In the late nineteenth century, the anthropology curators of the Smithsonian Institution consulted their cataloguing systems and storerooms, assessing specimens in order to determine which could be designated as duplicate specimens and exchanged with museums domestically and abroad. The status of ‘duplicate’ for specimens was contingent on conceptions of similiarity impacted by disciplinary classification praxis, with particular emphasis on object nomenclature and formal attributes. Using rattles from Haida Gwaii collected between 1881 and 1885 by James Swan for the Smithsonian Institution, (...) this article explores how anthropology curators designated rattles as exchangeable duplicate specimens. It considers cataloguing and spatial arrangements, as well as changing populations and formal characteristics of rattles, in order to explore how similarity was operationalized in the museum to produce duplicate anthropological specimens. (shrink)