Some assessment of reliability can be made by a direct comparison of temperatures determined independently using the three main sources of data; ESO, TCS and narrow band, to which may be added temperatures published using narrow band data. The relevant temperature ratios, ESO/TCS, TCS/Published NB are given in Table 5 (click here), together with standard deviations and the number of stars in common. The inference from these comparisons is that the temperature scales given by the three independent sources of data (narrow band, TCS and ESO) for overlapping stars are not significantly different.
| Class | No. of stars |  |  |
| III | 11 | 1.005 | 0.013 |
| V | 6 | 1.004 | 0.003 |
| Class | No. of stars |  | |
| III | 8 | 1.006 | 0.014 |
| V | 7 | 1.010 | 0.023 |
| Class | No. of stars |  | |
| All | 65 | 0.999 | 0.010 |
A further test of reliability is a comparison of the several relations
that have been derived for temperature versus the photometric parameters,
(B2-V1), (b-y), (B-V) and (V-K), corrected for extinction.
The result is
that the average of the ESO, class V, temperatures for these parameters
differs from the average of the TCS, Class V, temperatures by 0.28 per cent
(
per cent).
For Class III stars, the averages differ by
1.31 per cent (
per cent).
As these comparisons have been made using
pairs of star lists that have so few stars in common, it is not possible to
comment on the significance or otherwise of any difference between
temperatures derived from ESO and TCS data.