We would like to conclude with a comment on the group into which HD 50138 can be classified. It has usually been put either in the group of B[e] stars (Bopp 1993) or in the group of Herbig Be-Ae stars (Morrison & Beaver 1995). From the definitions of the groups (see Jaschek & Jaschek 1987), it is difficult to conclude that the object belongs clearly to either one of them. It has certainly many forbidden lines in emission, as required by B[e] stars, but contrary to other stars of the group, one perceives the photospheric spectrum of the underlying star. With regard to the membership in the group of Herbig Be-Ae stars, one of the original criteria was the existence of an associated nebulosity, which in HD 45677 is abs ent. It is thus not a Herbig Be-Ae star. Also Thé et al. (1994) do not include it in their catalogue of Herbig Ae/Be stars.
The fact that three stars analysed up to now in this series (MWC 51585, MWC 349A and MWC 645) do not show any absorption lines, whereas HD 50138 (as well as HD 45677) do show absorption lines, provides a clear indication of (at least) two subgroups of B[e] stars. Until many more stars of the group are analysed, it seems premature to call HD 50138 (or MWC 645) a typical B[e] star. Caution is thus recommended a Nnd one should not apply whatever is known on this particular star to other B[e] stars. The same applies of course to models.
Acknowledgements
It is a pleasure to thank Dr. D. Ballereau for permitting the use of the spectrum taken in Argentina. We also thank M. Ph. Moreau of the Haute Provence Observatory for the reproduction of the spectra. We thank the referee Prof. J.P. Swings for his careful reading of the paper.