This paper explores the possibility that dark matter particles are objects with a set of excitable levels, as for a string or drum. A change in excitation is associated with the absorption or emission of a photon, and the probability of such a change is dependent on the ambient photon energy density. Particle mass comes entirely from excitations. Such dark matter with excitable levels, or DMWEL, particles have a distribution of masses which freezes out to its present-day form during the early universe. In the particular model considered, consistency with Planck CMB anisotropy observations implies a cautious lower bound on the energy of the first excitable level of 2 to 200 eV for a set of reference parameters. While laboratory detection looks to require a considerable increase in the power of controlled light-sources, it might be possible in the near future to detect or constrain DMWEL further through improved anisotropy observations, closer determination of the CMB's maximum deviation from a blackbody shape, or from gravitational probes of the dark matter power spectrum. For certain parameters near to the boundary of the region allowed by CMB anisotropies, DMWEL dark matter is cold, but with a cut-off halo mass many orders of magnitude greater than that typical for WIMP cold dark matter.

This paper presents a kinetic theory model of gravitational collapse due to a small perturbation. Solving the relevant equations yields a pattern of entropy destruction in a spherical core around the perturbation, and entropy creation in a surrounding halo. This indicates collisional "de-relaxation" in the core, and collisional relaxation in the halo. Core-halo patterns are ubiquitous in the astrophysics of gravitational collapse, and are found here without any of the prior assumptions of such a pattern usually made in analytical models. Motivated by this analysis, the paper outlines a possible scheme for identifying structure formation in a set of observations or a simulation. This scheme involves a choice of coarse-graining scale appropriate to the structure under consideration, and might aid exploration of hierarchical structure formation, supplementing the usual density-based methods for highlighting astrophysical and cosmological structure at various scales.

We introduce a double power series method for finding approximate analytical solutions for systems of differential equations commonly found in cosmological perturbation theory. The method was set out, in a non-cosmological context, by Feshchenko, Shkil' and Nikolenko (FSN) in 1966, and is applicable to cases where perturbations are on sub-horizon scales. The FSN method is essentially an extension of the well known Wentzel-Kramers-Brillouin (WKB) method for finding approximate analytical solutions for ordinary differential equations. The FSN method we use is applicable well beyond perturbation theory to solve systems of ordinary differential equations, linear in the derivatives, that also depend on a small parameter, which here we take to be related to the inverse wave-number. We use the FSN method to find new approximate oscillating solutions in linear order cosmological perturbation theory for a flat radiation-matter universe. Together with this model's well known growing and decaying M\'esz\'aros solutions, these oscillating modes provide a complete set of sub-horizon approximations for the metric potential, radiation and matter perturbations. Comparison with numerical solutions of the perturbation equations shows that our approximations can be made accurate to within a typical error of 1%, or better. We also set out a heuristic method for error estimation. A Mathematica notebook which implements the double power series method is made available online.