Williamsella versus

Rioual, Patrick, Flower, Roger J., Chu, Guoqiang, Lu, Yanbin, Zhang, Zhongyan, Zhu, Bingqi & Yang, Xiaoping, 2017, Observations on a fragilarioid diatom found in inter-dune lakes of the Badain Jaran Desert (Inner Mongolia, China), with a discussion on the newly erected genus Williamsella Graeff, Kociolek & Rushforth, Phytotaxa 329 (1), pp. 28-50 : 42-44

Williamsella versus Fragilaria

Since both the genera Williamsella and Fragilaria have open girdle bands and a rimoportula at only one apex (which is not always the case for species included in Fragilaria , see below), Graeff et al. (2013) used other characteristics and characters to distinguish them.

i) The absence of spines and colonial formation:

Not all Fragilaria species form colonies and bear spines; there are numerous examples such as Fragilaria pennsylvanica Morales ( Morales 2003a), F. pectinalis (O.F. Müller) Lyngbye (1819: 185) as illustrated in Wetzel & Ector (2015), F. candidagilae Almeida, C. Delgado, Novais & S. Blanco (in Delgado et al. 2015: 4), F. boreomongolica Kulikovskiy, Lange-Bertalot, Witkowski & Dorofeyuk ( Kulikovskiy et al. 2010: 36). In their comprehensive revision of the needle-shaped group of fragilarioid diatoms, Lange-Bertalot & Ulrich (2014) also reported that several of the long, needle-like Fragilaria are spineless and have a solitary life-form including F. perdelicatissima Lange-Bertalot & Van de Vijver in Lange-Bertalot & Ulrich (2014: 19), F. saxoplanctonica Lange-Bertalot & Ulrich (2014: 30) , Fragilaria aquaplus Lange-Bertalot & Ulrich (2014: 32) and Fragilaria gracilis Østrup (1910: 190) .

More generally, within the Fragilariaceae the presence or absence of spines is not a stable character at the genus level. For example, in Stauroforma Flower, Jones & Round (1996: 53) , the spinose species Stauroforma exiguiformis

1 Here the nomenclature is complicated as species in the genus Synedra , as understood by Round et al. (1990) are now placed in Ulnaria . This is a nomenclatural act not a taxonomic one ( Williams 2011).

Flower, Jones & Round (1996: 53) can co-occur with the non-spinose species Stauroforma inermis Flower, Jones & Round (1996: 54) ( Flower et al. 1996). Species within Pseudostaurosira D.M. Williams & Round (1988: 276) generally possess spines but exceptions exist such as Pseudostaurosira parasitica (W. Smith) Morales (2003b: 287) . Another example was given by Flower (2005) who distinguished Pseudostaurosira microstriata var. spinosa Flower (2005: 65) from the non-spinose nominal variety, Pseudostaurosira microstriata (Marciniak) Flower (2005: 65) . Conversely, while species in Ulnaria are typically spineless, Ulnaria ungeriana (A. Grunow) Compère (2001: 100) have welldeveloped linking spines (Williams 1986, Siver et al. 2006). Even at the species level, the presence or absence of spines is not always a good criterion for taxonomic separation. For example, Staurosira dimorpha Morales (2010: 103) can form two sympatric variants, one with spines the other spineless. Interestingly, for this latter species the lack of spines does not prevent formation of colonies ( Morales et al. 2010).

ii) A preference for saline versus freshwater

To argue against the inclusion of their new species in the genus Fragilaria, Graeff et al. (2013) referred to Round et al. (1990) who stated that Fragilaria species are restricted to freshwater. This statement however, is not entirely correct as Fragilaria famelica (Kützing) Lange-Bertalot (1980: 749) is commonly reported from marine coasts ( Witkowski et al. 2000) and saline springs ( Wojtal 2013). Graeff et al. (2013) described Williamsella angusta from the main basin and the surrounding marsh of Blue Lake, Utah. There, the salinity varies from 5.5 ppt in the lake to 13.4 ppt in the marshes (see Table 1 in Graeff et al. 2013). From these values, Graeff et al. (2013) deduced that their new species was restricted to saline waters. For the taxon we found in the Badain Jaran lakes however, our data ( Fig. 76 ) suggest that it optimally grows in subsaline conditions, i.e. in a range of salinity of 0.5–3 ppt following the classification proposed by Hammer et al. (1983). The three lakes in which this fragilarioid diatom is most abundant have a salinity ranging between 1.3 and 2.0 ppt, although it was also found (0.2%) in one lake with much higher salinity (30.3 ppt).

iii) The external coverings of the areolae

In their description of Fragilaria, Round et al. (1990) wrote that the areolae of this genus have delicate, often disc-like, cribra (as illustrated in their fig. g, p. 347). A structure very similar to the one observed in W. angusta exists in Fragilaria longifusiformis P. Siver, E. Morales, B. Van de Vijver, M. Smits, P.B. Hamilton & H. Lange-Bertalot (2006: 185) , a species described from the south-eastern USA ( Hains & Sebring, 1981) but also reported in Europe ( Siver et al. 2006) and in Brazil ( Ludwig et al. 2015). Siver et al. (2006) went on to argue that the presence of these areolae coverings they called disc-like vola, was one of the important characters for attributing F. longifusiformis to the genus Fragilaria (instead of Synedra , i.e. Ulnaria ). Recently, Lange-Bertalot and Ulrich (2014) distinguished the subspecies F. longifusiformis ssp. eurofusiformis Lange-Bertalot and Ulrich (2014: 37) . SEM images of this subspecies indicate that external coverings of the areolae can also be present in this taxon (plate 18, fig. 3) but are often dissolved (plate 18, figs 6–8). External coverings of the areolae can be observed in other needle-like Fragilaria species such as F. tenera (W.Smith) Lange-Bertalot (1980: 746) as illustrated in Lange-Bertalot and Ulrich (2014, plate 3, figs 3–6) and in Almeida et al. (2016, figs 15–17), F. neotropica P.D. Almeida, E. Morales & C.E. Wetzel (2016: 171) (figs 43, 47, 49 in Almeida et al. 2016). Note that Almeida et al. (2016) used the term rotae when referring to these structures.

For Fragilaria species of smaller size, disc-like volae are clearly visible in SEM images of F. neointermedia Tuji & D.M. Williams (2013: 7 , fig. 37), F. capitellata (Grunow) J.B. Petersen (1946: 54) as illustrated in Tuji & Williams (2008, figs 47–49). Although they did not discuss the significance of these structures in their study, Wetzel and Ector (2015) also published SEM images that show external coverings of the areolae for F. vaucheriae (Kütz.) J.B. Petersen (1938: 167) , F. uliginosa Kulikovskiy, Lange-Bertalot, Witkowski & Dorofeyuk (2010: 37) , F. microvaucheriae Wetzel & Ector (2015: 282) and F. pectinalis , the later being the type species for the genus Fragilaria . Delgado et al. (2015) refers to rounded floating discs covering the areolae openings for F. candidagilae .

Interestingly, SEM illustrations of W. iraqiensis , the other species of Williamsella recently described from a spring-fed saline lake in Southern Iraq ( Al-Handal et al. 2016), also suggest that these external coverings are easily dissolved. They are missing from all the valves they illustrated except for one, where they occur on the apex (their fig.13).

In summary, it would appear that the disc-like external coverings of the areolae, usually referred as volae or rotae, seem to be a common feature of Fragilaria species. It is likely that difficulty concerning assessing the presence or absence of volae or rotae in some species is due to dissolution either in the field or as an artifact of the cleaning method as suggested by our observations (contrast Figs 42–44 taken from cleaned material derived from surface sediment samples with Figs 45–47 taken from untreated material observed on the glass fiber filters used for the chemical analyses of water samples). Molecular studies on Fragilariaceae also suggest that the various types of covering over the areolae do not correspond with monophyletic clades (Medlin et al., 2008, 2012).

In addition to the three characters discussed above, Al-Handal et al. (2016, see their Table 1) listed the presence of four plate-like plastids in Williamsella as another feature that differentiate this genus from the two plate-like plastids generally reported for the other Fragilarioid genera such as Ulnaria and Fragilaria . We found the same number and distribution of plastids in the Badain Jaran population (see Figs 1‒5 ) as the one shown by Al-Handal et al. (2016, Fig 7 ) for W. iraqiensis and this suggests that the two taxa are closely related. It is, however, not strictly correct that all Fragilaria species have two plate-like plastids as other configurations have been reported like in Fragilaria longifusiformis ssp. eurofusiformis Lange-Berlatot & Ulrich , which has undivided chloroplasts ( Lange-Bertalot & Ulrich 2014, plate 17, figs 5‒7) and in F. spectra , which has numerous and elongate plastids ( Almeida et al. 2016, figs 54, 55). In addition, the number and position of the plastids has not been reported for a majority of taxa currently considered to belong to Fragilaria . Therefore, while this character may be useful for the purpose of distinguishing species, it may not be appropriate at the generic level, at least until a more systematic study of the issue is carried out.

From the above, we conclude that erecting the genus Williamsella on the basis of the three characters presented in Graeff et al. (2013), i.e. colonial formation, habitat and areolae coverings and the additional character of the number of plastids ( Al-Handal et al. 2016), cannot be justified using these criteria. In addition, the structure of the cingulum (composed of several open copulae with a single row of areolae present along the advalvar edge) conforms to the description given for Fragilaria in Round et al. (1990).

One option is to emend the definition of both Fragilaria and Williamsella another is that W. angusta and W. iraqiensis should be re-assigned to the genus Fragilaria as it is currently defined. Based on the above, we propose the second option:

Figures

FIGURE 76. Distribution of Fragilaria crenophila var. sinensis in the 26 surface sediment samples included in the training-set and fitted generalized additive models of its responses along the salinity and total phosphorus gradients.

FIGURES 36–44. Scanning electron micrographs of Fragilaria crenophila var. sinensis var. nov. from Badain Jaran lakes. Fig. 36. External view of a whole valve. Fig. 37. Internal view of the whole valve. Fig. 38. External view of the centre of the valve. In the background, a detached valvocopula with a single row of areolae. Fig. 39. External view of apex depicting opposite striae, disc-like volae and the external opening of the rimoportula. Fig. 40. Internal view of apex showing the rimoportula. Fig. 41. Internal view of apex without a rimoportula. Fig. 42. Apex with external opening of the rimoportula (arrow). The areolae covering are partially dissolved. Fig. 43.Apex without rimoportula. Open copula (arrow). Fig. 43. Oblique view of an apex showing the ocellulimbus and in girdle view, components of the cingulum (labeled B1–B4) showing a single row of areolae and a scalloped pars interior. Scale bars = 10 μm (Figs 36, 37); = 2μm (Figs 38, 39, 42, 43); = 1μm (Figs 40, 41).

FIGURES 45–47. Scanning electron micrographs of Fragilaria crenophila var. sinensis from untreated material from Lake Shaobai Jilin, Badain Jaran Desert deposited on a GF filter (water sample collected in June 2007). Fig. 45. External view of the apex of a whole cell in girdle view. The cingulum is composed of four copulae (labeled B1–B4), each with a single row of areolae occluded with disc–like coverings. Fig. 46. Same as in Fig. 45 but with a cingulum composed of at least seven copulae (labeled B1–B7). Copulae with open ends (arrows). The areolae covering are partially dissolved. Fig. 47. Girdle view of central part of a broken cell showing disc–like coverings of the arolae and a cingulum composed of five girdle bands (B1–B5). Scale bars = 2 μm (Figs 46, 47); = 1 μm (Fig 45).

FIGURES 1–35. Light micrographs for the comparison of Fragilaria crenophila var. sinensis var. nov., Williamsella angusta Graeff, Kociolek & S.R. Rushforth and Fragilaria asiatica Hustedt. Figs 1–15. F. crenophila var. sinensis from lakes in the Badain Jaran Desert. Figure 12 is the holotype. Figs 16–25 W.angusta from Blue Lake, Utah, USA from isotype slide COLO 8508. Figs 26–35. F. asiatica from Tibet, from type slide KA–61. Figs 1–5, untreated material from epiphyton samples (Lake Shaobai Jilin, June 2007), bright field. Figs 6–35 cleaned preparations, taken in bright field and differential interference contrast. Scale bars = 10 μm.