In an earlier work we compiled a data base for six tracers of interstellar gas for the 467 early-type galaxies in the Revised Shapley-Ames catalog: H I, H-2, dust, optical emission-line gas, radio continuum emission, and X-ray-emitting gas. Here this data set is analyzed in an effort to understand the gaseous content by galaxy type and the relationship between gaseous components. Some new observations are presented for 21 cm H I emission (20 galaxies), CO (1-0) emission (18 galaxies), and 6 cm emission maps (11 galaxies). In elliptical galaxies the interstellar medium is dominated by the hot gaseous component, while in the Sa galaxies, cold interstellar material accounts for most of the gaseous mass; SO galaxies are intermediate, with similar amounts of hot and cold material. Normalized by optical luminosity, the H I and H-2 masses decrease so sharply from Sa- to E-type galaxies that almost no elliptical galaxies have detectable atomic or molecular gas. A similar but considerably less dramatic gradient is present between the normalized dust mass and galaxy type. A close relationship exists between the CO and IRAS fluxes, for 100-mu-m flux densities above 2 Jy. Whereas 6 cm emission is closely related to IRAS emission and star formation in spiral galaxies, the dispersion in this relationship becomes large for the earliest type galaxies, probably due to radio emission from low-luminosity AGNs. Radio emission is correlated with the presence of emission lines, both of which may be tracers of an active nucleus. In contrast to the cold interstellar components, the normalized X-ray emission is independent of galaxy type in E through SO/Sa systems, while the Sa and peculiar galaxies are slightly X-ray-poor. The X-ray luminosity is strongly dependent on the optical (blue) luminosity. The nature of this dependence is a function of the distance model. The simplest Hubble flow yields L(X) is-proportional-to L(B)1.75. We discuss an alternative distance model which yields L(X) is-proportional-to L(B)2.45, a relation which is in conflict with simple cooling flow models. The dispersion of the X-ray luminosity about this regression line is unlikely to result from stripping. The striking lack of clear correlations between hot and cold interstellar components, taken together with their morphologies, suggests that the cold gas is a disk phenomenon while the hot gas is a bulge phenomenon, with little interaction between the two. Then, the progression of galaxy type from E to Sa is not only a sequence of decreasing stellar bulge-to-disk ratio, but also of hot-to-cold-gas ratio.
