The Ca II resonance lines 
and 
and the Mg
II resonance lines 
and 
have been
traditionally used as diagnostics of luminosity and chromospheric activity
in cool stars (Linsky 1980 and the references therein,
Cassinelli & MacGregor 1986). Other useful indicators of
chromospheric gas include the Ly
and the Balmer lines of H I, the
infrared triplet lines (
8498, 8542, 8662) of Ca II and the
He I lines at 
5876 line and 
10830. The ultraviolet lines
have a limitation because of the low flux of cool stars in this part of the
spectral region, making it difficult to obtain spectra with high
signal-to-noise ratio. It is worth exploring lines, better placed in the
spectrum, that are sensitive to luminosity and other parameters. The Ca II
triplet lines at 
8498, 8542, 8662 are ideal candidates for
studying their sensitivity to various stellar parameters in cool stars.
These lines are due to transitions between the upper 
levels and the lower metastable 
levels. The same
upper levels connect to the ground state 
to give rise to
the Ca II H and K lines. Although the Ca II triplet lines have the lower
levels populated radiatively and are not collisionally controlled like the
Ca II H and K lines, they have been identified as very good luminosity
probes. These lines are strong, hence easy to observe even in faint stars.
They are relatively free from blends and are not contaminated much by the
telluric lines. On the contrary, the Na I lines at 
8183,
8195, which have also been used as luminosity probes in cool stars, have
several atmospheric absorption lines in their vicinity (Jones et al.
1984; Carter et al. 1986; Alloin & Bica
1989; Zhou 1991).
With the advent of the CCDs which are especially sensitive to the red, there
has been a spurt of studies of the Ca II triplet lines in
both stars and galaxies, to investigate, in particular, the behaviour of
their strengths as a function of surface gravity (which is directly
proportional to luminosity), metallicity and temperature and to use this
information in stellar population synthesis (Jones et al. 1984
(JAJ); Carter
et al. 1986; Alloin & Bica 1989; Diaz et al.
1989 (DTT) and Zhou 1991). JAJ observed a sample of 62
stars covering a range in spectral types from early B to mid-M and over 4
orders of surface gravity and a factor of 10 in metallicity. Examination of
their data led JAJ to suggest a single-valued relation between the Ca II
equivalent width (EQW) and the surface gravity for all stars in the spectral
range F through mid-M. DTT in their comprehensive study of 106 stars over
the spectral range F6 to M0 and of all luminosity classes found that the Ca
II lines exhibited a biparametric behaviour; there is almost no influence of
but there is dependence on gravity and metallicity. They
concluded that in the high-metallicity range, the Ca II triplet lines depend
on surface gravity only while in the low-metallicity range they depend on
metallicity as well.
There has also been a number of theoretical studies to understand the
effects of stellar parameters on the Ca II triplet strengths (Smith & Drake
1987, 1990; Erdelyi-Mendes & Barbuy 1991;
Jorgenson et al. 1992 (JCJ)). JCJ carried out detailed
non-LTE calculations of the Ca II EQWs over an extended span of parameters
and concluded that although the correlation between the Ca II EQW and 
is straightforward, the dependence on 
and metallicity is
more complex. Their theoretical work has shown that the line EQWs depend on
the response of the stellar atmospheric structure to changes in these
parameters and more importantly, the response to a change of a given
parameter is different for different parameter intervals considered of the
other parameters. Because of this complex behaviour of the lines, the
question of whether the nucleus of a galaxy or any composite stellar system is dwarf
enriched, for example, and/or metal rich has remained an open one.
In order to further explore the dependence of the Ca II triplet EQWs on
these stellar parameters, we had, in our previous study (Mallik
1994) observed the first two lines of the triplet, namely, the 8498
Å\
and the 8542 Å lines in 91 stars in the spectral range F8 - M4 of all
luminosity classes and over a range of metallicities [Fe/H] from -0.65 to
+0.60. A detailed analysis of these observations confirmed that the sum of
the equivalent widths (EQWs) of 
8498 and 
8542 strongly
correlates with surface gravity and also depends upon metallicity. As JCJ have
emphasised, it is very important for a
proper understanding of the behaviour of the Ca II triplet EQWs
that a sufficiently
large range in the values of parameters be covered.
We have in our current analysis enlarged
the earlier sample to 146 stars including more dwarfs, subgiants
and many more number of metal poor stars; some with a much lower metal content.
In particular, our sample now includes a few superluminous supergiants (of
luminosity 0-Ia) which turn out to be crucial in deriving
the behaviour of the Ca II EQWs. Also, since the 
8662 line
lies in a comparatively clean region of the spectrum with very few blends and
since it is possible to determine its EQW accurately,
observations of the
8662 line have also therefore been obtained for the entire sample.
Section 2
describes the observations and data reduction. The analysis of the Ca II triplet
observations is discussed in Sect. 3. Section 4 discusses chromospheric
activity as an important phenomenon influencing the line depth and hence the EQW
of the Ca II triplet line profiles. Conclusions are given in Sect. 5.