Next Contents Previous


This section ties together our standard picture of galaxy formation by hierarchical clustering (lectures by Shlosman, Scoville, Calzetti) with our School's subject of secular evolution (lectures by Athanassoula, Binney, Buta, Peletier, van Gorkom and me). For pedagogical reasons, it is useful to introduce this standard picture here. Shlosman (2012) provides more detail.

What is at stake for future work? We have a formation paradigm: quantum density fluctuations in non-baryonic dark matter form immediately after the Big Bang and then get stretched by the expansion of the Universe; gravity drives hierarchical clustering that causes the fluctuations to grow, separate, collapse and form galaxy halos; the baryons cool inside the halos to form stars and visible galaxies. Spirals form when halos accrete gas that dissipates and forms disks. Ellipticals form when galaxies collide and merge; then dynamical violence scrambles disks into ellipsoidal Es. This picture is well supported by theory and observations. What are the remaining puzzles – the cracks in the paradigm? They are short-cuts to progress.

First, what is understood? Hierarchical clustering of dark matter initial density fluctuations is nowadays calculated in exquisite detail (Fig. 77). Results are in excellent agreement with observations of large-scale structure.

Figure 77

Figure 77. The Millennium Simulation (Springel et al. 2005) is the iconic example of an n-body calculation of the formation of large-scale and galaxy-sized structures via hierarchical clustering of primordial quantum density fluctuations that have been stretched by the expansion of the Universe and increased in contrast by self-gravity.

Our job is illustrated in Fig. 78. Hierarchical clustering of dark matter is well understood (background image). More tricky is the physics of baryonic galaxy formation within dark halos. It is possible that all remaining problems with our formation picture on galaxy scales are problems of baryonic physics.

Figure 78

Figure 78. Theme of building a comprehensive picture of galaxy formation by studying the physics of baryonic galaxies as embedded in the dark matter hierarchy represented here by the Millennium Simulation. High-density environments are dominated by merger remnants – giant ellipticals in rich clusters. We understand them fairly well. Low-density environments are dominated by pure-disk galaxies such as M 101; we do not understand how they form. Bulge-dominated spirals like the Sombrero live in intermediate environments. Barred and other galaxies that undergo secular evolution also tend to live in intermediate-density environments.

The good news is that ellipticals are fairly well understood. Simulations of the hierarchical growth of galaxies suggest that they change back and forth between spiral and elliptical, depending on whether their recent history was dominated by a major merger or by cold gas dissipation (Fig. 79). Today's students did not live through the revolution in our understanding that resulted from Toomre's (1977a) introduction of mergers to our lexicon of galaxy formation. I therefore review this subject briefly.

Figure 79

Figure 79. Example of the evolution of a single galaxy by hierarchical clustering (Steinmetz & Navarro 2002, 2003). Colors denote old stars, young stars < 200 Myr old and cold gas (see keys). Scale bars are 5 kpc in all panels. Panel (a) shows the most massive progenitor at z = 4; it already contains both old stars and a gas disk. In panel (b), a classical bulge forms in a major merger at z ≃ 3 and then regrows a disk by later infall of cold gas. Panel (c): at z = 1.8, the galaxy looks like an early-type spiral with a dense bulge surrounded by a young disk. Panel (d): At z ≃ 1.6, tidal forcing by a companion shown in the z = 1.2 image triggers a bar. The satellite is accreted at z = 1.18, but the bar prominent in the young component survives for several more Gyr. Panel (e): At z = 0.7, the galaxy merges with another galaxy that has about half of its mass. The result is an elliptical galaxy at z = 0.27. This could accrete more gas and form a Sombrero-galaxy-like system, but it cannot get rid of its large bulge.

The first bad news is that we do not know how to form bulgeless galaxies. Continued bulge growth is inherent to the story in Fig. 79. Once you have a bulge, you cannot get rid of it. This problem was reviewed in Section 6.1.

The other bad news is that galaxy formation by hierarchical clustering of cold dark matter still has problems on the size scales of individual galaxies. Reviews of the subject run the gamut from very optimistic (Primack 2004) to sober (Silk & Mamon 2012) to very pessimistic (Kroupa 2012). This is a sign of a subject in flux – of cracks in the paradigm. These are opportunities.

The comfortable news is that we have a growing understanding of secular evolution in disk galaxies. It happens now in low- and intermediate-density environments. But finding pseudobulges in Virgo S0s shows that it had time to happen even in the progenitor environments of some present-day clusters.

8.1. Formation of ellipticals by major galaxy mergers

It is hard to describe to today's students what a revelation it was when Alar Toomre (1977a) presented his hypothesis that all ellipticals are created from progenitor disk galaxies by the dynamical violence of galaxy collisions and mergers. The feeling had been that stars are too small to collide, so interacting galaxies merely pass through each other. We missed two points. First, tidal effects are easily strong enough to scramble cold, rotating disks into dynamically hot ellipticals. Doing this work takes energy out of the orbits and soon causes the galaxies to merge. Second, large and massive halos of dark matter surround visible galaxies and give them much bigger collision cross sections than we thought when we saw only the visible stars. It is no accident that the merger picture became established soon after we realized that dark matter is real (Faber & Gallagher 1979). Mergers-in-progress turn out to explain a whole zoo of previously mysterious peculiar galaxies (e. g., Figs. 8082). Toomre's suggestion that mergers make all ellipticals is, as far as we know, exactly correct. And Toomre's (1980) additional hypothesis that mergers make classical bulges robustly looks to be exactly correct, too.

Figure 80

Figure 80. Formation of peculiar galaxies such as NGC 4676 ("the mice") by ongoing gravitational encounters. Such encounters explain most objects in (e. g.) Arp's (1966) Atlas of Peculiar Galaxies. (top) Hubble Heritage image. (bottom left) Our view of the initial conditions and (bottom right) the moment when the configuration matches the galaxies for an n-body simulation (Barnes 1998, 2004) of two infinitely thin disks (blue particles) embedded in spherical dark halos (red particles)./td>

I was in the audience for Toomre's (1977a) talk; we all remember it as a historic moment. But not everyone was immediately captured by the new ideas, and I confess that I was a partial agnostic for longer than most. Kormendy (1989) was a late paper that tried to keep merger enthusiasm from getting out of control. Its point is still correct – the sequence of increasing density in smaller ellipticals (Fig. 68) is a sequence of increasing dissipation with accompanying starbursts during formation. As emphasized all along by Toomre, merger progenitors generally contain gas; crunching gas likes to make stars, and so star formation is an integral part of spiral-spiral mergers.

By ~ 1990, we understood that ultraluminous IRAS galaxies ("ULIRGS") are prototypical dissipative mergers in progress (Joseph & Wright 1985; Sanders et al. 1988a, b; Sanders & Mirabel 1996; Rigopoulou et al. 1999; Dasyra et al. 2006a, b; see Dasyra et al. 2006c and KFCB Section 12.3.2 for reviews). Figure 81 shows the most famous example. Dust-shrouded starbursts generally dominate the far-infrared luminosity L > 1012 L (see Joseph 1999 and KFCB for reviews). Their structural parameters are consistent with the E fundamental plane (Kormendy & Sanders 1992; Doyon et al. 1994; Genzel et al. 2001; Tacconi et al. 2002; Veilleux et al. 2006; Dasyra et al. 2006a, b). Stellar velocity dispersions σ ≃ 100 to 230 km s-1 show that local ULIRGs are making moderate-luminosity ellipticals; i. e., the disky-coreless side of the E – E dichotomy in Figs. 4 and 60 (Genzel et al. 2001; Tacconi et al. 2002; Dasyra et al. 2006a, b; c).

Figure 81

Figure 81. Arp 220, the prototypical Ultraluminous Infrared Galaxy, an elliptical galaxy being formed by a dissipative merger accompanied by a dust-shrouded starburst. At left is a Hubble Heritage image. The HST NICMOS JHK image at right reveals two remnant nuclei separated by 0.98" ≃ 360 pc (Scoville et al. 1998).

Central to the rapid acceptance of the merger picture was the extensive observational evidence for mergers in progress that was published by François Schweizer (e. g., 1978, 1980, 1982, 1987, 1996). Figure 82 shows the particularly convincing example if NGC 7252 (the "atoms for peace" galaxy). It has both diagnostic tidal tails and "ripples" or "shells" in the light distribution that trace edge-on caustics of wrapped former disks. Shells, too, became standard merger diagnostics (Malin & Carter 1980, 1983; Schweizer & Seitzer 1988). Shells are seen in absorption as well as in stars; an example of the explanation of a previously mysterious peculiar galaxy as an S0 that contains an accreted and now phase-wrapped dust disk is NGC 4753 (Steiman-Cameron et al. 1992). The correlation that stronger fine structure such as shells is seen in ellipticals with younger stellar populations (bluer colors and stronger Hβ absorption lines) further supported the merger picture and filled in the evolution time sequence between mergers in progress and old, completely relaxed and phase-mixed ellipticals (Carter et al. 1988; Schweizer et al. 1990; Schweizer & Seitzer 1992). Compelling further support was provided both by detailed H i observations (e. g., Hibbard et al. 1994, 1995, 1996, 2001a, b – see and especially by n-body simulations (the master of the art is Josh Barnes 1988, 1989, 1992). Simulations further confirmed that mergers dump huge amounts of gas to galaxy centers, thereby feeding starbursts (e. g., Barnes & Hernquist 1991, 1992, 1996; Mihos & Hernquist 1994; Hopkins et al. 2009a). By the time of the reviews of Schweizer (1990, 1998); Barnes & Hernquist (1992), Kennicutt (1998c) and Barnes (1998), the merger revolution in our understanding of elliptical galaxies was a "done deal".

Figure 82

Figure 82. NGC 7252 is another prototypical merger-in progress that is making an elliptical galaxy (Schweizer 1982). Two tidal tails that point in roughly opposite directions and that have opposite velocity differences with respect to the systemic velocity are the simplest diagnostic signature of a merger in progress (see the seminal paper by Toomre & Toomre 1972). The essential point is that dynamical clocks run most slowly at the largest radii, so remnant tidal tails persist long after the main bodies of interacting galaxies have merged. For H i observations and n-body models of NGC 7252, see Hibbard et al. (1994, 1995, 1996).

A variant of the merger picture involves the observation that many high-z galaxies are dominated by 108 – 109 M, kpc-size star forming clumps (e. g., Elmegreen et al. 2005, 2007, 2008a, 2009a, b; Bournaud et al. 2007; Genzel et al. 2008; Förster Schreiber et al. 2009; Tacconi et al. 2010). Bournaud's galaxy UDF 1668 (Fig. 83) is remarkably similar to the initial conditions used by van Albada (1982) to simulate the collapse of lumpy initial conditions. He showed (Fig. 83 here) that relatively gentle collapses produce Sérsic-function profiles with n ≃ 2 – 4 like those in real classical bulges (e. g., Fisher & Drory 2008). This is confirmed in modern n-body merger simulations (e. g., Hopkins et al. 2009a).

Figure 83

Figure 83. Mergers of two galaxies that consist mostly of stars make Sérsic (1968) function remnants with indices n ~ 2 – 4. An early illustration of this is van Albada (1982), whose initial conditions look remarkably similar to the clumpy high-z galaxy UDF 1666 studied by Bournaud et al. (2007). It is an example of the clump instability picture discussed by Elmegreen et al. (2008b). Van Albada's initial conditions were parameterized by the ratio of twice the total kinetic energy to the negative of the potential energy. In virial equilibrium, 2T / W = 1. For smaller values, van Albada found that gentle collapses (2T / W = 0.5) make Sérsic profiles with n < 4, whereas violent collapses (2T / W ltapprox 0.2) make n gtapprox 4. This is a sign that the clumps discussed by Elmegreen et al. (2008b) merge to make classical bulges.

Elmegreen et al. (2008b) model gas-rich galaxy disks in the early Universe and find that they violently form clumps like those observed (Fig. 84). The clumps quickly merge and make a high-Sérsic-index bulge. It rotates slowly. Rotation velocities decrease with increasing distance from the disk plane. So these are classical bulges, and this is a variant on the merger picture. From many colorful conversations with Allan Sandage, I suspect that he would have welcomed this "ELS with lumps" picture (ELS = Eggen et al. 1962).

Figure 84

Figure 84. A variant of the merger picture involves high-z disks that are unstable to the formation of large clumps which quickly merge to form a classical bulge. It has ellipsoidal (not cylindrical) rotation (bottom). From Elmegreen et al. (2008b).

8.2. Mergers and secular evolution both happen in a hierarchically clustering Universe

"It is impossible to remove the problem of galaxy formation from its cosmological context of hierarchical clustering" (Jones 1992). Begun in papers like White & Rees (1978), our picture of hierarchical clustering has reached a remarkable level of sophistication. The Millenniumm Simulation is one example from a vast literature. I want to emphasize again that the merger formation of elliptical galaxies that was discussed in Section 8.1, the secular evolution of disk galaxies that was discussed in Sections 26, and the environmental secular evolution that was discussed in Section 7 all happen within the cosmological context of hierarchical clustering. Much work remains to be done in connecting the story of galaxy formation on the kpc scales of most studies of individual galaxies with the Mpc scales where n-body dark matter simulations are at their best. Current work is dominated by the complicated physics of baryons, including the effects of reionization, dissipation, star formation, energy feedback and active galactic nuclei. We like to think that galaxies are mature objects and that our job is to study galaxy evolution to see how they got that way. But ~ 2/3 of the baryons in the Universe do not yet live in galaxies or have not yet cooled and formed stars (e. g., Fukugita et al. 1998; Davé et al. 2001; Read & Trentham 2005). Galaxy formation is much less "finished" than we like to think! Our job is far from finished, too. For all of us students of galaxies, this is good news.


It is a great pleasure to thank Jesús Fálcon-Barroso and Johan Knapen for organizing this 2011 workshop, for inviting me to give the introductory and closing lectures, and for meticulously editing this book. I am also grateful to Jésus for his untiring efforts to make our visit to the Canary Islands such a pleasure. I also especially want to thank IAC students Judit Bakos, Javier Blasco Herrera, Santiago Erroz, Adriana de Lorenzo-Cáceres, Mireia Montes, Agnieszka Rys, José Ramón Sánchez-Gallego and Marja Seidel for their help with the workshop and for their kindness to Mary and to me. And finally, I would like to thank all of the above and all of the students for making this an extraordinarily pleasureable and productive experience. It was a pleasure to meet you all, and I hope to see you again often. Warmest best wishes from Mary and from me for successful and fulfilling careers.

Mary and I very much enjoyed our interactions with the other lecturers, Lia Athanassoula, James Binney, Albert Bosma, Ron Buta, Daniela Calzetti, Reynier Peletier, Nick Scoville, Isaac Shlosman and Jacqueline van Gorkom.

These lectures were prepared, delivered and in large part written up during two visits to the Max-Planck-Institut für Extraterrestrische Physik, Garching-bei-München, Germany and the Observatory of the Ludwig-Maximilians-Universität, Munich, Germany. It is a great pleasure to thank Managing Director Ralf Bender and the staff of both institutes for their wonderful hospitality and financial support. I also warmly thank Christa Ingram and Bettina Niebisch of the Max-Planck-Institut für Extraterrestrische Physik for transcribing my oral lectures. And I sincerely thank the many people who allowed me to reproduce figures (see captions).

As with all my papers, I owe a huge debt of gratitude to Mary Kormendy for her editorial help and for her patience and understanding during the preparation of this review.

This work makes extensive use of data products from the digital image database of the Sloan Digital Sky Survey (SDSS). Funding for SDSS and SDSS-II was provided by the Alfred P. Sloan Foundation, the Participating Institutions, the National Science Foundation, the US Dept. of Energy, the National Aeronautics and Space Administration, the Japanese Monbukagakusho, the Max Planck Society and the Higher Education Funding Council for England. The SDSS is managed by the Astrophysical Research Consortium for the Participating Institutions. They are the American Museum of Natural History, Astrophysical Institute Potsdam, University of Basel, University of Cambridge, Case Western Reserve University, University of Chicago, Drexel University, Fermilab, the Institute for Advanced Study, the Japan Participation Group, Johns Hopkins University, the Joint Institute for Nuclear Astrophysics, the Kavli Institute for Particle Astrophysics and Cosmology, the Korean Scientist Group, the Chinese Academy of Sciences (LAMOST), Los Alamos National Laboratory, the Max-Planck-Institute for Astronomy, the Max-Planck-Institute for Astrophysics, New Mexico State University, Ohio State University, University of Pittsburgh, University of Portsmouth, Princeton University, United States Naval Observatory and the University of Washington.

This research depended critically on extensive use of NASA's Astrophysics Data System bibliographic services. I also made extensive use of the NASA/IPAC Extragalactic Database (NED), which is operated by the Jet Propulsion Laboratory and the California Institute of Technology under contract with NASA. And I used the HyperLeda electronic database (Paturel et al. 2003) at and the image display tool SAOImage DS9 developed by the Smithsonian Astrophysical Observatory.

My work on secular evolution is supported by NSF grant AST-0607490 and by the Curtis T. Vaughan, Jr. Centennial Chair in Astronomy.

Note added November 2013: This review was finished and submitted in mid-June 2012. The astro-ph version is essentially identical to the published paper except for differences in spelling (British there and American here) and for improvements here in a few figures. The present version remains up to date except for one important new paper by Sellwood (2013). This is a broad review of the theory of "Secular Evolution in Disk Galaxies" that overlaps some of the present subjects but that also discusses many additional fundamental topics. It is an important complement to the present paper and to Kormendy & Kennicutt (2004).

Next Contents Previous