Recently, Chang et. al. (1992) and Carlsson, Rutten and Shchukina (1992) (CRS) demonstrated the non-LTE formation mechanism behind the 12 μm Mg I emission lines (6g–7h, 6h–7i) observed in the solar spectrum (Murcray et. al., 1981). CRS stress the generality of this mechanism showing that it is a natural consequence of the recombination flow from the large Mg II reservoir through the Rydberg levels of Mg I. We have noted the close parallel between Mg I in the solar atmosphere and Mg II in the atmospheres of B stars (where Mg III plays the role of the reservoir) and investigated the operation of this mechanism in high-ℓ infrared transitions of Mg II. We have employed a 58 level Mg II atom including all energy levels through n = 25 and a total of 491 linearized radiative transitions. The coupled equations of radiative transfer and statistical equilibrium were solved with the MULTI code in its local operator form (Carlsson, 1992).

Near-IR variable Be stars are identified from multi-epoch observations spanning 20 years using a statistical technique. In this manner, observations from different observing sites can be meaningfully combined and compared. A more thorough investigation of the incidence and properties of IR variability in Be stars as a class of stars is then possible.

The radiative transfer equation is solved for a generalised disc model of the circumstellar envelope around Ψ Persei to investigate the nature of the long wavelength turndown in the continuum spectrum of Be stars. The flux density and source dimensions at 15 GHz rule out truncation of the circumstellar envelope around Ψ Persei as the cause of the turndown.

Recently, Brα and Brγ emission was detected in the infrared spectrum of the B0.2V star τ Scorpii, without noticeable emission in Hα (Waters et al. A&A 272, L9-L12, 1993). Here we present simple HI recombination line calculations in the infrared and in the optical that demonstrate that there could be a class of B stars with low-density discs, a factor of 100 lower in density compared to normal Be stars, which may have escaped detection so far.

Be stars are defined to be non-supergiant early-type stars of spectral type B showing at times Balmer emission lines in their spectra. These stars often develop strong stellar winds considered to be variable in nature (Slettebak 1988) and have high rotational velocities compared to normal stars of similar spectral types. They also tend to show an excess amount of energy in the near- and far-infrared region compared to normal stars which is presumed to be due the surrounding material around the central star. Thus, the observed energy is a combination of that due to the stellar source and the surrounding material. Various attempts have been made to disentangle the stellar energy component from that of the circumstellar component in order to understand the nature, size and temperature of the envelope. These include:

a) Radius determination based on IR excess (Gehrz et al. 1974, Dachs and Hanuschik 1984; Waters et al. 1987),

b) Radius estimates from polarization and spectrophotometric data (Jones 1979),

c) Envelope dimensions derived from the width of shell absorption cores (Kogure 1969; Hirata and Kogure 1977),

d) Dachs et al. (1992) attempted to understand the physical properties, flow patterns and density distribution of the gas by a comparison of synthetic emission line profiles and empirical profiles measured for real Be stars.

Line emission in Be star spectra is accompanied by continuous emission both in the Balmer continuum and in the infrared spectral region, due to the same process that is responsible for Balmer line emission, i.e. to recombination radiation from ionized hydrogen in the extended circumstellar disks surrounding the hot central stars.

We sketch a new method for the accurate flux calibration and normalization of stellar spectra. This is of particular importance for the analysis of rapidly rotating early–type stars. Some preliminary log g determinations by profile fitting of H γ are presented.

Using published data and new data from mm observations, we calculate spectral indices α(12μ-25μ), α(25μ-1.1mm) and α(1.1mm-2cm) for Be stars. The index α(25μ-1.1mm), obtained for 8 stars, shows two characteristics: 1) for “normal” Be stars the index decreases from earlier types towards later types, i.e., later type Be stars tend to have shallower spectra in the IR-radio region than earlier types; 2) the index for shell stars appears to have smaller values than that for normal Be stars with same spectral types.

Circumstellar envelopes of young and evolved stars are responsible for many important phenomena concerning the exchange of matter, angular momentum, energy and maybe magnetic field between the core structure of stars and the interstellar medium. In particular, it is through them that matter enriched in heavy elements flows from evolved stars towards the interstellar gas, submitted to complex ordinary chemistry or photochemistry and condensation into solid particles.

Probing the atmospheres of Be stars by means of Optical Long Baseline Interferometry (OLBI) is a method still in its infancy. Published observational results in this field can be counted on one hand. In this paper we introduce the principles of OLBI as they apply to observations of Be stars and review results from existing long baseline interferometers. Regular operation of OLBI is most effective when carried out in conjunction with standard techniques such as photometry and spectroscopy. With the next generation of synthetic arrays combining sub-milliarcsecond resolution and a spectral resolution better than 10000, OLBI should be able to address the direct detection of pulsation and magnetic fields.

Eighteen bright Be stars have been observed polarimetrically and line- photometrically for searching for candidates for a direct measurement of the envelope diameters of Be stars by GI2T. I found that o Cas and 28 Tau are suitable in addition to γ Cas.

The recent progress in optical interferometry has made the direct study of the circumstellar matter around Be stars possible. The Hα emission region around γ Cas has been resolved with the I2T (Thom et al. 1986), GI2T (Mourard et al. 1989) and MkIII (Quirrenbach et al. 1993) instruments; the results support the basic picture of a rotating disk-shaped envelope (e.g. Poeckert and Marlborough 1978). Nearly spherical geometries can be ruled out. In this paper, first results from MkIII observations of a small sample of Be stars (see Table I) will be presented.

We discuss the effects of rotation on the structure of radiatively-driven winds. When the centrifugal support is large, there is a region, at low latitudes near the surface of the star, where the acceleration of gravity is larger than the radiative acceleration. Within this region, the fluid streamlines “fall” toward the equator. If the rotation rate is large, this region is big enough that the fluid from the northern hemisphere collides with that from the southern hemisphere. This produces standing shocks above and below the equator. Between the shocks, there is a dense equatorial disk that is confined by the ram pressure of the wind. A portion of the flow that enters the disk proceeds outward along the equator, but the inner portion accretes onto the stellar surface. Thus there is simultaneous outflow and infall in the equatorial disk. The wind-compressed disk forms only if the star is rotating faster than a threshold value, which depends on the ratio of wind terminal speed to stellar escape speed. The spectral type dependence of the disk formation threshold may explain the frequency distribution of Be stars. Observational tests of the wind-compressed disk model indicate that, although the geometry of the disk agrees with observations of Be stars, the density is a factor of 100 too small to produce the IR excess, Hα emission, and optical polarization, if current estimates of the mass-loss rates are used. However, recent calculations of the ionization balance in the wind indicate that the mass-loss rates of Be stars may be significantly underestimated.

We use a 2-D PPM code to simulate numerically the hydrodynamics of a radiation-driven stellar wind from a rapidly rotating Be-star. The results generally confirm predictions of the semi-analytic “Wind Compressed Disk” model recently proposed by Bjorkman and Cassinelli to explain the circumstellar disks inferred observationally to exist around such rapidly rotating stars. However, this numerical simulation is able to incorporate several important effects not accounted for in the simple model, including a dynamical treatment of the outward radiative driving and gas pressure, as well as a rotationally oblated stellar surface. This enables us to model quantitatively the compressed wind and shock that forms the equatorial disk. The simulation results thus do differ in several important details from the simple model, showing, for example, cases of inner disk inflow not possible in the heuristic approach of assuming a fixed outward velocity law. In addition, the disk opening typically has a half-angle of 2–4 degrees, somewhat larger than the ~ 0.5° predicted from the analytic model, and there is no evidence for the predicted detachment of the disk that arises in the fixed outflow picture.

We present theoretical line profiles and intensity maps from an axi-symmetric radiative wind model from a fast rotating Be star (Araújo & Freitas Pacheco, 1989; Araújo et al, 1993). The introduction of a viscosity parameter in the latitude dependent hydrodynamic code enables us to consider the effects of the viscous force in the azimuthal component of momentum equations. The line force is the same as Friend and Abbott (1986), but it is introduced a varying contribution of thin and thick lines from pole to equator by adopting latitude-dependent radiative parameters. The numerical calculation for parameters characteristic of early Be stars gives a density contrast between equator and pole of the order of 100. The total mass loss rates range from 10-8 to 10-7 solar mass per year. Furthermore, the “opening-angle” usually adopted in had-hoc models (Lamers and Waters, 1987; Waters et al., 1991) arises naturally as a result of our model. The velocity fields and density laws derived from the hydrodynamic equations have been used for solving the statistical equilibrium equations. By adopting the Sobolev approximation, we could easily obtain a good estimate of both electronic density and hydrogen level populations throughout the envelope (Stee and Araújo, 1994).

B[e] supergiants constitute one group of stars in the upper left part of the HR diagram. On the basis of their hybrid spectrum and their intrinsic polarization a two components envelope was suggested (Zickgraf & Schulte-Ladbeck, 1989): a “polar wind” and a denser “equatorial disc”.

Hot-star winds are believed to be driven by line-scattering of the star's continuum radiation. This review summarizes the physics of this line-driving process and the resulting basic wind properties. I also briefly discuss the generation of wind structure, both at small scale due to inherent instabilities in the line-driving, and at large scales due to various rotational effects.

We newly calculated the line radiative force with 520,000 atomic lines, which is twice as many as those of Abbott (1982), for OB supergiants. Our results are as follows. (1) The mass loss rates for O stars with Teff = 50,000K are seven times as large as Abbott's (1982) because of contribution from Fe iv lines. (2) Contribution from many weak lines increases the mass loss rates and decreases the wind velocities of OB stars within a temperature range of 10,000K ≤ Teff ≤ 30,000K. This result is qualitatively in accordance with the results from the recent observations of O stars. (3) The mass loss rates of OB stars depend on metallicity with Ṁ ∼ Z.

In the papers by Vilkoviskij (1981), Vilkoviskij and Tambovtzeva (1988) and Springmann and Pauldrach (1992) it was shown that ions can reach velocities relative to the wind plasma when the radiation pressure force exceeds the protons frictional force.

Litle is known about the wind of main sequence A7 to B4 stars. We here investigate the case of a radiatively driven wind for an A0 star and obtain stringent limits on the mass loss rate. We also show that frictional heating lead to the presence of a chromospheric region very near the photosphere.