Science historians Lorraine Daston and Peter Galison trace the emergence of scientific objectivity through a surprising archive: the illustrated atlases that trained scientists to see the natural world correctly. Beginning in the eighteenth century, they show how the ideal of objectivity - the effort to eliminate human subjectivity from scientific observation - gradually developed and competed with earlier ideals of truth-to-nature and later ideals of trained judgment. The book examines how atlas-makers represented plants, bones, crystals, and other objects, revealing changing assumptions about what constituted accurate representation and who should be trusted to make images. Daston and Galison demonstrate that objectivity is not a timeless scientific value but a historically contingent ideal with specific techniques, technologies, and practices. The invention of photography seemed to promise mechanical objectivity free from human bias, but scientists soon discovered that photographs required just as much interpretation as drawings. The book is lavishly illustrated, allowing readers to see the differences between representational strategies that embodied different epistemological ideals. For anyone interested in the history of science, visual culture, or the philosophy of knowledge, this work offers a pioneering study of how scientists have sought to separate reliable knowledge from human subjectivity.