Recent observations have shown that the intergalactic medium (IGM) is more transparent to Lyalpha photons close to Lyman Break Galaxies (LBGs) than at large distance from them, ie a proximity effect. Cosmological simulations including winds from LBGs have been so far unable to explain this trend. By coupling such simulations with the radiative transfer code CRASH, we investigate whether the addition of the ionizing radiation emitted by LBGs can increase the transmissivity by decreasing the neutral hydrogen fraction in the inner Mpc of the galaxy halo. The transmissivity as a function of distance is roughly reproduced only if LBGs are identified with dwarf galaxies (with masses < 10^9 solar masses), which are undergoing a vigorous (50 solar masses/yr) burst of star formation. Similar star formation rates in larger galaxies are not sufficient to overwhelm the large recombination rates associated with their denser environment. If so, photoionization partly reconciles theory with observations, although we discuss a number of uncertainties affecting both approaches.

Motivated by the relative lack of neutral hydrogen around Lyman Break Galaxies deduced from recent observations, we investigate the properties of the Lyalpha forest around high redshift galaxies. The study is based on improved numerical SPH simulations implementing, in addition to standard processes, a new scheme for multiphase and outflow physics description. Although on large scales our simulations reproduce a number of statistical properties of the IGM (because of the small filling factor of shock-heated gas), they underpredict the Lyalpha optical depth decrease inside 1 Mpc/h of the galaxies by a factor of ~2. We interpret this result as due to the combined effect of infall occurring along the filaments, which prevents efficient halo gas clearing by the outflow, and the insufficient increase of (collisional) hydrogen ionization produced by the temperature increase inside the hot, outflow-carved bubble. Unless an observational selection bias is present, we speculate that local photoionization could be the only viable explanation to solve the puzzle.

We investigate a new implementation of the Smoothed Particle Hydrodynamics technique (SPH) designed to improve the realism with which galaxy formation can be simulated. In situations where cooling leads to the coexistence of phases of very different density and temperature, our method substantially reduces artificial overcooling near phase boundaries, prevents the exclusion of hot gas from the vicinity of cold ``clouds'', and allows relative motion of the two phases at each point. We demonstrate the numerical stability of our scheme in the presence of extremely steep density and temperature gradients, as well as in strong accretion shocks and cooling flows. In addition, we present new implementations of star formation and feedback which simulate the effect of energy injection into multiphase gas more successfully than previous schemes. Our feedback recipes deposit thermal energy separately in cold dense gas and hot diffuse gas, and can explicitly reinject cold gas into the hot phase. They make it possible to damp star formation effectively, to reheat cold gas, and to drive outflows into the galaxy halo and beyond. We show feedback effects to be strongest in small mass objects where much of the gas can be expelled. After idealised tests, we carry out a first low resolution study of galaxy formation in a $\Lambda$CDM universe. Feedback results in substantial and mass-dependent reductions in the total baryonic mass gathered onto the final object as well as in significant modulation of the star formation history.

We study the evolution of ionization fronts around the first proto-galaxies by using high resolution numerical cosmological (Lambda+CDM model) simulations and Monte Carlo radiative transfer methods. We present the numerical scheme in detail and show the results of test runs from which we conclude that the scheme is both fast and accurate. As an example of interesting cosmological application, we study the reionization produced by a stellar source of total mass M=2 10^8 M_\odot turning on at z=12, located at a node of the cosmic web. The study includes a Spectral Energy Distribution of a zero-metallicity stellar population, and two Initial Mass Functions (Salpeter/Larson). The expansion of the I-front is followed as it breaks out from the galaxy and it is channeled by the filaments into the voids, assuming, in a 2D representation, a characteristic butterfly shape. The ionization evolution is very well tracked by our scheme, as realized by the correct treatment of the channeling and shadowing effects due to overdensities. We confirm previous claims that both the shape of the IMF and the ionizing power metallicity dependence are important to correctly determine the reionization of the universe.

We derive the expected Type II SN differential number counts, N(m), and Hubble diagram for SCDM and LCDM cosmological models, taking into account the effects of gravitational lensing (GL) produced by the intervening cosmological mass. The mass distribution of dark matter halos (ie the lenses) is obtained by means of a Monte Carlo method applied to the Press-Schechter mass function. The halos are assumed to have a NFW density profile, in agreement with recent simulations of hierarchical cosmological models. Up to z=15, the (SCDM, LCDM) models predict a total number of (857, 3656) SNII/yr in 100 surveyed 4' times 4' fields of the Next Generation Space Telescope. NGST will be able to reach the peak of the N(m) curve, located at AB approx 30(31) for SCDM(LCDM) in J and K wavelength bands and detect (75%, 51%) of the above SN events. This will allow a detailed study of the early cosmic star formation history, as traced by SNIIe. N(m) is only very mildly affected by the inclusion of lensing effects. In addition, GL introduces a moderate uncertainty in the determination of cosmological parameters from Hubble diagrams, when these are pushed to higher $z$. For example, for a ``true'' LCDM with (Omega_M= 0.4, Omega_Lambda=0.6), without proper account of GL, one would instead derive (Omega_{M}=0.36^{+0.15}_{-0.12}, Omega_{Lambda}=0.60^{+0.12}_{-0.24}). We briefly compare our results with previous similar work and discuss the limitations of the model.

Feedback effects due to massive stars and supernovae in the first objects are shown to strongly regulate both galaxy formation/evolution and the reionization process. Here we review the most important ones in some detail. We discuss how Type II supernovae can be used as tracers of the first objects and detected with NGST, for which we predict supernova number counts including the effects of gravitational lensing. Preliminary results on the formation of dust in the ejecta of supernovae of primordial composition are also presented. We finally turn to the consideration of the process of inhomogeneous reionization due to primordial stellar sources by means of high resolution numerical simulations, allowing for a self-consistent treatment of the above feedback processes. These simulations allow us to draw conclusions on the evolution and epoch of reionization and about the fate of reionizing objects. We conclude that a large fraction (~ 99%) of collapsed objects must be dark at redshift around eight.

We discuss some aspects concerning the formation and the impact of the first luminous structures in the universe (PopIIIs), with particular emphasis on their feedback effects on subsequent galaxy formation. We argue that supernovae in these objects might provide, if detected, important constraints for cosmological models. Some of these constraints come from the gravitational magnification of the supernova flux due to the intervening cosmological matter distribution which results in different predictions, for example, on rates, magnification probability, detection limits and number counts of distant supernovae. We show that NGST can tremendously contribute to such a study due to its advanced technological capabilities.

We study the gravitational lensing magnification produced by the intervening cosmological matter distribution, as deduced from three different hierarchical models (SCDM, LCDM, CHDM) on very high redshift sources, particularly supernovae in protogalactic (Pop III) objects. By means of ray-shooting numerical simulations we find that caustics are more intense and concentrated in SCDM models. The magnification probability function presents a moderate degree of evolution up to $z\approx 5$ (CHDM) and $z\approx 7$ (SCDM/LCDM). All models predict that statistically large magnifications, $\mu \simgt 20$ are achievable, with a probability of the order of a fraction of percent, the SCDM model being the most efficient magnifier. All cosmologies predict that above $z\approx 4$ there is a 10% chance to get magnifications larger than 3. We have explored the observational perspectives for Pop III SNe detection with NGST when gravitational magnification is taken into account. We find that NGST should be able to detect and confirm spectroscopically Type II SNe up to a redshift of $z \approx 4$ in the J band (for $T_{SN}=25000$ K); this limit could be increased up to $z\approx 9$ in the K band, allowing for a relatively moderate magnification. Possibly promising strategies to discriminate among cosmological models using their GL magnification predictions and very high-$z$ SNe are sketched. Finally, we outline and discuss the limitations of our study.

We review some of the possible observable effects (reionization, feedback on galaxy formation, supernovae and metal enrichment) of the first collapsed, luminous (Pop III) objects in the universe. We show that supernovae in these objects should be considerably magnified by the intervening cosmological matter distribution; the implications of this process are briefly discussed, anticipating some of the results of Marri & Ferrara (1998).