Next Contents Previous

11. STELLAR CONTENT RELATED TO TYPE: FORMATION AND EVOLUTION

One of the most remarkable features of the standard classification system is that the stellar content of galaxies varies systematically along the linear sequence from E to Sm. In the order Sa to Sd, the progressive changes are (1) increasing absolute luminosity of the brightest stars in regions of the spiral arms, (2) increasing percentage of mass in the form of gas and dust, (3) increasing sizes and numbers of H II regions in the spiral arms, and (4) progressively bluer integrated (B - V) and (U - B) colors, indicating progressively earlier type stars that contribute most of the light.

The correlations represent a physical result because the classification is made principally not on the resolution of the galaxy into stars, but rather on the character of the nuclear bulge, and on the presence, the shape, and the regularity of the arm structure.

Some of the questions raised by the correlation of content and form have been discussed by Sandage, Freeman, and Stokes (1970) in a review of Hubble's (1926) problem to find the true flattening of elliptical galaxies from the observed distribution of apparent flattenings. In that study, flattenings were compared for ellipticals and spirals, and we concluded with Hubble, Holmberg (1946), and de Vaucouleurs (1959a) that all spiral and S0 galaxies are flatter than the flattest elliptical.

Because flattening is a dynamical property that, in an equilibrium configuration, cannot change in times less than the relaxation time (~ 1012-1014 years), the difference in intrinsic flattening between E and S galaxies shows that one type cannot evolve into the other. A basic difference must then have existed between E and S systems at the time of their formation to cause such different forms now. An identification and an understanding of this difference is necessary at the outset of even the most elementary discussion of galaxy evolution.

Flattening can occur on a short time scale only if (1) the relaxation time itself is short under conditions that prevailed before the system formed into stars, (2) the gravitational potential energy of a stellar system is rapidly changing (cf. Lynden-Bell 1967), or (3) both. The difference in formation history of E and S galaxies must then be due to some difference, such as the angular-momentum distribution of the contracting protogalaxies, or any other agent that would control the initial rate of star formation during the collapse time of the protogalaxy toward the fundamental plane. Elliptical galaxies obviously did not collapse completely (they are not now highly flattened), and this shows that nearly all star formation took place in times short compared to the free-fall time (~ 109 years), leaving no gas to interact by gas-gas collisions to damp into a fundamental plane that is characteristic of spirals. On the other hand, the existence of such a plane in all spirals seems to betray a slower initial conversion of gas to stars. Furthermore, much of the gas has remained in the plane over the lifetime of the system.

The required rapid star formation in E systems and in the spheroidal components of spirals and S0 galaxies finds considerable direct support from observations. Baade's resolution of parts of the Local Group galaxies M31, M32, NGC 205, NGC 185, and NGC 147 into stars at MV appeq -3, B - V appeq + 1.5 at the same intensity level as the globular clusters embedded in their galactic halos shows the antiquity of these resolved subsystems (Baade 1944). The argument is made stronger by the lack of bright young blue supergiants in the spheroidal regions of S0, Sa, and Sb galaxies, and in E systems. Constraints on the age are stringent (Sandage 1969, 1971) from these data.

Considering the flattening and the stellar content data, and arguing from the relaxation problem of Lynden-Bell (1967), Sandage, Freeman, and Stokes reached the following conclusions.

1. Stars in the spheroidal component of all galaxies were formed very rapidly on a time scale comparable to the collapse time of the protogalaxy (ltapprox 109 years). The argument is independent of that used by Eggen, Lynden-Bell, and Sandage (1962) which was based on the orbital eccentricities of galactic stars. It depends here only (a) on the absence of a fundamental plane both in E galaxies and in the spheroidal subsystems of spirals, and (b) on the observation that the stellar distribution itself appears to be relaxed, presumably by Lynden-Bell's (1967) mechanism.

2. The halo stars were formed from matter having low angular momentum per unit mass, i.e., the spheroidal component, during the collapse time. Other matter with higher angular momentum collapsed to a disk.

3. The galaxy type was determined by the amount of free gas left over in the disk after collapse. No appreciable evolution along the Hubble sequence has occurred since the galaxies were formed.

4. The dominance of the disk in spiral and S0 systems betrays their mean angular momentum per unit mass, higher than that which exists in less-flattened E galaxies. Does the higher angular momentum slow the rate of star formation, keeping uncondensed matter (gas and dust) in reserve for new generations to form continuously in the dusty disks?

5. All stars in the spheroidal component of galaxies should show an age distribution DeltaT which is small compared to the total age T, such that DeltaT / T ltapprox 0.1. This expectation agrees with observational data for halo globular clusters in our own Galaxy, and with the uniformity of observed energy distributions I(lambda) for E galaxies and the centers of S0, Sa, and Sb systems.

Without collapse of matter from a wider configuration it is difficult to understand the presence of a dominant plane in spirals. But in ellipticals it is difficult to understand its absence without rapid and complete star formation on a time scale short compared with the free-fall time from the edges of a protogalaxy.

The comments in this last section may be more speculative than substantive, but they are discussed here because they follow naturally from certain characteristics of the Hubble classification. Because of this, it seems possible that the classification is fundamental in the sense discussed in Section 1 i.e., the connective parameters of the system appear to contain part of the physics of galaxies. That the Hubble system has this property would follow if ideas of this section could be substantiated. And likewise, that the Morgan nuclear sequence might also be a classification of the second kind would follow if the stellar component of the spheroidal bulge is understood in terms of stellar evolution of the composite H-R diagram (Chap. 2), as we now believe.

Next Contents Previous