5 Conclusions

In this work we have derived the H I-ionising background, resulting from the integrated contribution of QSOs and galaxies, taking into account the opacity of the intervening IGM. We have modelled the IGM with pure-absorbing clouds, with a distribution in column density of neutral hydrogen, $N_H{\sc i}$, and redshift, z, derived from recent observations of the Ly$\alpha$ forest (Kim et al. 2001) and from Lyman Limit systems. The QSOs emissivity has been derived from the recent fits of Boyle et al. (2000), while we have used the stellar population synthesis model of Bruzual & Charlot (2001) and a star-formation history from UV observations for the galaxy emissivity. Due to the present uncertainties in models and observations, we have used three values for the fraction of ionising photons that can escape a galaxy interstellar medium, $f_{\rm esc}=0.05$, 0.1 and 0.4, as suggested by local and high-z UV observations of galaxies, respectively.

The contribution of galaxies to the UV background is found to be comparable or larger than that of QSOs. This is consistent with other determinations of the galactic contribution to the background (Giallongo et al. 1997; Devriendt et al. 1998; Shull et al. 1999). Taking into account (in a rough form) the contribution of reemission from the IGM clouds (Haardt & Madau 1996; Fardal et al. 1998), we find that all models with can provide an ionising background within the limits measured from observations of the proximity effect at $z\sim 3$and not exceeding the upper limit for the local background at z=0.

The analysis of Kim et al. (2001) shows that the break in the evolution of the Ly$\alpha$ forest, ${\rm d}N/{\rm d}z$, occurs at $z\sim 1$. We used the evolution of the ionising background in our model to derive ${\rm d}N/{\rm d}z$, assuming that the formation of structure plays only a minor role (Davé et al. 1999). The rapid decrease of star-formation for can easily explain the observed break, while a QSO-only background would produce a break at $z\sim 2$ (Davé et al. 1999). For the ionising background to have an evolution similar to that of a galaxy-only model, high values of $f_{\rm esc}$ are needed. This result apparently pushes the galaxy contribution in an opposite direction with respect to the estimates of the absolute value of the ionising background. The flat evolution of ${\rm d}N/{\rm d}z$ at z<1 is much better reproduced adopting a $\Lambda$-cosmology ( $\Omega_{\rm m}=0.3$, $\Omega_\Lambda=0.7$) rather than a $\Omega _{\rm m}=1$ Einstein-De Sitter universe. However, such a result needs to be confirmed after properly taking the formation of structure into account.

A significant contribution from galaxies to the ionising metagalactic flux would correspond to a softening of its spectrum with respect to a purely QSO-dominated background. This effect would be particularly important at very high and very low redshift. Observations of the evolution of the Si IV to C IV ratio with redshift (Savaglio et al. 1997; Songaila 1998) seem to confirm the progressive softening of the UV background at z> 3.

To summarise our result, a galaxy-dominated background with is consistent with the estimates of $J(912~\mbox{\AA})$. In the hypothesis that the formation of structures plays a negligible effect, is needed in order to explain the observed ${\rm d}N/{\rm d}z$ of Ly$\alpha$ absorbers. The rapidly improving knowledge derivable from numerical simulations, the determination of the cosmological evolution of the Lyman forest, the proximity effect estimates of the ionising background, and the evolution of the intensity ratios of metal absorption lines will put soon constraints on the relative galaxy/QSO contribution to the UV background and, together with direct measurements of the Lyman continuum emission from galaxies, will make it possible to address the issue of the evolution of the escaping fraction of photons from galaxies as a function of z.

Acknowledgements

We are greatly indebted to F. Haardt (the referee) and P. Madau for providing us with results from their cosmological radiative transfer code prior to publication and for enlightening comments; and to S. Charlot for suggestions and help with the use of the latest version of the GISSEL code. We also acknowledge stimulating discussion with S. D'Odorico, B. Ciardi, A. Ferrara and A. Grazian. This work was partially supported by the Research and Training Network "The Physics of the Intergalactic Medium" set up by the European Community under the contract HPRN-CT2000-00126 RG29185.