Before describing the galaxy morphology-redshift relationship, and what it implies for understanding galaxy formation, I will review our current understanding of how galaxy structure correlates with physical properties of galaxies. It has been known for decades that, broadly speaking, galaxy morphology correlates with galaxy properties such as luminosities, sizes, gas content, colors, environment, masses, and mass to light ratios (e.g., Roberts & Haynes 1994). Generally, early-type galaxies (ellipticals) are larger, more massive, contain older stellar populations, and are found in denser areas than spirals. Later type galaxies, such as spirals, are bluer, contain younger stellar populations, more gas, and are less massive overall than the ellipticals. The detailed correlation between these physical properties and Hubble/de Vaucouleur classifications is however not strong. While in the mean properties change with Hubble type, there is significant overlap in any given property across the Hubble sequence 2.
Some quantitative measurements of galaxy structure, on the other hand, correlate strongly with physical properties, including: the current star formation rate, the stellar mass, galaxy radius, central black hole mass, and merging properties. It is impossible to describe all but a few of these correlations here. The first of these discovered is that the light concentration of an evolved stellar population correlates with its luminosity, stellar mass, and scale (e.g., Caon, Capaccioli & D'Onofrio 1993; Graham et al. 1996; Bershady et al. 2000; Conselice 2003). This was first noticed by the failure of the de Vaucouleurs r1/4 surface brightness profile to fit the surface brightness distributions of elliptical galaxies brighter than or fainter than MB ~ - 20 (e.g., Schombert 1986). It was realized later that the shape of the surface brightness profile for early types correlates strongly with its absolute blue magnitude (Binggeli & Cameron 1991; Caon et al. 1993). It was later shown that the general Sersic profile, with its concentration parameter n, gives a much better fit than the de Vaucouluer's profile for all early types (Graham et al. 1996). The central light concentration, as measured through the Sersic n index, or a non-parametric concentration index (C) (Bershady et al. 2000), also correlates with the mass of central black holes (Graham et al. 2001).
While the concentration of a galaxy's light profile correlates with its stellar mass, absolute magnitude, and size, in a sense revealing the past formation history of a galaxy, there are many indicators in the structures of galaxies for ongoing galaxy formation. For example, recently it has been shown that the clumpiness of a galaxy's light distribution correlates with the amount and the location of star formation (e.g., Takamiya 1999; Conselice 2003). The clumpiness is measured by quantifying the fraction of a galaxy's light in the rest-frame B-band in high spatial frequency structures. The ratio between the amount of light in these high spatial frequency structures and the total light gives a measure of the clumpiness, or star formation. This trend can be demonstrated by the strong correlation between the clumpiness index S (Conselice 2003) and H equivalent widths and colors of star forming galaxies. There is also a strong relationship between the dynamical state of a galaxy and the presence of a merger. Generally, merging galaxies are asymmetric, while non-mergers are not (Conselice et al. 2000a, b; Conselice 2003). This has been shown in numerous ways, including empirical methods (Conselice 2003) and the correlation of internal HI dynamics and asymmetries of stellar distributions (Conselice et al. 2000b).
As shown in Conselice et al. (2000a) asymmetric light distributions can also be caused by star formation. However, by decomposing light and kinematic structures in galaxies, it is possible to show that primary asymmetries are not the result of star formation, which forms in clumps (e.g., Elmegreen 2002), but from large scale lopsidedness on the order of the size of the galaxy itself (Andersen et al. 2001). Likewise, there is a strong correlation between the asymmetry parameter and the clumpiness parameter for normal star forming galaxies (Conselice 2003; Figure 1). Galaxies with high clumpiness values (S), which correlates with high amounts of star formation, have correspondingly higher asymmetry values. However, this correlation breaks down for systems involved in major mergers, such as nearby ultraluminous infrared galaxies (Figure 1). The nature of this deviation is such that a galaxy undergoing a merger has too high an asymmetry for its clumpiness, demonstrating that large asymmetries are produced in large scale features and not in clumpy, star formation like regions (Conselice 2003; Mobasher et al. 2004; Figure 1) 3.
Figure 1. The relationship between the asymmetry index (A) and the clumpiness index (S). Higher A and S values indicate galaxies that are more asymmetric and have a higher fraction of clumpy light, respectively. For normal galaxies (black squares, left panel) there is a strong correlation between A and S such that A = (0.35 ± 0.03) × S + (0.02 ± 0.01). The galaxies which deviate from this relationship are the ongoing major mergers, shown in the left panel as open circles. The right panel shows the deviation from the A - S relationship in units of the scatter of the asymmetry values of the normal galaxies (). Generally, only the mergers deviate from this relationship by more than 3. For a physical reasoning behind these correlations see the text and Conselice (2003).
2 Room prohibits a detailed discussion of all the problems with Hubble classifications. The fact that physical properties only correlate in the mean for a given Hubble type is only one of many issues. For a detailed discussion of this see Appendix A from Conselice (2003). It is still useful to separate in the broadest sense, ellipticals from spirals, as I do here, as these are fundamentally different galaxy types. Back.
3 The concentration index (C), asymmetry index (A), and clumpiness index (S) form the CAS morphology system described in detail in Conselice (2003). With these three parameters the major classes of nearby galaxies can be distinguished. Although this idea is not explicitly described in this review, it is used in papers described here, such as Conselice et al. (2004b) and Mobasher et al. (2004) to determine galaxy types. Back.