1994A&A...288..401M

The geometrical properties of E and S0 galaxies have been intercompared using the data collected in Paper III (Michard & Marchal 1994) for 108 RSA objects in a complete, luminosity and distance limited, sample. As the apparent flattening (largely an effect of projection along the line of sight), is a determining factor in the segregation between E and S0 objects, the working hypothesis has been made that an important bias is introduced in the recognition of the two classes. It is perhaps as well to assume that galaxies of both Hubble types belong, but for a small(?) minority, to a common population of objects with similar structures. This hypothesis receives strong support from the frequency-diagrams of the ellipticity ε_max, measured near its maximum or at the isophote of surface brightness V=21.5. The diagram for S0's alone cannot be generated by the random projection of any objects: it is clearly biased by the shift to the E type of many S0's of moderate inclination and relatively modest disk. This limits the significance of the same diagrams for E galaxies. The noted bias is much reduced if S0's and disky E's are considered together. Because of the strong outwards decrease of the ellipticities in disky E's and in the S0's with non-thin envelopes (thick disks and spheroidal haloes), the frequency diagrams of the ellipticities measured at the classical B=25, or at V=25, do not show the bias noted above for S0's. The lack of round E's requires the spheroidal components to be faintly triaxial, as recently emphasized by other authors. Our hypothesis is also supported by the overlap of E and S0 galaxies in ad hoc classification schemes of ellipticity profiles and envelope geometry, and in such correlation diagrams as: - the ellipticity in the envelope, i.e. near µ(V)= 25, against the intermediate maximum ellipticity - the extremum of the Carter's coefficient e₄ (or a₄ or c₄ in other similar works) against the maximum ellipticity - the disk extent, as far as it can be estimated without detailed modeling, against the maximum ellipticity. The "standard" structure of E-S0 objects includes a spheroidal, nearly oblate component, that cannot be of constant ellipticity: this parameter often increases outwards from its bulge value, then decreases again in the envelope. Disks of various brightness and extent, between dominating and vanishing, are the second essential component. Obviously, their relative importance is, together with the dust content, a physical criterion involved in the morphological separation between E and S0 objects, besides the inclination effect. In the present sample, the disk cannot be detected in nearly half of E classified objects, and is probably vanishing in about 40% of this class, or circa 17% of the total of studied early-type objects. This minority contains mostly boxy E's with non-standard ellipticity profiles, and, as is well known, no significant rotational support. From the convergent results of quantitative morphology and kinematics, the early-type galaxies should be segregated into the majority class of disky fast oblate rotators and a minority class of non-rotating ellipsoids.