The resolving powers of existing telescope arrays are quite different. For example the extended new MERLIN at 5 GHz readily yields maps of ~ 50 mas resolution compared with the VLA ~ 350 mas, WSRT ~ 4000 mas, EVN ~ 5 mas, global VLBI ~ 1 mas, etc. This means that images obtained with these instruments yield information on different scales that sometimes appear unrelated. However, the need to have structural information on intermediate scales demands that data from different arrays be combined to make a single image. This is particularly important in extra-galactic radio sources where the relationship between core-jet features on different scales (parsec and kiloparsec) and the connection between small-scale and extended features need to be established.
By combining uvdata from two or more arrays we can routinely make images that represent a balance between the capabilities of the combined arrays, (e.g. VLA + MERLIN, MERLIN+VLBI, VLA+WSRT etc) at cm wavelengths. In order to do this, one has to overcome a number of calibration and mapping difficulties (see Zhang et al. 1991; Akujor et al. 1992, in prep.).