Biodiversity is a key concept for the biological sciences. While it has its origin in conservation biology it has become useful across multiple biological disciplines as a means to describe biological variation. It remains, however, unclear what particular biological units the concept refers to. There are currently multiple accounts of which items of diversity the concept describes and how these are to be measured. I draw from the species concept debate to argue for a particular set of desiderata for “biodiversity” that is both principled and coheres with the concepts use. This is that biodiversity is a scientific concept that functions over multiple scales and that the units that it quantifies difference over should be theoretically fundamental, empirically tractable, and cross-system compatible. Given these desiderata I argue for a phylogenetic basis for biodiversity.