B. E. J. Pagel

We have extended our analytical chemical evolution modelling ideas for the Galaxy to the Magellanic Clouds. Unlike previous authors, we assume neither a steepened IMF nor selective galactic winds, since among the alpha-particle elements only oxygen shows a large deficit relative to iron and a similar deficit is also found in Galactic supergiants. Thus we assume yields and time delays identical to those we previously assumed for the solar neighbourhood. We include inflow and non-selective galactic winds and consider both smooth and bursting star formation rates, the latter giving a better fit to the age-metallicity relations. We predict essentially solar abundance ratios for primary elements and these seem to fit most of the data within their substantial scatter. Our LMC model also gives a remarkably good fit to the anomalous Galactic halo stars discovered by Nissen & Schuster (1997). Our models predict current ratios of SNIa to core-collapse supernova rates enhanced by 50 per cent and 25 per cent respectively relative to the solar neighbourhood, in fair agreement with ratios found by Cappellaro et al. (1993) for Sdm-Im relative to Sbc galaxies, but these ratios are sensitive to detailed assumptions about the bursts. We consider that the high SNIa/SNII ratios found in some late-type galaxies are related to their star formation history rather than to any differences in the IMF.

We model the evolution of abundances of oxygen, iron, α and r-process elements in the Galactic disc using the instantaneous and delayed production approximations which enable the problem to be handled analytically. In addition to using the new greatly improved data published by Edvardsson et al. and the oxygen abundance distribution function of G dwarfs in the solar neighbourhood as improved by Sommer-Larsen, we take into account new survey data on metal-deficient stars which show that the thick-disc population extends to arbitrarily low metallicities, meaning that a significant proportion of stars previously classified as belonging to the halo, on grounds of low metallicity, actually belong to the disc and should be modelled as such, without bringing in any prior phase associated with the halo. Comparison with theoretical yields from Type II and Type Ia supernovae shows good agreement, as has previously been found by Tsujimoto et al. using a different approach.