Do all BaII stars have white dwarf companions?

With the Hubble Space Telescope (HST) and the Goddard High Resolution Spectrograph (GHRS) we have observed four barium stars, three mild barium stars, and one weak G-band star in the ultraviolet spectral region. One barium star was observed with HST and the Space Telescope Imaging Spectrograph (STIS). The aim was to check the hypothesis that all these peculiar stars have white dwarf (WD) companions, which at their asymptotic giant branch phase transferred mass with peculiar element abundances to the present barium and CH peculiar stars. Assuming that the ultraviolet continua of the cool giants, including the barium stars, are generated in their chromospheres and that the relations between the continua and the emission lines created in the chromospheres and transition layers are similar in field giants and barium stars, we found that, indeed, most of our target barium and weak barium stars appear to have excess flux in the UV when compared to standard giant stars. For most of the stars the excess flux can be attributed to WD companions with temperatures between 10,000 and 12,000 K, if the WD mass is about 0.6 M.. Cooling times for the WDs were derived from their effective temperatures and model calculations by M. Wood. The calculated cooling times are longer than the lifetimes of the barium stars on the giant branch. For our target stars the mass transfer therefore happened while they were still on the main sequence. For two of the mild barium stars and one or perhaps two barium stars the derived cooling times for the WD companions come out to be longer than the total evolutionary times of the barium stars as calculated by Schaller et al. If our derivations are correct (the error bars are rather large) then either evolutionary models with larger convective overshoot have to be used for the barium stars or the cooling times of the white dwarfs have to be revised downward. Possibly an additional (as yet unknown) cooling mechanism has to be considered? The weak G-band star HD 165634, which has a carbon underabundance of about a factor of 10, also appears to have a WD companion. We discuss the implications of this very low carbon abundance.