We have studied the soft X-ray spectra of a heterogeneous sample of 75 Seyfert galaxies using data from the Einstein Observatory Imaging Proportional Counter (IPC). This is the largest such sample studied in soft X-rays and includes for the first time information about the soft X-ray spectra of type 2 Seyfert galaxies. The well-determined spectra in our sample (mostly high-luminosity Seyfert type 1's) are generally consistent with a single power-law index of α = 0.81 (with a corresponding uncertainty of 0.03 if the sample were homogeneous). Most AGNs may have "canonical" power-law spectra (α ∼ 0.7) in the harder, 2-20 keV band, but our survey shows via direct, object-by-object comparison that Seyfert spectra are systematically steeper in the soft X-ray band (0.2-4.0 keV). Furthermore, we find that soft X-ray spectral index and emitted luminosity are correlated, in the sense that higher luminosity, higher redshift AGNs tend to have steeper spectra. The lower luminosity, type 2 Seyferts in our sample tend to have, as a group, flatter power-law indices (α ∼ 0.5) than the type 1 Seyferts. Both these effects may be due to complex spectral structure unresolved by the IPC rather than differences in the shape of the power-law continuum. Most of the Seyfert galaxies in our sample have spectra consistent with no intrinsic absorption by cold gas. The low energy response of the IPC allows stringent upper limits to be placed on the absorbing column density, NH, typically a few times 1020 cm-2 for the brighter spectra. We also see a slight tendency toward an unphysical, negative, intrinsic absorbing column density (as was found by Wilkes and Elvis for quasars), which is probably indicative of soft excesses in the spectra at low energies. In some cases spectra do show significant positive intrinsic absorption. Despite our sensitivity to column densities in the range 1020-1021 cm-2, we find that NH is either consistent with the Galactic column or large and positive (≳ 1021 cm-2). The generally low intrinsic absorbing columns observed in our sample have important ramifications for studies of the reddening of Seyfert continua. Seyfert galaxies and quasars show a continuity of properties in the soft X-ray band. One example is the appearance of soft excesses (negative absorption) in both classes. Second, although the number of radio-loud objects in our sample is small, our Seyfert galaxy spectra are consistent with the correlation seen in quasars (Wilkes and Elvis) between radio-loudness and flatness of the soft X-ray spectrum. Third, our data are also consistent with the correlation found by Wilkes, Elvis, and McHardy for quasars (but see Boroson) between the strength of permitted optical Fe II lines and the steepness of soft X-ray spectrum, though uniform analysis of a larger sample is needed. The one difference between the Wilkes and Elvis survey and our own is that no quasar spectra were heavily absorbed. The flatness of our Seyfert 2 X-ray spectra (relative to the type 1's) may be a real difference (because of object type) or it may be a selection effect. We have spectra only for those type 2's that are brightest in X-rays, which tend to be radio-loud, and which therefore have flatter X-ray spectra. Alternatively, a partially covered or otherwise complex spectrum in the 0.2-4 keV band can produce an apparently flatter spectrum and lower luminosity when fitting a power law to the IPC data. Our low-resolution spectra are not inconsistent with an "obscured torus" model of Seyfert 2 galaxies (as published by Antonucci and Miller in 1985 and Krolik in 1988); however, we emphasize that the well-studied Seyfert 2 galaxy NGC 1068 has a very unusual (i.e., exceptionally steep) soft X-ray spectrum relative to our sample of type 2's.
