A phase-space distribution function of the steady state in galaxy models that admits regular orbits overall in the phase-space can be represented by a function of three action variables. This type of distribution function in Galactic models is often constructed theoretically for comparison of the Galactic models with observational data as a test of the models. On the other hand, observations give Cartesian phase-space coordinates of stars. Therefore it is necessary to relate action variables and Cartesian coordinates in investigating whether the distribution function constructed in galaxy models can explain observational data. Generating functions are very useful in practice for this purpose, because calculations of relations between action variables and Cartesian coordinates by generating functions do not require a lot of computational time or computer memory in comparison with direct numerical integration calculations of stellar orbits. Here, we propose a new method called a torus-fitting method, by which a generating function is derived numerically for models of the Galactic potential in which almost all orbits are regular. We confirmed the torus-fitting method can be applied to major orbit families (box and loop orbits) in some two-dimensional potentials. Furthermore, the torus-fitting method is still applicable to resonant orbit families, besides major orbit families. Hence the torus-fitting method is useful for analyzing real Galactic systems in which a lot of resonant orbit families might exist.

We discuss the possibility to probe the QCD critical point during the dynamical black hole formation from a gravitational collapse of a massive star, where the temperature and the baryon chemical potential become as high as T ~ 90 MeV and $\mu_B$ ~ 1300 MeV. Comparison with the phase boundary in chiral effective models suggests that quark matter is likely to be formed before the horizon is formed. Furthermore, the QCD critical point may be probed during the black hole formation. The critical point is found to move in the lower temperature direction in asymmetric nuclear matter, and in some of the chiral models it is found to be in the reachable region during the black hole formation processes.

The neutron capture cross section of the unstable s-process branching nucleus 185W has been derived from experimental data of the inverse 186W(g,n)185W photodisintegration taken with monochromatic photon beams from laser Compton scattering. The result of sigma = 553 +- 60 mb at kT = 30 keV leads to a relatively high effective neutron density in the classical s-process of N_n = 4.7 \times 10^8 cm^-3. A realistic model for the s-process in thermally pulsing AGB stars overestimates the abundance of 186Os significantly because of the relatively small neutron capture cross section of 185W.

Evolution of the statistical distribution of density field is investigated by means of a counts-in-cells method in a low-density cold-dark-matter simulated universe. Four theoretical distributions, i.e. the negative binomial distribution, the lognormal distribution, the Edgeworth series and the skewed lognormal distribution, are tested to fit the calculated distribution function, and it is shown that only the skewed lognormal distribution of second and third order can describe the evolution of the statistical distribution perfectly well from the initially Gaussian regime to the present stage. The effects of sparse sampling is also investigated and it is discussed that one should use a sample with number density of galaxies larger than $\sim 0.01h^3$Mpc$^{-3}$ in order to recover underlying density distribution.