Annu. Rev. Astron. Astrophys. 2006. 44: xxx-xxx Copyright © 2006 by Annual Reviews. All rights reserved

1. INTRODUCTION

Following Hubble (1936), we still classify galaxies as ellipticals, spirals and irregulars (see Sandage 2005, and references therein). This was an eyeball, purely morphological scheme; however, morphology correlates with the stellar population content of these galaxies, with typical ellipticals being redder than the others, and showing purely stellar absorption-line spectra with no or very weak nebular emissions. As a consequence, one often refers to early-type galaxies (ETG), even if they are color (or spectral type) selected rather than morphologically selected. Furthermore, the bulges of spirals of the earlier types show morphological as well as spectral similarities with ellipticals, and one often includes both ellipticals and bulges under the category of galactic spheroids.

Morphologically-selected and color-, or spectrum-selected samples do not fully overlap. For example, in a recent study of local (z ~ 0) ETGs from the Sloan Digital Sky Survey (SDSS) all three criteria were adopted (Bernardi et al. 2006), and the result is reported in Table 1 (M. Bernardi, private communication). ¹ From a sample including ~123,000 galaxies with 14.5 < r_Petrosian < 17.77 and 0.004 < z < 0.08, ETGs have been selected in turn by each of the three criteria (MOR, COL, SPE), and the resulting numbers of galaxies fulfilling each of them is given on the diagonal of the matrix. Out of the diagonal are the fractions of galaxies that satisfy two of the criteria, as labelled in the corresponding row and column. So, out of the morphologically-selected ETGs, 70% satisfy also the color selection, etc. The correlation between color and morphology selection persists at high redshift, e.g., at z ~ 0.7 about 85% of color-selected, red-sequence galaxies are also morphologically early-type, i.e., E/S0/Sa (Bell et al 2004a), an estimate broadly consistent with the local one, when allowing for the wider morphological criterion.

It is estimated that true ellipticals represent ~ 22% of the total mass in stars in the local universe, a fraction amounting to ~ 75% for spheroids (i.e., when including E0's and spiral bulges), whereas disks contribute only ~ 25% and dwarfs an irrelevant fraction (Fukugita, Hogan, & Peebles 1998). Although earlier estimates gave slightly lower fractions for stars in spheroids vs. disks (e.g. Schechter & Dressler 1987; Persic & Salucci 1992), it is now generally accepted that the majority of stars belong to spheroids. Figure 1 shows separately the mass functions of color-selected ETGs and of blue, star-forming galaxies, also based on the SDSS data (Baldry et al. 2004). Above ~ 3 × 10¹⁰ M red-sequence galaxies start to increasingly outnumber blue galaxies by a factor that exceeds 10 above ~ 3 × 10¹¹ M. Figure 2, drawn from the same mass functions, shows the contributions to the total stellar mass by red and blue galaxies in the various mass bins, along with the contributions to the total number of galaxies. Then, ETGs represent only 17% of the total number of galaxies in the sample, but contribute ~ 57% of the total mass. Moreover, 80% of the stellar mass in ETGs belongs to galaxies more massive than ~ 3 × 10¹⁰ M. Dwarfs are sometimes seen as the "building blocks" of galaxies, but at least in the present universe one can not build much with them.

Figure 1. The mass function of local (z ~ 0) early-type (red) galaxies and late-type (blue) galaxies from the Sloan Digital Sky Survey (Baldry et al. 2004). The solid lines represent best fit Schechter functions. For the blue galaxies the sum of two different Schechter functions is required to provide a good fit. Dotted lines show the mass functions from Bell et al. (2003).

Figure 2. The contributions to the total stellar mass and to the number of galaxies by early-type (red) and late-type (blue) galaxies in the various mass bins, as derived from the best-fit mass functions of Baldry et al. (2004) shown in Figure 1. The relative areas are proportional to the contributions of the early- and late-type galaxies to the total stellar mass and to the number of galaxies. (Courtesy of D. Thomas).

With spheroids holding the major share of stellar mass in galaxies, understanding their evolution - from formation to their present state - is central to the galaxy evolution problem in general. Historically, two main scenarios have confronted each other, the so-called Monolithic Collapse model (Eggen, Lynden-Bell & Sandage 1962, Larson 1974, Arimoto & Yoshii 1987, Bressan, Chiosi & Fagotto 1994), and the Hierarchical Merging model (e.g., Toomre 1977; White & Rees 1978). In the former scenario spheroids form at a very early epoch as a result of a global starburst, and then passively evolve to the present. If the local conditions are appropriate, a spheroid can gradually grow a disk by accreting gas from the environment, hence spheroids precede disks. In the merging model, big spheroids result from the mutual disruption of disks in a merging event, hence disks precede spheroids.

The two scenarios appear to sharply contradict each other, but the contradiction has progressively blurred in recent years. Evidence has accumulated that the bulk of stars in spheroids are old, and most likely formed in major merging events. In the hierarchical merging scenario (the only one rooted in a solid cosmological context) successive generations of models have struggled to increase their predicted stellar ages, so as to produce results resembling the opposite scenario. This review will not attempt to trace an history of theoretical efforts to understand the formation and evolution of ellipticals and spheroids. It will rather concentrate on reviewing the accumulating observational evidences coming from the stellar component of these galaxies. Other extremely interesting properties of ellipticals, such as their structural and dynamical properties, their hot gas content, central supermassive black holes, etc., will not be touched in this review, even if they are certainly needed to complete the picture and likely play an important role in the evolution of these galaxies. Also untouched are the internal properties of ETGs, such as color and line-strength gradients indicative of spatial inhomogeneities of the stellar populations across the body of ETGs. A complementary view of ETG formation based on the globular cluster populations of these galaxies is presented in the article by J. Brodie & J. Strader (this volume).

The "concordance cosmology" (_M = 0.3, = 0.7, H_o = 70) is adopted if not explicitly stated otherwise.

¹ Here, as well as through the whole paper, the definition of the specific criteria adopted by the various authors for their sample selections can be found in the original articles. Back.