Despite the sizable observational effort invested during the last decade, it is clear that much can be gained from a survey having greater sensitivity to the detection of emission lines. The sensitivity can be improved in at least four ways - by taking spectra with higher signal-to-noise ratio (S/N) and spectral resolution, by using a narrower slit to better isolate the nucleus, and by employing more effective methods to handle the starlight correction.
Using a double CCD spectrograph mounted on the Hale 5-m reflector at Palomar Observatory, high quality, moderate-resolution, long-slit spectra were obtained for a magnitude-limited (BT 12.5 mag) sample of 486 northern ( > 0°) galaxies (Filippenko & Sargent 1985, 1986; Ho, Filippenko, & Sargent 1995). The red camera covered the range 6210-6860 Å with ~ 2.5 Å resolution, while the corresponding values for the blue camera were 4230-5110 Å and ~ 4 Å. Most of the observations were obtained with a narrow slit (generally 2", and occasionally 1"), and the exposure times were suitably long (up to 1 hr or more for some objects with low central surface brightness) to secure data of high S/N. This survey contains the largest data base of homogeneous and high-quality optical spectra of nearby galaxies yet published. The selection criteria of the survey ensure that the sample is a good representation of the local (z 0) galaxy population (Fig. 2), and the proximity of the objects (median distance = 17 Mpc) enables fairly good spatial resolution to be achieved (typically 200 pc).
Figure 2. Distribution of Hubble types for the 486 galaxies in the survey. Ordinary (unbarred) galaxies are shown in the hatched histogram, and barred (SB and SAB) galaxies in the unhatched histogram. Classifications taken from de Vaucouleurs et al. (1991).
A common strategy for removing the starlight from an integrated spectrum is that of ``template subtraction,'' whereby a template spectrum devoid of emission lines is suitably scaled to, and subtracted from, the spectrum of interest to yield a continuum-subtracted, pure emission-line spectrum (e.g., Costero & Osterbrock 1977; Filippenko & Halpern 1984; Filippenko & Sargent 1988; Ho et al. 1993). In practice, the template is derived either from the spectrum of a different galaxy or from the spectrum of an off-nuclear position in the same galaxy. This approach, however, suffers from some limitations. For instance, the absorption-line galaxy chosen as the template may not exactly match the stellar component of the object in question; previous studies generally invested limited observing time to the acquisition of template spectra. In the case where an off-nuclear spectrum is used as the model, it may not be completely free of emission, and one cannot be sure that radial gradients in the stellar population are absent.
To perform the starlight subtraction in a more objective and efficient manner than has been done in the past, a modified version of the template-subtraction technique that takes advantage of the large number of absorption-line galaxies in the survey was developed (Ho et al. 1996c). Given a list of input template spectra and an initial guess of the velocity dispersion, the 2-minimization algorithm of Rix & White (1992) solves for the systemic velocity, the line-broadening velocity dispersion, the relative contributions of the various templates, and the general continuum shape. The best-fitting model is then subtracted from the original spectrum, yielding a pure emission-line spectrum. Figure 3 illustrates this process for the H II nucleus in NGC 3596 and for the Seyfert 2 nucleus in NGC 7743. In the case of NGC 3596, the model consisted of the combination of the spectrum of NGC 205, a dE5 galaxy with a substantial population of A stars, and NGC 4339, an E0 having a K-giant spectrum. Note that in the original observed spectrum (top), H, [O III] 4959, 5007, and [O I] 6300 were hardly visible, whereas after starlight subtraction (bottom) they can be easily measured. The intensities of both H and H have been modified substantially, and the ratio of the two [S II] 6716, 6731 lines changed. The effective template for NGC 7743 made use of NGC 205, NGC 4339, and NGC 628, an Sc galaxy with a nucleus dominated by A and F stars.
Figure 3. Illustration of the method of starlight subtraction. In each panel, the top plot shows the observed spectrum, the middle plot the best-fitting ``template'' used to match the stellar component, and the bottom plot the difference between the object spectrum and the template. In the case of NGC 3596 (a), the model was constructed from NGC 205 and NGC 4339, while for NGC 7743 (b), the model was derived from a linear combination of NGC 205, NGC 4339, and NGC 628.