The Li abundances for our programme stars have been derived from the Li I 6708 resonance line. We have measured the equivalent widths of the Li feature by direct integration and gaussian profile fitting. Both authors of this paper measured the equivalent width independently and we took the mean as the final value. Typically, the standard deviation of different measurements was 15 mÅ , which is the value that we took as error bar. We used NLTE curves of growth for the Li I line from Martín et al. (1994). The validity of these curves has been confirmed by the more recent work of Pavlenko et al. (1995). Effective temperatures for the IC 4665 stars were derived by two different methods: a calibration of B-V colour vs. (Arribas & Martinez Roger 1988), and a calibration of spectral type vs. (de Jager & Nieuwenhuijzen 1987). The B-V colours were taken from Prosser (1993) and they were corrected by the average cluster colour excess of E(B-V)=0.18. The spectral types were derived from comparison of our spectra with a battery of standards observed with the IDS (Montes et al. 1995b). We estimate that our spectral types are accurate to about half a spectral subclass. The temperature of P309 was estimated from the spectral type only, because the B-V calibration of Arribas & Martinez Roger did not include M-type stars.
The differences in the obtained from (B-V) and spectral type ranged from 30 to 310 K, with an average of 147 K. For the star P12 we found an anomalous large discrepancy between the colour and the spectral type. This star was classified by Prosser & Giampapa (1994) as G0V and they also gave a v sini =10km s . In contrast, we find a spectral type of K0V and a v sini =70km s . Our high v sini is consistent with the short rotational period found by Allain et al. (1996) of 0.60 days. We believe that there is a problem with the photometry and spectroscopy given by Prosser (1993) and Prosser & Giampapa (1994) for this star. Hence, we chose to ignore the photometry and used only our spectral type for deriving its .
Taking into account the maximum uncertainties in ( 150 K) and Li I equivalent widths ( 15 mÅ), they lead to an error bar in the Li abundances of 0.3 dex. This is similar to previous works on Li abundances in other clusters (e.g. García López et al. 1994). In Table 2 we give the NLTE Li abundances of our programme stars in the customary scale of log N(H)=12.
Table 2. H fluxes and Li abundances of IC 4665 stars
The H emission excess was measured using the spectral subtraction technique. Reference spectra were constructed for all the IC 4665 stars using main-sequence standards of the same spectral type, which were rotationally broadened and doppler shifted to give as good match as possible. We subtracted the reference spectra from the IC 4665 spectra and obtained the chromospheric contribution of H . For more details on this technique see Montes et al. (1995a). In Fig. 2 we display one example (P12) of the spectral subtraction method. In the subtracted spectra H shows up in emission and Li I in absorption, whereas the other photospheric lines are canceled. The equivalent width of H in emission were converted to flux using: a calibration of B-V vs. surface flux in the H region (Hall 1996); and a calibration of V-R vs. surface flux in the H region (Pasquini & Pallavicini 1991). We obtained the B-V and V-R colours from our spectral types and from dereddening the B,V photometry of Prosser (1993). The results obtained from both colours agreed quite well, and the final values adopted are given in Table 2. We note that for P27 the radial velocity changed from night to night and we also noticed variations in the H emission from logF(H )=5.63 to log F(H )=5.86. In all our spectra of P27 the Li I equivalent widths were similar except in that with the largest H emission, which presented a shallower Li I feature (equivalent width 110 mÅ). We suggest that P27 deserves monitoring of radial velocity, H and Li variations.
© European Southern Observatory (ESO) 1997
Online publication: July 3, 1998