Gas accretion is a vital support for galaxy evolution and the feeding of star formation. In recent years, the study of gaseous haloes surrounding disk galaxies has shown the presence of gas complexes, analogous to the galactic High-Velocity Clouds, that can be direct evidence of gas accretion. However, the accretion rates estimated from these features consistently gives values, which are one order of magnitude lower than what is needed to feed the star formation. This problem can be overcome if most of the accretion is ÒhiddenÓ and visible only indirectly through the effects it has on the kinematics of the halo gas. In this second scheme, the gas expelled from the disk through galactic fountains sweeps up ambient gas causing it to accrete. This model provides an explanation for the missing gas accretion and also reproduces the peculiar kinematics of the halo gas, in particular the vertical rotation gradient. I discuss these issues in relation to available data and to new investigations to be performed with current and future facilities.