The unusual mushroom-shaped HI cloud, GW 123.4--1.5, is hundreds of parsecs in size but does not show any correlations to HI shells or chimney structures. To investigate the origin and velocity structure of GW 123.4--1.5, we perform three-dimensional hydrodynamical simulations of the collision of a high-velocity cloud with the Galactic disk. We also perform a parameter study of the density, radius, and incident angle of the impact cloud. The numerical experiments indicate that we reproduce the mushroom-shaped structure which resembles GW 123.4--1.5 in shape, size, position-velocity across the cap of the mushroom, and the density ratio between the mushroom and surrounding gas. GW 123.4--1.5 is expected to be formed by the almost head-on collision of a HVC with velocity $\sim 100 \kms$ and mass $\sim 10^5 \Msun$ about $5 \times 10^7 \yr$ ago. A mushroom-shaped structure like GW 123.4--1.5 must be infrequent on the Galactic plane, because the head-on collision which explains the mushroom structure seems rare for observed HVCs. HVC-disk collision explains not only the origin of the mushroom-shaped structure but also the formation of a variety of structures like shells, loops, and vertical structures in our Galaxy.

High-resolution SiO observations of the NGC 1333 IRAS 4A star-forming region showed a highly collimated outflow with a substantial deflection. The deflection was suggested to be caused by the interactions of the outflow and a dense cloud core. To investigate the deflection process of protostellar outflows, we have carried out three-dimensional hydrodynamic simulations of the collision of an outflow with a dense cloud. Assuming a power-law type density distribution of the obstructing cloud, the numerical experiments show that the deflection angle is mainly determined by the impact parameter and the density contrast between the outflow and the cloud. The deflection angle is, however, relatively insensitive to the velocity of the outflow. Using a numerical model with physical conditions that are particularly suitable for the IRAS 4A system, we produce a column-density image and a position-velocity diagram along the outflow, and they are consistent with the observations. Based on our numerical simulations, if we assume that the initial density and the velocity of the outflow are $\sim 10 \cm3$ and $\sim 100 \kms$, the densities of the dense core and ambient medium in the IRAS 4A system are most likely to be $\sim 10^5 \cm3$ and $\sim 10^2 \cm3$, respectively. We therefore demonstrate through numerical simulations that the directional variability of the IRAS 4A outflow can be explained reasonably well using the collision model.

Thermal-gravitational instability (TGI) is studied in the protogalactic environment. We extend our previous work, where we found that dense clumps first form out of hot background gas by thermal instability and later a small fraction of them grow to virialized clouds of mass M_c >~ 6X10^6 M_sun by gravitational infall and merging. But these clouds have large angular momentum, so they would be difficult, if not impossible, to further evolve into globular clusters. In this paper, through three-dimensional hydrodynamic simulations in a uniformly rotating frame, we explore if the Coriolis force due to rotation in protogalactic disk regions can hinder binary merging and reduce angular momentum of the clouds formed. With rotation comparable to the Galactic rotation at the Solar circle, the Coriolis force is smaller than the pressure force during the early thermal instability stage. So the properties of clumps formed by thermal instability are not affected noticeably by rotation, except increased angular momentum. However, during later stage the Coriolis force becomes dominant over the gravity, and hence the further growth to gravitationally bound clouds by gravitational infall and merging is prohibited. Our results show that the Coriolis force effectively destroys the picture of cloud formation via TGI, rather than alleviate the problem of large angular momentum.

We study thermal-gravitational instability in simplified models for protogalactic halos using three-dimensional hydrodynamic simulations. The simulations followed the evolution of gas with radiative cooling down to T = 10^4 K, background heating, and self-gravity. Then cooled and condensed clouds were identified and their physical properties were examined in detail. During early stage clouds start to form around initial density peaks by thermal instability. Small clouds appear first and they are pressure-bound. Subsequently, the clouds grow through compression by the background pressure as well as gravitational infall. During late stage cloud-cloud collisions become important, and clouds grow mostly through gravitational merging. Gravitationally bound clouds with mass M_c > ~6 X 10^6 Msun are found in the late stage. They are approximately in virial equilibrium and have radius R_c = \~150 - 200 pc. Those clouds have gained angular momentum through tidal torque as well as merging, so they have large angular momentum with the spin parameter <lambda_s> ~ 0.3. The clouds formed in a denser background tend to have smaller spin parameters. We discuss briefly the implications of our results on the formation of protoglobular cluster clouds in protogalactic halos. (abridged)

We have performed two-dimensional hydrodynamic simulations of the collapse of isolated axisymmetric clouds condensing via radiative cooling in a primordial background gas. In order to study the development of the so-called ``shape-instability'', we have considered two types of axisymmetric clouds, oblate and prolate clouds of various sizes and with axial ratios of $0.5 \leq {R_{\rm c,R}} /{R_{\rm c,z}} \leq 2$. We find that the degree of oblateness or prolateness is enhanced during the initial cooling phase. But it can be reversed later, if the initial contrast in cooling times between the cloud gas and the background gas is much greater than one. In such cases an oblate cloud collapses to a structure composed of an outer thin disk and a central prolate component. A prolate cloud, on the other hand, becomes a thin cigar-shape structure with a central dense oblate component. The reversal of shape in the central part of the cooled clouds is due to supersonic motions either along the disk plane in the case of oblate clouds or along the symmetry axis in the case of prolate clouds. For a background gas of $T_h=1.7\times 10^6$K and $n_h=0.1 \cm3$ in a protogalactic halo environment, the mean density of the cloud gas that has cooled to $10^4$K increases to $100 n_h$ or so, in our simulations where nonequilibrium cooling is adopted and the background gas cools too. The spherical Jeans mass of such gas is estimated to be about $M_J \sim 5\times10^{7}\Msun$. In order for cloud mass to exceed the Jeans mass and at the same time in order for the thermal instability to operate, the initial cloud size should be around $1 - 1.5 l_{\rm cool}$ where $l_{\rm cool}$ is the cooling length.