Fig. 1: Schematic evolutionary paths of galaxies moving from the blue cloud to the red sequence (Faber et al. 2007), during gas-rich galaxy mergers (solid arrows). Dry galaxy mergers (open arrows) move galaxies along the red sequence and produce slow rotators

Fig. 2: 48 early -type galaxy velocity fields from the SAURON survey, ordered vertically, and from left to right, by increasing λ_R. Two general morphologies are seen: low level rotation, usually with a KDC or twist; and a well-ordered single-axis rotation field. The division of these two groups, which we refer to as Slow Rotators and Fast Rotators', occurs around a value of λ_R=0.1.

3. Limitations of the Original Survey
The galaxies in the original SAURON survey were selected as a representative sample, with a relatively small number of objects, and a wide range of masses and shapes of early-type galaxies. Specifically, galaxies were selected to be uniform in magnitude and ellipticity, but objects on the sky are not at all uniformly distributed in these quantities (Fig. 3). This selection imposes complex biases and makes it impossible to derive any robust statistical distribution of galaxy properties from the original SAURON sample, for comparison with simulations. The simple choice to have an equal number of E and S0 galaxies, although well-motivated at the time, becomes problematic after our finding that the morphological classification is nearly irrelevant and should be replaced by our kinematical one. Moreover, with only 48 galaxies, the statistical uncertainties are prohibitively large.

4. The ATLAS^3D Survey
The power of our kinematic classification based on λ_R is to be able to describe the red-sequence galaxy population in terms of their formation process. For comparison with theoretical predictions one needs to observe λ_R for a statistically significant, volume-limited sample of galaxies complete to some useful lower limit in mass. The main goal of the proposed survey is to determine how the star formation and feedback affect the transformation of galaxies, which move from the blue cloud to the red sequence during merging. This can be done by comparing the global properties of the galaxies in our sample with the predictions of numerical simulations. We will progress along the following main lines of research:

5. The ATLAS^3D Sample
The homogeneous, full-sky 2MASS survey (Skrutskie et al. 2006) allows for a robust sample selection using the K-band luminosity as a proxy for stellar mass. SAURON can measure reliable velocity dispersions in galaxies with M_K < -21.5 mag, and the Fe5270 absorption line (metallicity constraint) for galaxies with D < 42 Mpc. These simple selection criteria lead to a parent sample of 871 galaxies. The ATLAS^3D sample consist of the 260 early-type galaxies, extracted from this sample (Cappellari et al. 2011).