Open clusters offer a unique possibility to study the time evolution of the radial metallicity gradients of several elements in our Galaxy, because they span large intervals in age and Galactocentric distance, and both quantities can be more accurately derived than for field stars.

We re-address the issue of the Galactic metallicity gradient and its time evolution by comparing the empirical gradients traced by a sample of 45�open clusters with a chemical evolution model of the Galaxy.

At variance with previous similar studies, we have collected from the literature only abundances derived from high-resolution spectra. The clusters have Galactocentric distances $7 \la R_{\rm GC} \la 22$�kpc and ages from ~30�Myr to 11�Gyr. We also consider the α-elements Si, Ca, Ti, and the iron-peak elements�Cr and�Ni. Cepheids trace instead the present-day Fe�gradient in the inner parts of the disk.

The data for iron-peak and α-elements indicate a steep metallicity gradient for $R_{\rm GC}\la 12$�kpc and a plateau at larger radii. The time evolution of the metallicity distribution is characterized by a uniform increase of the metallicity at all radii, preserving the shape of the gradient, with marginal evidence for a flattening of the gradient with time in the radial range 7-12�kpc. Our model is able to reproduce the main features of the metallicity gradient and its evolution with an infall law exponentially decreasing with radius and with a collapse time scale of the order of 8�Gyr at the solar radius. This results in a rapid collapse in the inner regions, i.e. $R_{\rm GC}\la 12$�kpc (that we associate with an early phase of disk formation from the collapse of the halo) and in a slow inflow of material per unit area in the outer regions at a constant rate with time (that we associate with accretion from the intergalactic medium). An additional uniform inflow per unit disk area would help to better reproduce the metallicity plateau at large Galactocentric radii, but it is difficult to reconcile with the present-day radial behaviour of the star formation rate.

Our results favour a scenario where the Galactic disk is formed inside-out by the rapid collapse of the halo and by a subsequent continuous accretion of intergalactic gas