Based on a literature review and new morphological and molecular phylogenetic data, a summary of all described benthic Prorocentrum species is presented. Short descriptions of the 29 species are provided including information and illustrations of platelet pattern. Increasing morphological data of the same taxon from different sites has revealed unexpected morphological variability. The variability of characters used for taxonomy is discussed. Many of these may be more variable than previously thought or phenotypically plastic. We conclude that the following features are constant within a species: (1) asymmetry or symmetry of the theca in combination with (2) the shape of the periflagellar area visible on the right thecal plate, (a) arc-shaped, (b) wide or simply V-shaped, (c) narrow and deep V-shaped, and (d) linear. At the same time, we are now beginning to understand that the cell shape of some species is more variable than thought and not in others. The stability of pore patterns is not known to date nor is it known whether the presence of a starch sheath is a reliable taxonomic character. The morphology of the periflagellar area might be a very useful character. The unstable terminology to describe similar or identical structures has made it difficult to interpret the literature on these features. Therefore a unified revised terminology for morphological features, cell orientation, and platelet numbering are proposed. This review highlights uncertainty in species boundaries of several potentially toxin producing species. This includes the species Prorocentrum hoffmannianum, Prorocentrum belizeanum, Prorocentrum maculosum, and Prorocentrum faustiae. Some species are reported to have variable toxicity, e.g. P. belizeanum and Prorocentrum rhathymum. It is clear that several species complexes, containing genetically related cryptic species, are present. The species complexes proposed in the past and in this paper reflect the difficulties in distinguishing these species, which may be highly morphologically similar to one another, sometimes despite genetic differences. Additionally, the growing molecular data set has allowed us to evaluate current species concepts based on novel molecular genetic information. As some Prorocentrum species are potentially toxic and are included in monitoring programs, it is important to evaluate species boundaries as currently understood, and to determine our current knowledge gaps regarding the relationship between species and strain identification and toxin production. (C) 2013 Elsevier B.V. All rights reserved.