Our understanding of insect societies is rapidly expanding due to an emphasis on integrative approaches. Emerging tools enabling the molecular dissection of social behavior, together with novel hypotheses for the evolution of eusociality, are emblematic of this progress. However, an obstacle to a truly integrative approach remains, as social physiology--the basis of group-level coordination--has generally been neglected by geneticists. In this paper, we begin a synthesis of these fields by first reviewing three classes of social insect organization that mark major transitions in increasing social complexity. We then develop an expansion of the superorganism concept in order to place eusociality into a broad evolutionary context, and we also interpret current molecular and genetic work on the evolution of eusociality. The ground plan hypothesis proposes that eusociality arose via simple changes in the regulation of ancestral gene sets affecting reproductive physiology and behavior, and we argue that this hypothesis is explanatory for the evolution of division of labor (social anatomy) but not for the regulatory systems that ensure group-level coordination of action (social physiology), which we propose is dependent on previously unrelated traits that are brought together into novel genetic networks. We conclude with a review of recent work in sociogenomics that supports our hypotheses.