Every ten years United States congressional districts are drawn, physically constructing political representation based on where citizens live. Why is it done this way? Is territorial representation consistent with the broader normative ends of political representation for any large nation? And if not, how might random assignment provide a justifiable alternative to group representation? Using the case of the founding period of the United States and classic sources in western political theory to illustrate the argument, this book describes the conceptual, historical and normative features of the electoral constituency. Based on uncontroversial features of political legitimacy (like the idea that a representative should be accountable to those who elected him) it argues in favor of single member, electoral constituencies, each of which look like the nation they collectively represent. It argues that randomly assigning voters into permanent, national electoral constituencies would be worth considering in any large democratic nation.