The major developments in nuclear aerosol modeling following the accident to pressurized water reactor Unit 2 at Three Mile Island are briefly reviewed and the state of the art summarized. The importance and implications of these developments for severe accident source terms for light water reactors are then discussed in general terms. The treatment is not aimed at identifying specific source term values but is intended rather to illustrate trends, to assess the adequacy of the understanding of major aspects of aerosol behavior for source term prediction, and demonstrate in qualitative terms the effect of various aspects of reactor design. Areas where improved understanding of aerosol behavior might lead to further reductions in current source terms predictions are also considered.