This paper presents an experiment on the use of ontologies for the description of biodiversity datasets. Building upon the INSPIRE European recommendation as a way to model the BIOME ontology for the biodiversity domain, the proposed ontology combines concepts from INSPIRE, matching them against ones defined in the Dublin Core, FOAF and CERIF ontologies. To validate the ontology, a prototype called Dendro uses it to provide an environment where biodiversity metadata records are made available as Linked Open Data, allowing producers to describe their datasets in a way that captures the requirements of their domain and that can also be reused by other actors in the data exchange process.