Using external morphology of adults and tadpoles, osteology from high-resolution microcomputed tomography, vocalization analysis, and DNA sequence data, the identity of a reproducing Belgian population of invasive Xenopus at the current northernmost edge of the distribution of the genus in Europe was assessed. All data concur to an identification as Xenopus (Xenopus) laevis (Daudin, 1802). Genetically it is most closely related to populations of the Cape region in South Africa. No studies on the natural history of the Belgian Xenopus population and its impact on the local environment have been made to date.

Amphibians, France, freshwater biodiversity, invasive species, morphology, phylogeny, South Africa, systematics, taxonomy

The African clawed frog genus Xenopus Wagler, 1827 was recently revised by Evans et al. (2015) who described six new species and revalidated another, bringing the number of currently recognized species to 29, distributed among two subgenera. Their revision suggested that several populations might represent additional, undescribed species. Contrary to most species of Xenopus whose biology is poorly known, X. (Xenopus) laevis (Daudin, 1802) has been the subject of thousands of scientific and popular publications, because it is easy to breed in captivity and has been extensively used as an experimental model in biological and medical laboratories, and as a pet all over the world since the 1950s (Deuchar 1972; Cannatella and de Sá 1993; Gurdon and Hopwood 2000; Robert 2020; Fainsod and Moody 2022). Many individuals were released intentionally outside the species’ natural range or escaped from captivity, and the exceptional adaptability of this species, which originates from southern Africa, allowed it to establish reproductive populations in the wild in numerous countries in North and South America, in Asia, and in Europe (Peralta-García et al. 2014; Courant et al. 2017; Rödder et al. 2017; Mora et al. 2019; Wang et al. 2019). Within Europe, the northernmost populations were located in Wales and Lincolnshire in Great Britain, but they were recently extirpated, probably due to extreme weather conditions (Tinsley et al. 2015). The second northernmost populations within Europe were those of northern France, under study for two decades (Courant et al. 2017, 2018a, b, 2019), until the species was subsequently reported further north in a few Belgian localities along the French border (Gombeer et al. 2022; van Doorn et al. 2022).

External morphology is homogeneous among Xenopus species, and characters to differentiate them are often subtle (Evans et al. 2015). For instance, a reproductive population in Florida, first identified as Xenopus laevis by Hill et al. (2017) based on external morphology, was later shown, with the help of DNA sequence data and high-resolution microcomputed tomography, to belong to X. (Silurana) tropicalis (Gray, 1864) (Goodman et al. 2021), a species with a different body size and physiology, and hence potentially different invasion impacts and spreading abilities. DNA sequence analyses by De Busschere et al. (2016) demonstrated that the invasive French populations of Xenopus laevis involved two lineages, one from the southwestern Cape region, one from the northern area of South Africa (also see Furman et al. 2015). The expanding French populations of Xenopus laevis at the edge of the species’ distribution have shown morphological changes facilitating high dispersal abilities, in particular in the relative length of the hind limbs in both sexes (Courant et al. 2019), as well as physiological changes allowing to survive cooler weather conditions (Araspin et al. 2020; Padilla et al. 2020).

Considering that the systematics of the genus Xenopus is still in progress, that some introduced populations might disappear without having been accurately characterized, that expanding populations experience rapid morphological and physiological adaptations, and that a proper taxonomic identification allows to better understand the natural history, propagation risks and potential environmental impacts of an introduced population, it is important to voucher and document this northernmost, Belgian population. We do so hereafter, with the help of external morphology, osteology using high-resolution microcomputed tomography, DNA sequence data, and acoustics.

Using baited funnel traps and scoop nets, we (DB, JB, OSGP, and AV) collected Xenopus individuals in the afternoons of 8 and 15 September 2022 in a pond (50°45'19.7"N, 2°53'8.9"E) in Comines-Warneton village, Hainaut Province, Walloon Region, Belgium. The pond is located in an agricultural area between a corn field and a cow pasture. It is not connected to a stream, but located at a dozen meters from another pond, itself at a dozen meters from the Douve stream, an affluent of the Lys (or Leie) River, a left-bank tributary of the Scheldt (Escaut) River. The pond, partly shadowed by Salix trees and a Malus apple tree, is largely invaded by filamentous algae (Fig. 1). At the time of our visits, the pond’s maximum depth was approximately one meter.

Figure 1. 

Pond in Comines-Warneton village, Hainaut Province, Wallonia, Belgium, where Xenopus laevis were collected for the present study. Photograph by OSGP in September 2022.

Among the individuals collected, 11 were kept alive in two aquariums for behavioral observations. All others, i.e., 19 subadult and adult specimens and 14 tadpoles, were preserved as vouchers, euthanized with 10% ethanol following the procedure of the Institutional Animal Care and Use Committee (IACUC), fixed in 90% ethanol, and subsequently transferred into 70% ethanol for permanent storage in the herpetological reference collection of the Royal Belgian Institute of Natural Sciences (RBINS), Brussels.

The following external measurements (adapted from Tinsley 1973) were taken on the freshly preserved voucher specimens with a Mitutoyo Absolute Digimatic digital caliper to the nearest 0.02 mm: Body width: measured at widest point of abdomen in adults (in tadpoles, maximum width behind eyes); Eye diameter: horizontal diameter of the eye (circumorbital plaques not included); 5^th toe length: length of outer ventral surface of 5^th toe, from the base of the metatarsal to the toe tip; 1^st finger length: distance between base (between 1^st and 2^nd fingers) and tip of 1^st finger; Head width: measured along a line through the bases of the subocular tentacles; Hind limb length: distance between vent and tip of 5^th toe (limb extended as straight as possible); Internarial distance: distance between inner margins of nostril-bordering flaps; Interocular distance: minimal distance between eyes (eyes not including circumorbital plaques); Lower forelimb length: distance between outer angle of elbow to tip of 1^st finger; Nostril diameter: measured across long axis of nostril (including the bordering flap and papilla in adults); Snout length: perpendicular distance from the tip of the snout to a line through the subocular tentacles; Snout-vent length (SVL): measured from the snout tip to the anterior point of the vent; Tentacle length: measured from anterior base to tip of subocular tentacle; Tibia length: medial measurement along dorsal surface of tibiofibula.

Measurements specific to tadpoles include: Barbel length: length of extended barbel, from posterior base to tip (measured on the right side of the head except if the right barbel is damaged); Body length: measured straight along body-tail axis from snout tip to a point above the posterior extremity of the vent tube; Dorsal fin height: maximum height of the dorsal fin; Snout-eye distance: minimum distance between eye and snout tip; Snout-nostril distance: minimum distance between nostril and snout tip; Tail height: maximum tail height at the level of the posterior extremity of the vent tube or beyond; Tail length: measured from a point on the axis snout tip to tail tip above the posterior extremity of the vent tube to the tip of the tail; Tail muscle height: maximum height of the tail muscle at the level of the posterior extremity of the vent tube or beyond; Tail muscle width: width of the tail muscle above the posterior extremity of the vent tube; Total length: measured straight from snout tip to tail tip along body-tail axis (= sum of Body length + Tail length); Ventral fin height: maximum height of the ventral fin. The description of the tadpoles is adapted from Channing et al. (2012), Tapley et al. (2015) and Zahn et al. (2022); we follow Gosner (1960) to define their development stages. Paired measurements were taken on the right side.

The number of lateral line plaques of adults was counted on the right side between the posterior extremity of the eye and the vent; plaques outside the main plaque line are not counted. Sex of adults and subadults was determined based on the presence of a protruding cloaca (generally obvious in dorsal view) and the absence of nuptial pad on forearms (females), or on the presence of nuptial pads (even if sometimes the pads are not very contrasted in color, they are perceptible to the touch) on the fore arms and absence of protruding cloaca (males).

MicroCT scans were made at the RBINS facilities. Adult specimens were digitized using an EasyTom 150 (RX Solutions, Chavanod, France) with an aluminum filter at 10–30 W, 110 kV, 5.5–12.5 frames/s, 1440 projections per rotation, and 16.66 or 18.25 isotropic voxelsize for head scans, and 32.35 μm or 32.89 μm isotropic voxelsize for full body scans. Segmentation of the scans was done using Dragonfly software, v. 4.1 for Windows (Object Research Systems (ORS) Inc, Montreal, Canada, 2020). Visualization of the resulting meshes was done using GOM Inspect 2018 (Carl Zeiss GOM Metrology GmbH, Braunschweig, Germany).

Our voucher specimens were morphologically compared with literature data (Tinsley 1973, 1995; Slater et al. 2009; Evans et al. 2011, 2015, 2019; Hill et al. 2017 - taking into account the misidentification of Xenopus tropicalis as X. laevis in the latter reference; Porro and Richards 2017; Blackburn et al. 2019; Goodman et al. 2021; and references cited therein) and with preserved material housed in the collections of the RBINS, Brussels, and of the Royal Museum for Central Africa (RMCA), Tervuren, listed in the Appendix 1 (only a small part of the extensive Xenopus material housed in these two institutions is listed, according to the material we have personally revised).

Vocalizations of captive specimens were recorded with a Xiaomi Redmi Note 8T. The recorded vocalizations were imported in Raven Lite 2.0.4 (Ithaca, New York; The Cornell Lab of Ornithology). Within the software, the background noise was filtered. As the noise on the recording was continuously present on the entire recording, an adaptive filter was used to remove it. In the default settings of the adaptive filter, the broadband option was checked to remove a narrowband interference from a broadband signal. The filter order was set to 10 and the ALE (Adaptive Line Enhancer) delay to 1.

DNA-based species identification was done by extracting the DNA from six tadpole specimens using the Nucleospin Tissue Kit (Machery-Nagel), following the manufacturer’s protocols. Fragments of the mitochondrial genome were amplified; cytochrome oxidase I (COI) gene using the primer pair LCO1490 and HCO2198 (Folmer et al. 1994), cytochrome b (Cytb) gene using the primer pair CytbI and CytbII (Kessing et al. 1989) and a fragment of the large ribosomal subunit (16S) were amplified using the primer pair 16Sar and 16Sbr (Kessing et al. 1989).

PCR amplifications were performed in a total volume of 25 µl, containing 2 µl of DNA and 0.2 µM of each primer, and using 2× Qiagen® Multiplex PCR Kit with HotStarTaq® DNA polymerase with a final concentration of 3 mM MgCl₂. For all gene fragments, the PCR profile was 15 min at 95 °C followed by 35 cycles of 45 s at 95 °C, 45 s at 53 °C, and 60 s at 72 °C, with a final extension step of 10 min at 72 °C. All PCR products were purified using the ExoSAP-IT protocol (ThermoFisher) and were sent for bidirectional sequencing to Macrogen (Amsterdam, The Netherlands). Generated sequences were trimmed, corrected, and assembled using Geneious® v. 10.0.4 (Biomatters Ltd.). A consensus sequence was generated for each specimen. The generated COI sequences were only used for species validation and compared against GenBank using Geneious® v. 10.0.4 (Biomatters Ltd.) BLAST algorithm.

To allow a direct comparison, all Cytb and 16S sequences used by De Busschere et al. (2016) were downloaded and aligned using ClustalW in Geneious® v. 10.0.4 (Biomatters Ltd.). The alignments were checked for stop codons and trimmed to retain the 280 bp of the Cytb region and the 540 bp of the 16S region.

Haplotypes were determined by cutting alignments to equal length (final alignment size Cytb: 273 bp; 16S 544 bp) generating a Haplotype data file in DnaSP v. 6 (Rozas et al. 2017). Haplotype names agree with De Busschere et al. (2016). A minimum spanning haplotype network was made using PopART (Bandelt et al. 1999). All generated DNA sequences were deposited in GenBank (accession numbers: COI, OQ318517: OQ318522, Cytb, OQ343503: OQ343508 and 16S, OQ318448: OQ318453).

The combination of cloacal lobes not ventrally fused, absence of a claw on prehallux, absence of spicules on dorsum, presence of a dermal ridge along the 1^st toe from the prehallux, relatively long feet, relatively large eyes, more than a third of the eye covered by the lower eyelid, fused nasal bones, and the presence of vomer bones in the palate unambiguously shows that the Xenopus population from Comines-Warneton belongs to the subgenus Xenopus (see diagnostic characters in Evans et al. 2015), excluding the subgenus Silurana and the four species it currently contains (X. (S.) calcaratus Peters, 1875, X. (S.) epitropicalis Fischberg, Colombelli, & Picard, 1982, X. (S.) mellotropicalis Evans, Carter, Greenbaum, Gvoždík, Kelley, McLaughlin, Pauwels, Portik, Stanley, Tinsley, Tobias & Blackburn, 2015, and X. (S.) tropicalis (Gray, 1864)). Within the subgenus Xenopus, their dense marbled dorsal pattern readily separates the clawed frogs of Comines-Warneton from all species but X. borealis Parker, 1936, X. laevis, X. petersii Barboza du Bocage, 1895, X. poweri Hewitt, 1927 and X. victorianus Ahl, 1924, the latter three species being closely related to X. laevis but much smaller. The combination of the large size of adult males (SVL to at least 69.38 mm) and the absence of vomerine teeth distinguishes the clawed frogs of Comines-Warneton from all species of the subgenera Silurana and Xenopus except X. laevis (Evans et al. 2015, 2019; Porro and Richards 2017).

The tadpole morphology of many Xenopus species is unknown or fragmentary, often limited to a few Gosner’s (1960) stages. Still, a comparison among the known tadpoles shows strongly similar morphologies with few obvious differences (Channing et al. 2012; Evans et al. 2015; Tapley et al. 2015). However, the size, the low density, and the distribution of the melanophores on the dorsum and tail, as well as the relative shortness of the barbels agree with an identification as Xenopus laevis (Vigny 1979).

De Busschere et al. (2016) identified two phylogeographic lineages of South African origin invading France, one originating from northern South Africa and Rooikrantz Dam (SA5–SA6, referred to as the northern lineage), the other originating from the southwestern Cape (SA1–SA2). Despite the fact that 86% of the French specimens in the study of De Busschere et al. (2016) belonged to the northern lineage, sequences from the specimens from Comines-Warneton were most similar to sequences from Xenopus laevis from the Cape region (SA1–4, Figs 9–11).

Figure 11. 

Adult Xenopus laevis in a wetland (34°7'0.2"S, 18°22'39.3"E) at the Lake Michelle eco estate, Noordhoek, Western Cape, South Africa. Photograph by M. Burger.

Including our voucher series, plus the 11 specimens kept alive for behavioral observations and several specimens that escaped in situ, the sex ratio is approximately 1:1. In the natural environment of Xenopus laevis in South Africa (Fig. 12), females are often slightly more numerous than males but both sexes are sometimes found in equal proportions (Du Preez et al. 2005). However, it is known that males and females may respond differently to baited traps (Lobos and Measey 2002).

Figure 12. 

Wetland (34°7'0.2"S, 18°22'39.3"E) at the Lake Michelle eco estate, Noordhoek, Western Cape, South Africa, home to Xenopus laevis. Photograph by M. Burger.

While collecting Xenopus, two other amphibian species were observed in the pond of Comines-Warneton during our first visit: four juvenile Lissotriton vulgaris (Linnaeus, 1758) (Salamandridae) and two legged tadpoles and one just-metamorphosed individual of Pelophylax sp. (Ranidae). We observed no fishes in the pond, but we did observe high densities of insect larvae, including Odonata (Aeshnidae: Anax sp.; Coenagrionidae: Ischnura sp.; Libellulidae: Orthetrum sp., etc.), and aquatic Coleoptera (Dytiscidae: Agabus nebulosus (Forster, 1771), Hyphydrus ovatus (Linnaeus, 1761); Hygrobiidae: Hygrobia hermanni (Fabricius, 1775), etc.) and aquatic Heteroptera (Corixidae: Corixa punctata (Illiger, 1807); Notonectidae: Notonecta glauca (Linnaeus, 1758), etc.). During our visit to the pond on 10 October 2022, numerous adult Xenopus laevis were still actively going to the surface to breath, but not a single one was caught in our baited traps. This may indicate that, with the cooler weather of October and in preparation for the winter, they stop feeding or decrease their foraging activities.

It is difficult to assess when the pond in Comines-Warneton was first colonized by Xenopus laevis. The first observations of Xenopus in Belgium were made in 2006 in a pond (50°42'30.3"N, 2°52'43.3"E) in Le Bizet, and in a river directly connected to the Lys, the Rau d’Esseu (50°41'56.3"N, 2°54'28.0"E; van Doorn et al. 2022), at just 5.5 km S and 7 km S-SE, respectively, from the pond of Comines-Warneton (Fig. 13). The geographically closest confirmed established population of Xenopus laevis lives in a pond (50°40'13.4"N, 2°54'24.6"E; Fig. 14) in La Chapelle d’Armentières, northern France, at less than 10 km south of the pond in Comines-Warneton (Fig. 13). Given the relative isolation of the Belgian pond studied here, the abundance of Xenopus laevis in the pond, and the presence of tadpoles as well as subadults and large adults, it is certain that the pond has housed X. laevis for several years, and most probably a number of neighboring streams and ponds are also supporting them. The species is so secretive that the Western France population was discovered two decades after its introduction (Lorrain-Soligon and Secondi 2022). According to our discussions with local farmers, there is no indication that the propagation of Xenopus laevis from the closest locality in France was human-mediated. This implies that Xenopus laevis may have used the Lys and several of its affluents, at least the Douve and the Rau d’Esseu, to travel from France.

Figure 13. 

Map of Belgium showing the location of the pond in Comines-Warneton (largest, northernmost, red star), Hainaut Province, Wallonia. The three other red stars, smaller, show, from north to south, the locations of Le Bizet, Rau d’Esseu, and La Chapelle d’Armentières, respectively. Map by A. Tovar and JV.

Figure 14. 

Pond in La Chapelle d’Armentières, Nord Department, France, housing the northernmost French population of Xenopus laevis. Photograph by OSGP in September 2021.

Studies on sites in western France where Xenopus laevis is established suggest a negative impact on the native amphibian fauna and on the aquatic macroinvertebrate communities (Courant et al. 2018a, b). No studies on the natural history and the potential environmental impact of Belgian Xenopus laevis, at the northern edge of the genus’ distribution, have been made so far. The pond in Comines-Warneton, inhabited by at least two other amphibian species, is an ideal site for such studies.

We are grateful to Marc and Maxime Decroix (Mesen) and Bernard Devroede (La Howarderie, Le Bizet) for allowing access to their respective lands and ponds for African clawed frog surveys, and for useful information. We warmly thank Alejandra Tovar (RBINS) for preparing the map, and Marius Burger (North-West University, Potchefstroom) for providing photographs of Xenopus laevis and their biotope in the Cape. Steven J. R. Allain and an anonymous reviewer provided useful comments on a previous version of the manuscript.

Comparative material examined. The number of specimens is specified when it is above one. “DRC” = Democratic Republic of the Congo; “Prov.” = Province; “spec.” = specimens. Localities are given verbatim as appearing on the labels.

Xenopus (Silurana) mellotropicalis Evans, Carter, Greenbaum, Gvoždík, Kelley, McLaughlin, Pauwels, Portik, Stanley, Tinsley, Tobias & Blackburn, 2015: RBINS 13357–13358 (2 spec.), “Loango National Park, Ogooué-Maritime Prov., Gabon”.

Xenopus (Silurana) tropicalis (Gray): RBINS 165, “Bénué, Niger [sic; = probably Nigeria], Afrique occidentale”. RBINS 1811 (2 spec.), “Sénégal”. RBINS 13062 (2 spec.), “Tabou (4°25'N, 7°21'W), Ivory Coast”. RBINS 13028 (2 spec.) & RMCA 93-092-B-13–15 (3 spec.), “Lamin, western Gambia”. RMCA 93-092-B-24–27 (4 spec.), “Makumbaya, western Gambia”.

Xenopus (Xenopus) allofraseri Evans, Carter, Greenbaum, Gvoždík, Kelley, McLaughlin, Pauwels, Portik, Stanley, Tinsley, Tobias & Blackburn, 2015: RBINS 13359, “Loango National Park (2°20'37"S, 9°38'11"E), Ogooué-Maritime Prov., Gabon”. RBINS 13360, “Loango National Park (2°20'39"S, 9°38'12"E), Ogooué-Maritime Prov., Gabon”. RBINS 17089–17091 (3 spec.) & RBINS 17221, “Luki Reserve, Kongo Central, DRC”.

Xenopus (Xenopus) borealis Parker: RBINS 1094 (syntype), “Leikipia, 7–8000 feet, Kenya”.

Xenopus (Xenopus) gilli Rose & Hewitt: RBINS 7468, “Ottery, Cape Peninsula, South Africa”.

Xenopus (Xenopus) laevis (Daudin): RBINS 163, “Cap de Bonne Espérance” [= Cape of Good Hope, South Africa]. RBINS 163D, “Transvaal [= Transvaal Prov., South Africa]”.

Xenopus (Xenopus) muelleri (Peters): RBINS 9400 (9 spec.), “II/de/10, rivières à cours dénudé, Parc Nat. Garamba, Uele, Congo Belge [DRC]”.

Xenopus (Xenopus) poweri Hewitt, 1927: RBINS 4236 (7 spec.), “Mabwe, rive Est du lac Upemba, alt. 585 m, Parc Nat. de l’Upemba, Congo Belge [DRC]”.

Xenopus (Xenopus) parafraseri Evans, Carter, Greenbaum, Gvoždík, Kelley, McLaughlin, Pauwels, Portik, Stanley, Tinsley, Tobias & Blackburn, 2015: RBINS 17958–17959, “Ngoulmendjim (0°25'2.4"N, 10°34'4.7"E), Komo Department, Estuaire Prov., Gabon”.

Xenopus (Xenopus) pygmaeus Loumont: RBINS 18006, “Mukonoka River, Tshopo Prov., DRC.” RBINS 18204, “Mombongo, bank of the Congo River, Tshopo Prov., DRC”.

Xenopus (Xenopus) vestitus Laurent: RBINS 1811 (7 paratypes), “Rumangabo, Kinyantuku, Parc. N. Virunga, République du Zaïre [DRC]”. RBINS 1812 (paratype), “Rumangabo, Kinyantuku, Parc N. Virunga, Républ. Zaïre [DRC]”. RBINS 1813 (12 paratypes), “région Rutshuru, alt. ca. 1200 m, ex Parc N. Virunga, secteur sud, Républ. Zaïre [DRC]”. RBINS 1814 (paratype), “Rivière Bukengeri, alt. ca. 1200 m, région Rutshuru, ex Parc N. Virunga, secteur sud, Républ. Zaïre [DRC]”.

Xenopus (Xenopus) victorianus Ahl, 1924: RBINS 12675 (4 spec.), “Secteur Sud Rutshuru, Parc Nat. Virunga, Zaïre [DRC]”. RBINS 16835–16837 (3 spec.) & RBINS 17871, “Parc national de Rusizi, secteur Palmeraie, Burundi”. RBINS 16907, “Parc national de Rusizi, secteur Delta, Burundi”.

Xenopus (Xenopus) wittei Tinsley, Kobel & Fischberg: RBINS 1906 (37 paratypes), “Kalondo, alt. 1750 m, Lac Ndalaga, Mokoto, Parc Nat. Albert, Kivu, Congo belge [DRC]”. RBINS 1907 (4 paratypes), “Lac Magera, alt. 2000 m, Kivu, Zaïre [DRC]”. RBINS 1908 (4 paratypes), “Rivière Bishakishaki, alt. 2100 m, Kamatembe, Kivu, Zaïre [DRC]”. RBINS 1909 (9 paratypes), “Kanyabayongo, alt. 1750 m, Kabasha, Kivu, Zaïre [DRC]”.