2006A&A...456..451L

We present the second part of an optical spectroscopic study of planetary nebulae in the LMC and SMC. The first paper, Leisy & Dennefeld (1996A&AS..116...95L), discussed the CNO cycle for those objects where C abundances were available. In this paper we concentrate more on other elemental abundances (such as O, Ne, S, Ar) and their implications for the evolution of the progenitor stars. We use a much larger sample of 183 objects, of which 65 are our own observations, where the abundances have been re-derived in a homogeneous way. For 156 of them, the quality of data is considered to be satisfactory for further analysis. We confirm the difficulty of separating type I and non-type-I objects in the classical He-N/O diagram, as found in Paper I, a problem reinforced by the variety of initial compositions for the progenitor stars. We observed oxygen variations, either depletion via the ON cycle in the more massive progenitor stars, or oxygen production in other objects. Neon production also appears to be present. These enrichments seem to be explained best by recent models, some including overshooting, where fresh material from the core or from burning shells is brought to the surface by the 3rd dredge-up. All the effects appear stronger in the SMC, suggesting a higher efficiency in a low metallicity environment, either because the reaction itself is more efficient or because the increment is more visible when superposed on a low initial quantity. Neither oxygen nor neon can therefore be used to derive the initial composition of the progenitor star: other elements not affected by processing such as sulfur, argon or, if observed, chlorine, have to be preferred for this purpose. Some objects with very low initial abundances are detected, but on average, the spatial distribution of PNe abundances is consistent with the history of star formation (SF) as derived from field stars in both Clouds.