NEW MODELS FOR WOLF-RAYET AND O-STAR POPULATIONS IN YOUNG STARBURSTS

Using the latest stellar evolution models, theoretical stellar spectra , and a compilation of observed emission line strengths from Wolf-Raye t (W-R) stars, we construct evolutionary synthesis models for young st arbursts. We explicitly distinguish between the various W-R subtypes ( WN, WC, WO), whose relative frequency is a strong function of metallic ity, and we treat O and Of stars separately. We calculate the numbers of O and W-R stars produced during a starburst and provide detailed pr edictions of UV and optical emission line strengths for both the W-R s tellar lines and the major nebular hydrogen and helium emission lines, as a function of several input parameters related to the starburst ep isode. We also derive the theoretical frequency of W-R-rich starbursts . Our models predict that nebular He II lambda 4686 emission from a lo w-metallicity starburst should be associated with the presence of WC/W O stars and/or hot WN stars evolving to become WC/WO stars. In additio n, W-R stars contribute to broad components beneath the nebular Balmer lines; the broad W-R component may constitute several percent of the total flux in the line. We review the various techniques used to deriv e the W-R and O star content from integrated spectra, assess their acc uracy, and propose two new formulae to estimate the W-R/O number ratio from UV or optical spectra. We also explore the implications of the f ormation of W-R stars through mass transfer in close binary systems in instantaneous bursts. While the formation of W-R stars through Roche lobe overflow prolongs the W-R-dominated phase, there are clear observ ational signatures that allow us to distinguish the phases in which W- R stars are formed predominantly through single or binary star channel s. In particular at low metallicities, when massive close binaries con tribute significantly to the formation of W-R stars, the binary-domina ted phase is expected to occur at ages corresponding to relatively low HE equivalent widths. The observational features predicted by our mod els allow a detailed quantitative determination of the massive star po pulation in a starburst region (particularly in so-called ''W-R galaxi es'') from its integrated spectrum and provide a means of deriving the burst properties (e.g., duration and age) and the parameters of the i nitial mass function of young starbursts. The model predictions should provide the most reliable determinations to date. They can also be us ed to test current theories of massive star evolution and atmospheres and to investigate the variation in stellar properties with metallicit y.