Previous work suggested that a significant amount of iron may be dissolved in carbide precipitates in steels, an observation that has implications for the accuracy of the theoretical strengthening models applied to steels. In this study, a Philips EM 400T with STEM and EDAX facilities has been used to analyze small precipitates, extracted from vanadium-microalloyed steels on several types of replica. It was found that, in general, larger precipitates tended to show a low or negligible iron content, while smaller ones (∼5 nm diameter) apparently contained a significant proportion of iron. Experience with different replica preparation techniques and washing routines suggested that much of the apparent iron content in initial particle analyses was due to the presence of an etching residue. It was not possible, with use of the STEM, to identify the persistent high-iron content of the smallest precipitates as a coating or as an iron-rich particle core. An attempt was made to identify the source of the iron by comparing the analyses with models for a) an iron coating and b) an iron-rich core, for particles growing during aging. It is still uncertain from where the iron, present in the smallest precipitates, has originated.