Comment by mutatio

I think it's not about diversity, but lineage. The phenotype for "fish" is so tight and well defined; a salmon is closer related to a human in the tree of life than to a coelacanth even though both are categorised as "fish".