Redescription of Potamonautes sidneyi (Rathbun, 1904) (Decapoda, Potamonautidae) and description of a new congeneric species from KwaZulu-Natal, South Africa

Abstract A new species of freshwater crab, Potamonautes danielsi sp. n., is described from the southern region of the KwaZulu-Natal Province, South Africa. Potamonautes danielsi most closely resembles Potamonautes sidneyi which is re-described here, but can be distinguished by a suite of key morphological characters including carapace shape and width, slim pereopods, inflated propodi of the chelipeds, and the shape and terminal segment length:subterminal segment length ratio of the 1st gonopod. In a previous study (Gouws et al. 2015), a 9.2–11.8 % divergence was found in the mitochondrial COI and 16S genes of the Potamonautes sidneyi clade, allowing for the delineation of a new species. Despite the clear molecular distinction between the two species, it is difficult to separate them based on individual morphological characters, as there is a great deal of overlap even among key features. The new species is found in slow-moving mountain streams and pools at high altitudes between Umhlanga and Mtamvuna, in KwaZulu-Natal.


Introduction
Potamonautes sidneyi (Rathbun, 1904) was first described from "Natal, southern Africa" (presently KwaZulu-Natal Province, South Africa). The original description was sourced from the Muséum National d'Histoire Naturelle, Paris, allowing us to locate the syntypes. Two type specimens are known and are lodged at the Peabody Museum, Yale University (original catalogue number 1191). These were collected by Sarah Abraham in 1871, but no accurate locality data were provided. Rathbun (1904) listed another individual, collected from Port Natal (presently Durban), as P. sidneyi, but pointed out certain differences from the types, notably the lack of a concave ridge behind the eyes. Although the description (Rathbun, 1904) highlighted the main difference between P. sidneyi and P. perlatus (H. Milne-Edwards, 1837), it was brief and relatively vague with few quantifiable or measurable distinctions between the two species.
Potamonautes sidneyi is regarded as one of the most widespread potamonautid species, occurring from the eastern parts of South Africa, northwards to Zimbabwe and Malawi . Within the province of KwaZulu-Natal (KZN), P. sidneyi was thought to occur in the low-lying midlands regions from the Drakensberg to the coast and to inhabit the entire coastal zone from the Maputaland (northern KZN) to the northern border of Pondoland (northern Eastern Cape ending at the southern KZN border) Stewart 2001, Gouws et al. 2015). However, recent genetic analyses have shown that these KwaZulu-Natal populations include two distinct genetic lineages, i.e. a northern lineage and a southern lineage, with the divergences warranting recognition of these lineages as separate species (Gouws et al. 2015). The issue of which lineage corresponds to the described P. sidneyi was difficult to resolve due to various morphological similarities between the two lineages, the vagaries of the original species description and the lack of a precise locality for the type material (Rathbun 1904).
Examination of high resolution photographs of key diagnostic features, including the carapace, chelipeds and male gonopods of P. sidneyi type specimens alongside specimens collected from both lineages revealed that specimens from the northern Maputaland lineage likely represent P. sidneyi s. str. as they match the type specimens in terms of the following: the slim propodi of the chelipeds, the stout pereopods, the shape and terminal segment length:subterminal segment length ratio of the 1 st gonopod; and its larger size. The southern lineage, thus, represents a new species and is described in this paper by NP and GG while RP and NAFM contributed to the information on its ecology and natural history, and EL-W contributed to the redescription of P. sidneyi. The delineation of a new species at the northern border of the Pondoland region is significant, based on the distribution of the lineages revealed by Gouws et al. (2015), as it interrupts the unclear and often confusing transition between P. sidneyi and P. perlatus at locations where the two species overlap and are often morphologically indistinct (Barnard 1950, Gouws and.

Collection of crabs
Detailed photos of the original syntype and additional specimens of P. sidneyi were obtained from the Invertebrate Zoology Division at the Yale Peabody Museum (CT, USA) and the Muséum National d'Histoire Naturelle (Paris, France) respectively.
Crab specimens were collected from various localities around KwaZulu-Natal (Fig. 1). For the taxonomic description and morphometric analyses, crabs from localities 1 (Lake Sibaya), 2 (Mpophomeni Stream), 3 (Hluhluwe), 5 (Siyayi), 6 (Mhlanga), 7 (Oribi Gorge), and 8 (Mtamvuna) were used. The crabs from Mtamvuna, Mhlanga and Oribi Gorge were used to describe the new species, while the morphology and morphometric analyses of Potamonautes sidneyi were conducted using the syntype specimens, as well as crabs from Lake Sibaya, Mpophomeni Stream, Hluhluwe and Entumeni. Crabs were collected by hand or by net and preserved in 70% ethanol. Photographs were taken using a Canon Powershot G12 digital camera.

Morphological and morphometric analyses
For examination of P. danielsi type specimens, a pair of Vernier callipers was used to measure morphological variables. A Nikon SMZ25 microscope fitted with a Nikon Digital Sight DS-Fi2 camera was used for macro-examination and to take photos of gonopods and mouthparts. A Canon Powershot G12 was used to photograph the carapace and appendages.
For the redescription of P. sidneyi s. str., a lectotype was designated from the syntypes housed at the Peabody Museum.
Mandibular palp (Fig. 5C, D). Consisting of two segments; terminal segment smooth and undivided, with hirsute margins; dense tuft of long setae arising from base. Subterminal segment bulbous in appearance.
Pleopods (Fig. 5F, G, H, I). Gonopod 1 widest at base; both subterminal and terminal segments tapering, ending with sharp point. Medial margin fairly straight displaying extrusion near base; lateral margin concave relative to midline; both margins hirsute. Groove extending almost entire length of gonopod, visible on dorsal surface, lined with setae. Gonopod 1 terminal segment long (0.31 times length of subterminal segment), curving outwards (i.e. away from midline) when viewed dorsally. Gonopod 2 consisting of two segments. Distal segment very long (0.57 times length of basal seg- Figure 3. Regression analyses of morphometric measurements of P. danielsi sp. n. and P. sidneyi. Regression analyses of: A LogP2ML over LogP2MW B LogCRPW over LogCRPL; and C LogCWA over LogCL between the two species, P. danielsi sp. n. and P. sidneyi. All differences between regressions were statistically significant (p < 0.001).  (Rathbun, 1904). Male holotype CWW 47 mm (Yale Peabody Museum catalogue number 1191) A dorsal view B ventral view, and C cephalothorax, frontal aspect. Scale bar: 50 mm. Photos: Eric Lazo-Wasem. Variation. The species appears to be extremely variable, with the northern populations displaying a more pronounced scabrosity and granulation on the chelipeds and carapace with fine hairs in some cases. Pereopods range from slender to stout. The inflation of the propodi on the chelipeds varies too, with some specimens bearing slim chelipeds while others possess more inflated propodi similar to that of P. danielsi.
Live colouration. Usually a variety of brown, ranging from chocolate brown to light brown or beige. In some cases pereopods are lighter than the carapace itself, a feature seen more in northern populations.
Distribution. Known to extend from Lake Sibaya in the north to Durban along the coast (Fig. 1), based on recent sampling and Gouws et al. (2015). This species has also been recorded in the Mpumalanga Province.
Type locality. South Africa, Port Natal. Collected by Mme Sarah Abraham. The specimens were received by the Peabody Museum (Yale) in 1871, along with other material. There is a possibility that the crab was among material accumulated for years, before it was sent to Yale and therefore it is uncertain as to whether the specimens were actually collected in 1871. While Mme Abraham was known to reside in Maphumulo (approximately 55 km north of Port Natal), it is unknown where exactly the specimens were from.
Etymology. The species was named by Miss M.J. Rathbun in honour of Professor Sidney I. Smith from Yale University.
Remarks. Only two type specimens were indicated in the original description, i.e. one male and one female from Port Natal both reposited at the Yale Peabody Museum. The lectotype was chosen on the basis of its designation as the only male type specimen in the original description. Compared to the lectotype, the paralectotype appears to be the same on the basis of the carapace scabrosity and granulation, the stout pereopods and the slim propodus of the cheliped. Although two additional specimens were mentioned, they were not designated as types. The first, a single female belonging to the Muséum National d'Histoire Naturelle, was uncertainly classified as Thelphusa corrugata Heller (Milne-Edwards, 1869). Following examination, it resembles P. sidneyi s. str. based on the stout limbs and slender propodi of the chelipeds. Despite the wide distribution and lack of specific locality all three specimens represent the same species i.e. P. sidneyi s. str. Allotype: female, CL = 22.8 mm (Table 1), collection details as per holotype (SAMC A83488).

Potamonautes danielsi
Paratypes: (Table 1)  Diagnosis. Potamonautes danielsi exhibits smooth to lightly granulated carapace flanks and epibranchial regions compared to those of P. sidneyi s. str. which often display a pronounced scabrosity and heavy granulation. The anterolateral margin is straight and complete. Potamonautes daniesi has long slender pereopods and the propodi of the chelipeds are inflated in contrast to the stout pereopods and slim propodi of P. sidneyi s. str. Potamonautes danielsi has a low terminal segment length:subterminal segment length ratio of gonopod 1 compared to that of P. sidneyi s. str. with occasional faint granulation. Urogastric, cervical and intestinal grooves well-defined; cardiac and branchial grooves well-defined where attached to urogastric and cervical grooves, becoming poorly defined and faint towards edge of carapace. Epigastric lobes poorly defined above postfrontal crest; two slight indentations present, forked from midpoint of postfrontal crest. Postfrontal crest slightly granulated at branchial region, straight and distinct from epibranchial region to midpoint, curving downward at epibranchial region. Exorbital teeth present; no epibranchial teeth present. Flank of carapace scabrous, well-defined epimeral sutures dividing pterygostomial region from subhepatic and suborbital regions, well-defined pleural groove dividing subhepatic region from suborbital region.
Third maxillipeds (Fig. 7C, 8E). Filling entire buccal frame except for oval respiratory openings medially above maxilliped. Ischium scabrous, with wide groove running vertically. Flagellum on exopod very long, curved to form a loop.
Mandibular palp (Fig. 8C, D). Consisting of two segments; terminal segment smooth and undivided, with hirsute margins; dense tuft of setae emerging from base. Subterminal segment enlarged distally then compressed at joint with terminal segment.
Pleon (Figs 6B, 7B). Somites 1-6 four-sided with distally-rounded triangular terminal somite (telson). First 5 somites broad and short; somite 6 longer, about 1.6 times as wide as long, distal margins straight or slightly concave, lateral margins slanted towards medial line, lateral margins swell slightly at articulation with somite 5; telson terminally rounded, lateral margins concave, swell at articulation with somite 6; hirsute lateral margins Pleopods (Fig. 8F, G, H, I). Gonopod 1 widest at base; both subterminal and terminal segments tapering, ending with sharp point. Medial margin fairly straight; lateral margin concave relative to midline; both margins hirsute. Groove extending to almost entire length of gonopod, visible on dorsal surface, lined with setae. Gonopod 1 terminal segment short (0.21 times length of subterminal segment), curving away from midline when viewed dorsally. Gonopod 2 consisting of two segments. Distal segment very long (0.67 times length of basal segment), slim; basal segment with wide elongated base sharply becoming narrow around 0.4 of length; narrow process forming at this point leading to distal segment. Gonopod 2 fairly straight, barely curving outward when viewed ventrally; curving slightly inward towards medial line at tips of distal segment.
Variation. The species appears to be extremely variable, with the northernmost Mhlanga population more closely resembling P. sidneyi s. str. rather than the P. danielsi sp. n. holotype. The epibranchial corners of the Mhlanga type are scabrous and granulated. Granulation, however, is not as pronounced as in P. sidneyi s. str. and no fine hairs are observed on the carapace. The terminal segments of both gonopods in the Mhlanga type are also more curved (typical of P. sidneyi s. str.), as opposed to straight (typical of P. danielsi). The flagellum on the exopod of the third maxilliped is highly variable. In the Mtamvuna population, the flagellum is long and curves backward to form a loop in some specimens. In both the Oribi and Mhlanga populations this is not seen. Instead, the flagellum curves upwards, similar to the pattern observed in most other potamonautid species.
Live colouration. Variable. Carapace colour ranges from purple to reddish-brown to greenish-brown. Carapace and pereopods are fairly uniform in colour with tips of dactyli and chelipeds usually displaying a lighter orange colour.
Distribution. Currently known to extend from Mhlanga (Durban North) to the Mtamvuna River on the northern border of Pondoland (southern KZN), based on recent sampling and the results published in Gouws et al. (2015). Morphological examination of museum specimens shows that this species is also present in the Mpumalanga Province.
Etymology. The species is named after Professor Savel Daniels in recognition of his valuable contribution to knowledge of freshwater crabs in southern Africa.
Potamonautes danielsi shares outward similarities with P. perlatus (H. Milne Edwards, 1837), P. granularis Daniels, Stewart & Gibbons, 1998, P. sidneyi Rathbun, 1904, P. barbarai Phiri & Daniels, 2014, and P. barnardi Phiri & Daniels, 2014. All the above-mentioned species display an angular epibranchial corner with granulation or scabrosity and prominent postfrontal crests ( fig. 4A; cf. Daniels et al. 1998: figs 2A, 10, 11A;cf. Rathbun 1904: plate XIV fig. 5;cf. Daniels et al. 1998: fig. 2F). Additionally, these five species are typically widespread, large, robust species occurring from the middle to lower reaches of rivers. Potamonautes granularis differs from P. danielsi in that it consistently exhibits orange-tipped chelipeds, the branchial region is highly convex and the anterior margin curves heavily inwards at the midpoint (cf. Daniels et al. 1998: figs 2A, 10, 11A), while P. danielsi does not always have orange-tipped chelipeds, has a flatter branchial region and a fairly straight anterior margin (Fig. 7A). Potamonautes perlatus, P. barbarai, and P. barnardi are morphologically indistinct ) and all differ from P. danielsi slightly (Fig. 9). The main difference lies in the anterior margin, which is similar in P. sidneyi s. str., P. perlatus, P. barbarai and P. barnardi. When viewed dorsally, the anterior margin of P. danielsi sp. n. lies relatively straight with a slight forward projection medially (Fig. 9B). The anterior margins of P. perlatus, P. barbarai and P. barnardi all contain a concavity in the crest behind each orbit so that a wide "W" is formed (Fig. 9A). The orbits of P. danielsi are deeper set than the orbits of P. perlatus, P. barbarai and P. barnardi. Additionally, the epibranchial corner of the former is more angular while those of the latter species group are slightly more rounded. However, even here variation across populations makes it difficult to differentiate between these species based solely on morphology.
Potamonautes danielsi and P. sidneyi s. str. are difficult to distinguish based on morphology alone, as key characters often overlap. The type specimens of the two species exhibit marked differences i.e.: 1) P. danielsi sp. n. has a smoother or slightly granulated anterolateral margin with a smoother or scabrous epibranchial region ( fig. 7A), while P. sidneyi s. str. typically has a heavily granulated anterolateral margin with a highly scabrous branchial region (plate XIV fig 5-Rathbun, 1904; fig. 4A) and even bears fine hairs on the carapace in some populations; 2) the propodi of P. danielsi sp. n. are inflated (Fig. 8A, B) while those of P. sidneyi s. str. are slender (Fig. 5A, B); 3) P. danielsi sp. n. has a high terminal segment length: subterminal segment length of gonopod 2 (Fig. 8H, I) and a low terminal segment length: subterminal segment length of gonopod 1 (Fig. 8F, G), while P. sidneyi s. str. has a lower terminal segment length: subterminal segment length of gonopod 2 (Fig. 5H, I) and a higher terminal segment length: subterminal segment length of gonopod 1 (Fig. 5F, G); 4) P. danielsi sp. n. bears slim pereopods (Figs 7A, B) as opposed to the stout limbs of P. sidneyi s. str. (Figs  4A, B). However, across the known range of distribution these individual characters vary significantly, with one species resembling the other on occasions. Their separation, thus, requires the inclusion of the whole suite of key characters including the shape and width of the carapace, the inflation of the propodi, the slenderness of the pereopods, the terminal segment length:subterminal segment length ratio of gonopod 1, and the shape of gonopod 1. The granulation of the carapace alone is not a reliable distinguishing character to tell these two species apart. Following morphometric analyses, the new species was distinguished from P. sidneyi s. str. mainly by the carapace variables CWA, CL and CH, which contributed the most to distinguishing between the two forms in the discriminant analysis (Fig. 2). The classification functions for both species were as follows: Individuals were then reassigned to groups based on a priori probabilities, using these classification functions. Ninety-one percent (91%) of the P. danielsi sp. n. individuals and 95% of the P. sidneyi s. str. individuals were correctly classified, with only three and one individuals being reassigned to the other species, respectively. The following three regressions were used to support the distinction between the two species: A. P2ML/P2MW, B. CRPW/CRPL, and C. CWA/CL (Fig. 3a, b, c). Regression analyses showed that the two species are significantly distinct using these morphological regressions (P2ML/P2MW-SS = 0.62, df = 1, F = 581, p < 0.001; CRPW/CRPL-SS = 1.11, df = 1, F = 939, p < 0.001; CWA/CL-SS = 0.59, df = 1, F = 2923, p < 0.001).

Discussion
In a previous study (Gouws et al. 2015), a 9.2-11.8 % divergence was found in the mitochondrial COI and 16S genes of the P. sidneyi clade. Based on the genetic delineation of the two lineages, now P. danielsi sp. n. and P. sidneyi s. str., in Gouws et al. (2015), the distribution of P. danielsi sp. n. is currently known to encompass the coastal zone of southern KwaZulu-Natal, i.e., the northern Pondoland region. However, genetic analyses of recently collected specimens from a wider range of localities (Gouws, un-published data) suggest a larger distribution of the species. It is likely that it also occurs in Swaziland, but this is yet to be confirmed. With more extensive inland sampling it is possible that the discrete distributions of these species suggested in the earlier study (Gouws et al. 2015) may not be consistent. Furthermore, the huge morphological variation and overlap in key characters obscure purely visual differentiation between these two species in the field. A whole suite of key characters, and possibly even molecular analyses, might be necessary to tell them apart with reasonable confidence.
Phylogenetically, within the southern African potamonautid fauna, P. sidneyi s. str. belongs to the clade of large-bodied, robust freshwater crabs, including P. perlatus, P. granularis, P. barbarai and P. barnardi (Daniels et al. , 2015, that are mostly confined to the middle to lower reaches of rivers . The derivation of these species from a common ancestor explains their overall morphological similarities. Potamonautes danielsi, however, appears to belong to a clade that includes burrowing species such as P. isimangaliso, P. lividus and P. flavusjo (Gouws et al. 2015), all three of which are easily distinguishable from P. danielsi. While the phylogenetic placement of these taxa requires more rigorous testing, it would appear that the morphological similarity between P. danielsi sp. n. and P. sidneyi s. str. is not phylogenetically determined and may well reflect habitat similarities and environmental drivers, given their current distributions. This may support the lack of distinctive species-specific characters that appears to be a widespread trend for African potamonautids (Daniels et al. 2003, Jesse et al. 2010, Cumberlidge and Daniels 2014, Phiri and Daniels 2016. South Africa is fairly well-studied regarding the taxonomy of freshwater crab fauna ) and yet ongoing molecular work still yields novel undescribed species of Potamonautes, many of which are cryptic and thus easily mistaken for previously recorded species in the past . The large number of freshwater habitat types present through South Africa, ranging from the headwaters of the Drakensberg region to coastal freshwater seepage points, is associated with the high diversity and endemism of freshwater brachyurans found in the country (Darwall et al. 2009). Some species appear to be quite adaptable, inhabiting a variety of habitat types, such as P. perlatus which extends from the inland region of the Northern Cape Province to freshwater seepage barrage pools along the Eastern Cape coast (Perissinotto et al. 2014), withstanding a wide range of altitude, salinity and temperature throughout its distribution. Conversely, certain species occupy very specific niches, such as P. isimangaliso which is currently only known from ephemeral pans along the western shore of False Bay, iSimangaliso Wetland Park (Peer et al. 2015a).
Potamonautes danielsi sp. n. does not appear to be habitat-specific but seems to prefer purely freshwater habitats established in areas with summer rainfall (Fig. 10). In Mtamvuna and Oribi Gorge, specimens were found under boulders and logs in mountain streams (altitude = 140-150 m) flowing into or connected to the main rivers. The Umhlanga Reserve consists mostly of KwaZulu-Natal coastal belt (CB3 vegetation unit-Mucina and Rutherford 2006) with a small portion of Northern Coastal forest (FOz7 vegetation unit). While the latter is classified as least threatened, its location next to an endangered CB3 habitat type within a growing urban area means that the surrounding habitat is already heavily transformed. The crab populations in this reserve were found close to sea level in the coastal belt habitat, burrowing under dominant grasses and shrubs near a freshwater seepage area which is connected to the Mhlanga River Mouth. Both the Mtamvuna and Oribi Gorge reserves consist of Scarp Forest (FOz 5), which is well-known for housing many endemic tree species and forms a core component of the Pondoland Centre of Endemism (van Wyk and Smith 2001). These forests exhibit tall and well-developed canopy and understorey tree layers that provide moist damp areas during the rainfall season (spring-summer), with adequate shade and shelter for the crab populations dwelling in the streams.
The species is sympatric with Chiromantes eulimene (de Man in Weber, 1898) in all three habitats, although the latter prefers the reed-like habitat adjacent to the main river, while P. danielsi appears to prefer the slower-flowing streams running into the main river body.
A greater number of P. danielsi adults were found under boulders and detritus in the water, as opposed to in burrows, although no feeding behaviour was observed at the time of collection. Generally, the feeding ecology of all Potamonautes spp. is supposedly opportunistic and thought to shift with age. Gut content analyses and stable isotope analyses have been conducted on P. perlatus (see Hill and O'Keeffe 1992) and P. sidneyi (see Peer et al. 2015b), respectively. While adults of these two species are mainly herbivorous and detritivorous, juveniles appear to favour a carnivorous diet. This could possibly relate to the ontogenetic shift in habitats, where juveniles occupy the water body, while adults reside in burrows on the banks of streams or rivers. Juveniles encounter more potential prey items on the benthos than adults do near their burrows. Thus, the presence of P. danielsi adults in the water body means that a wider range of prey is available to them for consumption. However, the overall ecology of the southern African potamonautids remains highly understudied. As new species are described, interspecific ecological differences are becoming apparent highlighting the need for more ecological research in this field.
Currently, 20 species of Potamonautes have been described in South Africa with six additional new but undescribed species (Phiri and Daniels 2016). Daniels et al. (2014) highlighted the relatively poor exploration of high altitude mountainous freshwater habitats and predicted that future collections from these understudied areas will yield new undescribed species. This sheds new light on inland freshwater habitats, in terms of conservation, as highly endemic or specialist species are often a priority in earmarking areas for protection. As recent national biodiversity assessments regarding South Africa's inland water systems have highlighted the threat faced by these largely unprotected systems (Nel et al. 2011), exploring and documenting their rich biodiversity and specialised ecology should be prioritised. and measurement of specimen MNHN B3841. We would like to thank Mr A Bosman and Dr W Florence of the Iziko South African Museum (Cape Town, South Africa); Ms A Ndaba of the Ditsong National Museum of Natural History (Pretoria, South Africa); and Dr H James and Mr M Mlambo of the Albany Museum (Grahamstown, South Africa) who were all exceptional in providing smooth access to museum specimens. We also extend our gratitude to Mr W Coetzer from the South African Institute of Aquatic Biodiversity (Grahamstown, South Africa) for the editing of the map. This work is based on the research supported by the South African Research Chairs Initiative of the Department of Science and Technology (DST) and National Research Foundation (NRF) of South Africa. Any opinion, finding and conclusion or recommendation expressed in this material is that of the author(s) and the NRF does not accept any liability in this regard.