Taxonomic registers are key tools to help us comprehend the diversity of nature. Publishing such registers in the Web of Data, following the standards and best practices of Linked Open Data (LOD), is a way of integrating multiple data sources into a world-scale, biological knowledge base. In this paper, we present an ongoing work aimed at the publication of TAXREF, the French national taxonomic register, on the Web of Data. Far beyond the mere translation of the TAXREF database into LOD standards, we show that the key point of this endeavor is the design of a model capable of capturing the two coexisting yet distinct realities underlying taxonomic registers, namely the nomenclature (the rules for naming biological entities) and the taxonomy (the description and characterization of these biological entities). We first analyze different modelling choices made to represent some international taxonomic registers as LOD, and we underline the issues that arise from these differences. Then, we propose a model aimed to tackle these issues. This model separates nomenclature from taxonomy, it is flexible enough to accommodate the ever-changing scientific consensus on taxonomy, and it adheres to the philosophy underpinning the Semantic Web standards. Finally, using the example of TAXREF, we show that the model enables interlinking with third-party LOD data sets, may they represent nomenclatural or taxonomic information.