5 Conclusions

We have presented IR (JHK) photometry of the "representative sample'' of Be stars defined by  Paper I. We have derived a new technique for separating the effects of (i) emission of the circumstellar material, and (ii) interstellar reddening. The technique involves combining photometry of 3 IR filters, a general interstellar extinction law and an assumption (that we verify) that the colours of the circumstellar disc of a Be star can be related in a similar fashion.

By correlating our results with an independent measure of interstellar reddening (the equivalent width of the Na I 5890 Å line), we are confident that our method is valid. Using this technique we find that the disc emission makes a maximum contribution to the optical (B-V) colour of a few tenths of a magnitude.

We find significant correlations between the infrared excess emission from the disc and emission from a range of lines (H$\alpha$, Br$\gamma$, Br11, and Br18). Linear fits to these correlations have been derived. There is also a significant correlation between the $v \sin (i)$ (and also $\omega\sin(i)$) of the Be star and the infrared disc emission. These data present strong constraints for models of the structure of Be stars. Any model that purports to explain their disc structure must be able to reproduce the correlations presented in Figs. 5-9. In a future study we will contrast these observations with predictions from current models.

Acknowledgements

The TCS is operated on the island of Tenerife by the Instituto de Astrofisica de Canarias. Thanks to Chris Collins for useful discussions relating to the statistics of this paper. LH greatfully acknowledges PPARC funding.