astro-ph0005033/ms.tex
1: \documentstyle[epsfig]{mn}     % LaTeX A&A  Standard Fonts
2: 
3: \def\simless{\mathbin{\lower 3pt\hbox
4:      {$\rlap{\raise 5pt\hbox{$\char'074$}}\mathchar"7218$}}}   %< or of order
5: \def\simmore{\mathbin{\lower 3pt\hbox
6:      {$\rlap{\raise 5pt\hbox{$\char'076$}}\mathchar"7218$}}}   %> or of order
7: \def\Msun{{\rm M}_\odot}                                       % solar masses
8: \def\msun{{\rm M}_\odot}                                       % solar masses
9: \newcommand{\lsi}    {LS~I +61$^{\circ}$ 235}
10: 
11: \begin{document}
12: 
13: 
14: %
15:    \title[Global oscillations in LS I +61$^{\circ}$ 235/RX J0146.9+6121]
16:    {Correlated V/R and IR photometric variations in the Be/X-ray binary LS 
17: I +61$^{\circ}$ 235/RX J0146.9+6121}
18: 
19: 
20:    \author[Reig et al.]
21: {P. Reig$^{1,2}$, 
22: I. Negueruela$^{3}$, M.J. Coe$^{4}$, J. Fabregat$^{5}$,
23: A.E. Tarasov$^{6}$ and R. K. Zamanov$^{7,8}$\\
24: $^{1}$ Foundation for Research and Technology-Hellas. 711 10 
25: Heraklion. Crete. Greece.\\
26: $^{2}$ Physics Department. University of Crete. 710 03 Heraklion. Crete. 
27: Greece.\\
28: $^{3}$ SAX SDC, Agenzia Spaziale Italiana, c/o Telespazio, via Corcolle 19, 00131 
29: Roma, Italy.\\
30: $^{4}$ Physics \& Astronomy Department, Southampton University, 
31: Southampton, SO17 1BJ, UK.\\
32: $^{5}$ Departament d'Astronomia i Astrof\'{\i}sica, Universitat de 
33: Val\'encia, E-46100 Burjassot-Val\'encia, Spain.\\
34: $^{6}$ Crimean Astrophysical Observatory, 334413 Nauchny, Crimea, 
35: Ukraine.\\
36: $^{7}$ National Astronomical Observatory Rozhen, P.O. Box 136, BG-4700 Smoljan, 
37: Bulgaria.\\
38: $^{8}$ Departamento de  F\'{\i}sica, Universidad de Ja\'en,
39: C/Virgen de la Cabeza 2, 23071 Ja\'en, Spain.
40: }
41: 
42: 
43:    \date{Received ????????; accepted ??????????}
44: 
45:    \maketitle
46: 
47:    \begin{abstract}
48:    
49: We report on the long-term variability of the Be/X-ray binary LS I
50: +61$^{\circ}$ 235/RX  J0146.9+6121. New  optical   spectroscopic  and 
51: infrared  photometric observations  confirm the presence of global 
52: one-arm   oscillations in the circumstellar disc of the Be star and allow
53: us to derive a V/R quasi-period of 1240$\pm$30 days. Pronounced shell events,
54: reminiscent of the spectacular variations in Be stars, are also seen. We
55: have found that the $J$, $H$ and $K$ infrared photometric bands vary  in 
56: correlation   with  the spectroscopic V/R variations, implying that the
57: one-armed disc oscillations are prograde. The effect of the  oscillations
58: is not only seen in the H$\alpha$  line but also in the He I
59: $\lambda6678$  and  Paschen  lines.  Since  these lines are formed at
60: different  radii in the equatorial  disc of the Be star, such effect
61: confirms the global nature of the  perturbation. The Keplerian disc has
62: been found to be denser than the average of a sample of isolated Be stars,
63: which may be indicative of some kind of interaction with the compact
64: companion. Finally, from a {\em Rossi X-ray Timing Explorer}  observation
65: we derive a spin period of the neutron star of 1404.5$\pm$0.5 s. 
66: 
67: \end{abstract}
68: 
69: 
70:       \begin{keywords}
71:       		stars: individual: LS I +61$^{\circ}$ 235 --
72:                 binaries: general -- 
73: 		X-rays: stars --
74:                 stars: pulsars: individual: RX J0146.9+6121 --
75: 		stars: emission-line, Be
76:       \end{keywords}
77: 
78: 
79: \section{Introduction}
80: 
81: 
82: RX J0146.9+6121 is one of the slowest high mass X-ray pulsar systems
83: (Mereghetti et al. 1993; Hellier 1994). Its optical counterpart is the
84: $V=11.2$, B1III-V star  LS I +61$^{\circ}$ 235 located at an estimated
85: distance of 2.3$\pm$0.5 kpc (Motch et al. 1997; Coe et al. 1993; Reig  et
86: al. 1997a, hereafter R97). R97 derived the astrophysical  parameters of the
87: optical counterpart and  reported  V/R variations  with a quasi-period of
88: $\sim 3$ years. The line profile variability was attributed to the 
89: prograde  precession  of a one-armed  mode  confined in the Be star's
90: disc.  
91: 
92: Be/X-ray  binaries  comprise  approximately  70\% of the more  general
93: class of high mass X-ray  binaries,  the other  $\sim$~30\%  containing
94: evolved  (luminosity  class I and II) primaries.  In a Be/X-ray binary the
95: optical companion, a Be star, is characterised by an emission line
96: spectrum and an infrared  excess when compared to normal  B-type stars of
97: the same spectral  type. These two observational properties have their
98: origin in the cool gaseous quasi-Keplerian disc which lies on the equatorial
99: plane of the central star. A neutron star revolves  around the 10-20~$\msun$ 
100: primary in a rather  eccentric  orbit (e $\sim$  0.2--0.8) and
101: accretes material  expelled by the Be star from the disc -- in the form of a
102: low velocity, high density  equatorial wind -- giving rise to the X-rays. 
103: 
104: The physical properties of the Be star's disc in Be/X-ray binaries have
105: traditionally been considered to be the same as those of isolated Be
106: stars. Indeed, the long-term variability characterised as disc-loss phases
107: and V/R variations have been seen in both, isolated and Be/X-ray binaries
108: (Okazaki 1997; Negueruela et al. 1998).
109: However, it is not yet clear what role the compact companion in
110: Be/X-ray binaries may play in the onset and subsequent development of the
111: perturbation that gives rise to the asymmetric profiles or in the
112: formation and loss of the disc. There is growing evidence that the
113: circumstellar disc surrounding isolated Be stars and Be star in X-ray
114: binaries may not share, on average, the same physical properties (Reig et
115: al. 1997b; Negueruela et al. 1998; 1999)
116: 
117: In  this  paper  we  analyse  optical,  infrared  and  X-ray  data 
118: and search for correlations  between the  characteristics of the
119: radiation at  these  wavelengths. Our ultimate goal is to assess the
120: validity of the global one-armed oscillation model and whether the compact
121: companion has any effect on the V/R variability.
122: 
123: %----------------------------------------------------------------------------
124: \begin{table} 
125: \begin{center} 
126: \caption{Journal of spectroscopic observations and spectral parameters for
127: the H$\alpha$ line. Errors are $\simless$ 8\%}   
128: \label{sp} 
129: \begin{tabular}{lccccccl} 
130: \hline 
131: Date	&MJD	&Tele 		&Range  &EW	&$\Delta V$	&profile\\
132: 	&	&scope		&nm     &\AA	&km/s		&\\ 
133: 17-08-94& 49582 & BNAO$^a$	&650-662 &--8.4	 &252	&V$>$R  \\ 
134: 08-06-96& 50243	& BNAO$^a$	&650-662 &--9.6  &257	&V$<$R  \\ 
135: 09-06-96& 50244 & BNAO$^a$	&650-662 &--10.1 &258	&V$<$R  \\
136: 01-08-96& 50297	& INT$^a$	&610-700 &--10.1&316	&V$<$R  \\ 
137: 26-08-96& 50322	& BNAO$^a$	&650-662 &--10.2 &298	&V$<$R  \\ 
138: 13-11-96& 50401	& CAO   	&653-660 &--10.3	&259	&V$<$R  \\ 
139: 12-12-96& 50430 & CAO		&653-660 &--10.3	&251	&V$<$R  \\ 
140: 01-02-97& 50481	& INT$^a$	&580-740 &--11.9&288	&V$<$R  \\
141: 21-08-97& 50682	& INT$^a$	&610-700 &--12.1&242	&V$>$R  \\ 
142: 18-09-97& 50710	& CAO		&653-660 &--12.0	&223	&V$>$R  \\ 
143: 10-10-97& 50732 & BNAO$^a$	&650-662 &--10.7	&233	&V$>$R  \\ 
144: 28-10-97& 50750	& JKT		&638-673 &--11.8	&258	&V$>$R  \\ 
145: 10-02-98& 50855	& INT   	&640-680 &--11.2&235	&V$>$R  \\
146: 01-10-98& 51088	& BNAO$^b$	&650-670 &--11.0	&337	&V$>$R  \\ 
147: 04-10-98& 51091 & BNAO$^b$	&650-670 &--11.3 &333	&V$>$R  \\ 
148: 08-10-98& 51095 & WHT   	&649-661 &--11.5	&336	&V$>$R  \\ 
149: 22-07-99& 51382 & CAO   	&653-660 &--10.8 &298	&shell  \\
150: 25-07-99& 51385 & SKI   	&555-756 &--11.9 &312	&shell  \\ 
151: 26-07-99& 51386 & SKI   	&555-756 &--11.6 &325	&shell  \\ 
152: 19-08-99& 51410 & INT$^b$	&624-680 &--11.2&298	&shell  \\ 
153: 17-09-99& 51439 & BNAO$^b$  	&650-670 &--12.2 &286	&shell  \\
154: 19-09-99& 51441 & BNAO$^b$  	&650-670 &--11.7 &285	&shell  \\ 
155: \hline
156: \multicolumn{5}{l}{BNAO$^a$: ISTA 580$\times$400 + 632 l/mm} & \multicolumn{2}{l}{0.2 \AA/pix}  \\
157: \multicolumn{5}{l}{BNAO$^b$: Photometrics 1024$\times$1024 + 632 l/mm} & \multicolumn{2}{l}{0.2 \AA/pix} \\
158: \multicolumn{5}{l}{CAO: Electronix 1024$\times$260 CCD + 600 l/mm} &\multicolumn{2}{l}{0.06 \AA/pix} \\ 
159: \multicolumn{5}{l}{JKT: TEK4 + RBS + 2400 l/mm}  &\multicolumn{2}{l}{0.4 \AA/pix} \\ 
160: \multicolumn{5}{l}{INT$^a$: TEK3 + IDS + 235 mm + R1200 l/mm} & \multicolumn{2}{l}{0.8 \AA/pix} \\ 
161: \multicolumn{5}{l}{INT$^b$: EEV10 + IDS + 500 mm + R1200 l/mm} &\multicolumn{2}{l}{0.2 \AA/pix} \\ 
162: \multicolumn{5}{l}{WHT:	UES + 31 l/mm, R=54000} &\multicolumn{2}{l}{0.06 \AA/pix}\\ 
163: \multicolumn{5}{l}{SKI: ISA608 2000$\times$800 + 1300 l/mm} & \multicolumn{2}{l}{1 \AA/pix}
164: \end{tabular} 
165: \end{center} 
166: \end{table}
167: %----------------------------------------------------------------------------
168: %----------------------------------------------------------------------------
169: \begin{table} 
170: \begin{center} 
171: \caption{Journal of infrared observations}  
172: \label{irobs} 
173: \begin{tabular}{cccccccc} 
174: \hline 
175: Date	&MJD	&J	&H	&K \\ 
176: 13-01-96	&50096.42    &9.86$\pm$0.03   &9.58$\pm$0.02   &9.35$\pm$0.02\\ 
177: 19-03-96	&50162.36    &9.78$\pm$0.02   &9.60$\pm$0.01   &9.40$\pm$0.02\\ 
178: 28-07-96	&50293.68    &9.73$\pm$0.05   &9.51$\pm$0.06   &9.34$\pm$0.04\\
179: 29-07-96	&50294.65    &9.73$\pm$0.10   &9.51$\pm$0.09   &9.31$\pm$0.08\\ 
180: 18-07-97	&50648.65    &9.83$\pm$0.02   &9.56$\pm$0.02   &9.32$\pm$0.02\\ 
181: 19-07-97	&50649.66    &9.77$\pm$0.01   &9.52$\pm$0.01   &9.32$\pm$0.01\\ 
182: 21-07-97	&50651.66    &9.76$\pm$0.02   &9.53$\pm$0.02   &9.29$\pm$0.02\\
183: 17-06-98	&50982.72   &10.01$\pm$0.01   &9.81$\pm$0.02   &9.53$\pm$0.02\\ 
184: 27-10-98	&51114.42   &10.01$\pm$0.02   &9.69$\pm$0.02   &9.48$\pm$0.02\\ 
185: 28-10-98	&51115.67    &9.82$\pm$0.02   &9.66$\pm$0.02   &9.45$\pm$0.02\\ 
186: 28-10-98	&51115.67    &9.85$\pm$0.02   &9.65$\pm$0.02   &9.44$\pm$0.02\\
187: 26-07-99	&51386.72    &9.80$\pm$0.02   &9.58$\pm$0.02   &9.37$\pm$0.02\\ 
188: 27-07-99	&51387.66    &9.96$\pm$0.03   &9.67$\pm$0.02   &9.43$\pm$0.02\\ 
189: 29-07-99	&51389.63    &9.85$\pm$0.02   &9.61$\pm$0.02   &9.36$\pm$0.02\\ 
190: 30-07-99	&51390.66    &9.81$\pm$0.02   &9.58$\pm$0.02   &9.34$\pm$0.02\\
191: 01-08-99	&51392.64    &9.90$\pm$0.04   &9.63$\pm$0.03   &9.41$\pm$0.02\\ 
192: 02-10-99	&51454.49    &9.79$\pm$0.02   &9.54$\pm$0.02   &9.33$\pm$0.02\\ 
193: 03-10-99	&51455.60    &9.86$\pm$0.02   &9.58$\pm$0.02   &9.32$\pm$0.02\\ 
194: 04-10-99	&51456.56    &9.82$\pm$0.02   &9.57$\pm$0.02   &9.31$\pm$0.02\\ 
195: \hline
196: \end{tabular} 
197: \end{center} 
198: \end{table}
199: %----------------------------------------------------------------------------
200: %----------------------------------------------------------------------------
201:         \begin{figure*}
202:     \begin{center}
203:     \leavevmode
204: \epsfig{file=figures/serie.eps, width=14.0cm, bbllx=20pt, bblly=315pt,
205:   bburx=535pt, bbury=710pt, clip=}
206:  \end{center}              
207:         \caption{Evolution of the H$\alpha$ line profile over the 
208: period 1991 August 28 -- 1999 September 17 (MJD 48497--51439)}
209:         \label{vr}
210:         \end{figure*}
211: %----------------------------------------------------------------------------
212: 
213: \section{Observations and results}
214: 
215: 
216: \subsection{Optical data}
217: 
218: Optical  spectroscopic  observations were made from the 2.6m telescope at 
219: the  Crimean   Astronomical   Observatory   (CAO),  the 2m RCC telescope
220: of the Bulgarian National Astronomical Observatory "Rozhen" (BNAO), the
221: 1.3m telescope of the Skinakas Observatory (SKI), in Crete (Greece), the
222: 1.0m Jacobus Kapteyn  Telescope (JKT), the 2.5m Isaac Newton  Telescope 
223: (INT) in service mode and the 4.2m William Herschel  Telescope (WHT). The
224: last three telescope are located in the Observatorio del Roque de Los
225: Muchachos (La Palma, Spain). Table~1 shows the journal of the observations
226: and the instrument set-up. For  a description of the observations not
227: mentioned in the table see R97.
228: 
229: Fig \ref{vr} shows a selected sample of H$\alpha$ line profiles of LS I
230: +61$^{\circ}$ 235 covering the period 1991 August -- 1999 September. Part
231: of the new data (all  observations  taken after 1996  February  plus those
232: from Rozhen BNAO) are plotted together with some of the spectra presented
233: in R97. The evolution of the V/R ratio can be seen in Fig \ref{corr}$a$.
234: Different shapes of the line have been represented by different symbols.
235: Stars are used for $ V > R$ points, triangles for $V < R$ phases, dots for
236: single-peak lines and squares for shell profiles (i.e. when the central
237: absorption goes beyond the continuum). Note that the shell phase is very
238: brief and always occurs during the transition from $V > R$ to $V < R$. In
239: contrast, the transition from $V < R$ to $V > R$ is separated by
240: single-peak profiles. Strictly speaking there are not symmetric
241: single-peak lines in our observations (as observed in pole-on stars) since
242: one can always see flank inflections revealing a second peak.
243: Nevertheless, we will refer to as single-peak phase the profiles seen
244: during the transition  $V < R$ to $V > R$. By fitting a sine  function to
245: the V/R curve we refined the V/R quasi-period to a value of 1240$\pm$30
246: days.  
247: 
248: We have also measured the separation of the blue and red peaks by fitting
249: two Gaussian functions to the line profile (Fig~\ref{corr}$d$). The peak
250: separation gives a measure of the velocity field, assuming it to be
251: Keplerian. The blue-dominated profiles seem to sample a wider range of
252: velocities than red-dominated profiles: $V > R$ points spread over a
253: velocity range 220--360 km s$^{-1}$, whereas $V < R$ points distribute
254: around 250--320 km s$^{-1}$. Also, blue-dominated profiles with high
255: values of the peak separation occur only at the end of the $V > R$ cycle,
256: just before the shell phase. 
257: 
258: The density wave does not only affect the H$\alpha$  line but also the
259: Paschen series and He I $\lambda$6678 (Fig~\ref{global}). Since  these 
260: lines  are  formed  at  different   regions  inside  the equatorial disc
261: the perturbation  must extend over a very wide region, hence  confirming
262: its global  nature. 
263: 
264: %----------------------------------------------------------------------------
265:         \begin{figure}
266:     \begin{center}
267:     \leavevmode
268: \epsfig{file=figures/corr2.eps, width=7.0cm, bbllx=48pt, bblly=156pt,
269:   bburx=293pt, bbury=720pt, clip=}
270:  \end{center}             
271: 	\caption{Evolution of the V/R ratio ({\em a}), J band ({\em b}),
272: K band ({\em c}), J-K index ({\em d}), 
273: H$\alpha$ equivalent width ({\em e}) and peak separation ({\em f}).
274: A correlation between the V/R cycle and the infrared emission is apparent. 
275: The arrow in panel {\em (e)} indicates the X-ray outburst}
276:         \label{corr}
277:         \end{figure}
278: %----------------------------------------------------------------------------
279: 
280: 
281: \subsection{Infrared data}
282: 
283: The  infrared  observations  were taken with the  Continuous  Variable
284: Filter (CVF)  infrared  photometer,  using the 1.5-m Carlos  S\'anchez
285: telescope,  located at the Teide  Observatory,  in  Tenerife  (Spain). The
286: data were reduced following the procedure described by Manfroid (1993).
287: Instrumental values were transformed to the TCS standard system (Alonso et
288: al. 1998). Table~\ref{irobs} shows the results of the infrared
289: observations. For earlier observations the reader is referred to R97.
290: 
291: The  combination  of the old  and  new  infrared  data  yields  a very
292: interesting result, namely the correlation  between the V/R variations and
293: the infrared  magnitudes.  Fig~\ref{corr} shows the J and K light curves and
294: the evolution of the colour index J-K, covering the
295: period 1991 August to 1999 October.  The light curves were rebinned into 30
296: day bins. The errors were calculated from the photometric errors depending
297: on the number of points, $N$, in each bin: when the corresponding bin
298: contained only one point we simply took the observation error, if $1 < N
299: \leq 5$ then we defined the error as $\sum|x_i-x_m|/N$ and if $N > 5$ then
300: we considered the standard deviation  $\sqrt{(x_i-x_m)^2/N}$. 
301: 
302: As can be seen in Fig \ref{corr}, there is a distinct  modulation with an
303: amplitude, from maximum to minimum, of $\sim$ 0.3 mag.  The period of
304: this  modulation is  3.1--3.8  years, in  good  agreement  with the 
305: optical  V/R variability of 3.4 years. Interestingly, the infrared maxima
306: occur in coincidence with the optical ($\sim$ MJD 49250) and X-ray ($\sim$
307: MJD 50630) outbursts. On  the  other   hand,   the   infrared   colours 
308: do  not  change  drastically  over  the  period  of  the observations. 
309: That is, the slope of the infrared continuum remained the same over the
310: time covered by the observations.
311: 
312: \subsection{X-ray data}
313: 
314: LS I +61$^{\circ}$ 235 is  a persistent low-luminosity   Be/X-ray  
315: binary.  These systems   are   characterised   by long pulse periods, low
316: X-ray variability  ($L_{max}/L_{min}  \simless 10$) and low but 
317: permanent  levels of X-ray  emission  (Reig \& Roche 1999).  
318: 
319: RX J0146.9+6121 was observed with the RXTE {\em Proportional Counter
320: Array} (PCA) on March 21, 1998 (02:01--11:54 UT). Good time  intervals 
321: were defined by removing data taken at low Earth elevation angle ($<$ 
322: 10$^{\circ}$)  and during times of high particle background.  An  offset 
323: of only  0.02$^{\circ}$  between  the  source position and the pointing of
324: the satellite was allowed, to ensure that any possible short stretch of
325: slew data at the beginning and/or end of the  observation  was  removed. 
326: All five PCA units  were  functioning during the entire  observation.  The
327: total net exposure was 19223 s. Due to the relatively wide field of view
328: of the PCA instrument (1$^{\circ}$ FWHM) the nearby X-ray source 4U
329: 0142+61 (White  et al.  1996) also contributed to the total flux. After 
330: correcting for collimator  efficiency the contribution of 4U 0142+61 to
331: the total observed  flux in the energy  range  2--30 keV was estimated  to
332: be of $\sim$ 10\%.  Because of this no X-ray spectral analysis was
333: attempted.
334: 
335: The pulse period was  determined by  correcting  the data to the
336: solar system  barycentre  and using the epoch  folding  technique, i.e, we
337: folded the data over a range of periods  and  searching  for a maximum 
338: $\chi^2$ as a function of period.  The pulse period found was
339: 1404.5$\pm$0.5 s, which is virtually the same as the  1404.2$\pm$1.2 s
340: pulse period obtained in another RXTE  observation  nine months earlier
341: (Haberl et al. 1998). 
342:  
343: %----------------------------------------------------------------------------
344:         \begin{figure}
345:     \begin{center}
346:     \leavevmode
347: \epsfig{file=figures/global.eps, width=8.0cm, bbllx=20pt, bblly=310pt,
348:   bburx=488pt, bbury=712pt, clip=}
349:  \end{center}              
350: 	\caption{The V/R oscillations are not only seen in H$\alpha$ ({\em
351: left}) but also in He~I $\lambda6678$ ({\em middle}) and Paschen lines Pa11
352: ($\lambda8863$) and Pa12 ($\lambda$8750) ({\em right})}
353:         \label{global}
354:         \end{figure}
355: %---------------------------------------------------------------------------
356: 
357: \section{Discussion}
358: \subsection{Global $m=1$ oscillations}
359: 
360: In R97 we  investigated  the  different  models  that had been put forward
361: to explain the V/R  variability in Be stars and concluded that the model
362: which best accounted for the  observational  data in LS I +61$^{\circ}$ 235
363: was the Global  One-armed  Oscillation  model (Okazaki 1991, 1997;
364: Papaloizou et al. 1992).  This model  suggests that the  long-term  V/R 
365: variations  are caused by global $m=1$ oscillations in the cool equatorial
366: disc of the Be star.  In other words, an enhanced density perturbation
367: develops on one side of the disc, which  slowly  precesses.  The 
368: precession  time being that associated with the V/R  quasi-period.  The
369: density  perturbation is confined  within a few stellar radii in the disc
370: and the precession period turns out to be fairly insensitive to the size
371: of the disc (Savonije \& Heemskerk 1993).
372: 
373: One  prediction  of the model is that no  changes  in the slope of the
374: infrared   continuum  are   expected.  The  reason  is  that  the  V/R
375: variations are not the result of changes in the radial gradient of the
376: circumstellar  gas.  The slope of the infrared  continuum is a measure of
377: the radial density distribution but since matter in the disc does not
378: move  radially  no changes in the shape of the  infrared continuum are
379: expected.  This is exactly the behaviour that we find in the case of LS
380: I +61$^{\circ}$ 235.  While the  individual  infrared  photometric bands
381: changed  ($\Delta J \approx \Delta H \approx  \Delta K \sim 0.3$
382: magnitudes) the infrared colours remained unchanged (Fig.~\ref{corr}).
383: 
384: In  principle,  the issue of whether  the  motion of the  perturbation
385: occurs  in the  same  sense  (prograde  rotation)  or  opposite  sense
386: (retrograde  rotation) to the stellar rotation can be found out from the 
387: observations.  Telting et al.  (1994) realised that a prograde revolution
388: implies that the $V > R$ phase must be followed by a shell profile  and a 
389: similar profile but with much less pronounced absorption feature (or
390: possibly a single peak line if the inclination is low) during the
391: transition from $V < R$ to $V > R$.  These characteristic  line shapes
392: must translate into noticeable  photometric variations.  According  to 
393: Mennickent et al. (1997),  we should expect a minimum of brightness when
394: $V=R$ {\em prior} to the $V < R$ phase if the  motion  is  prograde  and
395: $V > R$  {\em after}  $V=R$  if retrograde.  In LS I+61 235 the minimum of
396: brightness  in the infrared photometric  bands occurred  during the shell
397: phase ($V=R$) before the $V < R$ phase began,  confirming the prograde
398: nature of the precession inside the disc.
399: 
400: However, models of one-armed global density waves  cannot reproduce 
401: strong  shell-non-shell transitions like the ones seen in \lsi\, which are
402: reminiscent of the so-called {\em spectacular variations} (Doazan et al.
403: 1983). Such shell events seem to be rare a phenomenon which have been
404: reported for only three Be stars: $\gamma$ Cas, 59 Cyg and Pleione (Hummel
405: 1998 and references therein), all of which are either binaries or 
406: suspected binaries. One possible explanation might be a {\em thick} disc in the
407: region where the perturbation lies.  When the perturbation is behind the
408: star and for the right inclination angle, no shell event is seen because
409: the disc in between the central star and the observer is thin and does not
410: occult the star nor the perturbation. The shell phase would occur
411: when the perturbation is at inferior conjunction since the thicker disc
412: would hide the central star from the observer. An alternative explanation
413: is given by Hummel (1998) who suggested a tilted or warped circumstellar
414: disc with precessing nodal line in addition to the density wave (see also
415: Porter 1998).
416: 
417: 
418: \subsection{A dense circumstellar disc in \lsi}
419: 
420: The works by Dachs et al. (1986) and Hanuschik et al. (1988) have shown
421: that the equivalent width of H$\alpha$ line emission for Be stars
422: increases with the effective disc radius.  Since for rotationally
423: dominated  profiles $\Delta V/(2\,v\,\sin{i})\;$ can be regarded as a
424: measure  of the radius of the H$\alpha$ emitting region (Huang 1972), we
425: expect a correlation between the peak separation and the H$\alpha$
426: equivalent width. Hanuschik et al. (1988) derived the following law
427: 
428: \begin{equation}
429: \log \left(\frac{\Delta V}{2 v \sin i} \right)= a \log (EW(H\alpha)) + b
430: \end{equation}
431: 
432: \noindent where $v \sin i$ is the projected rotational velocity and
433: EW(H$\alpha$) is given in anstrongs. $a$ and $b$  are related to the
434: rotational law index $a=-j/2$ ($j$=0.5 for Keplerian rotation and $j$=1
435: for conservation of angular momentum) and with the disc electron density,
436: respectively. A least square fit to the LS I +61$^{\circ}$ 235 data gave
437: $a=-0.23\pm0.10$, i.e. $j\approx 0.5$ and $b=+0.1\pm0.1$. These values are
438: to be compared with the average values $a=-0.4$ and $b=-0.1$ found by
439: Hanuschik et al. (1988) for a sample of 26 isolated Be stars.  The higher
440: value of $b$ in \lsi\ implies a denser disc than those of isolated Be
441: stars. 
442: 
443: The main (and probably only) difference between a Be/X-ray binary and an
444: early-type isolated Be star is the presence of a neutron star in the
445: former. It then seems natural to attribute the dissimilarity in the
446: properties of the circumstellar envelopes to the influence of such compact
447: companion. The neutron star trims the disc to a certain radius and
448: prevents its free growth, making it denser. Disc truncation has been
449: suggested in other Be/X-ray binaries like V0332+53 (Negueruela et al.
450: 1999). This idea  would support the hypothesis that the neutron star 
451: plays a fundamental role in the evolution and properties of the equatorial
452: disc in Be/X-ray binaries, as proposed by Reig et al. (1997b). 
453: 
454: 
455: \subsection{X-ray/optical/IR correlations}
456: 
457: In R97  the  observation  of an optical  outburst around 1993 
458: September-October (MJD $\sim$ 49260) was reported. During the outburst the
459: H$\alpha$ equivalent width, EW(H$\alpha$),  changed by $\sim$ 10 \AA, in
460: about 270 days, decreasing to  pre-outburst  values  ($\sim  -8 $\AA) in
461: about the same amount of time (Fig \ref{corr}$c$).  This increased 
462: coincided with a single-peak  phase of the H$\alpha$  profile.  The
463: question of whether this  outburst was an  isolated  event  or was
464: associated with the V/R cycle remained open due to the short coverage of
465: the data.  The new observations show no new EW(H$\alpha$)  maximum.  After
466: the  outburst  the  EW(H$\alpha$)  increased  slowly  up to a level of
467: $-12$ \AA, considerably lower than the peak of 1993 September.  The new
468: single-peak  phase should have occurred during 1997 March-May. 
469: Unfortunately,  the star was too close to the Sun to be observed. 
470: However, we notice that EW(H$\alpha$) seems to have reached a maximum
471: value just before and after the period  when  we  would  expect  the 
472: single-peak   phase  (around  MJD 50500). Thus, we are inclined   to think
473: that the higher EW(H$\alpha$) associated with single-peak profiles
474: reported in R97 may   reflect  the  fact  that the  fluxes  of both
475: components are adding up and are not affected by the  absorption feature 
476: present  in the other  profiles.  In this  context  then, the increase in 
477: EW(H$\alpha$) is real and  would be an event related to the motion of the
478: density pattern in the disc, rather than an isolated episode.
479: 
480: In 1997 July (MJD 50634) \lsi\ underwent a small X-ray outburst 
481: (Haberl et al. 1998).  The X-ray  luminosity increased by
482: nearly a factor of five in one week reaching  3.45  $\times$ 10$^{35}$ erg
483: s$^{-1}$ in the energy range 0.5--10 keV. This outburst is marked with and
484: arrow in Fig~\ref{corr}$c$. Interestingly, the outburst occurred at the
485: time of the expected infrared and EW(H$\alpha$) maxima.   One is then
486: tempted to attribute this correlated X-ray/optical/IR behaviour to the
487: high density perturbation where most of the Balmer emission is formed. If
488: the inclination of the system is less than 90$^{\circ}$,  when the
489: high-density part of the equatorial disc is behind the star it offers the
490: largest geometric area (especially so if the disc is thicker in this
491: region) and the highest optical and infrared emission. If the neutron star
492: happens to be close to the Be star it will accrete from this high-density
493: material and the X-ray emission will be enhanced. New optical and X-ray
494: observations around the next expected maximum are needed to solve this
495: issue.
496: 
497: 
498: \section{Conclusion}
499: 
500: Optical  spectroscopic  observations confirm  the  presence  of  global  
501: one-armed oscillations  in  the circumstellar  disc  of \lsi.  These 
502: oscillations  manifest themselves as quasi-periodic  variations in the
503: shape of the H$\alpha$ line, whose  asymmetric  double peak profile 
504: alternates  between red- (V$<$R) and  blue-dominated  (V$>$R)  emission.
505: The system also goes through Be-Be shell transitions, which might indicate
506: an asymmetric vertical structure of the disc in the form of a thick  or a
507: tilted disc. The V/R quasi-period is determined to be $\sim$ 1240$\pm$30
508: days. We have found a correlation between the infrared emission and the
509: V/R variations.  This is the first time that such  correlation is reported
510: in a Be/X-ray  binary. From the pattern traced by the IR light curves in
511: relation to the V/R ratio we conclude that the one-armed disk oscillations
512: are prograde. The Be star's disc in \lsi\ is found to be denser than 
513: isolated Be stars, which may be connected to the presence of the neutron
514: star.  From a {\em Rossi X-ray Timing Explorer}  observation we derive a
515: spin period of the neutron star of 1404.5$\pm$0.5 s. 
516: 
517: %The V/R variability is not exclusive of the H$\alpha$  line but is also
518: %present in the He I  $\lambda6678$  and  Paschen lines.  Since  these 
519: %lines  are  believed  to form at very  different depths in the Be star's 
520: %circumstellar  envelope we conclude  that the oscillations are affecting a
521: %large portion of the circumstellar  disc, i.e.  they are indeed global.
522: %The Be star's disc in \lsi\ is found to be denser than  isolated Be
523: %stars.  From a {\em Rossi X-ray Timing Explorer}  observation we derive a
524: %spin period of the neutron star of 1404.5$\pm$0.5 s. 
525: 
526: \subsection*{Acknowledgements}
527: 
528: We thank Chris  Moran for  providing  us with one of the INT  spectra and
529: Dr E. V. Paleologou for helping us with the spectroscopic observations at
530: Skinakas Observatory. Skinakas Observatory is a collaborative project of
531: the University of Crete, the Foundation for Research and Technology-Hellas
532: and the Max-Planck-Institut für Extraterrestrische Physik. P.  Reig 
533: acknowledges  support  via  the  European  Union Training and Mobility of
534: Researchers Network Grant  ERBFMRX/CT98/0195. IN is supported by an ESA
535: external fellowship. R. Zamanov acknowledges support from  Direcci\'on
536: general de relaciones culturales y  cient\'{\i}ficas, Spain. Some
537: observations were taken as part of the ING service observing programme. We
538: are grateful to the referees, Dr D. Baade and Dr T. Rivinius, for useful
539: comments.
540: 
541: \begin{thebibliography}{}
542: \bibitem{} Alonso A., Arribas S., Mart\'{\i}nez-Roger C. 1998, A\&AS 131, 209
543: \bibitem{} Coe M.J., Everall C., Norton A.J., Roche P., Unger S.J., Fabregat J., 
544: Reglero V. \& Grunsfeld J.M., 1993, MNRAS, 261, 599
545: \bibitem{} Dachs, J., Hanuschik, R., Kaiser, D. \& Rohe, D. 1986, A\&A, 159,
546: 276
547: \bibitem{} Doazan V., Franco M., Rusconi L., Sedmark G., Stalio R., 1983,
548: A\&A, 128, 171
549: \bibitem{} Haberl F., Angelini L. \& Motch C., 1998, A\&A, 335, 587
550: \bibitem{} Hanuschik R.W., Kozok J.R., Kaiser D., 1988, A\&A 189, 147
551: \bibitem{} Hellier C., 1994, MNRAS, 271, L21
552: \bibitem{} Huang, S. 1972, ApJ, 171, 549
553: \bibitem{} Hummel W., 1998, A\&A 330, 243
554: %\bibitem{} Hummel W. \& Vrancken M., 1995, A\&A, 302, 751.
555: %\bibitem{} Hummel W., Hanuschik R.W, Vrancken M., 1998, in Kaper L., Fullerton 
556: %A.W., eds., Cyclical Variability in Stellar Winds, Proceedings of the ESO 
557: %Workshop held at Garching, Germany, 14-17 October 19. Springer-Verlag, 1998, 
558: %Berlin, p.343
559: %\bibitem{} Hummel W., Hanuschik R.W., 1997, A\&A 320, 852
560: %\bibitem{} Israel G.L., Mereghetti S. \& Stella L., ApJ, 1994, 433, L25
561: \bibitem{} Manfroid J. 1993, A\&A 271, 714
562: \bibitem{} Mennickent, R. E., Sterken, C. \& Vogt, N. 1997, A\&A, 326, 1167
563: \bibitem{} Mereghetti S., Stella L. \& De Nile F., 1993, A\&A, 278, L23
564: \bibitem{} Motch C., Haberl F., Dennerl K., Pakull M.\& Janot-Pacheco E.,
565: 1997, A\&A, 323, 853
566: \bibitem{} Negueruela I., Reig P., Coe M.J. \& Fabregat J., 1998, A\&A,
567: 336, 251
568: \bibitem{} Negueruela I., Roche P., Fabregat J., Coe, M. J., 1999, MNRAS,
569: 307, 695
570: %\bibitem{} Negueruela I. \& Okazaki A.T., 1999, A\&A, in press
571: \bibitem{} Okazaki A.T., 1991, PASJ 43, 75
572: \bibitem{} Okazaki A.T., 1997, A\&A 318, 548
573: \bibitem{} Papaloizou, J.C., Savonije, G. J., Henrichs, H. F, 1992, A\&A, 265, 
574: L45
575: \bibitem{} Porter J.M., 1998, A\&A 336, 966
576: \bibitem{} Reig P., Fabregat J., Coe M. J., Roche P., Chakrabarty D., Negueruela 
577: I \& Steele I., 1997a, A\&A, 322, 183, (R97).
578: \bibitem{} Reig P., Fabregat J. \& Coe M.J., 1997b, A\&A, 322, 193
579: \bibitem{} Reig P. \& Roche P., 1999, MNRAS, 306, 100
580: \bibitem{} Savonije G.J. \& Heemskerk M.H.M., 1993, A\&A, 276, 409
581: %\bibitem{} Stella L., Israel G.L. \& Mereghetti S., 1998, in{\em The Many Faces 
582: %of Neutron Stars}, Kluwer Academic Publishers. Ed Buccheri et al.
583: \bibitem{} Telting, J.H., Heemskerk M.H.M., Henrichs H.F. \& Savonije G.J.,
584: 1994, A\&A, 288, 558
585: \bibitem{} White N.E., Angelini, L., Ebisawa K., Tanaka Y. \& Ghosh P., 1996, 
586: ApJ, 463, L83
587: 
588: \end{thebibliography}
589: 
590: 
591: 
592: \end{document}
593: