A NEW METHOD FOR ANALYZING HORIZONTAL BRANCH MORPHOLOGY AND MASS-LOSS

In this paper we discuss the effects on the horizontal branch morpholo gy of possible star-to-star variations in various parameters of the st ars in globular clusters. These parameters include the mass loss effic iency, the mass of the core at the helium core flash, the metallicity, and the initial mass. We have developed a method that allows us to co mpute many evolutionary tracks quickly and accurately, and we are ther efore able to study the effects of a wide range of changes in the para meter space. Our method relies on observed relations between luminosit y and effective temperature, and on existing model computations of the core mass and luminosity at the helium core flash. The mass, M(RGB), of the stars at the red giant branch is determined from a method that is to a large extent independent of computed models. With M(RGB) known , we can determine the value of the mixing length parameter, alpha, th at is required (for a given grid of models) in order to fit observed c olour-magnitude diagrams. By applying this method to observations of g lobular clusters (GCs) of a wide range of metallicities, we find that alpha (on the RGB) is independent of metallicity (but its value is dep endent on the code used). For the extensive grid of Sweigart & Gross ( 1978) we find its value to be alpha = 1.4+/-0.1. The variation in colo ur of the HB stars in a given GC can be well explained as a spread in the ratio of total mass to core mass. It is demonstrated that the mass differences between RGB and HB stars, as well as the mass spread on t he HB, can be explained as a consequence of normal Reimers-type mass l oss. We show that star-to-star variations of the mass loss parameter, eta, between eta = 0.0 and eta = 0.7 can explain the HB morphology in typical GCs while at the same time being fully consistent with observa tions of field stars (e.g., those on which Reimers' mass loss law is b ased), as well as with the RGB morphology of the GCs. We predict that the variation in T(eff) near the red giant branch tip (RGT) for stars with different eta will be nearly unobservable and stars with differen t mass loss efficiency will reach approximately the same luminosity at the RGT. If, as an alternative, star-to-star variations in the mass, M(c)RGT, of the core at the helium core flash should explain the HB mo rphology (e.g., due to variations in the core rotation velocity), this variation must be +/-0.05 M. in order to explain the typical spread i n mass observed in GCs. Such variations in M(c)RGT, are only consisten t with the observed RGB in GCs if a few of the stars that are usually believed to be AGB stars are in fact RGB stars with ''delayed'' core f lash. Other scenarios that could explain the average mass difference b etween stars at the HB and the RGB are either unable to explain the sp read in HB mass, or lead to contradictions with observations of the RG B stars. Such scenarios include star-to-star variations in the mixing length parameter, the value of Z, and the initial mass on the RGB, as well as variations in possible mass loss caused by the He core flash. Due to the high flexibility of our method to continuous variations in a wide range of parameters, it is also suitable for estimating for exa mple the likelihood of forming hot subdwarfs in single-star evolution, the possible contribution of HB-manque star to the UV light in galaxi es of various metallicities, and the conditions for forming precursors of white dwarfs.