This article is about a period of technology transfer – the late 1910s and 1920s – when wartime aerial reconnaissance techniques and operations were being adapted to a range of civilian uses, including urban planning, land use analysis, traffic control, tax equalization, and even archaeology. At the center of the discussion is the ‘photomosaic’: a patchwork of overlapping aerial photographs that have been rectified and fit together so as to form a continuous survey of a territory. Initially developed during the First World War to provide coverage of fronts, photomosaic mapping was widely practiced and celebrated during the postwar years as an aid to urban development. The article traces both the refinements in photomosaic technology after the Armistice and the rhetorical means by which the form’s avant-garde wartime reputation was domesticated into an ‘applied realism’ that often effaced its site-specific perspective, its elaborately rectified optics, and the oppositionality of both its military and civilian uses. The article has a broader theoretical aim as well. Classic statements of both structuralist and post-structuralist spatial theory (Barthes and de Certeau are the primary examples here) have produced an ossified geometry wherein the vertical is the axis of paradigm, top-down strategy, and manipulative distance and the horizontal the axis of syntagm, grassroots tactics, and resistant proximities and differences. In its close study of the technology and rhetoric surrounding interwar photogrammetry, the article provides an example of how one might reverse the long-standing misrecognition of high-altitude optics as effacing time, difference, and materiality – and what it might mean to view such optics as, instead, a resource in turning from abstract toward differential conceptions of both aerial photography and our theoretical habits. This turn I call ‘applied modernism’, a term that accesses both the wartime photomosaic’s affiliation with avant-garde painting and its insistence that portraits of the total are always projections from partial, specific vantages.