The dependence of the
on the
(at given gravity and chemical composition) governs
the appearance of a star in the CMD.
This is best shown by the colour-
relation for any pair of
pass-bands. The colour is simply given by
with and
17W, 21W, 25W, 30W, 33W, 43W, 45W, 55W, 60W,
70W, 81W, 85LW
for the WFPC;
and
15F, 17F, 22F, 33F
for the FOC, with obvious meaning of the symbols. In all cases
.
Table 2: Model Galaxies: ,
and Age are in Gyr, the galaxy
mass is in
The relations for each group are shown in the series of Figs. 3 (click here)
through 10 (click here) limited to the case of solar composition and
gravity typical of main sequence stars (log g=5).
It is apparent that colours involving one or two UV pass-bands
do not possess a monotonic relation with . With the standard
Johnson system, it is natural to see colours becoming redder as
decrease, as long as the colours are defined as bluer magnitudes minus
redder magnitudes. With the HST system the colour becomes bluer again when the
temperature is cooler than a certain value that depends on the pass-band
(hereafter the colour turnover). The turnover temperature somewhat
depends on the gravity and chemical composition because the spectral
energy distribution varies with them. Finally, the colour turnover
does no longer occur for the visible-red pass-bands.
The turnover is caused by the combination of two effects. First the UV pass-bands are affected by the the so-called visible/red-leak problem, i.e. long tails of significant transmittance at wavelengths much longer than the nominal peak value (cf. Figs. 1 (click here) and 2 (click here)). For very hot stars, the amount of flux gathered in the visible/red-leak region is small and the effect of this on the colour is negligible. As the star gets cooler the amount of flux falling into the visible/red-leak region becomes more and more important and so its effect on the colour. The other effect is that the various UV filters have visible/red-leaks that intersect each other. For instance the visible/red-leak of the filter 15W is higher than the visible/red-leak of the filters 21W and 25W, whereas the visible/red-leak of the filter F30W is higher than the previous ones. This anomalous behaviour tends to disappear going to filters of longer and longer peak wavelength. Analogous considerations hold for the UV filters of the FOC.
The obvious consequence of the colour turnover is that one cannot
determine a unique value of the temperature from the colours at least
as long as some UV bands are involved. Although the combination of a
UV pass-band with one, whose peak is at much longer wavelength can
improve upon this point of ambiguity, there are cases in which the
effect is still there (see for instance the case of or
).
The effect of the colour turnover on the observed CMD is easy to foresee and is discussed in the next section by means of select isochrones.
Finally, we notice that colours like
,
,
, and
with
, respectively,
show pronounced dips at
which are caused by the secondary
peaks in the functions
pointed out in the
previous sections.
There is a final remark to be made. The colour- relations for
the WFPC2 system obey the basic condition that at
all
the colours are zero or nearly zero (see Figs. 3 (click here) through 8 (click here)),
which is typical of the ground based
photometric systems (Johnson for instance). This means that the calibration
of the
has been properly made. In the case of the FOC
system, the above condition is no longer always satisfied, cf. Figs. 9 (click here)
and 10 (click here). At
the colours may be significantly
different from zero. Similar behaviour is also visible in the FOC colours
calculated by Yi et al. (1995). This implies that the corresponding
requires further calibration. Although this point can be
easily fixed up by adding suitable shifts to the magnitudes or suitably
correcting
the
, we prefer to present results for
the original
.