Recent observational and theoretical developments in the study of novae, particularly dwarf novae, are discussed. Mechanisms promoting mass transfer include (i) nuclear evolution or (ii) envelope instability of the red star and (iii) gravitational radiation of orbital angular momentum. Growing observational evidence against (ii) is supported by recent theoretical work on the medium and long term response of stellar radii to mass loss. Mechanisms (i) and (iii) may operate alone or in concert, depending on the circumstances.

The first half of the paper contains a discussion of the chemical composition of the envelopes of supernovae for the period close to light maximum. The principal conclusions are: The abundance of hydrogen in the envelopes of type I supernovae is low, much lower than that of nitrogen; the abundance of oxygen and carbon is also noticeably lower than that of nitrogen; it seems that there is plenty of helium and metals in these envelopes. The information for type II supernovae is more limited. But it is quite certain that the abundance of hydrogen in the envelopes of these stars is much higher than in the envelopes of type I supernovae.

In the second half of the paper the problem of supernova remnants is discussed, the circumstellar shells around supernovae (which according to S. van den Bergh and M. Peimbert are ejected from the star before its explosion) are also included. The discussion of this problem permits to confirm again the idea that there is a very close similarity between supernovae and novae. To be more exact there are reasons to suggest that supernovae as well as novae are double star systems, that they are relatively ‘old’, that they are quite peculiar objects and that they are not the final stage of evolution of ‘normal’ stars.

Tayler: I wish to provoke some discussion between observers and theoreticians. Theoreticians require supernovae to produce large amounts of heavy elements and that very large mass loss should occur. However, it is not clear that observers ever see a very high loss of mass or very large amounts of heavy elements. Is this a serious discrepancy?

The spectrum of the supergiant FG Sge has been studied from a series of high dispersion 120-in. coudé spectrograms obtained during the interval 1969–72, thus continuing the work of Herbig and Boyarchuk (1968). The star, of effective temperature about 6500 K in 1972, is cooling at the rate of 250 K yr-1; it is known to have ejected a still visible planetary nebula some 6000 years ago (Flannery and Herbig, 1973). Abnormally strong absorption lines of Y ii, Zr ii, Ce ii, La iiand other s-process species began to appear in the spectrum of the central star some time after 1967 and have progressively strengthened. Present abundances per gram of these elements are about 25 times the solar value. There is little doubt that the atmosphere of the star has been enriched in these elements during the past seven years, but the rate of enrichment now appears to be slowing down.