Abstract
Contemporary philosophical debate on realism revolves around the interpretation of theories well confirmed by experiment. This paper seeks to rebalance the debate by focussing attention upon experimental practice itself. It argues that the production of observation reports entails the interactive stabilisation of three elements: material experimental practice, instrumental modelling of that practice, and phenomenal modelling of the material world. The entanglement of these three elements is exemplified in a historical case study. Such entanglements block correspondence realism and point, instead, to an anticorrespondence realism which requires a constructivist analysis of scientific practice for its full articulation.