In part I we adapted the recent analytical mean-field theory of polymer adsorption by Semenov et al. to match the numerical lattice theory of Scheutjens and Fleer (SF). Here we calculate explicitly the contributions of trains, loops, and tails to the adsorbance and their size distributions. We choose conditions in the so-called plateau region, which is most relevant from an experimental point of view. We thus use the "plateau approximation", where the two simultaneous equations that determine the proximal length p and the distal length d can be uncoupled. The variations with bulk concentration, chain length, and surface affinity compare quite nicely with the SF numerics. Both a good solvent (at the mean-field level) and Theta solvent are considered; the agreement is better in the former case.