The effects of rapid rotation on the emergent energy distribution, line profiles, atmospheric motions and polarization are discussed. A simplified explanation of some of the effects is presented. Results of detailed radiation transfer calculations are briefly reviewed. The rotation can lead to circulation and turbulent motions in the photospheric layers which could affect the outflow from the Be stars. The rotation rates actually observed in the Be stars are sufficiently below the critical rate that many of the effects predicted by the plane parallel atmosphere calculations should be small. Nevertheless, the models are useful and necessary for estimating rotation speeds from lines that are widely separated in wavelength. The rapidly rotating photospheric models predict far too small an infrared excess, as well as too small an intrinsic polarization. The explanation of these observations requires that geometrically extended envelopes be considered. Theoretical models for the intrinsic polarization are critically discussed. It is stressed that polarization is a powerful diagnostic for determining the asymmetrical structure of the outer atmospheres of the Be stars.

Rotation has two main qualitative effects on the structure of a star: it distorts the shape and it reduces the total luminosity. The luminosity depends only on the total angular momentum but the shape, and so the observable properties, depends on the internal distribution of angular momentum. The expected distribution, even for main-sequence stars, is an unsolved problem. Redistribution of angular momentum can occur as a result of large-scale circulation currents, instabilities and magnetic stresses and there may be no steady state. If shear instabilities are dominant, then stars may rotate nearly uniformly. Studies of non-radial oscillations might perhaps eventually lead to a direct probing of the internal angular momentum distribution by the methods now being used to study the internal rotation of the Sun. Differential rotation near the surface would probably give rise to turbulent motions, which could generate the mechanical energy flux which seems to be a necessary input to the winds in hot stars.

Recent observational work on spectra obtained at optical and near-infrared wavelengths is reviewed for “ordinary” (non-supergiant non-peculiar) Be and Be-shell stars, with particular emphasis on comparison between high-resolution spectral measurements and current Be star models. Emission-line profiles are interpreted in terms of geometry, dimensions and dynamics of a dense (N ≃ 1011 cm-3), cool (T ≃ 104 k) line-emitting region in the circumstellar envelope, separated from the thin, hot expanding stellar wind region. Present observations support a rotating-disk model for the cool, line-emitting region. Indications pointing towards structuring of cool circumstellar envelope regions are reported. Significant spectral variations are detected both on short (less than one day) and long time scales (months to years). Typical cycle times for large-amplitude V/R variations and intervals between shell episodes depend on spectral type, increasing with decreasing effective temperature of the central stars and ranging from a few years for early-Be type stars to a few decades for late Be stars.

We review the recent results on flux distribution of Be stars from 320 to 850nm. It appears that data are available for only 10 to 15 % of the objects recognized as “normal” Be stars. The main current problems appear to be the variability of the continuous radiation, and the correction for interstellar reddening. Various attempts to solve these matters are discussed. We stress the need for fairly high (lnm) resolution spectral scans obtained simultaneously with data in other spectral ranges, repeated at appropriate intervals of time. Such requirements make it necessary to restrict the observing programmes to a limited sample of carefully chosen objects.

Kogure and Suzuki (1984) have noticed that EW Lac (HD217050, B2 III pe) has revealed the long-term variation notably different before and after about 1976. Before 1976, the profiles of spectral lines were sufficiently symmetric and stable, suggesting a circular disk or ring for the structure of envelope. In contrast, since 1977, EW Lac entered a phase of anomalous V/R variation.

We have obtained about 180 coudé spectra in 16 years by 1985 with the 188 cm reflector at the Okayama Astrophysical Observatory. Most of them were measured using the PDS densitometer at the Kwasan Observatory, University of Kyoto for the radial velocities and line profiles. A part of results were reported by Suzuki and Kogure (1986).

At first we examine the main characteristics of 3 well-known shell stars (EW Lac, ζ Tau, 48 Lib) before and during their variable shell phase (long time scale V/R and RV cycles). For that we have investigated a) recent studies of our Meudon working group, based on long sequences of spectroscopic observations (Hubert et al. 1983, 1986a; Mon 1984) b) previous data found in the literature c) additional information by a reexamination of line profiles obtained from our material. For ζ Tau and 48 Lib the selected cycles have been chosen among the most representative ones in the history of their respective shell episode. For EW Lac, which entered an active phase in 1977, we have studied the only recently observed cycle of shell feature variations. A summary of the 3 V/R shell stars characteristics during 1) the pre variable shell phase 2) the variable shell phase is reported elsewhere (Hubert et al. 1986b); we only give here general trends of it, Table 1.

Photometric periodicities with two unequal maxima and minima reported for Be stars are shown to result from a fundamental and overtone associated with rotation of an atmospheric bulge. It is the form of the scattering function of the free electron that produces the double periodicity; the stellar geometry — co-latitude of the bulge and inclination of the stellar rotational pole — controls the amplitudes of the two components. The form of the light curves is consistent with the oblique rotator model.

Linear optical polarization of 88 Her has been measured in V spectral region during the period 1974–1985. The mean annual values of the intrinsic polarization parameters are presented. The polarization percentage changes from 0.15%(1976) to 0.56%(1979). Small values of the polarization percentage correspond to the period when the envelope effect is negligible. The maximum of polarization percentage has been found during the early period of strong shell phase. The polarization position angle varies between 53 and 83 degrees.

All programme stars including standard stars were observed with the standard RI filters of the Johnson system with a 20-arcsec diaphragm for 4 nights during the 1982–83 observing season. A photoelectric photometer attached to the 1.2 meter reflecting telescope of the Japal— Rangapur Observatory was used for these observations. A dry-ice cooled EMI 9558B photomultiplier tube was used as the detector and the photo-current was recorded by means of a GR 1230A DC-amplifier and a Honeywell Brown chart recorder.

The extinction coefficients for R and (R-I) were determined from the observations of standard stars taken from the list of Johnson (1966) and also from the observations of an early and a late spectral type star on each night and all the observations were corrected for atmospheric extinction. The instrumental magnitudes of a programme star were transformed to the standard system through normal procedures.

During two observation periods in Nov./Dec. 1981 and April/ May 1985, absolute spectral energy distributions of 41 southern and equatorial Be stars were photometrically measured at 10 Å spectral resolution in the wavelength range between 3200 Å and 8500 Å, using the scanning spectrophotometer attached to the 61 cm telescope of the University of Bochum at La Silla/Chile. Spectral fluxes of the program stars have been determined with an error of 0m.03 or less.

For a number of bright southern Be stars, high-resolution, high S/N spectroscopy has been performed for Hα, Si II λ6371 and four Fe II emission lines. With the exception of four stars, both Hα and weak emission line profiles exhibit resolved double-peak structure. Peak separation increases with stellar projected rotational velocity. In many cases, Hα profiles (but not Fe II lines) show resolved fine-structure (inflections in the flanks) indicating a two-component structure of the disk. Envelope models with arbitrary density and rotational velocity laws have been fitted to the Fe II line profiles with satisfactory results. It is found that the innermost Hα disk component also emits the Fe II lines and has an average outer radius of 5 r*; density falls off as r-l.5 or more rapidly. The outer disk component emits the faint and narrow secondary component of Hα lines with average outer radius 20 r*. The envelopes of the four stars with a strong, dominating single peak both in Hα and Fe II either are dominated by strong radial motions or, more likely, have elliptical disk-like structure.

Absolute spectrophotometry data have been obtained for a sample of Be stars. These data were obtained with the York Observatory 60 cm telescope and vidicon spectrophotometer. The data have been corrected for differential atmospheric extinction, instrumental response, and interstellar extinction and calibrated in terms of absolute flux. Absolute Hα and Hβ fluxes have been determined.

It is well known that the intrinsic colors of Be stars are not the same as those of “normal” B stars and they are often redder. So interstellar absorption of Be stars determined from photometric indices may be overestimated. In order to avoid this, a method often used to determine the interstellar absorption of Be stars is based on the ultraviolet interstellar absorption bump at 2200 Å (Dorschner 1975; Nandy et al. 1975, 1976). However some peculiar Be stars exist whose circumstellar matter contributes to the 2200 Å bump (Savage et al. 1978). The aim of this work is to test the deredenning method from the 2200 Å bump for “classical” Be stars, in determining the interstellar absorption in the region of the Be star, using the surrounding normal stars.

Within the framework of a coordinated campaign of observations in optical and IR of a sample of 21 southern Be stars, we report some results from spectra taken at ESO La Silla Observatory with 1.5m telescope equipped with IDS and the Boller & Chivens grating spectrograph used at 220 A/mm reciprocal dispersion. The spectra, taken in March 1985, range from about 4250 to 7000 A. The equivalent widths of the Balmer lines, either in emission or in absorption, have been computed. Fig. 1 shows the relationship between the logarithm of the absolute value of the Hα, Hγ equivalent widths ratio and the logarithm of the luminosity of the stars in solar units.

The first indication that Be stars have an excess near-IR radiation, compared to normal stars of the same spectral type was found by Johnson et al. (1966). Johnson (1967) noticed that in his sample of 85 early type stars all the Be stars and shell stars had an excess in K-L and he concluded that this is due to IR emission from circumstellar shells. Woolf et al. (1970) suggested that the IR excess of Be stars might be due to free-free emission in the ionized circumstellar envelope which also produces the Balmer emission lines, but their observations at 5 and 10 µ could not rule out the possibility that circumstellar dust contributed to the excess. The observations by Allen (1973) of a large number of Be stars up to a wavelength of 3.5 µ could not make the distiction between free-free or dust emission either.

We have observed 103 bright Be stars in the near infrared up to 10500 A with a dispersion of 230 and 50 A mm-1. The observations were performed with a Reticon (1024 diodes) attached to the ROUCAS spectrograph at the 193 cm telescope of the Haute Provence Observatory.

In this spectral range, the Be stars are characterized by the lines of HI (Paschen series), 0I(7772-74-75, 8446 A), CaII(8542, 8662, 8498 A), FeII(7712, 9997 A) and NI(8686-83-80, 8719-12-03, 8629 A).

On our spectra, the CaII triplet is always blended with P13, P15, P16, and only the enhancement of these lines permits to conclude to the CaII presence. The 0I 8446 A is perturbed by the Paschen lines P17 and P18 (low dispersion spectra) and P18 (high dispersion spectra). The NI 8686-83-80 A lines also perturb the P13 profile.

Spectra of northern Be stars in the near infrared region have been already described by some authors (see Slettebak, 1979 for a brief review). This paper presents general results for southern Be stars obtained from more than 100 spectrograms in the region λλ7750-9000 Å the spectra were taken in four nights of February 1985 at the 1.5m ESO telescope with the reticon instrumentation; the dispersion is 58 Åmm-1. All the available southern Be stars (right ascension between 4h and 17h) listed in the Bright Star Catalogue (Hoffleit and Jaschek, 1982) were observed. Table 1 lists the number of stars vs. spectral type (columns 1 and 2) and the number of stars which display emission at the Paschen series, 0I λ8446 Å and Call triplet (columns 3 to 5).

Infrared observations represent a powerful tool in understanding the physical conditions of the circumstellar matter around Be stars. Previous IR studies have been reported by different authors (i.e. Gehrz et al. 1974, Persi et al. 1982, Ashok et al. 1984, and Waters, 1986).

For this purpose, we have undertaken an IR study of 21 southern Be stars including JHKLM photometry, CVF spectra at resolution R=100 around the IR HI emission lines Brγ, Brα, and Pfγ and IRAS data at 12 and 25 microns taken from the Point Sources Catalogue. The IRAS flux densities were corrected following the procedure described in the IRAS Explanatory Supplement.

The near-IR observations were obtained in March 1985 at the lm ESO (La Silla, Chile) telescope, equipped with the InSb Spectrophotometer. In the next section we discuss the nature of the observed IR excesses, and from the measured Brα/Brγ intensity line ratios for three Be stars, we derive the mass-loss rates.

The mass loss rates of a large sample of Be stars derived from the UV and the IR are compared. The IR rates were derived using a simple equatorial disc model, and are typically a factor 100 larger than those derived from the UV. In terms of mass fluxes, the mass flux in the polar regions (derived from the UV observations) is about a factor 103 smaller than the mass flux in the equatorial regions. The dependence of MIR and Muv on stellar luminosity is studied. It is shown that MIR depends weaker on L than Muv. This suggests that two different mechanisms are responsible for the mass flux in polar and equatorial regions.