Cocconeis molesta Kütz., C. diaphana W.Sm. and C. dirupta W.Greg. (Bacillariophyta): type material, ambiguities and possible synonymies

F.T. Kützing introduced Cocconeis molesta with only an uninformative description and a poor illustration: C. molesta has small, oblong valves and is an epiphyte. Another species, Cocconeis diaphana, described by William Smith, is said to have larger valves than C. molesta, with frustules that are relatively oblong. Smith described two forms: one with a distinct fascia on its raphe valve (var. β), the other without this feature. A third species, Cocconeis dirupta was described by Gregory, who expressed doubts that it differed from C. diaphana. Finally, Cocconeis molesta var. crucifera Grunow was first introduced in Van Heurck’s Atlas but was subsequently treated by Van Heurck as a synonym of C. molesta. No previous account has examined the type material of these species. In this paper, we undertake that task and examine type slides and raw material in order to discriminate these different taxa. We conclude by recognizing three species: Cocconeis molesta Kütz., C. diaphana W.Sm. and C. dirupta W.Greg. Cocconeis diaphana var. β is considered to be a synonym of C. dirupta and C. molesta var. crucifera is considered to be a synonym of C. molesta. Lectotypes are designated for C. diaphana and C. dirupta.


Introduction
The diatom genus Cocconeis Ehrenb. (Ehrenberg 1837: 173) is diverse, comprising numerous small species, several of which are very similar and difficult to identify. Thus, there is a need to re-examine the type material in order to improve, clarify and expand the original descriptions and to determine whether or not these taxa are synonymous. Unfortunately, most diatom species are identified by comparing specimens collected from places other than the type locality. However, in the case of Cocconeis, studies on the type material have been increasing (Jahn et al. 2009;Romero 2011;Riaux-Gobin et al. 2014a, 2014bRomero & Riaux-Gobin 2014), particularly the detailed examination of raw type material with a scanning electron microscope (SEM) (e.g. Romero & Riaux-Gobin 2014).
Herein, we focus on Cocconeis molesta Kütz., Cocconeis diaphana W.Sm., Cocconeis dirupta W.Greg. and some other taxa pertaining to the same group, all of which are small and very similar in morphological characteristics (under LM), and may have led to taxonomic confusion in the past (see Discussion).

Cocconeis dirupta W.Greg. and allied taxa
Cocconeis molesta Kütz. was introduced along with a poor illustration and very succinct description, which mentioned only the small size and oblong shape of the frustule and that it was epiphytic on Callithamnion cruciatum (C.Agardh) Nägeli (Kützing 1844: 71, pl. 5, fig. 7, reproduced in Fig. 1).
Another species with a relatively oblong frustule but with larger dimensions than C. molesta was subsequently described as Cocconeis diaphana W.Sm. (Smith 1853: 22, pl. 30, fig. 254, reproduced in Figs 2-3), and included two varieties; one without a distinct fascia and the other with a conspicuous fascia (β), with the latter accompanied by a laconic note stating "β. Nodule dilated into a stauros". These forms will be cited hereafter as var. diaphana and var. β. Two varieties, corresponding more or less to var. diaphana and var. β, were later described as C. diaphana var. amygdalina Grunow ex Cleve (Cleve 1895) without a fascia, and C. diaphana var. dirupta (W.Greg.) Rabenh. (Rabenhorst 1864) with a distinct fascia, respectively.
Lastly, Cocconeis molesta var. crucifera Grunow ex Cleve was invalidly described in Van Heurck (1880-1885 and afterwards classified as a synonym of C. molesta by Van Heurck (1896: 291, pl. 29, fig. 823). Van Heurck (1896) examined C. molesta type material, from the isotype present in his collection (Kützing 259) and noticed the stauros on the raphe valve and produced drawings identical to those first illustrated as C. molesta var crucifera in Van Heurck (1880-1885. Most of the above cited taxa (see further comments below), have been described with a fascia (or stauros), particularly on their raphe valve, and have poor original descriptions. Although several studies have contributed to a better understanding of these taxa, unresolved problems still remain, while other studies have provided revised descriptions that did not include observations on the types (e.g. about C. dirupta: Hustedt 1931Hustedt -1959Foged 1978;Kobayasi & Nagumo 1985;Sar et al. 2003;Riaux-Gobin et al. 2011;Lobban et al. 2012;about C. diaphana: Álvarez-Blanco & Blanco 2014;about C. molesta var. crucifera: De Stefano et al. 2000;Sar et al. 2003).
We here examine several original materials that were cited in the protologues in order to discriminate different taxa, clarify their definitions and resolve their nomenclatural issues, including synonymies, confirmation of types and designation of a lectotype for Cocconeis diaphana.

Material and methods
Materials used in this study derived from several sources: (1) Friedrich Traugott Kützing collection,no. 259

Terminology and abbreviations
For the description of the frustule, terminology follows Anonymous (1975), Ross et al. (1979) and Round et al. (1990). As previously proposed, in particular by Riaux-Gobin et al. (2013), we designate the valve with a raphe as the raphe valve (RV) and the valve without a raphe as the sternum valve (SV).
Since some of the original raw material used by F.T. Kützing has been found, whereas the same was not possible for W. Smith, R.K. Greville and W. Gregory, scanning electron microscope (SEM) observations have been made on only one type material. Nevertheless, modern light microscopy allows the striation and other valve features to be observed on the type slides, so that the original description can be emended.

Results
Class

Original description
Kützing (1844)  The frustule is small (ca. 16.4 µm long, 9.7 µm wide) and strongly arched (RV concave, SV convex). The RV striation is hardly discernible, with ca. 30 striae in 10 µm (Fig. 18, arrowhead). The SV areolae are arranged along axial rows in a zig-zag pattern (22-23 rows in 10 µm, Figs 16,17,20). The helictoglossae are close to the margin and deflected in opposite directions (Fig. 15), the raphe is straight. The SV sternum is narrow and straight, with a small oblong central area. The RV fascia is narrow and half a valve in length.

SEM examination of the Herbarium sheet BM 000905975
The material is very poor and only a broken SV of Cocconeis molesta was observed, with the following morphometrics and features: estimated valve length (> 17 µm), SV stria density (

Notes
When W. Smith introduced C. diaphana, he also proposed an unnamed variety β characterized by "the nodule dilated into a stauros". This variety, mostly represented in the Sidmouth gathering, though originally a syntype must be excluded from this species. It will be here treated as belonging to

Description of specimens from "Jersey, Pontac", after LM examination of the type material
BM 23161 (Figs 21-26) contains oblong-elliptical, thinly silicified and diaphanous valves, 33-42 µm (38 µm ± 2.5 σ) long, 21-26 (23 µm ± 2.2) wide, l/L 1.65 ± 0.2, with 26 ± 2.2 SV striae in 10 µm, 25 ± 0.7 RV striae in 10 µm, and ca. 23 axial SV rows in a zig-zag pattern (Fig. 26). The SV sternum is straight and narrow, the central area very reduced (Figs 23-24). The RV has clearly identifiable striae, with 25 areolae in 10 µm. The RV fascia is short and quite high, with the aspect of an elliptic central area more than a real fascia. The helictoglossae are straight and off the margin (Fig. 24). As this slide (BM 23161) contains the specimen best matching the original description, we here designate it as the lectotype of C. diaphana W.Sm.

Remark
Thanks to the kindness of Bart Van de Vijver we have received several SEM photographs of the isolectotype of Cocconeis diaphana deposited in the Van Heurck collection in Meise (BR). This material is epiphytic on marine Rhodophyceae, and is present on a mica labelled "Cocconeis diaphana n.sp., Jersey, Aug. 14. 1852", preserved in the W. Smith collection, vol. I, p. 33. These SEM photographs (Figs 39-44) are a perfect match for the LM figures in this paper and support our emended description.
Valves round-elliptical, relatively small. The SV is coarsely striated and punctuated, usually with no axial rows of areolae. The SV sternum is large and composed of two lanceolate parts (Figs 27-29, 33), one median SV stria lacking on one side (Fig. 27, arrow) or both (Fig. 33). The RV has coarsely punctuated striae, strongly radiate. The RV fascia is narrow and wide (often more than half a valve in length). The raphe is often slightly sigmoid with helictoglossae deflected towards opposite sides (Fig. 34). Slide BM 19589 only shows the variety β.
The two examined slides of "C. diaphana Sidmouth" (VI-45-B10/BR and BM 19589) show valves with similar characteristics (33)(34), close to those of Cocconeis dirupta W. see below). Therefore, var. β is here excluded from C. diaphana and considered as a synonym of C. dirupta.

Type material
Lectotype, designated here: slide BM 1420, labelled "Arran 56" in Greville collection, the only syntype found in BM.

Notes
Several localities (Glenshira sand, Mr Miles's Corallina gathering and several dredgings) cited by Gregory in the original description could contain syntypes, but only the one designated as lectotype was found in BM. According to W. Gregory "C. diaphana […] which was found occurring with C. dirupta […] may perhaps be an imperfect form of C. dirupta […] or possibly [...] the lower valve' of the latter", with some further notes that "it is, however equally probable that these forms belong to different species".
Figs 35-38. Cocconeis dirupta W.Greg. from BM 1420. 37. Lectotype illustration. SV sternum lanceolate with apices slightly bent in opposite directions (36), one SV stria lacking on one side (35, 36, arrow), RV striae strongly radiate, RV fascia narrow and extended (38, arrows), helictoglossae deflected in opposite directions (37, arrowheads) and raphe slightly sigmoid (37). Scale bars = 10 µm. The complete frustules appear obscure while the separate RV and SV appear translucent, which probably justifies W. Gregory's hesitation about the existence of a second taxon (C. diaphana W.Sm.) in his material. On the other hand the R.K. Greville drawing (Gregory 1857: fig. 25) probably superimposed the two valves. Contrarily to the latter drawing (Gregory 1857, fig. 25 reproduced in Fig. 4), the fascia does not reach the valve margin in the many specimens observed in this study.
The slide BM 1420 from Arran 56 (Gregory in Greville collection) is designated here as the lectotype, since nothing else has been traced that could be considered as original material studied by W. Gregory.

Ambiguities and synonymies
The type of Cocconeis molesta has unique features and morphometrics (as discussed above) that cannot be confused with those of C. diaphana. Cocconeis molesta has small dimensions, a wide fascia, RV striae that are hardly discernible (ca. 30 in 10 µm), and helictoglossae close to the margin and bent in opposite directions.
Nevertheless, several recent papers refer to Cocconeis molesta as C. cf. molesta (e.g. Bruder & Medlin 2007;Baldi et al. 2011) which shows the difficulty in accurately defining this taxon. Although Cocconeis molesta is rarely illustrated in publications, ambiguities have been noticed when analyzing some of the images: da Silva (1946, fig. 91) presents a drawing of a large RV fascia; an SV is seen in the LM (at http://www.diatomloir.eu/Site%20Diatom/Sardiadeux.html), with a much bigger one corresponding to the RV; Majewska et al. (2014, pl. 2, fig. 13) showed the species in SEM with morphometrics (from fig. 13: ca. 57 µm long, 30 µm wide, with ca. 18 striae in 10 µm) but it might be Cocconeis diaphana var. diaphana.
On the other hand, the taxon Cocconeis molesta var. crucifera has been abundantly illustrated in both LM and SEM (e.g., Kobayasi & Nagumo 1985;De Stefano et al. 2000;Sar et al. 2003) though the type material for this taxon has apparently not been checked nor compared with that of C. molesta.
Therefore, it is very likely that the images and illustrations in these references actually belong to Cocconeis molesta, with C. molesta var. crucifera being a synonym.
The specimens of Cocconeis diaphana from "Jersey, Pontac" (Figs 21-26) and from 'Sidmouth' (Figs 27-34) obviously pertain to two distinct taxa. Smith (1853) originally stated that there were two forms, but without indicating their respective origins. The present study shows that the specimens found in the "Jersey, Pontac" sample do fit the concept of Cocconeis diaphana var. diaphana (oblong shape, reduced RV transverse central area), while the specimens from "Sidmouth" (with a round to sub-orbicular shape and a wide fascia) belong to the unnamed variety β which was later included as a synonym of C. dirupta by several authors (e.g., Cleve 1895; Hustedt 1931Hustedt -1959. Cocconeis diaphana var. β and C. dirupta have very similar features (Figs 27-38), except for the general shape of the valve illustrated in the original drawings, with C. diaphana var. β being more elliptical. The conspicuous striae characterizing C. dirupta were not illustrated in the original drawing of C. diaphana var. β (Fig. 2), but the two taxa may still be conspecific.
In the original drawings and descriptions of Cocconeis dirupta and C. diaphana var. β there was no indication of a sigmoid raphe, which slightly contradicts our own observations on the original material of C. dirupta and C. diaphana var. β (as discussed above).
Recently, Álvarez-Blanco & Blanco (2014) proposed that C. dirupta and C. molesta var. crucifera are synonymous with C. diaphana, not by examining the type material, but based on the fact that the raphe is straight in all of the original illustrations of these taxa. However, after type examinations, Cocconeis dirupta has a slightly sigmoid raphe. Furthermore, Cocconeis dirupta, C. molesta and C. diaphana var. diaphana seem distinct taxa, while C. molesta var. crucifera is herein considered a synonym of C. molesta.

Allied taxa
Some small Cocconeis species that have been recently described (Suzuki et al. 2001(Suzuki et al. , 2008(Suzuki et al. , 2012 from the Japanese coast, appear to share similarities with C. molesta and C. diaphana, particularly concerning their SV (on LM): e.g. C. churalis Hid.Suzuki, C. nagumoi Hid.Suzuki, C. shikinensis Hid. Suzuki and C. baikalensis (Skvortzov & Meyer) Skvortzov. Nevertheless, these taxa have no fascia on their RV and possess several features which allows them to be easily differentiated in the SEM. Amongst the newly described taxa from Japan, Cocconeis tortilis Hid.Suzuki (Suzuki et al. 2014) has some morphological affinities with C. molesta, except for the RV fascia reaching the valve margin and the SV sternum being larger than in C. molesta. The spiral aspect of Cocconeis tortilis may be due to the ecology of the taxon (found as an epiphyte on Codium intricatum Okamura). Suzuki et al. (2014: 223) also stated that Cocconeis molesta var. crucifera (synonym of C. molesta) has no stauros, but a wide fascia.

Conclusions
Our study stresses the importance of checking all possible original materials (especially types) in order to make an improved species definition. Thus, the name Cocconeis molesta var.  (1895)].
Similarly, our examination of the original material of Cocconeis diaphana W.Sm. allowed us to point out the differences between the two syntypes, which indeed belong to different taxa. Lectotypes have been designated here for C. diaphana and C. dirupta.
From this study, three species have been recognized: Cocconeis molesta Kütz., C. diaphana W.Sm. and C. dirupta W.Greg. Cocconeis diaphana var. β is considered to be a synonym of C. dirupta and C. molesta var. crucifera a synonym of C. molesta var. molesta.
(CNRS-EPHE, CRIOBE) is acknowledged for funding the CORDIA program (COral Reef DIAtoms) in the South Pacific.