Excessive proliferation of new dinoflagellate generic names: the case of Torquentidium (Ceratoperidiniaceae, Dinophyceae)

New genera of unarmoured dinoflagellates represented by a single or very few species have largely increased in the last two decades. An example is the new genus name Torquentidium recently proposed for species that were transferred into Pseliodinium in 2018. The SSU- and LSU rRNA gene sequences of the type species of Pseliodinium and Torquentidium only differed by 0.3% and 2%, respectively, and they clustered together as a monophyletic group with high support. Based on these tree topologies and/or the sequence divergences, each dinoflagellate species would be classified in its own genus. The morphological character proposed to distinguish Pseliodinium and Torquentidium is that the cingulum encircled the cell 1 and 1.5 times, respectively. This kind of arbitrary numeric morphometric character, proposed one century ago, is currently considered invalid for the generic split of unarmoured dinoflagellates. For example, Polykrikos geminatum (with ~1.5 cingular turns) clusters between species of Polykrikos (1 turn).


Introduction
The dinokaryotic dinoflagellates have been traditionally divided into armoured (thecate) and unarmoured (naked or athecate) dinoflagellates. The robust cell covering of the armoured dinoflagellates have facilitated the studies, and the arrangement of the thecal plates is a useful diagnostic character for the classification at the genus level. The study and classification of the unarmoured dinoflagellates is more difficult due to the cell delicacy, the lack of thecal plates, and the polymorphism of some species. The classical generic classification was based on morphometric characters such as the degree of cingular displacement (the distance between both ends of the cingulum) used for the classical macro-genera Gymnodinium F. Stein and Gyrodinium Kofoid & Swezy. However, this character varies intraspecifically (Gymnodinium fusus F. Kofoid & Swezy), and the molecular data have revealed that other morphological characters (i.e., the shape of the apical groove) are more useful for the classification at the genus level (Daugbjerg et al., 2000). The type of cell covering is other diagnostic character supported by the molecular data, and species with smooth and striate cell surface never cluster together (i.e., Gymnodinium and Gyrodinium). That character is easier to observe during the routine light microscopy observations than the apical groove. Cochlodinium F. Schütt, other classical macro-genus, was a pool of species where the cingulum forms a descending left spiral of 1.5-4 turns around the cell (Kofoid & Swezy, 1921). The value of cingulum encircling the cell 1.5 times is arbitrary, and other species characterized by 1 and 1.5 cingular turns were placed in Gyrodinium (G.

Schütt/Gyrodinium falcatum
flavescens Kofoid & Swezy, G. melo Kofoid & Swezy). More than one turn of the cingulum is also a character of the warnowiid dinoflagellates (Erythropsidinium P.C. Silva, Warnowia Er. Lindemann, etc.), but this character is not used for the generic classification in that family (Kofoid & Swezy, 1921). Molecular data confirmed that the genus Cochlodinium pooled unrelated species that shared more than one cingular turn around the cell. Cochlodinium geminatum (F. Schütt) F. Schütt characterized by 1.5 cingular turns was transferred into Polykrikos Buetschli (Qiu et al., 2013). Consequently, the number of turns of the cingulum appear as an invalid diagnostic character at the genus level as this trait is variable among the species of a single genus. Reñé et al. (2013, their figure 5) reported that the sequences of the taxon formerly known as Gyrodinium fusus or Gyrodinium falcatum (with a single turn of the cingulum around the cell) clustered 4 between species with ~1.5 turn (Cochlodinium cf. helix, C. cf. convolutum, Cochlodinium sp.1) as a monophyletic group with strong support (bootstrap support, BP 99%; posterior probability, PP 1). Again, one or more turns of the cingulum around the cell is not a valid diagnostic character for the generic split.
When the molecular data of the type species of Cochlodinium, C. strangulatum F. Schütt, was available (Gómez et al., 2017), the sequences of Cochlodinium species were dispersed into four distinct clades. This supported that a cingulum with more one turn appeared repeatedly in the evolution. The type species of the genus Pseliodinium Sournia, P. vaubanii Sournia, is a junior synonym of Gymnodinium fusus/Gyrodinium falcatum, and Pseliodinium appears as the earlier available genus name for the members of this monophyletic group. New combinations in the genus Pseliodinium were proposed for   There is no unique number of sequence difference indicating that two DNA sequences belong to the same or distinct taxa. However, we can expect some coherency, 9 especially when the topology of the phylogenetic trees shows a monophyletic group for Pseliodinium and Torquentidium with strong support (Reñe et al., 2013;Gómez et al. 2017;Gómez, 2018;Hu et al., 2019;Shin et al., 2019). A recurrent criticism is that each organism has a distinct evolutionary rate, but at least we can assume similarities between the members of the Ceratoperidiniaceae. The LSU rRNA gene sequences of

Morphological support
There is no molecular support for the generic split, and the morphological character used for the distinction of Pseliodinium and Torquentidium based on a numeric morphometric value is a step back to the past. In the classical classifications, Kofoid & Swezy (1921) used arbitrary morphometric values for the generic split. For example, a cingular displacement lower or higher than 20% of the cell length for Gymnodinium and Gyrodinium, respectively; the relative height of the episome (Amphidinium Claparède & J. Lachmann); or the number of turns of the cingulum (Cochlodinium) that were later demonstrated to be unsupported by the molecular data (Daugbjerg et al., 2000;Gómez et al., 2017). Shin et al. (2019)  Cochlodinium species which cingulum encircles 1.5 times the cell, while the other species of Polykrikos has one cingular turn (Qiu et al., 2013). If the criterion by Shin et al. (2019) is followed, Polykrikos geminatum should be placed into an independent monotypic genus, and the genus Polykrikos split into at least four genera.
For most of the dinoflagellates, the cingulum that harbours the transversal flagellum encircles the cell with one turn. The transversal flagellum is responsible for the locomotion, and species with a longer cingulum have higher swimming speeds (Jeong et al., 1999;Sohn et al., 2011). A longer cingulum results when both ends of the cingulum are more distant (higher cingular displacement), and/or the cingulum describes more than one turn around the cell. In coastal waters, bloom-forming species have developed different strategies to increase the swimming speed for predator avoidance or migrations such as the formation of colonies and/or the increase of the length of the cingulum with more than one turn (i.e., species formerly classified in Cochlodinium such as

Synonymy of Pseliodinium pirum and Cochlodinium convolutum
The molecular data support the markings in the cell covering as a generic diagnostic character in unarmoured dinoflagellates. For example, the clades of Gyrodinium or Torquentidium also contains species with longitudinal striae. The truth is that nobody has ever observed a photosynthetic unarmoured dinoflagellate with longitudinal striae enclosed in a hyaline membrane. There are records in the literature of cells identified as Cochlodinium/Pseliodinium pirum, but they are illustrated with a smooth surface (Kofoid & Swezy, 1921;Gárate-Lizárraga, 2014;Gómez, 2018;Hu et al., 2019). Shin et al. (2019, Table 2) compared the morphological characters of T. convolutum and T. pirum. The only remarkable difference was that T. pirum is covered by longitudinal striae, citing Schütt (1895) and Kofoid & Swezy (1921) as data sources. However, Kofoid & Swezy (1921, p. 375) clearly stated for their records of Cochlodinium pirum: "We find no striae on our specimens on close examination". Kofoid & Swezy (1921, their figure 101) illustrated of C. pirum with a smooth surface (see the original illustrations in the Fig. 3). Kofoid was a splitter taxonomist that described the species ignoring the intraspecific variability and using morphometric numeric values as a diagnostic character based on the measurements of a single or very few individuals. For example, Kofoid & Swezy (1921) used in their dichotomous key used criteria such as "Displacement of girdle 0.82 transdiameter" for the split of Cochlodinium convolutum and other species, while their measurements were restricted to four individuals that co-occurred with their observations of C. pirum. Kofoid & Swezy (1921, p. 376) reported for C. pirum: "It is 14 very near C. convolutum, but has less torsion, 0.5 instead of 0.6, and less displacement of the girdle, 0.5-0.6 instead of 0.6 total length, and is more ochraceous and less greenish in color". Obviously, the colour is a poor diagnostic character, especially when you are observing moribund cells after net sampling and transport to the lab, and a difference of torsion (0.5 versus 0.6) is even difficult to note in this polymorphic taxon. The data evidence that there are no morphological differences between C. pirum and C. convolutum, and both species are synonyms. Cells of C. pirum with a surface striation do not exist. The molecular data do not evidence a high species richness in this monophyletic group. At the present, sequences from different geographical origins identified as Cochlodinium cf. helix (Reñé et al., 2013), Pseliodinium pirum (Hu et al., 2019) and Torquentidium convolutum (Shin et al., 2019) correspond to a single species, and the topology of this monophyletic group remains invariable since Reñé et al. (2013).
In the etymology of the genus name Torquentidium, one of the accepted meanings of the stem -torquĕo, torquenti-is torment, torture. In Torquentidium convolutum, the meanings of both genus and epithet names are 'twist' or 'turn'. Shin et al. (2019, p. 13) did not report the etymology of the suffix -dium-. It is uncertain whether the aim of the authors was to use -dinium-(meaning again 'turn around' or 'twist', referring to dinoflagellate) that is the most common suffix for dinoflagellates. Anyway, the genus name will be commonly misspelled as 'Torquentidinium'. So much 'twists' should not Type species: Pseliodinium vaubanii Sournia Other species: