Appendix A: Long slit spectroscopy  

NGC 128 and the template stars HR 459 (K2III), HR 6884 (K0III), and HR 8610 (K2II), have been observed with the 3.6 m ESO telescope equipped with EFOSC in three different nights of June 1990.

The detector was a RCA CCD $640\times1024$ pixels, used in binned mode. The pixel size of 30 $\mu$m maps $0\hbox{$.\!\!^{\prime\prime}$}68$ on the sky. The grism gives a dispersion of $\sim$2 Ang/pxl in the wavelength range 5000-7000. The $3\hbox{$.\mkern-4mu^\prime$}6\times 1\hbox{$.\!\!^{\prime\prime}$}5$ spectrograph slit was aligned with the major and minor axes of the galaxy.

In the first night we got an exposures of 1 hour along the major axis in the South direction, placing the center of the galaxy at one end of the slit in order to cover as much as possible the whole galaxy. In the second night we put the slit in the North direction, but unfortunately the weather was cloudy and with bad seeing ($1\hbox{$.\!\!^{\prime\prime}$}7 - 2\hbox{$.\!\!^{\prime\prime}$}1$). The third night we got a single centred spectrum of the minor axis.

After the standard procedures of bias and dark subtraction and flat-fielding of the raw data, we corrected the distortion pattern by deriving a line-by-line wavelength calibration from the comparison spectra.

The science spectra have been conservatively adaptive-filtered following the procedures described by Richter et al. (1992). This technique allows to obtain reliable rotation curves (RCs) and velocity dispersion (VD) profiles extended up to $\sim$2 mag fainter in surface brightness.

In the spectra are clearly visible the absorption features of the MgI triplet ($\lambda\sim$ 5200) and the Na D-band around $\lambda\sim$ 5900. The analysis of the spectra is based on the Fourier Correlation Quotient method (FCQ), as developed by Bender (1990). We essentially used the same procedures described with more details in D'Onofrio et al. (1995).

Unfortunately the low S/N ratio and the unfavourable spectral resolution hampered the acquisition of an accurate broadening function and we were not able to check the kinematical signatures of eventual subcomponents. We could only verify that the change of the template stars introduces no differences in the final RCs and VD profiles.

In our long slit spectra of NGC 128 we recognized five emission lines, [NII] ($\lambda$6548), H$\alpha$ ($\lambda$6563), [NII] ($\lambda$6583), [SII] ($\lambda$6717) and [SII] ($\lambda$6731). The most intense one is [NII] ($\lambda$6583), while H$\alpha$ is attenuated by the neighbouring absorption lines.