Ontologies are a tool for Knowledge Representation that is now widely used, but the effort employed to build an ontology is still high. There are a few automatic and semi-automatic methods for extending ontologies with domain-specific information, but they use different training and test data, and different evaluation metrics. The work described in this paper is an attempt to build a benchmark corpus that can be used for comparing these systems. We provide standard evaluation metrics as well as two different annotated corpora: one in which every unknown word has been labelled with the places where it should be added onto the ontology, and other in which only the high-frequency unknown terms have been annotated