If the damped Ly- systems are the progenitors of galactic disks, then our findings restrict scenarios suggested for disk formation. For example if the disk-like structure indicated in the case of PKS0458-02 is a generic feature of the damped systems, then their detection out to z = 2.8 indicates that galactic disks have formed by t (0.9 - 2.1) h-1 Gyr. This pushes the epoch of disk formation back to the time when globular clusters first formed, and consequently argues against a significant age gap (Mihalas and Binney 1981) between the disk and the halo. Furthermore, the early formation of disks restricts models in which disks form slowly, on a Hubble time scale (Gunn 1982b).
If, in addition, the radii of the high-z disks are large compared to RHo, then a natural sequence of events is one in which the protogalaxy first collapses to a giant disk which subsequently contracts in the plane. Schiano, Wolfe and Chang (1990) recently ran a series of hydrodynamical calculations to check the plausibility of this scheme. The protogalaxy was simulated by axially symmetric spheroids of gas and dark matter in which the ratio of masses, Mdm / Mgas 10. The gas and the dark matter obtain their spin from tidal torques induced by neighboring galaxies (Peebles 1969; Fall and Efstathiou 1980). Initially both spheroids are hydrostatically supported by the random motions of isothermal density distributions which fall off like r-2 beyond the core radius, rc. The gas collapses because it cools on a time-scale short compared to the free-fall time. Figure 3 shows the density contours in a meridional slice of the spheroid at early stages of collapse. The important point is that collapse to a large disk is possible by z 3 in most cosmological models.
Figure 3. Logarithmic density contours (cgs units) and velocity fields in the r-z plane for a model with rc = 3 kpc. Elapsed times are (a) 0.30, (b) 0.49, (c) 0.70, (d) 1.03 Gyr.
In the second phase of the calculation we examined the contraction of the centrifugally supported disk that formed out of the initial collapse shown in Figure 3. We considered a mechanism which is a dynamical consequence of the infall of mass shed by halo stars (Ostriker and Thuan 1975). When the infall increases the mass of the disk by a factor, f, the radius of the disk must contract by the same factor: this occurs because the radius, R, of the disk times M(R), the mass within R, is an adiabatic invariant. Our numerical calculations showed that the disk did contract by a factor f in about 1 Gyr, although the contraction was accompanied by axisymmetric instabilities. The rapid shrinkage of the disk implies that the detection rate of damped Ly- systems at z 1.5 should be much less than the rate reported in our survey, even after cosmological effects are taken into account.
This work was supported by NSF grant AST 84144-14.