Visual and Near-Infrared Imaging of Ultraluminous Infrared Galaxies: The IRAS 2 Jy Sample

New near-infrared and visual images at 2.2 microns and 6550 A are presented for 46 galaxies having infrared luminosities of L_IR_ > 8.5 x 10^11^ L_sun_, 60 micron flux densities greater than 1.94 Jy, and declinations greater than -35^deg^. These galaxies make up a significant fraction of a complete, northern hemisphere sample of ultraluminous infrared galaxies. Visual and/or near-infrared imaging data now exist for 56 ultraluminous infrared galaxies out to nearly 50000 km s^-1^. Of these 56 galaxies, 53(95%) show evidence for current or past interactions. Among these systems, there are a large variety of visual morphologies, including strongly interacting pairs with apparent tidal tails, as well as single, distorted galaxies with close double nuclei. There are three galaxies which, to the limits of the imaging data, do not appear to have suffered a recent interaction or merger. Approximately 47% (25/53) of the interacting systems have double nuclei, with projected nuclear separations ranging from 0.3 to 48 kpc. Seven systems have nuclear separations larger than 10 kpc. If the 53 interacting galaxies are viewed as stages in the evolution of pairs of interacting spiral galaxies to a single, luminous AGN or starburst, the present imaging data can be used to estimate the lifetime of the bright infrared phase. Including only those sample galaxies with morphological evidence for interactions, we calculate a lower and an upper limit to the lifetime of the ultraluminous infrared phase of the sample as a whole to be 2 x 10^8^ and 2 x 10^9^ yr, respectively. Comparison of these dynamical estimates to models of the mergers of gas-rich galaxies and the rates at which fuel is exhausted by starbursts or AGN suggests the lifetime of the ultraluminous phase lies much closer to the smaller of these two values. Selecting galaxies based upon luminous infrared activity clearly biases the sample towards merging galaxies with small physical separations. However, the existence of pairs with large separations indicates that the ultraluminous phase may in some cases start early during the merger process. Alternatively, these systems may contain unresolved third nuclei responsible for triggering the ultraluminous activity. We briefly compare our results to recent models of merging spiral galaxies.