We have obtained extensive spectroscopic observations of Nova Cen 1986 at different phases of the event.

From the observed Hα/Hβ line ratio we derived a colour excess EB-V = 0m.54. Using the empirical relationship between the fading time and the absolute magnitude at maximum light as well as the width of the Nal interstellar lines, we estimated for the nova a distance of 1.2 ± 0.2 kpc. The full-width of non-blended lines at the nebular phase indicates an expansion velocity of 1400 ± 120 km s−1

The temporal behaviour of the Hβ emission suggests that the electron density varies with time as Ne ∝ t−3, i.e., the shell evolves more or less homologously. The electron temperature was estimated from the [NII] and [OIII] line ratios. In the O+2 region the electron temperature is in the range 14100 ± 1200 K. The N+ region has a lower temperature, T ≃ 8000 K around day 142 after maximum light. The N+ temperature increases in this region as the shell evolves and attains the same values observed in the O+2 region around days 300-470.

Spectroscopy of the nebular (as opposed to the stellar) remnants of old novae can be a powerful diagnostic of physical conditions in, and mode of excitation of, nova shells. While optical spectroscopy of nova shells has been widely used in this way, few are sufficiently well resolved, and sufficiently bright, to be observed with IUE.

Combined optical and ultraviolet spectroscopy of only two nova remnants–those of GK Per and DQ Her–have been reported to date. Ferland et al. (1984) demonstrated that the shell of DQ Her is extremely cool (Te ≃ 500 K) and is excited by a hot central source. Bode et al. (1988) showed that the ultraviolet and optical lines in the spectrum of the GK Per nebula can best be understood in terms of shock excitation, consistent with its observed properties at radio wavelengths (Seaquist et al. 1989).

We present CCD spectra of novae PW Vul, QS And, Herculis 87 and V1819 Cyg obtained in the wavelength range 680 to 1080 nm at the Cassegrain focus of the 193–cm telescope at the Observatoire de Haute–Provence (CNRS, France). The reciprocal dispersion is 23 nm/mm.

Novae ejecta pass through four distinct phases of evolution of the emission-line spectrum, caused by different ionization characteristics of the shell. These include a neutral (I), an auroral (II), a coronal (III), and a nebular (IV) phase. Photoionization from the contracting photosphere of the hot white dwarf is the source of the ionization, including the highly ionized coronal phase. Changing emission line ratios in certain novae that develop dust are caused by condensation of grains from the gas, and can be used to determine the composition of the dust. In V1370 Aql, substantial silicate grain formation appears to have taken place, probably within the ionized gas.

The disks referred to in the title of this paper are specifically those present in cataclysmic variables in which the accreting white dwarf has a relatively weak magnetic field (≪ 1 MG). Such systems are classified either as nova-like variables or as dwarf novae, and are of interest here because they are believed to be novae in quiescence (Ritter and Livio discuss this point elsewhere in this volume).

This review aims to do two things: i) to summarise what has been learned about the winds associated with non-magnetic cataclysmic variables both from observation and from numerical modelling, and ii) to outline ideas about the nature of the mass loss mechanism. By contrast with the certainty that nova outflows are the consequence of thermonuclear runaway, it shall be seen that the fundamental cause of mass loss from cataclysmic variables remains obscure. An earlier review of this subject is by Cordova and Howarth (1986). Also of interest are some sections of the monograph on dwarf novae and nova-like variables by LaDous (1989).

During nova outbursts a large part of the envelope mass is blown off. This mass ejection is caused mainly by the radiation driven wind and by the Roche lobe overflow. Theoretical studies on the acceleration mechanisms are briefly reviewed. Finally, I present theoretical light curves for U Sco based on the steady wind approach.

Evidence is given that classical novae have optically thick winds for a certain time after optical maximum. Multifrequency observations especially of FH Ser indicate acceleration of the wind at very large optical depths, probably by radiation pressure of an object above the Eddington limit. This result though fairly convincing needs checking by proper radiative transfer calculations. A way in which it might be produced is mentioned as is the influence of the wind on the structure of the ejected envelope.

The formation and evolution of dust grains following a nova outburst are reviewed, and various observational methods of constraining the nature of nova dust are described. Despite recent advances in understanding grain formation and evolution in classical nova systems, there remains a number of outstanding problems, the solutions to which are likely to come from studies of other aspects of the nova outburst.

Continuum and line blanketing models for the photospheres of novae in the early stages of their outbursts are presented. The expanding envelopes are characterized by a very slow increase of density with decreasing radius which leads to very large geometrical extensions and large temperature differences between the inner and outer parts. The spectra show a large IR excess and a small Balmer jump which may be either in absorption or in emission. For the parameters considered (Teff = 104, 1.5 × 104, 2 × 104K, R = 1011 cm, solar composition), most lines are in absorption. The effects of both modifications in the temperature structure (e.g. by heating from shock fronts) and changes in the abundances of the heavy elements on the emergent spectra are briefly discussed.

The exact value of the accretion rate in novae is very important to the theory of thermonuclear runaways and nova statistics. Estimates of the typical accretion rates in nova and novalike objects are 1018g/s (1.6 10−8 M⊙/yr). Patterson (1984) estimated that if the above high accretion rate prevails the expected number of novae according to Bath and Shaviv (1976) exceeds the observed number. Prialnik, Livio, Shaviv and Kovetz (1982) have shown that no thermonuclear runaway can be obtained if the accretion rate is as high as few 10−9 M⊙/yr. Actually, the very strong thermonuclear runaways (TNRs) are obtained only for very low accretion rates namely, 10−10 to 10−11 M⊙/yr. A way out from this inconsistency was suggested by Shara, Livio, Moffat and Orio (1986) in the form of the hibernation model. In this model the high accretion rate is found only shortly before and soon after the nuclear runaway and the nova spends most of the time in a state of very low accretion rate - so low that most novae are not observed at all. The high accretion rate, according to this theory, is a precursor to the runaway. The hibernation model is very attractive in spite of a meager observational proof so far and theoretical justifications. Yet, it is one of the most promising ideas in this field. The hibernation scenario claims that the accretion rate is a function of time. It is high just prior to the runaway and it continues to be high for some time just after the runaway. It then declines to very low values. From the point of the model for the thermonuclear runaway the problem is how the star reacts to a temporary high accretion rate that comes after a long period of low accretion rate (Shaviv and Starrfield, 1987). Thus, even if the accretion rate is low during most of the time the phase of high accretion can still affect the outcome of the thermonuclear runaway.

During the evolution of a classical nova outburst, there are four times at which a significant infra-red (IR) flux is expected. The first Isothermal Dust Phase has been observed in a number of novae and analyses of this phase give valuable information on the properties of the emitting dust. In order of time from visual maximum, the three later phases are due to IR emission from X-ray heated grains, fine-structure line emission from a cold phase of the nova nebula, and grain cooling from shocked gas at the interface between the nova ejecta and the interstellar medium. In the next sections, theoretical estimates of the peak IR luminosity, flux distribution and time of peak emission for each of these phases are discussed.

Some post-novae, e.g. DQ Her, CP Pup, T Aur, have cool envelopes, Te < 1000K. As has been shown by Ferland et al. (1984) this effect is primarily due to large overabundances of the CNO elements which at low densities can very effectively cool off the nebular gas through infra-red lines.

We have constructed an extensive grid of photoionization models in order to study the exact conditions required for the electron temperature to drop to such low values. As can be seen in Figs. 1 to 4, our results can be summarized as follows.

1. 1- The abundance of oxygen which is the dominant coolant in the conditions considered has to be at least 10 times the solar value (with C and N accordingly overabundant).

2. 2- The gas density, n, has to be below 1000cm−3.

3. 3- The effective temperature of the ionizing source, Teff, has to be below 2 105K.

4. 4- The ionization parameter U = QH / (4πr2nc) (QH: number of ionizing photons emitted per second by the central source, r: radius of the envelope) has to be between a certain range, typically 10−2 < U < 10−3, whose exact limits, however, depend on other parameters such as Teff and the abundances.

The chemistry of the formation of nucleation sites in the ejecta of novae has been investigated. Previous authors (e.g. Clayton and Wickramasinghe 1976) have noted that grain growth is rapid once nucleation sites have formed but the assumption is generally made that the ejecta is in LTE during the nucleation phase. This is obviously not the case as the ejecta is illuminated by an intense radiation field so that the lifetimes of the larger species are of the order of one second, even the ejecta kinetic temperature is only 1000-2000 K. We propose that nucleation sites form by heterogeneous chemical reactions in a highly non-equilibrium chemistry.

Preliminary results of 1– and 2– dimensional hydrodynamical calculations of the common envelope phase in very slow classical novae are presented. We show that frictional deposition of orbital energy and angular momentum into the envelope can potentially induce mass loss. Specifically, we find that despite rapid initial spin–up of the envelope, ejection of mass in the orbital plane continues at a substantial rate.

It is well known from observations that some novae show infrared continuum excesses indicating dust heating by the strong ultraviolet radiation field. It seems, however, that dust formation in nova shells depends on some yet unknown intrinsic parameters since the characteristic infrared excesses are not observed in all objects. To study the conditions under which dust nucleation may occur in nova shells, as a first step we investigate the time dependend ionization and temperature structure.

Non–LTE effect and formation of the Balmer line Hα in the optically thick nova wind expected from the radiation–pressure–driven continuous ejection model (cf. Bath and Shaviv, 1976; Ruggles and Bath, 1979) was investigated by use of the pure–hydrogen spherical model atmosphere, in which constant radial expansion velocity, inverse square law of density distribution, and planar grey temperature distribution were assumed. The coupled equations of radiative transfer and statistical equilibrium on the four levels plus continuum model atom (in which Hα line and Balmer–, Paschen–, and Brackett– continua are explicitly treated) were solved by the complete linearization procedure based on the comoving frame method (cf. Mihalas and Kunasz, 1978). We calculated four models as results of the combination of the surface density ρsurf (at ᵀRos = 10−13) (10−12 g/cm3, 10−13 g/cm3) and the expansion velocity Vexp (0 km/sec, 1000 km/sec). We assumed 1012 cm for the photospheric radius rph (at ᵀRos = 1), 10000 K for the effective temperature appearing in the temperature formula, and 100 km/sec for the microturbulent parameter. The low density models have larger extension (rsurf/rph ≃4.1) than the high density ones (rsurf/rph ≃1.7).

The HeI triplet to singlet line ratio Λ = F(λ5876)/F(λ6678) in symbiotic stars and novae is often much lower than the planetary nebulae value of ≈ 3.5, which is also the low density value Λ0 predicted by recombination theory.

Fig.1 shows that the dereddened HeI line ratios ^ are different for S type and D type symbiotic stars. The spectrophotometric data are mainly from Allen (1984, personal communication) and Blair et al.(1983). The S and D type classification divides the symbiotic stars according to the IR continuum into objects with a stellar continuum or dust emission. Practically all D types have low nebular densities Ne ≲ 107cm−3 and a rich spectrum of forbidden lines, while in S type objects these lines are suppressed by collisional deexcitation due to the high electron densities Ne ≳ 107cm−3. The observed HeI line ratio ^ in S type symbiotics, which is systematically too low when compared with the theoretical low density value Λ0 = 3.5, can therefore be attributed to the high nebular densities in these objects.

Theoretical modeling of novae in outburst predicts that they should be active emitters of radiation at soft X-ray wavelengths twice during their outburst. The first time occurs very early in the outburst when only a very sensitive all sky survey will be able to detect them. This period lasts only a few hours for the very fastest novae. They again become bright in X-rays late in the outburst when the remnant object becomes very hot and is still luminous. Both simulations and observations show that novae can remain very hot for months to years. It is important to observe them at these late times because a measurement both of the flux and temperature can provide information about the mass of the white dwarf, the turn-off time scale, and the energy budget of the outburst.

In this paper we explore to what extent the TNR model of nova outbursts and our current concepts of the formation and secular evolution of cataclysmic binaries are compatible. Specifically we address the following questions: 1) whether observational selection can explain the high white dwarf masses attributed to novae, 2) whether novae on white dwarfs in the mass range 0.6M⊙ ≲ M ≲ 0.9M⊙ can occur and how much they could contribute to the observed nova frequency, and 3) whether the high mass transfer rates imposed on the white dwarf in systems above the period gap can be accommodated by the TNR model of nova outbursts.

The cataclysmic binaries are products of the nonconservative evolution of close binaries with large initial mass ratios of components. The accretors in cataclysmic binaries can be helium or carbon-oxygen or oxygen-neon-magnesium white dwarfs. Their annual birthrates are ~0.005, ~0.005, and ~0.00005 respectively. In one-zone approximation of a thin accreting shell we estimate the critical masses of hydrogen and helium shells and recurrence time scales of thermonuclear runaways.