As noted in the lower right, arrows convey intensity of citation from one field to the next. The structure is very similar to that revealed by eigenfactor.org's plots (previously blogged here; apparently at least one of the authors is involved in the eigenfactor project) albeit with some additional and rather useful dimensionality.
Or particular interest is the social sciences, which Rosvall and Bergstrom map out in detail:
Note how high the flow intensity within economics (circle darkness) is compared to flow outside (circle border), as well as the relative patterns to other fields, almost all of which cite economics more than economics cites them. The pattern is similar though not quite as extreme for psychology. There are a variety of reasons one could imagine for this (rigor of work, differences in strategic citation behavior, range of topics studied, etc. etc.) but it's a pretty nice bit of academic-social network structure to ponder on its own.