2 Selection of the sample

We have searched the literature in order to collect a sample of giant stars representative of the different populations in our Galaxy. The main criteria imposed on the search,without order of priority, were determined by the final aim of the programme (i.e. photometric calibration of the effective temperature scale) and were that the stars should:
a) have published UBV(RI) and uvby photometry,
b) have determined or estimated metal abundance,
d) be preferably close to the solar neighbourhood, in order to avoid large interstellar reddening, which biases the analysis of temperatures,
d) have (at least tentative) spectral type determinations, and
e) be contained in the input catalogue of Hipparcos.
Among the different works considered for the extraction of stars, a main contribution has been taken from Anthony-Twarog & Twarog (1994), where a calibration of metallicity, and an accurate estimation of the reddening based on the reddening maps of Burstein & Heiles (1984) is presented for a sample of more than 100 stars. The selection of stars of this catalogue has the further advantage that their Strömgren photometry is homogenized. We have also included in our sample most of the giant stars contained in the Catalogue of Spectroscopic Abundances (Cayrel et al. 1992) observable from the Observatorio del Teide ($-30^{\circ}$ < $\delta$ < $+64^{\circ}$). From the extended work on photoelectric UBVRIJHK photometry of globular cluster red giants summarized by Frogel et al. (1983a,b) we have included 137 stars from various globular clusters (M 3, M 67, M 15, M 13, M 71 and 47 Tuc) measured by Cohen et al. (1978) and Frogel et al. (1979, 1981). We have also included 22 giants measured by Carney (1983a), 27 giants from Arribas & Martínez-Roger (1987b) and 39 giants from Selby et al. (1988).

In summary, the sample consists of approximately 400 stars. The metallicities which will be adopted in the present study are obtained from the Catalogue of Spectroscopic Abundances (Cayrel et al. 1992), and from revised photometric calibrations which are scaled to the spectroscopic values compiled in that catalogue. The uncertainty in [Fe/H] for the stars in the sample ranges from 0.15 to 0.3 dex.
The optical photometry which will be used in the forthcoming calibrations was obtained from different sources.
The UBV(RI) photometry has been taken mainly from the compilations by Mermilliod & Mermilliod (1994); Lanz (1986); Morel & Magnenat (1978), the Hipparcos Input Catalogue (Turon et al. 1993); Carney (1983a,b); Jasniewicz (1982) and Johnson et al. (1966). For the remainder of the stars the sources were the Bright Star Catalogue (Hoffleit & Jaschek 1982). The relations provided by Bessel (1983) were used to transform RI photometric data from the Cousins and Kron systems where necessary.
Strömgren photometry was extracted mainly from the homogenized catalogue of Hauck & Mermilliod (1990).
The JHK photometry was specifically measured in this programme for nearly 75% of the stars in the sample; for the remainder of the stars the sources were those quoted above.
Apart from unavoidable contamination, several subgiants have been purposely included in the sample to extend the range of surface gravity in the final calibration. Possibly there are also a small number of horizontal branch stars among the bluest stars of the sample.
A crude estimate of the features of the total sample is revealed in Fig. 1, where we show the location of the stars in the (B-V) : (U-B) diagram. The dispersion above the average line of solar abundance stars is caused by the blanketing effect associated with the metal content.

 

Figure 1: (B-V) vs. (U-B) for the stars of the sample. The lines show the average fiducial lines for giant stars. The open pentagons show the mean integrated colours of a sample of globular clusters as taken from the data compiled by Peterson (1986)