Now that it has been established that any differences between temperatures found using TCS, ESO, narrow band and narrow band published filters, are reasonably small, we combine together all of the various measured temperatures for each star to give the results listed in Table 1 (click here). In this section we relate these temperatures to stellar photometry parameters. We have so far neglected the small additional effect of metallicity in these representations of temperatures and a small correction term for this will be discussed later in Sect. 9.
For this purpose we divide the stars of Table 1 (click here) into the three luminosity classes V, IV and III, and consider them separately, beginning with class V. Further, because we wish to compute polynomial coefficients of good accuracy for use in deriving temperatures from a database of photometric parameters, we choose for this purpose only the measures from Table 1 (click here) that we judge on independent evidence to be the most reliable. This segregation is based on the extent of the agreement among measures using different filters, agreement with previous measures using the IRFM, and the existence of a complete MK spectral type. These stars have temperatures accurate to between 0.5 per cent and 0.8 per cent in Table 1 (click here). Also, as our wish is to derive an accurate representation of stellar temperatures by a polynomial and a small correction factor for the effect of metallicity, we include at this stage only those stars with a measured metallicity that is less than 0.2 dex. In the following sections, we present separate regression analyses using data for the (B2-V1), (B-V) and (b-y) photometric parameters, all corrected for interstellar extinction, for the three luminosity classes V, IV and III separately, and for the two temperature regions, T < 6000 K and T > 6000 K, where this is possible, and including the restrictions described above.
Table 6 (click here) presents the
polynomial coefficients for relations between
temperature and the photometric quantities
, 
and 
,
derived in the way described above for these classes of stars that have
very small metallicities. Standard deviations in the percentage differences
between measured and calculated temperatures are also given. These
standard deviations may be taken as a guide to the reliability of
temperature determinations for the selected stars, but it is likely that
the deviations also include a small spread in the determinations of each
photometric parameter, together with a real spread in stellar temperatures.
Figure 3 (click here) shows a
typical plot of temperature versus 
for class V stars,
but the best parameter for all class V stars is apparently .
For class IV stars, regression constants and standard deviations for
class IV stars are given only for temperatures greater than 5000 K.
The restrictions described above have reduced the number of available
stars so much that a division into two temperature ranges is not practical.
However, a plot of temperature against
shows that a discussion of
the whole range of temperature for these stars is reasonable, and all
available stars have been used for the calculation of regression
coefficients presented in Table 6 (click here).
The standard deviation of temperatures,
of 0.91 per cent, is satisfactory. Regression constants for class III
stars, using the above restrictions, are given for temperatures in the
two ranges T < 6000 K and T >6000 K.
The quantity , where K is measured on the Johnson scale,
is especially useful because it usually yields a standard deviation
about the regression line that is smaller than those shown by other
photometric parameters.
| p | A | B | C |  | |
| Class V | (B2-V1) | 7995 | -7246 | 3284 | 1.52 |
 | (b-y) | 9865 | -13319 | 7483 | 1.30 |
| (B-V) | 8515 | -5270 | 1440 | 1.72 | |
| Class V | (B2-V1) | 7749 | -7911 | 7856 | 1.20 |
 | (b-y) | 9034 | -12562 | 11813 | 1.09 |
| (B-V) | 9134 | -8600 | 5398 | 1.32 | |
| Class V | (V-K) | 8921 | -2848 | 441 | 1.30 |
 | |||||
| Class IV | (B2-V1) | 7949 | -6944 | 3407 | 0.91 |
 | (b-y) | 9030 | -11396 | 8313 | 1.10 |
| (B-V) | 9078 | -6545 | 2298 | 1.43 | |
| Class III | (B2-V1) | 5504 | -190 | -1198 | 1.01 |
|
| (b-y) | 4507 | 3213 | -4454 | 1.00 |
| (B-V) | 4842 | 1413 | -1426 | 0.91 | |
| Class III | (B2-V1) | 7581 | -5927 | 3898 | 1.20 |
|
| (b-y) | 8613 | -10332 | 8613 | 0.99 |
| (B-V) | 8667 | -6383 | 3145 | 1.36 | |
| Class III | (V-K) | 8305 | -2036 | 224 | 1.38 |
|
| |||||