Discussion and conclusion

With our new observations we cannot confirm the small CaiiK emission noted by DLS. If we accept that the occurrence of a shock wave in the atmosphere of $\rho$Puppis is real, we can confirm its transient nature. Because, in our previous study (PaperI), we have not detected any spectral signature induced by a shock wave within the profile of H$\alpha$ and some metallic lines, we can suggest that, when a shock is present within the atmosphere, its intensity must be small. Both from observations and pulsation models, it is known that the reproductibility of the dynamical motion of the different atmospheric layers is not the same over a few successive cycles (see e.g., Fokin et al. [1998]). Especially the shock intensities are strongly dependent on the dynamical state of the atmosphere. Consequently, it is quite plausible that the visibility of the weak shock detected by DLS in $\rho$Puppis is extremely variable and that its detection requires observations during several cycles and not one or two as done up to now. This means that new long time observations are necessary to determine the exact occurrence of the shock wave in $\rho$Puppis.

As shown by Fracassini et al. ([1983]), the permanent UV Mgii emission has a larger strength near maximum light (at phases $\sim0.96$-0.12) i.e., when a shock wave induced by the classical $\kappa$-mechanism, is crossing the photospheric layers (see e.g. Fokin & Gillet [1997]). Because this emission is permanent, we must accept that a chromosphere exists in $\rho$Puppis and that the shock passage enhances the emission. Moreover Fracassini et al. ([1983]) put into evidence a variable violet-red asymmetry in k1 and k2 Mgii components. Nevertheless, the quality of their observations is not high enough to confirm the presence of a blue asymetry which is symptomatic of outward propagating waves in the chromosphere. Future UV-observations are necessary to verify this important point which would give us an independent confirmation of the presence of the shock wave.

From the repeated observations of the UV Mgii lines of several classical Cepheids made by Schmidt & Parsons ([1982]), the pulsation appears to affect the mechanism of chromospheric heating. Moreover, these authors conclude that the mechanical energy of the pulsation seems to be at the origin of the chromospheric activity in this class of variable stars. Consequently, a strong coupling between the pulsation and the chromospheric state might take place and its understanding is certainly the basic step to access to a quantitative modelling. Although the calculation of a pulsation chromospheric model as done by Sasselov and Lester ([1994]) is out of the scope of this paper, we suggest that, if future observations confirm the small CaiiK emission, a theoretical approach of this problem would be required.