The Utrecht Orbiting Ultraviolet Stellar Spectrophotometer S59 on board the ESRO TD-1A satellite has observed ultraviolet spectra of about 200 stars in the wavelength regions 2060–2160 Å. 2490–2590 Å and 2770–2870 Å with a resolution of 1.8 Å (cf. de Jager et al., 1974). The spectra are analyzed in order to find UV criteria for stellar classification. Particular attention has been given to Fe III lines since no strong lines of this ion occur in the visible part of the spectrum. As one of the most striking results, it is found that the feature at 2078 Å, which is mainly due to Fe III, is very sensitive to luminosity.

Of the nearly 300 stars which have been observed in the ultraviolet with Copernicus, 60 have been chosen for publication of their complete intermediate-resolution spectra, which consist of scans made with photomultiplier U2 at a nominal resolution and step length of 0.2 Å. The spectra cover the wavelength range 1000 to 1450 Å and are expressed in the form of direct numerical tabulations, compressed-scale plots, and synthetic photographic spectrograms. These three modes of presentation are expected to satisfy the needs of various types of research on stellar spectra, ranging from detailed and rather specialized analyses of line profiles, which require accurate numerical results, to broad, comparative studies of various qualitative features over different spectral types, for which the plots or photographs are best suited. From the qualitative comparisons one might expect to synthesize more exacting criteria for spectral classifications of normal stars, using ultraviolet instead of visible spectra. The catalogue has been submitted for publication in the Astrophysical Journal Supplements.

The suitability of Copernicus U2 spectrometer scans (resolution 0.2 Å, spectral range roughly 1000–1450 Å) for purposes of spectral classification is discussed. The main conclusion is that while many features in the spectra complicate the situation (numerous interstellar features and those related to the phenomenon of mass loss), the general behavior of the lines in the spectra of most of the 22 stars studied is entirely consistent with the presently accepted classifications of these objects. Whether a significant increase in the accuracy of determination of all of the relevant classification parameters can be obtained from this material is not yet certain. It is emphasized that study of the high-resolution Copernicus scans should be of great value for forthcoming classification programs necessarily utilizing spectral data of considerably lower resolution.

Selected regions of the ultraviolet spectrum of Beta Lyrae were observed in 1973 and 1974 at the epochs of the two eclipses and of the two quadratures with the Copernicus Princeton University spectrometer. The results were published in the Astrophysical Journal (Hack et al. 1975, 1976). Here we summarize the results of a third series of observations with the same instrument covering the whole period of 12.93 days. The following spectral regions were observed in June 1975 during 13 consecutive days: λ 1036–1060; 1300–1326; 1398–1416; 2050–2098; 2580–2632; 2777–2812 in the low resolution mode; λ 1172–1177 and 2795–2799 in the high resolution mode. The main results are the following: 1) The radial velocity curve for λ 1175 C III is almost 180° out of phase with the orbital velocity curve of the primary star and the mean velocity is - 240 km s−1; K = 70 km s−1. This suggests that λ 1175 C III is formed in an expanding envelope associated with the secondary. The mass ratio can be determined: m2 /m1 = 2.7 −1.1/+0.2 .2) The continuum variation confirms the results obtained with 0A0-2. At λ 2100 we observe the light of a continuum indicating that the temperature of the body eclipsing the primary star is lower than that of the primary and of about 9000 K. The light curves at λ 1900, 2100 can be explained with the contribution of Fe III emission in the circumbinary plasma. On the contrary, the depth of the secondary minimum at λ 1311–1326 and at λ 1410 cannot be explained with the Si III and Si IV emissions. A possible explanation is the presence of a hot spot in the region of the eclipsing body (probably a disk surrounding the companion) where the stream from the primary impinges on the disk. The hot spot can also explain the behavior of the infrared light curves observed by Jameson and Longmore (1976).

The Mg II resonance lines at 2795.523 and 2802.698Å and their respective subordinate lines at 2797.989 and 2790.768Å are probably among the most prominent and interesting spectral features in the ultraviolet; they are perhaps the most significant in the mid-ultraviolet. They are also observable in one form or the other in stars of practically all spectral types. We shall discuss relatively high (about 0.4Å) resolution observations of these features.

In the early spectral types both the resonance and subordinate lines are seen primarily as absorption lines. An interstellar absorption of varying strengths is superimposed on the photospheric absorption of the resonance lines. The strengths of the photospheric resonance and subordinate lines increase from 0 to B, e.g., Lamers et al. (1973) and Kondo et al. (1975). The subordinate lines begin to merge with the resonance lines in late-B stars. In mid-A type stars, the resonance and subordinate absorption strengths become maximum. In F-type stars, the photospheric absorption strenghts continue to decrease. Chromospheric emissions become definitely detectable in F-type stars (Kondo et al. 1972). In a G2 V star, the sun, the chromospheric emission is fairly prominent at the core of relatively weak photospheric absorption. In K and M-type stars, this region is presumably dominated by the chromospheric emissions of Mg II resonance lines with the photospheric absorption becoming negligible in late-K stars; the only extant observation in this region is that of e Eri (K2 V) (McClintock et al. 1975).

We have had today for the first time a confrontation of the views of many astronomers about the use of ultraviolet spectroscopic and photometric data in the field of spectral classification.

Dr. Jaschek recalled us that much work has still to be done from the ground since only 14% of the stars up to the tenth visual magnitude have M.K. spectral types. The current accuracy is around one tenth of a spectral type and 0.6 of a luminosity class. Jaschek notes that the luminosity classification from spectral inspection is superior to what is obtained from photometric data, a fact which is not surprising if we recall that lines are much more sensitive to pressure effects than the continuum. However, the introduction of a third parameter seems necessary, although it is not clear what this parameter should be.

Heterogeneities in the solar atmosphere exist on many different length scales ranging from values as large as the solar radius (~106 km) down to features which are identifiable only by interferometry (~102 km). Rather than simply cataloguing the observed parameters of each and every known type of heterogeneity, I would like to concentrate on a few types of heterogeneities, with a view to identifying the information which is currently available concerning the physical mechanisms responsible for creating the in homogeneities. It is only if we can first identify the physics of each type of heterogeneity that we can hope to take even the first step towards predicting how each particular heterogeneity should scale to other stars. Since the present session is a joint discussion among mainly stellar astronomers, I feel that this approach is probably the most favorable method to present some of the large amount of information now available on solar features. Of course we expect that our solar information will be of most use to stellar astronomers in interpreting observations of stars which have similar spectral types to the sun. Nevertheless, we hope that nature will be kind enough to allow us to scale at least some of our information over a non-negligible area in the H.R. diagram.

The observations of spectrum-variability and light-variability of Ap stars are reviewed. It is shown that these variations are interpretable as due to the changing aspect of the spotted surface as the star rotates. It is stressed that we understand fairly well the geometry of the phenomenon but the physics is very far from being understood.

Magnetic Ap stars are probably those where the presence of a spotted surface is very evident. Their spectrum-variability (profiles, line-intensity and radial velocity), light-variability and magnetic field variability, all occurring with the same period, are explained in a simple way if we assume that these variations are due to the changing aspect of the spotted surface as the star rotates. The oblique rotator model was proposed by Babcock in 1949 and by Stibbs in 1950 and was worked out in great detail by Deutsch (1954). This model allows us to explain the magnetic field variation from some + 1000 to some - 1000 gauss in a few days; it explains the crossover effect, the line-width versus period relations, the line - intensity and radial velocity variation, and in part also the light curves. The main objection against the oblique rotator hypothesis was the supposed existence of many irregularly variable magnetic stars. However, the large number of observations accumulated in the last twenty years indicates that probably all magnetic Ap spectrum-variables are regular variables with periods which are generally of a few days, but includes a small group of long period variables (100 days up to 23 years for HD 9996). The light variability, which is the quantity measurable with the highest precision, has often remained undetected, because the amplitude is always small, in many cases few hundreths of magnitude.

The first detection of spots on dwarf M stars was by Kron on YY Geminorum in 1948. Since then phenomena attributable to starspots have been found by numerous authors beginning with Kraft and Krzeminski on the star now known as BY Draconis and on CC Eridani by Evans. There is a close correlation between spottedness, Balmer emission and flaring in dwarf M stars. Evans appears to have been the first to attempt to analyse the small range light variations on spotted stars in terms of modulation caused by axial rotation. In comparison with the solar case one must assume that the spots are very large and long lived. There is also an ambiguity in dealing with very large dark spots as between a star with a dark area or a cool star with a bright area (the zebra effect— is a zebra a black animal with white stripes or a white animal with black stripes?). It is difficult to arrive at a solution for the spot parameters which is unique. In particular several analyses which have appeared have assumed that maximum light corresponds to the immaculate star whereas in fact the maximum light in the variation corresponds to the immaculate star dimmed by the light of any constant star abstraction.

There is no generally accepted definition of AR Lac Stars, and the term RS CVn stars is used interchangeably or to refer to a particular subgroup. For the purposes of this discussion I use the term AR Lac stars to refer to detached close binaries showing Ca II emission in at least the cooler component outside eclipse, the hotter component being a main-sequence or subgiant star of spectral type F or G. Most of the systems show irregularities in their light curves as well as period changes. In order to determine whether a system is detached, one must know both the mass ratio and the relative radii. The determination of minimum masses is a fairly straightforward spectroscopic task, and provisional values are available for 22 of the systems, two or possibly three of them being non-eclipsing. All but 3 (AD Cap, RT Lac, RV Lib) have masses of the two components within 30% of each other. Because of appreciable irregularities in the light curves, the radii are subject to considerable uncertainty even when photometry of good precision is available. Nevertheless the 9 systems with very provisional radii all appear to be detached. These all have mass ratios near unity. We may assume, as a working hypothesis, that the other systems with mass ratios near unity are also detached and hence also belong in the AR Lac group. Most of the data referred to are to be found in IBVS 1083.

Resolved images of the disks of the largest stars observed with the largest telescopes can be constructed using the class of techniques called speckle imaging. The observations must be made with narrow passbands (~ 10 nm), short exposures (~ 20 ms) compensation for atmospheric dispersion, high magnification and good signal-to-noise ratio. One specific technique applied to a Ori (Lynds et al., 1976) shows slight but apparently real differences in the images of the disk corresponding to low and high opacity in the stellar atmosphere which we interpret as due to temperature differences. There are also significant differences in the star’s diameter and/or limb darkening at the two different opacity wavelengths.

There are several observations that suggest the existence of surface heterogeneities on the T Tau-type stars. Firstly Dr. Ismailov from Shemakha has found strong variations of emission line profiles for 5-10 minutes: 2-component profiles become 3 or 1 component and vice-versa. The results were confirmed recently in the Sternberg Institute. Secondly, variations of polarization degree and relative intensities in the IR calcium triplet were found when the stellar brightness were being constant. Then, the relative intensities of the Ca triplet are not equal to what must be expected for both optically thin or thick but homogeneous emission regions. On the basis of the data Petrov and Sheberbakov from Crimean Observatory have proposed that T Tau stars have active regions with strong magnetic fields similar to stellar ones: they can give noticeable heterogeneities for calcium emission, for polarization effects in dust envelopes and local processes for profile variations. The existence of rather strong and local magnetic fields on T Tauri stars may give a key for explaining the mysteries of these variables. Small variations of strong magnetic fields on stars with a convective zone can give a new mechanism for stellar variability. As it is known, only a quarter of the energy flux that reaches the bottom of the convective zone, is emitted from a sunspot in form of light and the main part of the energy escapes in form of hydromagnetic waves. Then, the sunspot exists if the field is not less than 1 kilogauss. Therefore the field acts as a relay: if the field is low we have light, if the field is strong we have mainly hydromagnetic waves. The transition from low to high field may occupy only a small part of the whole field range. The field variations on T Tau stars can give strong brightness variations without additional energy sources but only by magnetic relay operation. I hope that this scheme gives a possibility to explain a set of features of T Tau stars (such as their high ratio of chromospheric to photospheric radiation, IR excesses, the nature of their flares) as well as main features of FU Ori-type flares.