1 Introduction

Planetary Nebulae (PN) are the final evolutionary stage of intermediate mass stars, and their chemical abundances have been extensively used to probe the chemical evolution of the Galaxy and the Magellanic Clouds. In these objects, abundances of elements like He, C and N were probably modified at the surface of the progenitor, due to dredge-up episodes prior ejection of the outer envelope. On the other hand, abundances of Ne, S and Ar represent the conditions of the interstellar medium at the epoch of the formation of the parent star.

Oxygen deserves a special attention since it is one of the key elements used to trace the chemical evolution of stellar systems. However, concerning changes in the oxygen abundance along the stellar evolution of the progenitor, we don't have a satisfactory answer yet. If only CN-cycle products are dredge-up, oxygen remains unchanged, but some reduction is expected if the ON-cycle is operative. Theoretical estimates predict a maximum (uncertain) reduction of about 0.11 dex in the surface abundances of massive progenitors (Renzini & Voli 1981). The analyses of the C/N and C/O ratios in galactic supergiants suggest that these stars have dredged-up essentially CN-cycle products and that oxygen remained unchanged (Luck & Lambert 1985). Concerning PN, different data sets (Köppen et al. 1991; Kingsburgh & Barlow 1994; Costa & de Freitas Pacheco 1996) indicate that the oxygen abundances of massive and young planetaries (type I objects) agree with abundances derived from HII regions, consistent with the conclusion derived from galactic supergiants, namely, oxygen was not affected. These first considerations may give some support to the scenario in which oxygen would not be altered by mixing episodes and its abundance reflects the pristine chemical composition of the medium from where PN have been formed out.

Abundance studies performed in the early eighties by Torres-Peimbert (1984) seem to suggest an anti-correlation between the N/O ratio and the oxygen abundance. This anti-correlation would be restricted primarily to type I planetaries, since non-type I objects would show a positive correlation, according to other investigations (Torres-Peimbert & Peimbert 1977; Peimbert & Serrano 1980). With a sample covering a broader abundance range, Kaler et al. (1990) found that also non-type I planetaries have a negative N/O vs. O/H correlation. These observations were interpreted by Peimbert (1985) as an evidence that most of the nitrogen in type I planetaries is of secondary origin, coming from oxygen rather than carbon. Such an anti-correlation was noted to be present not only in galactic PN but also in LMC and SMC data (Aller 1983; Henry et al. 1989).

In the past years, an observational program aiming to obtain chemical abundances of southern PN, including also objects in the Magellanic Clouds, has been carried out at the University of S. Paulo. Chemical abundances for 74 galactic PN included in our survey were summarized by Costa & de Freitas Pacheco (1996), and results for 23 PN in the LMC were reported by de Freitas Pacheco et al. (1993a,b). In the present work we report additional abundance data for 23 PN in the SMC as a part of our survey, and then review the problem of the N/O vs. O/H correlation. For galactic PN we use our sample constituted by 74 objects (Costa & de Freitas Pacheco 1996), covering an oxygen abundance range of about two orders of magnitude 7.08 $< \varepsilon(\rm O) <$ 9.11 (where $\varepsilon(\rm X)$ = log(X/H) + 12). For the LMC we used data by de Freitas Pacheco et al. (1993a,b) and for the SMC, the present results, both supplemented by literature data. The plan of the paper is the following: in Sect. 2 we summarize observations, data reduction and physical properties for the 23 PN in the SMC; in Sect. 3 we discuss the N/O vs. O/H relation for galactic and Magellanic PN, and in Sect. 4 we present our main conclusions.