Type:— INDIAN OCEAN, La Grande Coulée, Ile de la Possession, Crozet Archipelago (sample BM290, coll. date 4.I.1998, leg. B. Van de Vijver), holo- BR-4717! (Meise Botanic Garden, Belgium), iso- slide 405! (University of Antwerp, Belgium). The holotype is represented by Fig. 94.
PhycoBank registration:—http://phycobank.org/103147
Synonym:— Eunotia paludosa sensu Van de Vijver et al. (2014)
To exclude from synonymy:— Eunotia paludosa Grunow 1862
LM (Figs 85–119): Frustules rectangular in girdle view with barely to very slightly concave ventral margins, frustule width 3.5–6.0 µm (Figs 85–87). Valves weakly arched with more or less convex dorsal margins. Ventral margins moderately concave in all cell cycle stages down to the shortest valves. Apices distinctly narrowed, strongly protracted, subcapitate and dorsally reflexed in all stages. Valve dimensions (n=50): length 17–44 µm [up to 70 µm in other populations on the sub-Antarctic islands (Van de Vijver et al. 2014)], width 2.0–2.5 µm (occasionally up to 3.5 µm), length-to-width ratio 17–20. Terminal raphe nodules close to the poles. Striae 18–20 in proximal parts, up to 22 in 10 µm in the distal parts. Areolae not discernible in LM.
SEM (Figs 120–126): Striae uniseriate throughout, composed of small, rounded areolae, ca. 50 in 10 µm (Figs 120–122). Mantle striae ventrally composed of up to 6 small, rounded areolae near the valve middle, only 2 near the apices (Fig. 120). Spines lacking (Figs 121, 122). Externally, raphe branches curving from valve mantle onto the valve face (Figs 120, 122). Terminal raphe fissures extending rather long onto the valve face in a distinct pore, more than halfway to the dorsal margin (Figs 122, 124). Single rimoportula present at one of both poles, external opening distinctly visible between the smaller areolae (Fig. 123, arrow). Internally, rimoportula located close to the helictoglossa (Figs 125, 126). Helictoglossa prominent at both poles (Fig. 125). Girdle composed of 2–5 open, perforated copulae, including the valvocopula (Fig. 121).
Etymology:—The specific epithet insularum is a plural genitive in Latin indicating here the sub-Antarctic islands where this species is found.
Distribution and ecology:—Surprisingly, the new taxon appears to be a distinct, yet up until recently unidentfied Eunotia species. After revision of its taxonomic identity on all sub-Antarctic islands in the southern Atlantic and Indian Ocean, it becomes clear that E. insularum occurs over almost the entire sub-Antarctic region, encompassing the Atlantic and Indian Oceans. Currently, the new species was observed on seven archipelagos and islands, such as the Falklands/ Islas Malvinas (reported as E. pseudopaludosa in Jüttner & Van de Vijver 2018), South Georgia, Iles Crozet, Iles Kerguelen, and Heard Island (on the latter 4 localities reported as E. paludosa; Van de Vijver & Beyens 1998, Van de Vijver et al. 2001, 2002, 2004, Van de Vijver et al. 2014). However, the species seems absent, so far, from all neighbouring continents and the entire Holarctic realm.
Eunotia insularum was frequently and abundantly observed (as E. paludosa) from wet acid soils and submerged to wet terrestrial mosses and bog ponds, mostly in peat-dominated valleys. A particular feature of all sub-Antarctic localities is the complete absence of Sphagnum species, replaced by mosses such as Drepanocladus uncinatus (Hedw.) Warnst. in the peat formation on these islands. On Ile Amsterdam, the most northern of the sub-Antarctic islands in the southern Indian Ocean, E. insularum was found associated with Sphagnum -dominated bog ponds, as this is the only island with this kind of vegetation (Van de Vijver et al. 2008, Flatberg et al. 2011, Chattová et al. 2021).