A STUDY OF METAL ABUNDANCE PATTERNS IN COOL WHITE-DWARFS .3. COMPARISON OF THE PREDICTIONS OF THE 2-PHASE ACCRETION MODEL WITH THE OBSERVATIONS

We compare the predictions of our two-phase accretion model with all t he observations currently available on the metal-line phenomenon in co ol white dwarfs. The model is based on a simplified picture of the int erstellar medium, but makes specific predictions as to the presence of metals in the photospheres of these stars. We find that the gross obs erved metal abundance patterns are very well accounted for within the framework of the model. For many individual stars, it is also possible to infer directly the accretion rates necessary to maintain the photo spheric metal abundances. These rates agree remarkably well with the e stimates of the accretion rates from denser patches of the interstella r medium which we used in our idealized numerical simulations, thus pr oviding an important self-consistency check. While we have assumed a p riori that metals accrete in solar proportions, our final analysis sug gests the possibility of mild differentiated accretion for the metals. We also compute explicitly the detection probability for metals in co ol white dwarfs and find an excellent agreement with the observed freq uency of calcium-line objects in the known sample of cool He-rich whit e dwarfs. We discuss the question of traces of hydrogen in these stars . We find that hydrogen, like the metals, must be accreted. However, h ydrogen accretion must be reduced by factors of 10(2)-10(7) with respe ct to metal accretion to account for the observations. The idea of scr eening mechanisms thus surfaces again; hydrogen screening must, howeve r, operate over the full range of effective temperature 20,000 K great er than or similar to T(eff) greater than or similar to 5000 K. We als o discuss the case of carbon in cool He-rich white dwarfs. Although so me carbon, along with metals and hydrogen, must also accrete onto thes e stars, the accretion model cannot explain the peculiar carbon abunda nce pattern observed in such objects. The diffusion-dredge-up pollutio n model is to be preferred for that element. Furthermore, there is a h int that carbon accretion is also reduced with respect to metal accret ion. This suggests that further progress could be made by studying the process of fractionated accretion.