It is fitting that, in opening this Symposium with a consideration of the T Tauri stars, we note that this month marks the tenth anniversary of the September 1945 issue of the Astrophysical Journal, in which appeared the remarkable pioneering paper by Alfred H. Joy that initiated the study of emission-line stars associated with nebulosity. This contribution opened a new approach to the study of the relationship of stars to their environment. Today, ten years later, the T Tauri stars and their interaction with nebular material form a topic whose significance we may not fully appreciate and whose opportunities have been as yet only superficially exploited.

The RW Aurigae type was first introduced by P. P. Parenago [1] in 1932 for the designation of irregular variable stars that belong to spectral class G and show extremely rapid (up to 1 magnitude in several hours) and large (up to 3 or 4 magnitudes) variations of their brightness. Many stars of that type were found to be connected with dark diffuse nebulae. Some of them, as for instance R Monocerotis, RY Tauri, and R Coronae Australis, were also associated with bright and dark comet- or fan-shaped nebulae, often themselves variable.

First of all it will be necessary to say a few words on terminology. Different names are now in use either for the same group of variables or for different variants of a rather heterogeneous class. Let me list the following:

• RW Aurigae stars,

• RR Tauri stars,

• T Tauri stars,

• Nebular type variables,

• Orion type variables,

• Nova-like variables,

• Main sequence variables.

As is well known, very rapid and non-periodic changes in brightness have been discovered in several late-type dwarf stars in recent years. In the vicinity of the sun, within a radius not exceeding 10 parsecs, nine or ten such objects have been found and named ‘flare’ stars because of their extraordinarily rapid variations. The prototype of these flare stars is UV Ceti. For the purpose of the present discussion, we shall call these objects ‘classical’ flare stars.

Observational evidence of transient flares which produce conspicuous changes in the total brightness and in the spectra of dwarf stars is limited, at present, to a small number of faint late-type stars. The outbursts take place with extraordinary rapidity and the duration is usually only a few minutes or hours. Accompanying spectral changes such as veiled and fuzzy absorption lines, greatly increased intensity of the bright lines of hydrogen and helium, and the appearance of an emission continuum shortward of λ 3750, have been reported.

Dr A. H. Joy has shown that the star AE Aquarii is a spectroscopic binary. Seventy hours of photo-electric observations have revealed light variations accompanied by frequent explosions.

A programme of spectroscopic study of sub-luminous stars has been in progress at Palomar, using the coudé spectrograph. Various types of objects have been studied and some preliminary results are available. Certain other programmes will, when completed, be relevant to the subject-matter of this Symposium. A survey of possible velocity variations in sub-dwarfs of spectral types A to G is in progress. So far I can report that no emission lines have been found, and that there are very few spectroscopic binaries in this group. In another programme, the depth of the late-type absorption lines in the composite spectra of SS Cygni stars will be used to estimate the luminosities of the hot components near their minima. Spectra of old novae are also being obtained with a dispersion of 38 å/mm. Plates obtained thus far show that complex structure still exists in the emission bands.

It has been recognized in recent years that extremely rapid light-variations occur in certain stars. Flares lasting only a few minutes have been observed in a number of dMe stars[1]. Flaring has also been observed in the W Ursae Majoris star U Pegasi[2]. Rapid variations in light have been found in some T Tauri stars[3], in the U Geminorum variable AE Aquarii[4,5], and in the short-period eclipsing binary UX Ursae Majoris[6, 7, 8]. Observations during the past two years at the Mount Wilson Observatory have shown that extremely rapid and apparently continuous variations in light occur also in many of the old novae and in a number of related stars.

The present communication is devoted to some important problems connected with the interpretation of spectroscopic phenomena observed during the outbursts of novae.

In order to discover the cause of the outbursts of novae, a study of the physical processes taking place in novae during such explosions is not sufficient. A number of other, and no less important, properties and peculiarities of novae must also be investigated. Such properties are, for instance: the absolute magnitudes at light maximum and minimum, the temperatures, the light curves and their variations, the position of the novae on the Hertzsprung-Russell diagram, their spatial distribution in the Galaxy, and the mutual relations between the novae and other objects of similar nature (such as the recurrent novae, super-novae, nova-like stars, and the nuclei of planetary nebulae). A hypothesis of the origin of the outbursts of novae should be founded upon all these data as well as the other properties of novae.

The phenomena taking place as a result of a nova outburst indicate that a shock wave is propagated to the surface of the star. It may be the result of an explosion in the interior of the star. The hypothesis that peripheral thermo-nuclear explosions cause the outbursts of novae, and perhaps also those of nova-like stars, was advanced in 1947[1]. By ‘peripheral’ we mean some spherical layer which we shall name the A layer.

The physical state of the planetary nebulae has been investigated sufficiently well, but the main question of their origin and evolution still remains open.

In the period following the papers on instability among the hot stars of low luminosity, Dr L. Gratton presented a brief account of work in progress on η Carinae by Dr Platseck, Miss Ringuelet, and himself at Córdoba and La Plata. The bright line spectrum of η Carinae is being studied in detail; one interesting observation is that the emission lines of Ti 11 are quite weak as compared to their intensities in novae of comparable excitation. The Balmer lines are also faint. The hydrogen emission lines consist of fairly wide structures upon which are superimposed narrow emission cores, as well as absorption features. The La Plata workers find that the narrow emission lines originate in the nuclear star, while the bright lines that arise from the nebulosity in the immediate vicinity are very wide, their breadths corresponding to a velocity range of 500 to 1000 km./sec. The radial velocities vary greatly from one part of the nebula to another. Gratton and his colleagues believe that η Carinae is actually a member of the Carina O-association upon which it is seen superimposed. The corrected distance modulus of this complex is 12.3 magnitudes, with an uncertainty of less than half a magnitude. It follows that η Carinae at its maximum in 1843 had an absolute magnitude almost certainly exceeding −13, and even now is a fairly luminous object of M = −5 or −6. The La Plata astronomers believe it to be improbable, therefore, that η Carinae is a nova-like variable or a super-nova, but think rather that the object is to be regarded as similar to S Doradus or Hubble's variables in extragalactic nebulae. They comment on the probable membership of both S Doradus and η Carinae in O-associations, and suspect that this connexion may be of great evolutionary significance.

In 1945 we showed[1] from a study of six Wolf-Rayet stars that their spectrophotometric temperatures, when allowance is made for interstellar reddening, are still far below the temperatures as determined by the Zanstra method. It was demonstrated that these spectrophotometric temperatures (which are only slightly in excess of those of Ao stars) agree perfectly with N. Kosyrev's[2] theory of extended photospheres. Later Petrie[3] confirmed our observational results, though he did not compare the two sets of results for individual objects. This same result was recently obtained by Andrillat[4].

This paper consists of two parts which, at first sight, are quite detached from each other: the pulsating variables of spectral types B to F, and the close spectroscopic binaries of the types of β Lyrae, UX Monocerotis and U Cephei. I shall show that there exists a connexion between these groups which, though still rather nebulous (in both senses of this word), promises to yield interesting results.

In the discussion of the papers of Section III, Dr L. Gratton stated that in the case of β Lyrae and other close binaries, one may expect that the radii of the components are considerably larger than those of single stars of the same mass. This would cause the central temperatures to be lower and, in turn, the luminosities to be smaller. It might even turn out, according to Gratton, that in close binaries the generation of energy might depend on the p–p reaction rather than on the C–N cycle, notwithstanding the large masses of such stars.

The possibility of discovering phenomenologically similar objects either located in different stellar systems, or in totally different (according to their origin and age) parts of some complex stellar system (as, for instance, our Galaxy) is of extreme importance. The detection of such objects permits us to confirm that, in spite of different initial conditions and evolutionary paths, stars of quite different origin pass in the course of their evolution through the same stages.

In modern astronomy we acquire fundamental knowledge concerning the dimensions and the constitution of the exterior layers of giant stars by the study of eclipsing systems. The red giants which form part of the systems of VV Cephei and ζ Aurigae have been studied thoroughly from the changes that the spectra of these systems undergo when the light of the bright eclipsed star passes through the envelope of the rarefied giant. At the present time it is clear that stars having extended envelopes are not rare. Thus, for example, F. I. Loukatskaya[1] (at the Principal Astronomical Observatory of the Academy of Sciences of the Ukrainian S.S.R.) has studied the eclipsing star AW Pegasi in different parts of the spectrum; she has shown that this star possesses a semi-transparent envelope that produces a considerable part of the eclipse. It turns out besides that certain spectral lines, observed in the spectrum of the bright star, come from the semi-transparent envelope of the less luminous companion.

I should like to open my subject by attempting to answer the following question: how many parameters are necessary and sufficient for a complete specification of the form of both components in close binary systems? Their shape should, in principle, be specified by the nature of forces acting on their surfaces, and (provided that the free period non-radial oscillations of both components are short in comparison with that of their orbits) their distortion should be governed by the equilibrium theory of tides. The level surfaces of constant density then coincide with those of constant potential, and the boundary of zero density becomes a particular case of such equipotentials.