Within the genus Actinella F.W.Lewis, 71 taxa have been reported, including 57 taxonomically accepted species names and 3 accepted varieties. In addition, there are 5 species names that are currently of uncertain status and 6 names that have not been verified to date (Guiry and Guiry 2023; Kociolek et al. 2023). Synonymy seems to be quite rare in this genus. Following Algaebase (Guiry and Guiry 2023), only 2 species names and 1 variety name were considered as synonyms.

From the African continent and Indian Ocean islands, new species of the genus Actinella were described from Madagascar and reported from Mauritius, e.g. Actinella candelabrum (Manguin) Kociolek & K.Rhode (as Asterionella candelabrum Manguin), its form baculata Manguin, A. madagascariensis (Manguin) Kociolek & K.Rhode (as Asterionella madagascariensis Manguin), and A. bourrellyi Kociolek & K.Rhode (synonym of Asterionella madagascariensis var. minor Manguin) (Manguin 1949; Kociolek and Rhode 1998). Several other Actinella taxa have been reported from these two islands, such as A. australis (Manguin) Kociolek, K.Rhode & D.M.Williams (as A. punctata var. australis Manguin) (Manguin 1949).

Reports on Actinella from sub-Saharan continental Africa are rather limited. Actinella brasiliensis Grunow and A. pliocenica Hérib. & Perag. were reported from Sierra Leone (Carter and Denny 1982 and Woodhead and Tweed 1956, respectively). From the same country, Woodhead and Tweed (1957) described Actinella africana Woodhead & Tweed. Fusey (1964, 1966) mentioned the presence of Actinella eunotioides Hust., a species discovered in the Amazon basin in Brazil (Hustedt 1952), in the Central African Republic and described a new variety, A. eunotioides var. minor Fusey, and a new species, A. undulata Fusey. Other Actinella reported by Fusey (1964, 1966) from the Central African Republic, a country located in the northern part of the Congo Basin, include A. australis (as A. punctata var. australis Manguin) and A. fuseyi Schoeman (as A. robusta Fusey nom. illeg.). All these taxa were rare except for A. undulata which was “dispersed without being rare”. Actinella punctata F.W.Lewis was reported by Stoyneva-Gärtner and Descy (2020) in samples from the Congo River, although the identity of this tropical African taxon needs to be verified, as A. punctata was described from North America (Lewis 1864). Okito et al. (2021) also reported an Actinella sp. on herbarium material kept at the National Herbarium of the Congo at the Institut National pour l’Étude et la Recherche Agronomique (INERA), Yangambi (YBI), Democratic Republic of the Congo (D.R. Congo). More information on Okito’s observations of this taxon on Nymphaea lotus L. can be found in his master’s thesis (Okito 2019).

Our own observations of material collected in tropical Africa, including samples collected during the Boyekoli Ebale Congo 2010 expedition (e.g. Cocquyt and Taylor 2015) and in the framework of several projects thereafter (e.g. Cocquyt et al. 2016; Karthick et al. 2016), confirmed the presence of representatives of the genus Actinella in the Congo Basin. Besides valves resembling A. punctata, another taxon was observed that morphologically resembles A. eunotioides var. minor. Based on light microscope (LM) and scanning electron microscope (SEM) investigations, this taxon is here described as a species new to science: Actinella kufferathiana Cocquyt & J.C.Taylor, sp. nov.

Materials studied for the present study were collected between 2013 and 2017 in small rivers and streams in the Yangambi Biosphere Reserve, in the Tshopo Province, D.R. Congo.

CCA 3282: D.R. Congo, Libongo River, tropical rain forest, benthos (sand?) in the stream bed, 0°48’11.59”N and 24°31’53.80”E, collected by Myriam de Haan on 14 Nov. 2013.

CCA 4378: D.R. Congo, Moni River, just downstream from the campsite in the tropical rain forest, sand from the stream bed, 0°50’47.29”N and 24°32’36.63”E, collected by Nils Bourland on 5 Jul. 2017.

CCA 4379: D.R. Congo, Moni River, just upstream from the campsite in the tropical rain forest, sand from the stream bed, 0°50’47.76”N and 24°32’8.34”E, collected by Nils Bourland on 5 Jul. 2017.

CCA 4385: D.R. Congo, Ngima River, stream flowing through fields just outside the tropical rain forest, sand from the stream bed, 0°50’34.55”N and 24°29’15.14”E, collected by Nils Bourland on 5 Jul. 2017.

CCA 4386: D.R. Congo, Moni River, just downstream from the campsite in the tropical rain forest, sand from the stream bed, 0°50’47.29”N and 24°32’36.63”E, collected by Andreas Popelier on 18 Jul. 2017.

The samples were fixed in situ using a 20% v/v final concentration ethanol solution (Taylor and Cocquyt 2016). Part of these materials were cleaned with hydrogen peroxide (30%), rinsed four times with distilled water and mounted in Naphrax®. Light microscopy (LM) was carried out with an Olympus BX51 microscope equipped with Nomarski differential interference contrast optics (DIC) and an Olympus UC30 digital camera. For scanning electron microscopy (SEM), aliquots of the oxidized suspensions were filtered through 1 μm isopore® polycarbonate membrane filters (Millipore). Filters were air-dried, mounted on aluminium stubs using carbon tape, and sputter coated with a platinum layer of 20 nm and studied with a JEOL JSM-7100FLV Field Emission Scanning Electron Microscope (SEM) operating at 1 kV and 6.9 mm working distance.

Samples, cleaned materials, permanent microscope slides and SEM stubs are housed in the herbarium of Meise Botanic Garden (BR), Belgium.

Diatom composition for each sample was studied in LM by counting the first 500 valves observed. This allowed us to calculate the relative abundance of each taxon, including the new Actinella species.

A morphologically closely related species to Actinella kufferathiana sp. nov. is the tropical South American taxon Actinella eunotioides. Described from Jurucui Lake in Brazil, A. eunotioides was observed in the plankton (Hustedt 1952). The species is mentioned in the catalogue of plants and fungi from Brazil (Eskinazi-Leça et al. 2010) with a distribution of “native, not endemic; Southeast (RJ); Atlantic forest; epicontinental”. The taxon observed in our material from D.R. Congo differs from A. eunotioides mainly in that the dimensions of the valves are much smaller, and in that A. eunotioides has well-developed spines on the junction of the valve face and valve mantle, which are also of higher density (Table 1). The stria density is also sightly coarser. Moreover, the raphe endings are comma-shaped in A. eunotioides (Kociolek et al. 2001), while only curved at the head pole and slightly inflected on the valve face in A. kufferathiana. The valve length of A. eunotioides var. minor, a taxon described from the Central African Republic (Fusey 1964, 1966) on the other hand, corresponds with the smallest valves observed of the new taxon but the valves of this variety are narrower and the striae are denser. The ventral margin of A. kufferathiana is slightly concave while the illustration of A. eunotioides var. minor given by Fusey (1966) shows a slightly convex ventral side. Moreover, in the original description of the var. minor the absence of spines is mentioned. Although spines are present in A. kufferathiana, these are small and not always discernible in LM.

A few South American Actinella taxa such as A. pereunotioides Metzeltin & Lange-Bert., A. disjuncta Metzeltin & Lange-Bert., A. pseudohantzschia Metzeltin & Lange-Bert., and A. lange-bertalotii Kociolek show morphological similarities to A. kufferathiana. The valves of A. pereunotioides, described from Venezuela, South America (Metzeltin and Lange-Bertalot 2007), are significant larger (66–113 µm vs 23.5–75.5 µm) and more slender with a width of 4–5 µm (Table 1). Actinella disjuncta can easily be confused with A. eunotioides. The valve length of A. disjuncta (Metzeltin and Lange-Bertalot 2007) is similar to A. kufferathiana but the width is much smaller, 3.5–5.0 µm vs 5.0–7.0 µm, and the stria density much higher (Table 1). Moreover, the ventral valve margin is almost totally straight, lacking the slight expansion near the headpole as present in A. kufferathiana. The raphe endings are comma-shaped in A. eunotioides (Kociolek et al. 2001). Actinella pseudohantzschia, on the other hand, has broader valves (7.0–9.0 µm vs 5.0–7.0 µm), while the valve length and stria density are rather similar to A. kufferathiana (Table 1). The density of the spines is much coarser in A. pseudohantzschia compared to the new taxon (4–6 vs 12–15 in 10 µm).

Many other Actinella taxa described and reported from South America have a typical broadly rounded headpole and narrowly rounded footpole, e.g. A. brasiliensis Grunow, A. guianensis Grunow, or an undulating dorsal and ventral margin such as in A. mirabilis (Eulenst. ex Grunow) Grunow. These valve outlines are totally different from those of the taxon here described from the Yangambi Biosphere Reserve in the Congo Basin.

Actinella modesta Gerd Moser, Lange-Bert. & Metzeltin, described from New Caledonia (Moser et al. 1998) also shows some morphological resemblance with A. kufferathiana. In general, the valves of this taxon are longer (50–90 µm), and there is a marked difference in width (3.0–4.0 µm vs 5.0–7.0 µm) and the striation is denser (Table 1). Other Actinella taxa described from the Australian continent (Sabbe et al. 2001) have a different valve outline and are thus morphologically not closely related to A. kufferathiana. They differ among others in the valve shape and in the shape of the head pole which can be more clavate with much broadly rounded head pole or sub-rostrate to capitate.

Actinella sp. 2 (Siver et al. 2015: plate 1, figs P–Q) mentioned among the Actinella species described from fossil material from North America (Siver et al. 2015), is a small club-shaped species with some morphological resemblance to A. kufferathiana. This taxon, however, is much smaller (valve length 10–14 µm, width 2.5–3.5 µm vs 23.5–75.5 µm and 5.0–7.0 µm respectively) and has a higher stria density (20–22 striae in 10 µm mid-valve measured in Siver et al. (2015: plate 1, figs P–Q) as not given in the publication, vs 8–13 in 10 µm). The headpole of this Actinella sp. is more broadly rounded than in the described new species, and not protracted. This latter feature can also be the case in the smaller valves of A. kufferathiana.

There could be some doubts as to whether the structures at the edge of the valve face and mantle observed on Actinella kufferathiana are spines. According to Spaulding et al. (2021) “A spine is a pointed, silica extension of the valve, shorter than a seta. Spines may be solid or hollow, very long or tiny, single or many, and they may arise from different points on the valve in different taxa.” We suggest that the structures observed in the new species are all broken spines, except for the ones near the apices. When looking more closely, we can see that some of these structures are not broken and indeed represent spines (Fig. 3D short arrow). The larger apical spine is in most of the observed valves also broken. But why are most of them broken? This species was found in a river with sand bed and friction of the sand on the valves could be a possible explanation. Later erosion during sample storage could also be a possible explanation, but seems less likely.

Actinella kufferathiana was never common in the samples studied. In most cases it was present with less than 1% of the total diatom community which was dominated by species of Eunotia Ehrenb., and small naviculoid taxa, such as Nupela Vyverman & Compère and Mayamaea Lange-Bert., with an abundance of 37.6% and 50% respectively in sample CCA 4386. Among the Eunotia species, E. rudis Cocquyt & M.de Haan, described earlier from the Libongo river in the Yangambi Biosphere Reserve (Cocquyt et al. 2016) was observed. It has also been observed from the Moni and Loeo Rivers, and rarely some valves in the Isalowe River, close to one of its sources. Both taxa, A. kufferathiana and E. rudis, as well as Cavinula lilandae Cocquyt, M.de Haan & J.C.Taylor (Cocquyt et al. 2013), another species described from the Yangambi Biosphere Reserve, and also observed in the studied samples, are a typical component of the diatom flora of the rather pristine, sandy small rivers and streams of the Yangambi Biosphere Reserve in the Congo Basin.