Energy distributions from 0.32 to 1.06 μ for field Be stars having p designations in their spectral types show that stars with Balmer discontinuities in emission do exist. Furthermore, the filling in of the Balmer continuum near the discontinuity is usually observed. The continuum emission showing a peak near 1 μ is a common feature in the energy distributions presented here.

Energy distributions for a number of bright field Be stars are also presented and discussed.

The classical rotational hypothesis for Be stars is reviewed and discussed. Methods of measuring rotational velocities are considered and the particular difficulties that Be stars pose in this respect are pointed out. Several tests of the rotational hypothesis are discussed and shown to support this idea. New observations of Balmer emission-line widths in the spectra of 46 Be stars show a correlation with rotational velocity, also in support of the rotational hypothesis. The very large emission widths found for some Be stars are discussed and possible reasons for the broadening are presented.

Evidence of anomalous C – N abundances in the atmospheres of six rapidly rotating B-type stars has been found from an LTE analysis of ultraviolet line spectra obtained by Copernicus. Such anomalies are in the direction (carbon depletion, nitrogen enhancement) predicted by Paczyński (1971), but the extent of the anomaly grows toward earlier spectral type, contrary to Paczyński's predictions.

Oe stars are earlier type analogues of the better known Be class. These stars have relatively narrow emission in the hydrogen lines and sometimes in the helium lines, which often appears to be double. In several Oe stars, the emission is intermittent. Other properties of the class include relatively broad absorption lines and luminosities near the main sequence. Line profiles in three representative Oe stars are presented and discussed. The similarity of the spectra of Oe and Be stars suggests that the formation mechanism for the emitting region, whatever it is, must extend to the mid O-type stars, at least.

Photometric data for the hydrogen lines in 80 southern and 55 northern Be stars give evidence of a constant ratio for the emission in α and β in stars of similar spectral type. In an (α, β) diagram the intersection of this emission-line ratio with the standard relation for non-emission stars gives lower values than those corresponding to standard stars of the same spectral type. This may be due to rotational effects. A correlation of the variation in the β index with the V magnitude is also given.

During the past few years the search for radio stars has been both intensive and extensive, and Be stars are among the various types of objects which have been examined. A large number of Be stars have not been detected, and to date no classical Be star has been found to emit detectable amounts of radio emission. However, following the tradition of radio astronomy to emphasise the abnormal or extreme cases, radio emission has been observed which is associated with a few peculiar Be stars.

The history of ultraviolet studies of Be stars is barely ten years old. However, in the last decade, twelve major space experiments have observed Be stars in the ultraviolet region of the spectrum, and ultraviolet data for over 30 Be stars are now available in the literature. Table I shows some of the characteristics of the experiments. They include two rockets, five astronomical satellites, three manned satellites, and one planetary probe. Except for the rocket experiments, they are primarily survey instruments, which have provided ultraviolet data on early-type stars in general, and these data have proved to be extremely useful as standards of comparison for Be stars. Of the twelve spacecraft, two are presently operating: Copernicus and ANS. The two experiments complement one another very nicely in that the ANS experiment can obtain absolute continuous flux distributions, while the Princeton experiment can obtain high-resolution line spectra for the brighter Be stars.

Ultraviolet spectra of intermediate resolution have been obtained with Copernicus of twelve objects classified as Be or shell stars, and an additional 19 dwarfs of spectral classes B0-B4. Some of these spectra show marked asymmetries in certain resonance lines, especially the Si iv doublet at λ 1400 Å, indicating the presence of outflowing material with maximum velocities of nearly 1000 km s−1. Direct evidence for mass loss at these velocities is seen for the first time in dwarf stars as late as B1.5. Later than B0.5, the only survey objects showing this phenomenon are Be stars. Among the stars considered there is a correlation between the presence of mass-loss effects and projected rotational velocity, suggesting that the UV flux from B1-B3 dwarfs is sufficient to drive high-velocity stellar winds only if rotation reduces the effective gravity near the equator. The role of mass-loss in producing the Be star phenomenon and the effects of rotation on mass loss are discussed.

The intensities of the very strong lines of C IV λ 1549 Å and Si IV λλ 1394, 1403 Å observed in spectra obtained with Skylab experiment S019 provide a sensitive discrimination of spectral type between B0 and B2. Eye estimates of the Si IV/C IV ratio are tabulated for 33 B0-B2, class III-V stars of which 11 are emission-line stars. Seven of the emission-line stars show significantly smaller ratios than normal stars of the same MK class. The most outstanding examples are 60 Cyg, o Pup, η Cen, and Ara.

By using several ultraviolet spectrograms in the λ 2100–2800 Å wavelength range, the absorption features and the flux values of ζ Tau have been determined and compared to different early type stars.

The main part of this paper is concerned with observations obtained using the Princeton ultraviolet telescope-spectrometer on the Copernicus satellite through the Princeton guest investigator programme. These observations provided high resolution ultraviolet spectra of the bright southern shell star η Cen, which are now being analysed in detail. A short description of the new η Cen spectra and a progress report on the data analysis are presented below. In the second part of the paper we describe visible-region spectra of the northern Be stars γ Cas and φ Per obtained with a cross-dispersed echelle spectrograph on the 36-inch Yapp telescope at the Royal Greenwich Observatory.

The spectra of the ‘pole-on’ Be stars υ Cyg and μ Cen have been observed in the region λλ 1000–1450 Å with the Princeton Ultraviolet Spectrometer on board the Copernicus satellite. The data include scans of intermediate resolution (0.2 Å) with U2 and high resolution scans (0.05 Å) of selected suspected shell lines with U1. The spectra of υ Cyg and μ Cen are compared with a complete intermediate resolution Copernicus scan of the equator-on Be star φ Per. The ultraviolet spectra of υ Cyg and μ Cen appear to be identical to those observed in non-Be stars of comparable ground-based spectral type. Neither emission nor shell features are observed. However, numerous strong shell features are seen in φ Per. An upper limit on the column density of hydrogen in our line of sight is computed for υ Cyg from the observed strength of a weak feature at λ 1130.4 Å identified as an Fe III shell line. Estimates for the dimensions of the circumstellar envelope are then obtained. Some implications of the strengths of the Fe III shell lines at λ 1130 Å observed in υ Cyg, μ Cen, φ Per, and other Be stars are discussed.

As the title indicates two topics will be tackled in this presentation: spectroscopy and photometry. Let us therefore define first what ‘infrared’ means for these two subjects. I chose the following definitions:

1. – photographic infrared spectroscopy: wavelengths Hα ≤ λ < 1.2 μ

2. – near infrared photometry: wavebands: 1.6 μ ≤ λ ≤ 20 μ

We searched for the general properties of the Be stars with Paschen emission lines. First, we obtained a relation between the infrared excess of the Be stars and the presence of emission in the Paschen lines. Until now, these emission excesses could be related to no physical characteristics of classical Be stars. Then, the measures of Paschen decrements of 12 stars whose spectral types range from B0e to B5e allowed us to check several theoretical calculations about the formation of emission lines in the envelopes of Be stars. Thus we can see the prominent part played by the electronic collisions in the Sobolev theory. Indeed, only with calculations taking the electronic collisions into account, can we obtain theoretical values agreeing with both measured Paschen and Balmer decrements for the hottest stars of our sample. However, no theoretical values agree with the observed decrements for the cooler stars.

Faint emission-line stars found on Hα surveys are usually classified as planetary nebulae or Be stars on rather arbitrary criteria. A comparison of spectra of such stars with their optical and infrared properties suggests that a next step in their classification can be made quite rapidly by photometry. In particular the more unusual specimens can be readily isolated.

A review of polarization produced in the extended circumstellar disks about Be stars is given. While variability of the polarization and its peculiar wavelength dependence in wide-band continuum measurements are discussed, emphasis is placed upon the recently discovered polarization effects in the emission lines and upon a discussion of models. Scattering in the photosphere of a star, even one rigidly rotating near breakup velocity, does not explain the observations. Better fits to the observations are obtained by scattering from electrons in a flattened gaseous disk extending some 10 stellar radii from the star with electron temperatures Te ≅ 10 000 K and electron densities Ne ≅ 1012 cm−3. In some stars the emission flux in Hα and Hβ appears to be partly polarized.

We have made linear polarization measurements of several Be and shell stars across the Hα and Hβ lines, all of the stars exhibiting polarization in the continuum, the emphasis here being on measurements made of ζ Tau, 48 Per, φ Per, and γ Cas. Three types of results ensue: some stars show no significant change of polarization across the Balmer features (e.g., 48 Per, X Per); some stars show a reduced polarization across the features (e.g., γ Cas, ζ Tau) indicating the presence of intrinsic polarization; some stars show a change in the degree of polarization but with a marked rotation of the direction of vibration (e.g., φ Per, 48 Lib) which can be attributed to a combination of non-aligned intrinsic (circumstellar) and interstellar polarizations. Interpretations of these results are discussed, and we demonstrate the potential power of line profile polarimetry/photometry as an important new method for separating intrinsic and interstellar polarization effects, thus enabling polarization observations to be used as a constraint on models of Be stars.

The polarization of 48 Be stars has been measured in two bands near Hα with the aim of determining the relation between intrinsic polarization and v sin i. A technique developed by Poeckert (1975) is used to remove the effect of interstellar polarization. It is found that intrinsic polarization depends strongly on v sin i; stars with low v sin i having little or no polarization. We have calculated the i dependence of linear polarization for a disk model envelope and find that the polarization is proportional to τe sin2i when the disk is optically thin (τe is a characteristic electron scattering optical depth). A comparison of the observed relation between intrinsic polarization and v sin i, and that predicted for the disk model is illustrated. We find that an envelope with an electron density of ≤5 × 1011 cm−3 can account for the degree of intrinsic polarization observed in all the program stars. The fact that stars of low v sin i have little intrinsic polarization is evidence for the assumption that these stars are seen pole-on and that the envelopes around these stars are axi-symmetric. No apparent difference between pole-on stars and extreme Be stars was obtained.

The current state of understanding of line formation processes in expanding atmospheres is reviewed, and the successes and limitations of current computational techniques are summarized. Some results for differential rotation are also given, although very little work has been done in this area. Special attention is given to the severe difficulties that are encountered in inferring the structure of rapidly expanding or rotating atmospheres from observed line profiles because of the failure under these conditions of the Eddington-Barbier relation in integrated light; the value in this respect of continuum and interferometric observations is emphasized.

In order to understand the Be star phenomenon, and move in directions away from ad hoc models, detailed calculations have been made that treat nonlinear hydrodynamics self-consistently with non-L.T.E. radiative transfer.