Very recently the MACHO Collaboration published detailed data for 45 double-mode Cepheids in the Large Magellanic Cloud. Here we compare these data with new stellar pulsation models. We conclude that the new period ratios are in good agreement with a metal content Z = 0.005 – 0.01 in the Large Magellanic Cloud.

Compilation of the list of Cepheids which are members in binary and multiple systems is in progress. During this project SB-nature of a number of Cepheids was revealed. The frequency of binaries among Cepheids exceeds 60 per cent.

Cepheid PLC relations are usually constructed using mean light. We show that the use of maximum and mean light leads to a distance estimator for the LMC with 30% less scatter than existing relations.

We have undertaken a programme to calibrate the Cepheid PL relation zero-point by obtaining distances of Cepheids in open clusters and associations via the visual surface brightness technique. Results are now available for four stars (SZ Tau, CF Cas, CV Mon and DL Cas) and others are currently under analysis. Preliminary results suggest the ‘ZAMS-fitting’ distances to the host clusters are systematically smaller than those we derive from Cepheid surface brightnesses.

We have examined the uncertainties involved in using the visual surface brightness technique on Galactic Cepheid variables. The random error in a single Cepheid distance measurement is well determined to be ±8%. An upper limit to the systematic uncertainty is shown to be ±6% in distance. These combine for a single Cepheid to a typical uncertainty of ±10% and, for samples larger than ten Cepheids, to a typical uncertainty of less than ±6% in distance.

We compare the best determined Baade-Wesselink (BW) period-luminosity (PL) relation for 100 galactic Cepheids to the PL relation derived for 32 open cluster and association Cepheids by the ZAMS-fitting method. Eighteen stars in common lead to the conclusion that BW and ZAMS-fitting distances can agree to better than 0.1 mag, after proper allowance of systematic effects in both methods.

There is often some dispute in the literature as to whether a particular Cepheid (S Vul, for example) belongs to Type I or Type II. JHK photometry has been used to calculate Baade-Wesselink radii for a number of stars always, sometimes or occasionally stated to be Type II Cepheids. Radial velocities have been taken from the literature.

The (K, J – K) radius of SW Tau is 8.30 ± 0.34. This is about half of the expected radius for a Type I Cepheid at this period and normal for Type II.

For V553 Cen, the relative phasing of radial velocities and infrared photometry has to be done by choosing a relative phasing where the phase shift between the radius displacement curve integrated from the radial velocities and the photometric radius displacements is minimized, and where the surface brightness coefficient A (Balona 1977) is normal for a star with the colour indices of V553 Cen. Uncertainty in the reddening has a negligible effect on the radius solution obtained in this way. I have also considered the degree to which the K, J – K and K, V – K solutions give similar answers, as the relative phasing of the optical photometry and the radial velocities was more straightforward. The (K, J 73x2013; K) radius is 11.2 ± 0.4, normal for a Type II Cepheid.

New physical parameters of selected Cepheid variables classified as CW or CEP (Type II) in GCVS are presented. The analysis is based on published light and radial velocity curves. Two of the stars turned out to be classical Cepheids.

From radial velocity measurements obtained with a cross-correlation technique, the variation of turbulence during the pulsation cycle is studied for a sample of 40 Cepheids. We will propose a new way to separate classical and s-Cepheids. More complete results will appear in a forthcoming paper (Bersier & Burki 1995)

The radial velocities have been measured with the spectrometer CORAVEL (Baranne et al. 1979), whose cross-correlation function (CCF) is fitted with a Gaussian, giving the radial velocity Vr, the width σobs the depth H and the continuum, normalised to 1. The pulsation broadens the lines and thus also the CCF. In the Gaussian approximation one can write

where σobs is the observed width, σinst is the instrumental width, σpuls is the additional width caused by the pulsational velocity field and σres contains all the other effects (turbulence, rotation, magnetic field, etc.). To be less affected by the noise in the data, a Fourier series has been fitted to each curve of σobs. With numerical simulations, one is able to synthesise the additional Doppler width due to pulsation, with a high accuracy. The instrumental width being well known for CORAVEL, the computation of σres is then straightforward. One then has a curve in phase for σres. From this curve, we determined the maximum residual broadening σmax (observed at or very close to minimum radius), and the width σo that the star would have if it did not pulsate. As shown by Bersier & Burki (1995), σo is slightly higher than the mean value of σres.

There has been a number of observational programmes that have endeavoured to investigate the atmospheric velocity fields in Cepheids (e.g., Sanford 1956, Wallerstein et al. 1992, Butler 1993). These studies measured the radial velocities of lines of different strength, excitation and ionisation potential as these provide an indication of line formation at different levels in the atmosphere. From these measurements, the presence of velocity gradients can be inferred, but determination of the magnitude of such gradients requires knowledge of the spectral line depth of formation. Through dynamical modelling we are endeavouring to ascertain what is actually being measured in the above observational programmes.

The problem of deriving the centre-of-mass velocity of a radially pulsating star is reexamined. New observations of line asymmetry and Non-LTE radiation hydrodynamics point to a systematic effect of about 1 km s−1 in Cepheids.

In the past years, in a series of papers (Antonello & Poretti 1986, Antonello et al. 1990, Mantegazza & Poretti 1992, Poretti 1994) the Fourier decomposition was successfully applied to the light curves of galactic Cepheids with P < 8 d. The separation of these Cepheids into two groups was evident. The first, more numerous (195 stars after the latest inclusion) group is constituted by the classical Cepheids, i.e., the Cepheids which follow the Hertzsprung progression: they occupy a narrow region in the ϕ21 – P plane and are also characterised by a R21 ratio larger than 0.20. The second group, containing 30 objects, is constituted by Cepheids which deviate from the Hertzsprung progression, describing a ‘Z’ crossing the classical sequence in the ϕ21 – P plane; their R21 values are smaller than 0.20. The relation between stars of the upper and lower sequence of the ‘Z’ is confirmed by the ϕ31 – P plane, where these stars describe an unique sequence. As first suggested by Antonello & Poretti (1986), the splitting of Cepheids into two groups can be explained by a different pulsation mode: while the classical Cepheids are pulsating in the fundamental radial mode, the other Cepheids are pulsating in the first overtone radial mode.

We present a progress report on a project to investigate the differences in absolute physical parameters for a sample of five equal period SMC Cepheids from a Baade-Wesselink analysis of the stars.

We have determined new Cepheid distances on a homogeneous system to nine nearby galaxies with Cepheid observations in the V bandpass, using the extensive calibration of the galactic Cepheid PL(V) relation of Gieren, Barnes & Moffett (1993). Absorption corrections to the observed distance moduli were taken from the RC3 catalog (de Vaucouleurs et al. 1991). For most of the galaxies, the galactic PL(V) relation of GBM gives a very good fit to the Cepheid data.

V and I band CCD photometry of the young clusters NGC 330 (SMC), NGC 1850, NGC 2058 and NGC 2065 (LMC) has been used to find variable stars in and around the clusters. The variables detected are mostly Cepheids and long-period variables (LPVs). Here we outline some results of these studies.

[Fe/H] values of cluster RRab variables were calculated from the Fourier parameters of their light curves using a recently published method. The average cluster metallicities obtained this way agree with direct observations within the error range.

Period behaviour of 62 RR Lyrae stars in the M15 globular cluster has been investigated. About one half of the sample (30 stars) exhibited linear period change. The remaining 32 variables can be characterized by abrupt or erratic changes in their periods.

The distribution of the RR Lyrae stars in the globular cluster ω Centauri is studied, using the data published by Martin in 1938. Comparing the accumulated numbers of variables within the angular distance, r, from the centre of the cluster, it is shown that the known RRc variables seem to be less concentrated towards the centre than the RRab stars. 6-10 RRc stars are missing within r = 2–3 arcmin, if the two distributions are actually identical, as expected from evolution models. We conclude that the central region may contain several RRc stars that have not yet been discovered. This could be due to their relatively low amplitude, since it is very difficult to find low-amplitude variables by means of photographic photometry in the crowded, central region of ω Cen with large and varying corrections due to unresolved background light.

B and V CCD frames of the globular cluster NGC 7006 have been obtained to study the period change rates and determine photometric colours for the RR Lyr variables.

We compare observed fluxes from the ultraviolet (IUE) through J and K with recent Kurucz model atmospheres to determine a temperature for the F5 lb supergiant α Per. The two most important advances in this study as compared with previous work are the use of well calibrated ultraviolet fluxes and the use of models with an appropriate microturbulence.