The la-type supergiants of spectral type B to G are probably all variable in radial velocity and luminosity; in addition, on the average, the amplitude of the variations increases with increasing intrinsic luminosity and are larger for the earliest and latest spectral types (Abt, 1957; Maeder & Rufener, 1972; Burki et al., 1978; Grenon, 1993). The variations are not strictly periodic, but can be frequently described by a characteristic time Tc, i.e., some kind of ‘most probable periodș (e.g., Sterken, 1977; Rufener et al., 1978; Percy et al., 1979; van Genderen et al., 1992).

It is possible to define a Tc-luminosity-colour relation for the supergiant stars (Maeder & Rufener, 1972; Burki, 1978). However, this relation exhibits a large dispersion, due to the imprecise luminosity and colour determinations for these stars and, essentially, to the fact that each supergiant can show very different Tc values. Good determinations of >Tc require very long-term, continuous, photometric monitorings.

The wind-formed regular structures revealed in early supernova (SN) spectra can be a result of the radial pulsation of the precursor.

High-resolution time-series infrared spectra have been observed for a total of 20 Long Period Variables. The stars can be placed in four groups with distinct kinematic behaviour.

We describe a survey of suspected small-amplitude red variables, and long-term (up to 10 years) monitoring of about two dozen such variables.

We report on pulsational variability found in our decade-long study of UV and optical RV, light, K-line data sets for α Ori, α Sco A, and α Her A. As summarized by Smith et al. (1990), our radial velocities are derived from McMath observations of five metallic lines near Hα plus Hα itself. Optical light curves were obtained by small photometric telescopes, and K-line emissions from the Mt. Wilson K-line project. UV continua at 3000Å and k/h emission indices were obtained with the IUE LWP camera.

Light curve analysis of the carbon variable RY Dra and the oxygen-rich variables TX Dra and AF Cyg is summarised. In each case, two shorter periods were found (which may be connected with the radial pulsation of the stars) and also a long term variation in the mean brightness. In the case of RY Dra, the amplitudes of the shorter periods both decreased while the amplitude of the long secondary variation was found to be increasing. The light curves and frequency spectra of TX Dra and AF Cyg can be interpreted as a result of mode switching. These stars pulsate with two different periods and there are time intervals when the longer period disappears and the shorter one remains dominant, and it lasts for many cycles. We concluded that the double and multi-periodic red semiregular variables may be very important from both theoretical and observational point of view, because, based on the ratio of the periods, we can identify the pulsational modes in these variables.

Semiregular variables of type SRa and SRb are studied by fitting combinations of blackbodies to new near infrared and IRAS data.

The RV Tauri stars are semiregular pulsating variables located in the brightest part of the Cepheid II instability strip. They have a characteristic light curve of alternating deep and shallow minima. A subset of the RV Tauri stars (the RVb subclass) exhibit long-term (500 to 2600 day) light and radial velocity variations. Although it is well established that the short-term variations are due to pulsations, the long-term behaviour is not well understood.

BVRI photometry and high-resolution spectra of U Mon (the brightest member of the RVb subclass) were obtained at the Mt John University Observatory (MJUO) between 1990 Aug and 1994 May. The light and colour curves obtained clearly show the long-term variation in U Mon (Fig. 1(a) and (b)). The reddest colours occur slightly later than the long-term minimum in the light curve. The short-term light and colour variations are ‘damped’ at the long-term minimum.

Well-sampled light curves have been obtained for 1612-MHz hydroxyl masers in Mira and OH/IR stars monitored with the Hartebeesthoek radio telescope. Pulsation modes are investigated via the ratios of the observed periods in the maser variations.

We have surveyed the inner part of the Galaxy for OH 1612 MHz maser emission from OH/IR stars. The full survey is expected to give around 2000 detections. Here we report initial results for the Bulge region between |l| < 10° and |b| < 3°.

The observations discussed here were taken with the Australia Telescope Compact Array in Oct-Nov 1993. For each field we used a total integration time of 8 min. spread over an hour-angle range of 12 hours in intervals of an hour. Interference in the 1612 MHz band is a major concern in this program - it is ubiquitous for the shorter baselines. Another concern is the beam-pattern resulting from taking short cuts spread over 12 hours, with a one dimensional array. This causes sidelobe patterns which can easily lead to spurious detections. In addition to these problems, the size of the data set calls for special reduction techniques. We have developed a non-interactive method that is not I/O intensive and is well suited for use on supercomputers. To identify the OH/IR stars, all fields are ‘skimmed’ at successively weaker levels, subtracting point sources at the positions of the highest peaks in the fields, until we reach the (known) 3 σ level. Our method reduces the data reduction time by a factor of eight whilst giving comparable results to conventional cleaning methods. The disk space required is reduced by a factor of 100.

I consider maser emission in the H2O spectral line λ = 1.35 cm from long-period variable stars. Some models for H2O maser variability are suggested.

We have collected long data series of some SR and Mira stars from data banks. In many cases the light curves show strong fluctuation or variation in the period, phase and amplitude. The detection of the significant changes of these parameters is important from the point of view of pulsation theory and stellar evolution. We present here the analysis of V Boo. The used method is the Fourier and wavelet analysis in order to point out the modulations in the frequency and amplitude.

We investigate non-LTE effects in the hydrogen recombination zone of Long-Period Variables of Mira type. We propose that the Balmer hydrogen emission lines could be formed in this deep layer. We show how the spectral cyclic variations of Miras could be mimicked with this interpretation. As a result the radius of a Mira in the Balmer transitions could be much smaller than its radius in the molecular absorptions, which are formed in higher layers.

The American Association of Variable Stars Observers is the world’s largest organization of variable star observers. Its services to the astronomical community and database on pulsating variables are described.

In order to exploit the imminent HIPPARCOS data we have developed a new Maximum-Likelihood method to determine luminosity calibrations and kinematical properties with as bias-free as possible estimates. In most cases the samples are inhomogeneous, containing stars of several types. The method includes a way to separate the different groups. We applied the method to Mira variables contained in the HIPPARCOS Input Catalogue.

Some years ago, Willems & de Jong (1988) noticed that many carbon stars display an excess of emission at 60 µm and explained it by the presence of a fossil dust shell, containing only cold dust. This detached dust shell would be the result of an interruption of the mass loss, consequence of a thermal pulse. Detached shells around C stars have actually been mapped in the CO lines (Olofsson et al. 1992), and at 60µm (Waters et al. 1994). In 1992, Zijlstra et al. found about 100 M stars displaying an excess of emission at 60 µm, and proposed that interruptions of the mass loss due to thermal pulses is a general phenomenon on the AGB. This assumption is now supported by the theoretical calculations of Vassiliadis & Wood (1993). Here we present a detailed study of the 100 M stars of Zijlstra et al. in order to test the previous assumption.

We have carried out a search for obscured AGB stars in the LMC, observing fields in the vicinity of faint IRAS sources. About 25 heavily reddened AGB stars were found, with bolometric magnitudes close to the classical AGB limit of Mbol = −7.1. 10−µm. photometry for 5 sources show that these are oxygen-rich; they are similar to the Galactic OH/IR stars. Mass-loss rates vary between 5×10−4 and 2×10−6 M⊙ yr−1. We find no evidence that the mass-loss rates vary with luminosity. Neither do the mass-loss rates differ in a clear systematic way between the SMC and the LMC. The expansion velocities appear to be slightly lower in the LMC than in the Galaxy. All of the luminous stars for which periods are available have luminosities significantly above the period-luminosity relation derived for optical LMC Miras.

The evolution of low and intermediate mass (1-8 M⊙) stars along the Asymptotic Giant Branch (AGB) is ruled by processes of mass loss, causing the whole convective envelope to be gradually ejected into space. If the stellar mass is sufficiently high (M ≥ 1.5 M⊙) the envelope itself becomes enriched in nucleosynthesis products (carbon and s-process nuclei) and the star evolves into a C-rich phase. AGB stars are hence surrounded by O-rich or C-rich envelopes, opaque at optical wavelengths, which are best studied through direct imaging in the infrared (IR).

Interest in the hydrogen-deficient carbon (HdC) stars, and the R Coronae Borealis (RCB) stars in particular, has been rekindled by the realisation that these stars provide the potential for testing models in a number of areas. These include: pulsation theory; grain formation; grain evolution; stellar nucleosynthesis and extended stellar atmospheres.

Observational work on these objects has increased dramatically over the last 10 years. This has included both spectroscopic and photometric projects, see Lambert & Rao (1994) and Lawson et al. (1990).

We have continued to observe a number of the cool HdC stars in the Galaxy and the LMC and now have light curves exceeding 2000 days in duration for many of these objects. In particular, we are observing the RCB stars to investigate both the low amplitude, pulsation-related variations present at maximum light and the photometric behaviour during the large amplitude declines.

The observed correlation between pulsational phase in RCB stars and the timing of their declines shows empirically that the stellar pulsations and dust formation are intimately connected (e.g., Lawson et al. 1992). However, the nature of this relationship and the process of dust formation itself are not well understood. We have shown that it is likely that dust is forming in close proximity (< 2 stellar radii) to the RCB star photosphere, based on time scales for acceleration of the dust, eclipse of the chromospheric region, and dispersal of the dust (Clayton et al. 1992, 1993; Whitney, Balm & Clayton 1993; Whitney et al. 1993). The temperature at which amorphous carbon forms can be as high as 4000 K, and can occur in conditions far removed from thermodynamic equilibrium, as long as a mechanism exists to contain carbon atoms within a given volume. A likely form of carbon condensate is fullerenes such as C60 (Whitney, Balm & Clayton 1993). Shocks in the stellar atmosphere due to the pulsations may provide such a mechanism for containing the carbon. However, no spectroscopic signature related to the dust formation has been seen. We have searched unsuccessfully for IR emission bands of C60 (Clayton et al. 1995a).