For a long time there was not a clear interpretation of the fact that the emission observed at eclipse is systematically too high compared with predictions based on disk observations. This discrepancy has been attributed to the hydrogen density, the atomic abundances, or the multiplet excitation temperatures, the physical cause being assumed to be some inhomogeneities. We have shown (Liege 1978, under press) that it can be attributed to the microturbulence. The radiative energy absorbed in the doppler width of the absorption profiles of lines, produces high excitation temperatures. The relative abundances of Ba II, Sr II, Ti II with these excitation temperatures are in agreement with the photospheric abundances.

Assuming photospheric abundances, we get the hydrogen density versus the altitude while the observations give ne ne/Te3/2. The solution of these two equations, taking account of non-LTE for hydrogen but without the assumption of hydrostatic equilibrium, gives a chromospheric model. The model obtained is very near the usual quiet chromospheric models based on disk data.

This work confirms the validity of our assumption of the influence of microturbulence on the chromospheric emission observed at eclipse.

The velocity of microturbulence is frequently determined from Fe I lines. Unfortunately classical gf-values of this element have excitation dependent errors. In the absolute curve of growth analysis of some F-type stars with new gf-values, the author found that a great part of the depth dependence (i.e. excitation dependence) of microturbulent velocity, which had been derived by many authors with old gf-values, is the consequence of these errors.

The errors of old gf-values increased with excitation potential and the errors were compensated by adopting velocities of microturbulence decreasing with excitation potential. On the other hand, gf-values of ionized elements (e.g. Ti II and Fe II) are not changed as much as those of Fe I, accordingly the ionization dependence can also be eliminated.

Existence of discrete structures of velocity and magnetic field in stellar atmospheres follows from the Vlasov’s integral equation

where is a statistical function of the distribution of the elements of matter having masses and velocities, v, dT is elementary phase volume, F is functional which arrange the ties between macrovelocity, magnetic field, temperature, etc., providing self containment of cosmic plasma.

The circumstellar lines due to A lΠ - X1Σ transition of CO has been identified for the first time in the spectra of an early type star. The spectrum, taken from the International Ultraviolet Explorer between 1150 - 1900 Å, of 9 Cep, a B2Ib star with E(B-V) = 0.47, shows not only the interstellar absorption lines of CO but also absorption features shifted towards short wavelength relative to the interstellar lines. The amount of shifts towards short wavelength from the 1 - 0, 2 - 0, 3 - 0, 4 - 0, & 6 - 0 interstellar absorption bands of CO has been found to be fairly close and corresponds to a velocity of 450 km sec-1. All attempts to identify these set of lines in a consistent way with atomic or ionic lines have yielded essentially negative results. This has led us to tentatively identify these lines due to circumstellar ∞ around 9 Cep. An estimate of rate of mass loss from the circumstellar ∞-column density leads to a value of Ṁ ≥ 10-11 M⊙ yr-1.

The method of establishing of internally consistent abundance-oscillator strength scales by using solar fraunhofer lines is elaborated and investigated.

The error of internal accuracy should not exceed 0.05 - 0.06 dex. The absolute accuracy depends on the accuracy of “reference” gf-values.

The oscillator strengths for about 800 Fe I lines are obtained. The comparison of the results for 19 lines common in our and Blackwell et al. (1976) investigations gives the difference log gfblack - log gfauth = Δ = -0.044 ± 0.010. The accidental part of the difference actually determines the internal accuracy of the obtained oscillator strengths.

The Kurucz and Peytremann (1975) oscillator strengths for Fe I lines are analysed. Large systematic errors depending on gf and excitation potential are revealed and investigated. For some lines those errors may change the true values of gf by two orders of magnitude.

We extend to the lower main sequence stars the analysis of convection interacting with rotation in a compressible spherical shell, already applied to the solar case (Belvedere and Paterno, 1977; Belvedere et al. 1979a). We assume that the coupling constant ε between convection and rotation, does not depend on the spectral type. Therefore we take ε determined from the observed differential rotation of the Sun, and compute differential rotation and magnetic cycles for stars ranging from F5 to MO, namely for those stars which are supposed to possess surface convection zones (Belvedere et al. 1979b, c, d). The results show that the strength of differential rotation decreases from a maximum at F5 down to a minimum at G5 and then increases towards later spectral types. The computations of the magnetic cycles based on the αω-dynamo theory show that dynamo instability decreases from F5 to G5, and then increases towards the later spectral types reaching a maximum at MO. The period of the magnetic cycles increases from a few years at F5 to about 100 years at MO. Also the extension of the surface magnetic activity increases substantially towards the later spectral types. The results are discussed in the framework of Wilson’s (1978) observations.

Asymmetries in photospheric lines were described briefly in my earlier contribution to these proceedings and in more detail in the January 15, 1980 issue of the Ap. J. The cores of the stronger lines in α Boo (K2III) show a blue shift, quite opposite to the effect seen in the solar lines. Suspecting this difference to be due to the shift in ionization with effective temperature, two other stars were measured, one slightly hotter than α Boo (β Gem KOIII) and one slightly cooler (β UMi K4IV). The preliminary results seem to confirm the effective temperature dependence since no asymmetries were found for β Gem and stronger blue-shifted core asymmetries were found for β UMi.

A non-LTE computation of multiplet 38 of Ti I (11 lines) has been undertaken in order to determine small and large scale unresolved motions contributing to the Doppler broadening of solar lines at the centre of the disk (vertical motions) and at the edge of the disk (horizontal motions).

The abundance of Titanium and the total Doppler velocity (all scales) are determined by fitting observed and computed profiles of weak unsaturated lines (W < 12 mÅ) of the multiplet. Then saturated lines having the same lower level as the weak lines are computed for a variety of partitions of the total kinetic energy between the small scale and the large scale modes going from 0% small scale to 100% small scale. Oscillator strengths with internal accuracies of about 2% from Wahling (1977) have been used. The location of the observed profile (taken from Delbouille et al. for the centre of the disk and from Brault and Testerman, KPNO for μ 0.2) among the computed profiles yields the partition of the energy between the two modes.

The computations done so far with 4 levels and continuum give less than 20% of the energy in the small scale mode. Further computations with more levels are needed to establish this ratio with better accuracy. The total energy for vertical motions has a root square velocity of about 1.4 km/sec whereas the same quantity is 2.2 km/sec for horizontal motions.

Magnetic flux tubes are usually envisioned as small discrete structures sparsely distributed throughout an otherwise uniform “intertube” medium. An important distinction between the flux tubes and the surrounding atmosphere is the presence of a strong chromospheric temperature inversion at high pressures in the flux tubes, while the intertube component has only a mild, low pressure chromosphere, if any at all. This implies that the flux tubes will be enormously brighter in conventional chromospheric diagnostics than the intertube component. However, the unresolved magnetic elements cover perhaps only 10% of the “quiet” Sun at chromospheric heights. Consequently the intense K and k emission cores are severely diluted. The net result is a weak emission reversal that is not characteristic of either the flux tube or intertube chromosphere. Even in the thermal microwave continuum longward of 100 μm, the flux tubes can contribute significantly to low spatial resolution spectra. Consequently, the spatially averaged microwave emission is also not characteristic of either of the distinct components.

If magnetic flux tubes are indeed the dominant class of atmospheric inhomogeneity in the Sun and other cool stars, then single-component interpretations of spatially unresolved data can be completely misleading, especially for inferring important auxiliary quantities, for example chemical abundances, line broadening parameters and chromospheric energy budgets. In the latter case, chromospheric radiative cooling rates derived from empirical mean models could be overestimated by up to an order of magnitude.

Let me start by saying that we have heard some excellent review papers and I doubt very much that I can contribute much of importance in addition to what we have heard already.

1 also wish to apologize to all those who have made important contributions during this meeting and who will not be mentioned in this 30-minute summary. If I wanted to mention everybody I would just have enough time to read the program.