The absorption calcium triplet spectrum

We have measured the equivalent width of the two main CaT lines at 24#24Å, according to the method described by Díaz, Terlevich & Terlevich (1989, hereafter DTT). The nuclear spectrum (central 0.72 arcsec) is plotted in Fig.  for reference; the three CaT lines, MgI 25#25, and two of the hydrogen Paschen lines in absorption are identified. The weakness of Pa14 indicates a negligible contamination of the CaT lines by Pa16, Pa15 and Pa13, that fall in their red wings. The results of the pixel to pixel equivalent width measurements along the slit, for the three position angles, have been combined and are shown in a single Fig. . The lefthand side of this figure, (-10 arcsec to the nucleus) corresponds to the NE-SE quadrant on the eastern side of the galaxy. The righthand side of the figure (from the nucleus to 10 arcsec) corresponds to the NW-SW quadrant on the western side. There is a clear systematic trend in this combined figure, that can also be appreciated (albeit with a worse signal-to-noise ratio) along the three individual PA. Outside the circumnuclear ring, ew(CaT) takes a low value of 26#26 Å; at and within the ring, 5 arcsec east and 4 arcsec west, ew(CaT) jumps to a higher value of 7 Å; and closer in, within the partially resolved nucleus, ew(CaT) rises to a value between 8 and 9 Å. These are three distinct regimes clearly present in the circumnuclear region.

The interpretation of ew(CaT) as a function of a stellar population depends on the metallicity (DTT; see the most recent work by García-Vargas, Mollá & Bressan, 1998, and references therein). We can have an indication of the metallicity by measuring the ew(MgI). According to DTT this line is sensitive to metallicity and to effective temperature, but not to gravity. In the nucleus of NGC 6951, ew(MgI)27#27 Å, that indicates (see fig. 8 of DTT) the production by cool ( 28#28 K), high metallicity stars (solar or higher). For this metallicity range, ew(CaT) is a function mainly of gravity and so we can now interpret the measurements in terms of stellar populations.

The lowest values of ew(CaT), present mainly outside the ring, are due to an old stellar population of giants, with an age around 0.5 to 1 Gyr. The highest nuclear values, ew(CaT)29#29 Å indicate the presence of an important population of red supergiant stars (RSG), that dominate the luminosity at these wavelengths, with an age around 10 to 20 Myr. In the circumnuclear region, ew(CaT) takes intermediate values; it is significantly higher than in the outer parts, but the actual value belongs to a range that can be explained by either giants or RSG. Our interpretation is as follows. The measured equivalent widths in the circumnuclear and nuclear regions are a lower limit to the actual values, for two reasons. First, these equivalent widths are diluted by the continuum of the young stellar population responsible for the ionization of the gas; we cannot precisely gauge with our present data set how important is this dilution, for which we would need spectroscopy in a longer wavelength range, including the ultraviolet, but we do see the blue stellar knots so that some dilution by them must be taking effect. Second, at least outside a radius of 2 arcsec, we can distinguish two distinct stellar kinematical components; although we cannot measure independently their contribution to the luminosity or to the ew(CaT), it is at least possible that one of these two distinct populations (one that can be identified with the bar population) contributes to dilute the other population (whose kinematics is dominated by rotation, see below). Thus, if dilution is taken into account it is likely that the circumnuclear values of ew(CaT)30#30 Å are in fact larger, so that they also reflect a population of RSG within the ring.