7 Summary

The IRFM has been applied to a sample of approximately 500 giant stars later than F0, which cover the metallicity range (0.5, -3.5). Near-IR monochromatic fluxes have been used in order to derive T_J, T_H, T_K and T_L' for each star. The uncertainties in the input parameters needed to apply the IRFM and the induced errors on the three/four temperatures derived have been computed. The consistency of the temperatures derived in the three different bands is fairly good above 4000 K. The final temperature for each star in the sample has been derived considering the mean of T_J, T_H, T_K and T_L' weighted with the inverse of their errors. From the analysis of the systematic errors associated with the uncertainty in the absolute flux calibration in the near-IR, the expected indeterminacy of the zero point of the temperature scale should be around 1%. However, the good agreement between the IRFM and direct temperatures of a sample of 18 giant stars suggests a lower uncertainty. The internal estimated precision for the final temperatures, considering the effect of accidental errors, is around 1.5%. The comparison with other works also based on semi-empirical and empirical methods shows slight discrepancies which may be explained by considering the internal errors affecting the determination of temperatures.

Acknowledgements

We are grateful to the referee, Dr. Lynas-Gray, for his careful assessment of the paper and for many comments and suggestions which have certainly improved the final version. We are also grateful to Terry Mahoney for correcting this paper for English and style. We have made use for this research of the SIMBAD database, operated at CDS, Strasbourg, France.