Discovering diatom species: is a long history of disagreements about species-level taxonomy now at an end?

Background – Now and in the past, species discovery in diatoms begins, and often ends, with a survey of morphological variation to find breaks in the variation pattern that allow diagnosable entities to be defined and named. For this process to be effective, it needs to be informed by an understanding of the mechanisms that generate variation and many mistakes were made in the early 19 th century because of poor knowledge of the diatom life cycle and phenotypic plasticity; some taxonomically important life-cycle characteristics were not properly documented until 1932. Acceptance of the Darwinian view of species as taxa arbitrarily set along a continuum of divergence was accompanied in the late 19 th and early 20 th century by description of many varieties and forms; most recently described taxa, on the other hand, have been species. The neo-Darwinian emphasis on reproductive isolation as an important factor in speciation, introduced during the 'New Synthesis' of the 1940s, did not become influential in diatom taxonomy until the 1970s. It has since been a source of controversy, some seeing it as having no place in taxonomy, others regarding it as a useful aid to the detection of species boundaries, alongside character-based approaches, both morphological and molecular. Review – This paper discusses changes in how species have been discovered and circumscribed in diatoms, and seeks to establish whether there is a basis for consensus in future work in this field. Conclusion – Whereas morphology is currently still the primary tool for discovering diatom species diversity, molecular methods may be more cost-effective in future and are the only practical means of exploring cryptic (including pseudocryptic) diversity, which appears to be widespread. By treating species as separately evolving metapopulation lineages, as recommended by de Queiroz, different approaches can be accommodated (including tests of reproductive compatibility), providing a framework within which conflicting results can be analysed and reconciled