This paper examines the transformation of local uncertainties encountered by working scientists into global certainty, or `scientific facts'. It discusses six mechanisms by which scientists transform local uncertainty: attributing certainty to the results of other fields; substituting processual for production evaluations in the face of technical failures; ideal type substitutions; shifting clinical and basic evaluation criteria; ad hoc generalizing of case studies; and the subsuming of epistemological questions in internal debates. The data are drawn from a study of late nineteenth-century British neurophysiologists (surgeons, neurologists, pathologists, physiologists). The approach is drawn from the sociology of work.