This section compares the efficiency of 
with
the criteria introduced by former authors
when using the 
system or other photometric systems.
The lower efficiency of p in the
intermediate
and late regions is paralleled by previous authors.
Maitzen (1975)
showed using spectrophotometry how the depth of the 
feature decreases towards the cooler, i.e.
Sr stars.
and 
were found to be correlated except for
the Sr stars.
Maitzen & Vogt (1983) also found
that
the correlation between the
Geneva's parameter
and the index was lower for the Sr stars than
for the Si and Cr stars. In particular, the criterion of
mag in detecting CP2 stars is only able to classify 36% of the Sr
stars sample used by the authors. In our case, the smaller size of CP2
samples (only 93 and 70 CP2 stars for the intermediate and late regions,
respectively), hindered the analysis.
So, the following comparison is restricted to the stars in the early region.
system
Figure 6: [m1] - [u-b] diagram for the stars in the early region.
Dots represent normal stars and full circles CP2 stars.
The solid line is the adopted standard relation for normal stars and
the dashed line indicates the threshold above which 50% of the
CP2 stars are placed
A study of the location of CP2 stars in a colour-colour diagram was carried out
by Cameron (1967). On a c1-m1 diagram, certain regions corresponding
to high m1 values are mainly occupied by Ap stars.
Maitzen (1976) and
Adelman et al. (1995) showed that there is a relation between
, using (b-y) as free parameter, and the strength of the
feature.
Figures 6 (click here) and 7 (click here) show our samples on the [m1]-[u-b] and
[c1]-[u-b] diagrams. In the same way as with the p coordinate,
we fitted standard relations to normal stars in each diagram,
we considered peculiarity thresholds
for which 50% of CP2 stars are separated and we computed the percentage of
contamination of normal stars. The contamination ratios are of 16% and
17%. So, these diagrams are less effective
than the p-[u-b] diagram.
Figure 7: The same as Fig. 6 but for the [c1] - [u-b] diagram
On the other hand, a star with an 
mag, computed by using
the calibrations for normal stars
(Crawford 1978), was considered as peculiar
by Philip et al. (1976).
In our CP2 sample, there are only 16 stars (3%) with such a colour excess,
while all normal stars have a reddening higher than this. These 16 stars
have a 
so they are also classified as peculiar
with our peculiarity level of 
.
photometry
Maitzen (1976) introduced a photometric index (the a index) to measure the
intensity of the depression at ,
by comparison of the flux at the center with the mean of
the flux at each side of the depression.
The index is the difference between the measured a index for a
given star and the a index of the normality line for the same temperature.
For the CP2 stars this index is significantly positive and
all the stars with 
mag are considered as
peculiar stars.
We compiled the published a photometry in the literature:
Maitzen 1976;
Maitzen & Vogt 1983; and the series of papers Photometric Search for
CP2 Stars in Open Clusters
(Maitzen 1993; Maitzen et al. 1988 and
references therein)
and built a sample of
1703 stars with measured , 930 having complete
photometry, too. The stars were treated with the same classification
algorithm and their colour excess was computed. There are 577 stars
belonging to the early region and 453 of them have 
mag.
Comparison of the defined by the Eqs. (1) and (2) with
the measured
shows a high correlation between both indices (Fig. 8 (click here)a).
The 
threshold is able to detect 60% of the stars with
mag
with a contamination of about 5% of stars with 
mag.
Although the index is more efficient than the index,
the correlation between them demonstrates that
is also a good indicator of peculiarity.
On the other hand, the correlation between and
is better than the
correlation between and 
or
, (Figs. 8 (click here)b and 8 (click here)c).
The contaminations in these cases are three times higher.
Figure 8: a) Correlation between and defined by
Eqs. (1) and (2).
b) Correlation between and .
c) Correlation between and .
Horizontal and vertical dashed lines correspond to the threshold
values of either index for the detection of CP2 stars
The Geneva photometric system is capable of detecting CP2 stars through
the depression (North & Cramer 1981).
Two indices, and 
, were defined with this
purpose. is defined as the measured (V1-G) for a given
star less the (V1-G) for a normal star with the same (B2-V1)
colour.
is a combination of the filters U, B1, B2, V1 and G
(Cramer &
Maeder 1979, 1980) and has the advantage of being reddening free, although it
seems that it is slightly sensitive to certain systematic
effects which are not clearly understood
(North & Cramer 1981).
Figure 9 (click here) shows the comparison between the
and the .
The photometric data in the Geneva system were extracted from
Rufener (1988).
There are 1235 normal stars and 396 CP2 stars
with measurements in both the Geneva and the photometric
systems. The stars with mag
are considered peculiar stars. This criterion recognizes 71% of CP2
stars with a contamination of only 1%.
So, the Geneva system is more
efficient than the Strömgren-Crawford system,
since it gives a direct measurement of the flux depression at
.
is not a reddening free index either and the very
reddened CP2 stars are not detected
(Hauck & North 1982).
Figure 9: Correlation between defined by
Eqs. (1) and (2) and . Dots are normal stars
and full
circles are CP2 stars from
Renson et al. (1991) catalogue.
Horizontal and vertical dashed lines correspond to the threshold
values of both indices to recognize CP2 stars
For an analysis of the correlation between and
peculiarity indices see Maitzen & Vogt (1983).