Table 1 (click here) lists the final temperatures (denoted TEMP) derived for the chosen stars. The column S shows the source of infrared data as described in the Table 1 (click here) caption. Where data have been obtained at one observatory only, these final temperatures are the means of several determinations based on either J,K,L filters or J,H,K,L filters. Table 3 (click here) shows the dependence of determined temperatures on the three infrared filters used at the two observatories. In all cases, the agreeement between the stellar temperatures determined using the three filters is close to one per cent. We show later that temperatures derived from TCS narrow band measures agree well with published temperatures of Blackwell & Lynas-Gray (1994), obtained using very similar filters. Because the TCS narrow band temperatures and published narrow band temperatures have equal weight, their averages are presented in Table 1 (click here) as final temperatures. Estimates of the percentage error of all temperature determinations are tabulated in the column PE of Table 1 (click here). These estimates are based on considerations concerning the accuracy of absolute fluxes and the agreement between temperatures derived using the various filters, and on the relations between temperatures and various photometric indices discussed below. No account has been taken of uncertainties due to difficulties in the theoretical modelling of stellar atmospheres discussed for example by Mégessier (1994). In addition, important stellar data such as metallicity may either be unavailable or inaccurate. Because of these considerations, the reliability of such estimates of absolute accuracy may be regarded as questionable, but we suggest that they give at least a useful indication of the relative accuracies of the determined stellar temperatures presented in Table 1 (click here). Unfortunately, it is not possible to make a definitive comparison between temperatures determined using interferometrically measured angular diameters and integrated fluxes because such determinations are most often made using low temperature giant stars, and the IRFM technique (as presented in this paper) is not suited to such low temperature stars because infrared fluxes at longer wavelengths would be needed.
| Filter ratios | Temperature ratios | ||
| ESO | NB | TCS | |
| J/K |  |  |  |
| H/K |  |  | |
| L/K |  |  |  |