We present an Analytic Model of Intergalactic-medium and GAlaxy evolution since the dark ages. AMIGA is in the spirit of the popular semi-analytic models of galaxy formation, although it does not use halo merger trees but interpolates halo properties in grids that are progressively built. This strategy is less memory-demanding and allows one to start the modeling at redshifts high enough and halo masses low enough to have trivial boundary conditions. The number of free parameters is minimized by making the causal connection between physical processes usually treated as independent from each other, which leads to more reliable predictions. But the strongest points of AMIGA are: i) the inclusion of molecular cooling and metal-poor, population III (Pop III) stars, with the most dramatic feedback, and ii) the accurate follow-up of the temperature and volume filling factor of neutral, singly, and doubly ionized regions, taking into account the distinct halo mass functions in those environments. We find the following general results. Massive Pop III stars determine the IGM metallicity and temperature, and the growth of spheroids and disks is self-regulated by that of massive black holes developed from the remnants of those stars. Yet, the properties of normal galaxies and active galactic nuclei appear to be quite insensitive to Pop III star properties owing to the much higher yield of ordinary stars compared to Pop III stars and the dramatic growth of MBHs when normal galaxies begin to develop, which cause the memory loss of the initial conditions.

Comparing the excursion set and CUSP formalisms for the derivation of the halo mass function, we investigate the role of the mass definition in the properties of the multiplicity function of cold dark matter (CDM) haloes. We show that the density profile for haloes formed from triaxial peaks that undergo ellipsoidal collapse and virialisation is such that the ratio between the mean inner density and the outer local density is essentially independent of mass. This causes that, for suited values of the spherical overdensity $\Delta$ and the linking length $b$, SO and FoF masses are essentially equivalent to each other and the respective multiplicity functions are essentially the same. The overdensity for haloes having undergone ellipsoidal collapse is the same as if they had formed according to the spherical top-hat model, which leads to a value of $b$ corresponding to the usual virial overdensity, $\Delta_{vir}$, equal to $\sim$ 0.2. The multiplicity function resulting from such mass definitions, expressed as a function of the top-hat height for spherical collapse, is very approximately universal in all CDM cosmologies. The reason for this is that, for such mass definitions, the top-hat density contrast for ellipsoidal collapse and virialisation is close to a universal value, equal to $\sim$ 0.9 times the usual top-hat density contrast for spherical collapse.

We establish a one-to-one correspondence between virialized haloes and their seeds, namely peaks with a given density contrast at appropriate Gaussian-filtering radii, in the initial Gaussian random density field. This fixes a rigorous formalism for the analytic derivation of halo properties from the linear power spectrum of density perturbations in any hierarchical cosmology. The typical spherically averaged density profile and mass function of haloes so obtained match those found in numerical simulations.