We have presented a photometric monitoring of twelve oxygen-rich Miras
over a 4-year
period. The properties of these southern hemisphere variables were rather
poorly known before this study. This work has considerably improved the set of
available LPV light-curves in visible colours:
between 53 and 140 
measurements are presented for each star
covering between three and seven successive cycles.
This allowed us to reconstruct the light-curves
in all the colours.
We then derived the spectral type of these Long Period Variables at all
the epochs of the observations.
Spectral type versus phase variations over a whole cycle are therefore given
for the first time together with the UBVRI light-curves. New photometric
parameters (visible and spectral type extrema, spectral type amplitudes, etc.)
have been derived. We have also proposed a period for V Cha (
days) and W Cha is confirmed as being an oxygen-rich Mira.
Moreover the repeatability of the visible and colour light-curves during the
successive cycles has been discussed: strong variations in all the
filters are reported some of them over quite a short term. We thus confirm
that rapid variations are certainly real in Miras. We finally have
derived the distance of these twelve LPV and shown that rather good
estimates can be obtained from UBVRI data at maximum light.
This paper shows once again that broad-band photometry is a powerful tool for studying Long-Period Variables. Furthermore these cool stars emit the largest part of their energy in the infrared. Simultaneous observations from the U filter up to the infrared could help in the modelling of these stars as Le Bertre already did for R For (1988). Furthermore the cycle-to-cycle variations are smaller at longer wavelength where intrinsic properties of Miras can then be derived more accurately. Future monitoring of LPV should thus cover a larger and redder spectral domain than the one studied here. In another respect we have seen that the real maximum of luminosity has to be observed to determine the distance of these stars. This and the cycle-to-cycle variation studies at all wavelengths require collection of data over a large number of successive cycles and with a good sampling in time during each period (i.e. at leat two or three points per month). Such programs would require a large amount of telescope time and emphasize the utility of telescopes dedicated to stellar photometry.
Acknowledgements
This work has been achieved thanks to the help of P. Bouchet (ESO) and of other observers at La Silla (W. Brandner, J.F. Claeskens, E. Grebel, T. Lehmann, E. Gosset, etc.). We thank L.A. Willson and M.O. Mennessier for their comments on the manuscript and N. van der Bliek for her help when reducing the spectra. We are also grateful to the referee T. Le Bertre for his constructive suggestions and to N. Ryde for comments and improvement of english. Informations from SIMBAD data base at the CDS, Strasbourg, France, were used for the preparation of this paper. Part of this work has been supported by a grant from the European Space Agency awarded to P. de Laverny.
Table 2: V magnitudes, 
colours and derived spectral types
Table 5: Comparison of the Mira distances calculated in Table 4 with
previous and independent determinations. Distance of other Miras calculated
from UBVRI data found in the literature are also determined and compared
to other determinations
