Standard questions in the theory of administrative law involve the allocation of power among legislatures, courts, the President, and various types of agencies. These questions are often heavily informed by normative commitments to particular allocations of governmental authority among the three branches of the national government. These discussions, however, are incomplete because agencies are typically treated as unitary entities. In this Article, we examine a different question: how does administrative law allocate power within agencies? Although scholars have sometimes cracked open the black box of agencies to peer inside, their insights are localized and confined to particular contexts. We will generalize the idea, attempting to show that administrative law allocates power both horizontally and vertically within agencies and offering some hypotheses about the nature of the resulting effects. Horizontally, administrative law directly or indirectly determines the relative influence within agencies of various types of professionals—lawyers, scientists, civil servants, politicians, and others. Vertically, administrative law directly or indirectly determines the relative influence within agencies of appointed agency heads, midlevel bureaucrats, and line personnel. This perspective illuminates several of the most puzzling judicially developed principles and doctrines of administrative law, including the doctrines surrounding Chenery, Chevron, Mead, and Accardi, as well as agency structures and procedures established by statute or executive order. The internal allocation perspective offered here both improves upon and critiques existing justifications for these developments and in that sense points the way toward a superior understanding of administrative law.