New and interesting Surirella taxa (Surirellaceae, Bacillariophyta) from the Congo Basin (DR Congo)

Two new diatom taxa belonging to the genus Surirella, S. ebalensis and S. congolensis, are described from material of the Congo Basin, downstream Kisangani, DR Congo. The fi rst taxon is a small, rather common species in the studied material; the second a somewhat larger diatom that was only sporadically observed. The morphology of both taxa is examined with light and scanning electron microscopy. The differences between the new and other closely related taxa such as S. agonaensis and S. bonsaensis, and S. takoradiensis, S. tenuissima and S. pseudotenuissima, respectively, are discussed.


Introduction
Algal reports from for the Democratic Republic of the Congo (DR Congo), formerly Belgian Congo and Zaire, are very limited (e.g., Duvigneaud & Symoens 1949;Hustedt in Huber-Pestalozzi 1949;Kufferath 1956a, b;Compère & Symoens 1987, 1988;Compère et al. 1989;Golama Swana Kaketa 1992, 1996and Compère 1995).During the Belgian-Congolese Boyekoli Ebale Congo 2010 Expedition, covering 250 km of the Congo River and some of its tributaries between Kisangani and Bumba (see Nagy et al. 2011;Virgilio et al. 2011), samples for diatom studies were taken in an unexplored, from an algal point of view, part of the Congo Basin.Some extra samples were taken in this region during two smaller expeditions in the frame of the COBAFISH project.In the sampled area, where people from rural villages live predominantly from subsistence-based agriculture (closely associated with rivers) and fi sheries, pristine habitats are rather scarce.The area is characterized by many oil palm and rubber plantations, coffee, cacao, and tea and cotton cultivation, generally farmed by smallholders.The nearest area from where results on diatom studies have been reported in the past is close to Kisangani (Golama Swana Kaketa 1992, 1996;Compère 1995).
From samples taken during the aforementioned Boyekoli Ebale Congo 2010 Expedition, a new diatom, Cavinula lilandae Cocquyt, M. de Haan & J.C. Taylor was recently described (Cocquyt et al. 2013).The present paper deals with the description of two other new taxa, both belonging to the genus Surirella Turpin.The fi rst taxon is a small, rather common species in the studied material sampled in the Congo Basin during the 2010 expedition and during two smaller expeditions in November 2012 and September 2013.The second taxon is a somewhat larger diatom that was only sporadically observed in the studied material.

Material and methods
Material from the Congo Basin, downstream of Kisangani, was collected in May 2010 during the Boyekoli Ebale Congo 2010 Expedition.Supplementary samples were taken in November 2012 and September 2013 on two smaller expeditions in the frame of the COBAFISH project.For the description of the new taxa discussed in the present paper the following samples were examined: -CCA 2070: Lomami River, 0.49339° N and 24.16960° E, epiphytic on Nymphaea lotus Linnaeus; collected by François Darchambeau and Ernest Tambwe on 24 Nov. 2012; temperature 25.7 °C, pH 6.06, conductivity 22.2 μS cm -1 , 0.060 mg l -1 NH 4 , 0.005 mg l -1 NO 2 , 0.446 mg l -1 NO 3 , 0.086 mg l -1 soluble reactive phosphorous (SRP).
In contrast to the samples from the Boyekoli Ebale Congo 2010 Expedition which were fi xed in situ with 30% formalin, the samples collected on later expeditions in 2012 and 2013 were preserved in ethanol (20% vv.).Part of these samples were oxidized with hydrogen peroxide (30%), rinsed fi ve times with distilled water and embedded in Naphrax® mounting medium.Light microscopic (LM) investigations were carried out with an oil immersion objective (100×) using an Olympus BX 51 microscope, equipped with Nomarski differential interference contrast optics (DIC) and an Olympus UC30 digital camera.For scanning electron microscopy (SEM) aliquots of the oxidized material were fi ltered through 0.3 μm Millipore® fi lters and then in turn fi xed to aluminum stubs, air-dried and sputter-coated with 50 nm of gold.SEM studied were carried out both at the Botanic Garden Meise (Belgium) and the North-West University (South Africa) with a JEOL-5800LV operating at 25 kV and an FEI Quanta 2000 fi eld emission scanning electron operating at 10 kV respectively.
The raw and the cleaned samples and the permanent slides for diatom investigation are housed at the herbarium of the Botanic Garden Meise (BR).A copy of each permanent slide will be deposited at the Centre de Surveillance de la Biodiversité (CSB) or at the Herbarium of the University of Kisangani, DR Congo.
Water temperature, conductivity and pH were measured using an YSI ProPlus multiprobe fi eld meter equipped with a pH sensor 1001 ProSeries.Samples for nutrient analyses (NH 4 , NO 3 , NO 2 and SRP) were fi ltered and acidifi ed in situ, and kept frozen until spectrophotometric analysis in the laboratory of the Unité d'Océanographie Chimique, Département d'Astrophysique, Géophysique et Océanographie, University of Liège, Belgium.

Etymology
The specifi c epithet refers to the Lingala word for river which was used in the naming of the expedition, Boyekoli Ebale Congo 2010 (the Study of the Congo River 2010), during which numerous algal samples were collected.

Type material
Holotype Slide BR 4398 from sample CCA 2070, Botanic Garden Meise, Belgium (BR).The valve representing the holotype is here illustrated in Fig. 1A.

Type locality
Oriental Province, DR Congo, Lomami River (0.49339° N and 24.16960° E).Epiphyton on Nymphaea lotus; collected by François Darchambeau and Ernest Tambwe on 24 Nov. 2012.Figs 2-5) Axial area narrow to absent near the poles.In the middle of the valve the axial area becomes broader and then larger in the depressions than on the top of the porcae (transapical valve undulations).Porcae perpendicular to the axial axis, becoming radiate near the poles and reaching the axial area but not in the extension of each other near the axial area.Striae biseriate, seldom with uniseriate parts near the axial area (Figs 2C, 4B), about 30 in 10 μm.Near the fenestrae the striae often become triseriate, sometimes quadriseriate (Fig. 2D), and also present for a short distance on the raphe canal just above the fenestrae.Striae composed of 80-110 areolae in 10 μm.Numerous blunt spines scattered over the valve face (Fig. 2B-F).On the top of the porcae these spines can become large, up to 1 μm (Fig. 4A-B).In the depressions the spines look more like elongated granules.Another kind of ornamentation is also present on the valve face: small granules located on the striae, composed of an areola that seems to secrete silica on the outside of the valve face (Fig. 4C-D).Sometimes all areolae of almost the entire stria have these granules on the valve face.On the inside the elongated granules are characterized by a perforation of the wall (Fig. 4D) that has no rimmed margins as the normal areolae do (Fig. 5A).Blunt spines are also present on the keel; the spines orientated towards the valve face are of different length and arranged on a single transapical row (Fig. 2B).These spines are often transapically elongated near their base, the shorter ones more triangular in shape (Fig. 4B).On the side towards the mantle the spines on the keel are often bifurcated; they are draped over a large part of the indented mantle side, partially covering the fenestrae (Fig. 3C).The mantle face near the edge of the mantle at the junction with the valvocopula is smooth and bears a single row of small areolae and oval silica plaques (Fig. 3C).At the apical pole these silica plaques are replaced by around 10 short spherical shaped silica elements (Fig. 2F).Bi-to triseriate striae are present on the indented mantle face at the level of the fenestrae bars.Fenestrae are broader than the alar canals; 30 fenestral bars in 10 μm.External raphe endings are straight, not enlarged at both poles.Internally the raphe is continuous at the head pole while at the base pole the raphe slit is internally interrupted at different levels.Both slits end in a structure resembling a reduced helictoglossa, lacking the typical lip-like structure, and located farther away from the raphe ending (Fig. 5B-D).The girdle is composed of two bands.The open valvocopula band is not smooth but bears tiny granules (Fig. 3C); the second band with a broad ligula and the same tiny ornamentation except on the part of the ligula that is covered by the valvocopula (Fig. 3A-B).

Diagnosis
Valves heteropolar, lanceolate with broadly rounded apical pole and subacute base pole, becoming gradually narrower from the apical pole to the base pole; larger valves slightly constricted mid-valve.

Etymology
The specifi c epithet refers to the Congo River and the Democratic Republic of the Congo where the samples were collected.

Type material
Holotype Slide BR 4399 from sample CCA 2071, Botanic Garden Meise, Belgium (BR).The valve representing the holotype is here illustrated in Fig. 6G.

Morphological observations on scanning electron microscopy (Figs 8-9)
External view Axial area narrow.Striae biseriate, continuing on the keel also above the fenestrae and on the mantle side.Raphe endings near the apical pole straight not enlarged (Fig. 8B-C), near the base pole straight and slightly bent towards the valve face (Fig. 8D).Striae composed of round areolae, in 10 μm, biseriate and continuing on the raphe sternum.Fenestrae open and divided by fenestral bars, openings between the fenestral bars located in the prolongation of the striae.No ornamentation such as spines or silica granules present on the valve face, except for silica plaques present at the edge of the valve mantle at the place where the valvocopula is attached to the valve (Fig. 8E).

Internal view
Raphe continuous on the apical pole (Fig. 9A) and interrupted near the foot pole with straight and slightly expanded raphe endings (Fig. 9B-C).
Girdle composed of several open, unornamented bands.

Discussion
Surirella ebalensis sp.nov.resembles S. agonaensis Foged described from Ghana in West Africa (Foged 1966;Cocquyt & Kusber 2010).The shapes of the valves differ slightly, but the valve width is distinctly narrower in S. ebalensis sp. nov. (16.1-19.4μm) compared to the broader S. agonaensis (27.5-30.0μm).The valve length also is smaller: 40.3-64.4versus 78-80 μm, respectively.The alar canals are denser in Surirella ebalensis sp.nov.than in S. agonaensis: 2.2-3.0 compared to 1.4-1.6.Moreover, the wing projection is very distinct in the new species compared to the slightly distinct projection in S. agonaensis.Besides these differences observed in LM, the SEM study of S. ebalensis sp.nov.showed the presence of draped, often forked silica spine-like structures extending on the keel at the mantle side partially covering the fenestrae (Fig. 3C), these structures are absent in S. agonaensis (Cocquyt & Kusber 2010).The spines of variable length on the top of the porcae (Fig. 4A-B) as present in S. ebalensis sp.nov., have not until now been observed in S. agonaensis, where silica granules are the only ornamentation on the valve face.
The closely related S. bonsaensis Foged, also decribed from Ghana in West Africa (Foged 1966;Cocquyt & Kusber 2010) differs in valve shape, which is broadly linear and isopolar, and in larger valve dimensions (length 70-80 μm, width 24.5-26.0μm).There exists a small overlap in number of alar canals in 10 μm: 2.2-3.0 in S. ebalensis sp.nov.and 2.0-2.4 in S. bonsaensis.As in S. agonaensis the wing projection in S. bonsaensis is also less pronounced than the distinct projection in S. ebalensis sp.nov.. Moreover the striation is coarser (25 in 10 μm in S. bonsaensis, 30 in 10 μm in S. ebalensis sp.nov.) and in LM no spines or other ornamentation of the valve face is visible.
Spines on the keel in other Surirella taxa were also observed in material from Sulawesi (Bramburger et al. 2006), e.g., Surirella alata Hustedt, S. fi mbriata Hustedt and S. tubicola Bramburger & Hamilton.These spines are of different length than in the Congo material studied, but the spines in our material are often much larger.In the Sulawesi material spines were also observed on the valve face, for example in S. tenacis Bramburger & Hamilton.Both are hollow structures, different from the fl at full spines found in for example S. nervosa (A.W.F.Schmidt) Mayer.Some of the spine-like structures present on the valve face resemble the four spines present on the valve face of S. quadridentis Bramburger & Hamilton.However, these four spines of S. quadridentis resemble more the spine-like structures of Cymatopleura calcarata Hustedt: also hollow structures (Cocquyt 2001) which are probably processes and not true spines.More studies, including internal valve views and cell wall ontogeny are needed to understand these spine-like structures.
Surirella congolensis sp.nov. is closely related to S. takoradiensis Foged, described from Ghana in West Africa (Foged 1966;Cocquyt & Kusber 2010).Although both taxa are heteropolar and about the same length (18.7-33.0compared to 20.0-27.0 (38.0) μm for S. takoradiensis) the valve shape differs: lanceolate in S. congolensis sp.nov.while ovate to oval-elliptical in S. takoradiensis, slightly constricted in the larger valves compared to no constriction at all, subarcuate compared to rounded base pole.Valves are narrower in S. congolensis sp.nov.(5.5-6.9 μm, seldom up to 7.9 μm) than in S. takoradiensis (7.0-9.0 (10.5) μm), and the alar canals in 10 μm are mostly denser, 5.5-6.0 and 4.0-6.0,respectively.The fenestrae are lower in S. congolensis sp.nov.than in S. takoradiensis where the fenestrae are almost rectangular in girdle view.Surirella congolensis sp.nov. is also related to Surirella pseudotenuissima Leclercq, described from Sumatra by Leclercq (1983) using material of the Friedrich Hustedt Diatom Collection (BRM X4/75, picture of the type in Hustedt 1938, tab. 43, fi g. 8).Although S. pseudotenuissima is a tropical species it was reported from temperate regions in Europe (Belgium) by Denys (1985).It is a heteropolar species with a similar valve length of 25.0-57.0μm, but much broader with a width of 9.5-13 μm.The alar canals (3.4-5.0 in 10 μm compared to 5.5-6.0 in 10 μm in S. congolensis sp.nov.) as well as the striae (23-26 striae in 10 μm compared to about 30 in 10 μm in S. congolensis sp.nov.) are coarser.The new taxon is also related to Surirella tenuissima Hustedt which has heteropolar valves with a length of 17.0-38.0μm, a width of 6.0-11.0μm and 4.0-7.0alar canals in 10 μm, which are smaller than the fenestrae (Hustedt in Huber-Pestalozzi 1942).Surirella tenuissima is a littoral species reported from South America and Indo-Malaysia.In Sulawesi and the Philippines (Leyte) it was also found in the plankton (Hustedt in Huber-Pestalozzi 1942).

Fig. 2 .
Fig. 2. Surirella ebalensis sp.nov., type material from sample CCA 2070, Lomami River, DR Congo, SEM.External view.B, C, F = detail of the raphe keel with blunt spines orientated towards the valve face and which are draped over a large part of the indented mantle side.A. Overview.B-C.Detail of the valve ornamented with silica granules and blunt spines.D. Detail of the apical pole, showing the curved raphe endings.E-F.Detail of the foot pole showing the straight raphe endings.F. Short spherical shaped silica elements near the pole (arrow).Scale bars: A = 10 μm; B = 4 μm; C, F = 2 μm; D-E = 1 μm.

Fig. 3 .
Fig. 3. Surirella ebalensis sp.nov., type material from sample CCA 2070, Lomami River, DR Congo, SEM.External view.A-B.Detail of the girdle showing a part of a girdle with ligula (arrow) and the neighbouring interrupted band.C. Detail of the valve mantle with the draped silica spines (arrow) and the silica plaques (arrow) near the edge of the mantle and the valvocopula.D. Detail of the girdle band near the pole.Scale bars = 2 μm.

Fig. 4 .
Fig. 4. Surirella ebalensis sp.nov., type material from sample CCA 2070, Lomami River, DR Congo, SEM.External view.Details of the various types of spines on the valve face and the keel.A-B.Detail of the valve surface with the biseriate striae (arrow) becoming sometimes uniseriate near the axial area.C-D.Section of the valve face showing the simple perforation of the silica wall at the areolae.Scale bars: A-B = 2 μm; C-D = 1 μm.

Fig. 5 .
Fig. 5. Surirella ebalensis sp.nov., type material from sample CCA 2070, Lomami River, DR Congo, SEM.Internal view.A. Detail of the multiseriate striae, and the perforation of the elongated granules without rimmed margin (arrow).B, D. Detail of the continuous raphe near the apical pole with a reduced helictoglossa (arrow).C. Detail of the straight not expanded raphe endings at the foot pole and their reduced helictoglossae (arrows).Scale bars: B = 2 μm; A, C-D = 1 μm.
DiagnosisValves slightly heteropolar, broadly linear to slightly cuneate with broadly rounded poles and slightly constricted margins in the larger valves, smaller valves not constricted.Apical pole broader than base pole.Length: 39.4-64.5 μm; width: 16.1-19.4μm, in the constricted part 18.7 μm.Wing projection very distinct, alar canals smaller than fenestrae, 2.2-3.0 in 10 μm.Fenestral bars distinct in LM, 30 in 10 μm.Striae indistinct in LM, 30 in 10 μm.Valve face covered with numerous silica granules and spines.Spines also present on the keel.Porcae perpendicular to the axial axis, reaching the axial area and becoming radiate near the poles.