We have observed the Ca II triplet lines in 146 stars to understand the
interplay of such parameters as temperature, gravity and chemical
composition in determining the strengths of these lines. Our observations
span all luminosity types, a range in spectral types from F8 to M4 and in
[Fe/H] from -3.0 to +1.1. The detailed empirical analysis has stressed on
the importance of considering a sufficient range in 
, 
and [Fe/H] to cover all stars that contribute to the integrated light of the
calcium lines in a stellar system and has shown a more complex dependence of
the EQW of the Ca II lines on the above parameters than previously found. The
salient conclusions are:
a) The Ca II triplet EQWs are indeed sensitive to surface gravity. Assuming
it to be a single valued linear relation, the observers in the past have
used it to determine the luminosity of stars in galaxies. However,
this relationship is
nonlinear over the 
range covered and is also influenced by 
and [Fe/H] in complicated ways.
b) The dependence of the EQW on surface gravity is steeper for metal rich stars than for metal poor stars. This fact can be used to determine the luminosity of the dominant population from the Ca II spectra in high metallicity composite stellar systems. It is not possible to derive useful information from the Ca II spectra in low metallicity stellar systems unless the effective temperature of the stars contributing to the light is known independently.
c) The Ca II triplet EQWs are only mildly dependent on effective temperature. And
even here, giants and dwarfs respond differently to changes in 
.
d) Dwarfs and giants respond differently even to metallicity. The sensitivity of Ca II EQW on metallicity is strong for supergiants and weaker in giants and dwarfs, even though the range in [Fe/H] covered in the latter is much larger. Even with [Fe/H] ranging only from -0.88 to +0.50 in supergiants, the correlation appears to be nonlinear.
e) It is found that stars of the same temperature, luminosity and metallicity may have varying central depths and this has to do with the chromospheric activity varying among these stars. The shallower the lines in the spectrum of a star, the more active the chromosphere of this star is. Chromospheric emission fills in the absorption profile and raises the core.
The Ca II triplet line strength and shape thus depend not only on luminosity and metallicity but also on chromospheric activity. To be able to use the Ca II triplet lines as a tool for discriminating stellar populations in galaxies, one has to understand the complex behaviour of the Ca II triplet lines with respect to all these parameters. Only when the sensitivity to each parameter is independently determined, can one exploit this information in the understanding of the stellar content of galaxies. The Ca II triplet lines as a potential probe for the stellar population synthesis should be viewed in the light of this complex dependence.
Acknowledgements
I especially wish to thank the referee for several useful comments and suggestions that have helped improve paper.