1619: 53643.233 & 0.241 & 2 & 11.141 & 0.014 \\
1620: 53646.201 & 0.395 & 2 & 11.288 & 0.010 \\
1621: 53650.388 & 0.613 & 2 & 11.181 & 0.008 \\
1622: 53651.352 & 0.663 & 1 & 11.154 & \nodata \\
1623: 53652.397 & 0.717 & 2 & 11.185 & 0.003 \\
1624: 53655.185 & 0.862 & 2 & 11.101 & 0.026 \\
1625: 53657.186 & 0.966 & 2 & 11.175 & 0.018 \\
1626: 53658.132 & 0.015 & 4 & 11.318 & 0.016 \\
1627: 53660.178 & 0.121 & 4 & 11.070 & 0.003 \\
1628: 53663.209 & 0.278 & 2 & 11.121 & 0.001 \\
1629: 53664.219 & 0.331 & 2 & 11.205 & 0.004 \\
1630: 53665.227 & 0.383 & 2 & 11.329 & 0.004 \\
1631: 53666.232 & 0.435 & 2 & 11.144 & 0.003 \\
1632: 53669.062 & 0.582 & 4 & 11.104 & 0.006 \\
1633: 53670.051 & 0.633 & 2 & 11.121 & 0.007 \\
1634: 53672.047 & 0.737 & 1 & 11.099 & \nodata \\
1635: 53673.020 & 0.787 & 3 & 11.095 & 0.021 \\
1636: 53675.047 & 0.893 & 2 & 11.139 & 0.002 \\
1637: 53676.180 & 0.951 & 4 & 11.158 & 0.006 \\
1638: 53677.196 & 0.004 & 2 & 11.349 & 0.010 \\
1639: 53679.157 & 0.106 & 2 & 11.150 & 0.007 \\
1640: 53680.269 & 0.164 & 2 & 11.100 & 0.002 \\
1641: 53682.075 & 0.257 & 4 & 11.136 & 0.022 \\
1642: 53683.264 & 0.319 & 2 & 11.195 & 0.011 \\
1643: 53685.105 & 0.415 & 2 & 11.186 & 0.001 \\
1644: 53688.131 & 0.572 & 2 & 11.145 & 0.001 \\
1645: 53689.107 & 0.622 & 2 & 11.133 & 0.000 \\
1646: 53691.197 & 0.731 & 1 & 11.225 & \nodata \\
1647: 53692.077 & 0.777 & 2 & 11.159 & 0.001 \\
1648: 53694.082 & 0.881 & 2 & 11.189 & 0.001 \\
1649: 53695.082 & 0.933 & 2 & 11.168 & 0.008 \\
1650: 53698.036 & 0.086 & 2 & 11.082 & 0.011 \\
1651: 53700.083 & 0.192 & 2 & 11.087 & 0.001 \\
1652: 53701.047 & 0.242 & 2 & 11.120 & 0.002 \\
1653: 53703.075 & 0.347 & 2 & 11.326 & 0.002 \\
1654: \bottomrule
1655: \end{tabular}
1656: \end{table*}
1657:
1658: \begin{table*}[!t]\centering
1659: \tablecols{5}
1660: \setlength\tabnotewidth{0.60\textwidth}
1661: \setlength{\tabcolsep}{0.9\tabcolsep}
1662: \small
1663: \caption{Average photometric visual magnitudes: SWOPE and SMARTS}
1664: \begin{tabular}{lrrrr}
1665: \toprule
1666: $<MJD>$ & $<phase>$ & Num. & $<m_v>$ & s.d. \\
1667: \midrule
1668: 53704.072 & 0.399 & 2 & 11.255 & 0.004 \\
1669: 53706.066 & 0.503 & 2 & 11.109 & 0.017 \\
1670: 53707.092 & 0.556 & 2 & 11.120 & 0.004 \\
1671: 53710.030 & 0.709 & 3 & 11.109 & 0.013 \\
1672: 53711.043 & 0.761 & 2 & 11.102 & 0.021 \\
1673: 53712.096 & 0.816 & 2 & 11.086 & 0.006 \\
1674: 53714.134 & 0.922 & 2 & 11.121 & 0.001 \\
1675: 53715.171 & 0.975 & 2 & 11.182 & 0.020 \\
1676: 53716.027 & 0.020 & 2 & 11.309 & 0.004 \\
1677: 53719.040 & 0.176 & 2 & 11.110 & 0.001 \\
1678: 53720.049 & 0.229 & 2 & 11.171 & 0.001 \\
1679: 53722.022 & 0.331 & 2 & 11.259 & 0.011 \\
1680: 53724.045 & 0.436 & 2 & 11.160 & 0.011 \\
1681: 53725.024 & 0.487 & 2 & 11.115 & 0.023 \\
1682: 53727.029 & 0.591 & 2 & 11.149 & 0.000 \\
1683: 53752.092 & 0.892 & 2 & 11.142 & 0.005 \\
1684: 53756.059 & 0.098 & 2 & 11.105 & 0.001 \\
1685: 53757.027 & 0.148 & 2 & 11.108 & 0.008 \\
1686: 53758.025 & 0.200 & 4 & 11.135 & 0.019 \\
1687: 53759.027 & 0.252 & 3 & 11.138 & 0.035 \\
1688: 53958.306 & 0.596 & 1 & 11.115 & \nodata \\
1689: 53960.339 & 0.701 & 2 & 11.120 & 0.004 \\
1690: 53965.295 & 0.958 & 2 & 11.188 & 0.002 \\
1691: 53966.310 & 0.011 & 2 & 11.400 & 0.001 \\
1692: 53968.322 & 0.116 & 2 & 11.143 & 0.001 \\
1693: 53970.319 & 0.219 & 2 & 11.151 & 0.003 \\
1694: 53976.356 & 0.533 & 2 & 11.169 & 0.009 \\
1695: 53983.365 & 0.896 & 9 & 11.183 & 0.011 \\
1696: 53987.332 & 0.102 & 2 & 11.155 & 0.001 \\
1697: 53989.342 & 0.207 & 2 & 11.168 & 0.006 \\
1698: 53991.336 & 0.310 & 3 & 11.168 & 0.023 \\
1699: 53993.339 & 0.414 & 3 & 11.244 & 0.008 \\
1700: 53995.325 & 0.517 & 2 & 11.186 & 0.004 \\
1701: 53998.274 & 0.670 & 2 & 11.126 & 0.008 \\
1702: 53999.263 & 0.722 & 2 & 11.177 & 0.021 \\
1703: 54004.245 & 0.980 & 2 & 11.239 & 0.002 \\
1704: 54007.391 & 0.143 & 2 & 11.091 & 0.008 \\
1705: 54009.379 & 0.247 & 2 & 11.116 & 0.014 \\
1706: 54011.350 & 0.349 & 2 & 11.316 & 0.003 \\
1707: 54013.302 & 0.450 & 2 & 11.177 & 0.039 \\
1708: 54017.288 & 0.657 & 2 & 11.120 & 0.007 \\
1709: 54051.014 & 0.408 & 2 & 11.217 & 0.025 \\
1710: 54052.012 & 0.460 & 2 & 11.155 & 0.011 \\
1711: 54053.041 & 0.513 & 2 & 11.140 & 0.012 \\
1712: 54054.033 & 0.564 & 2 & 11.106 & 0.006 \\
1713: 54056.247 & 0.679 & 1 & 11.124 & \nodata \\
1714: \bottomrule
1715: \end{tabular}
1716: \end{table*}
1717:
1718: \begin{table*}[!t]\centering
1719: \tablecols{5}
1720: \setlength\tabnotewidth{0.60\textwidth}
1721: \setlength{\tabcolsep}{0.9\tabcolsep}
1722: \small
1723: \caption{Average photometric visual magnitudes: SWOPE and SMARTS}
1724: \begin{tabular}{lrrrr}
1725: \toprule
1726: $<MJD>$ & $<phase>$ & Num. & $<m_v>$ & s.d. \\
1727: \midrule
1728: 54057.223 & 0.730 & 2 & 11.083 & 0.000 \\
1729: 54059.177 & 0.832 & 2 & 11.084 & 0.001 \\
1730: 54061.123 & 0.933 & 2 & 11.122 & 0.008 \\
1731: 54062.195 & 0.988 & 1 & 11.262 & \nodata \\
1732: 54066.001 & 0.186 & 4 & 11.085 & 0.001 \\
1733: 54068.153 & 0.297 & 2 & 11.130 & 0.007 \\
1734: 54070.125 & 0.400 & 4 & 11.224 & 0.015 \\
1735: 54074.173 & 0.610 & 2 & 11.124 & 0.001 \\
1736: 54080.161 & 0.921 & 2 & 11.148 & 0.001 \\
1737: 54082.233 & 0.028 & 2 & 11.275 & 0.004 \\
1738: 54084.189 & 0.130 & 2 & 11.120 & 0.001 \\
1739: 54086.101 & 0.229 & 2 & 11.151 & 0.019 \\
1740: 54088.193 & 0.338 & 2 & 11.274 & 0.004 \\
1741: 54089.195 & 0.390 & 2 & 11.315 & 0.002 \\
1742: 54090.200 & 0.442 & 2 & 11.156 & 0.001 \\
1743: 54092.199 & 0.546 & 1 & 11.150 & \nodata \\
1744: \bottomrule
1745: \end{tabular}
1746: \end{table*}
1747:
1748: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
1749:
1750: \end{document}
1751:
1752: