How are the objects of science constituted? The objects of a science may be as apparently straightforward as plants in botany, or as complex as the experimental systems of the modern biosciences. Only rarely do scientists work with "raw," unprocessed natural objects; never do they study all the possible objects even of a restricted domain of nature or society. By the early seventeenth century natural philosophers and mixed mathematicians had begun to substitute mathematical idealizations for real bodies in motion; naturalists of the same period replaced the many individual plant specimens they observed with a composite illustration intended to represent the "type" of the species; the scientific laboratory creates effects and compounds under deliberately artificial conditions. Since the seventeenth century, the natural sciences have been extremely fertile in creating new kinds of scientific objects through mathematical, experimental, and visual techniques. More recently, the human sciences have presented equally revealing examples of how scientific objects emerge, and sometimes how they disappear: the market in economics, mortality in human demography, society in sociology, the self in psychology. All of these entities may have existed since ancient times, but not as objects of sustained scientific inquiry, much less as the raison d'être of whole disciplines. The constitution of scientific objects through selection, representation, idealization, manipulation, and stabilization is a process that is at once central to and yet invisible within the history, philosophy, and sociology of science. This project was the first to address the topic directly, with examples as diverse as anomalies in natural philosophy, the aether in physics, value in economics, experimental systems in microbiology, and culture in anthropology.