Description of a new species of Potamonautes MacLeay, 1838, from the iSimangaliso Wetland Park, South Africa

Abstract A new species of freshwater crab, Potamonautes isimangaliso sp. n., is described from the western shores of False Bay, Hluhluwe, within the iSimangaliso Wetland Park, South Africa. While bearing a superficial resemblance to Potamonautes lividus, the new species has been found to be genetically distinct, diverging from the former by 7.4–7.8% in mtDNA. Potamonautes isimangaliso most closely resembles Potamonautes lividus, but is distinguished by a unique suite of carapace characters, colouration, and size. The new species also lives in close association with oxygen-poor, fresh ephemeral pans, while the habitat of Potamonautes lividus is well above the surface water line of the closest water body. An updated identification key for the Potamonautes species of South Africa is provided.


Introduction
Freshwater crabs play a key role in ecosystems by serving as an important food source for larger taxa, and recycling nutrients through detritivorous feeding habits (Cumberlidge 2009). They link terrestrial and aquatic habitats by moving between the two systems, and are considered bioindicator species of environmental change in some habitats (Schuwerack et al. 2001). Potamonautes is the only genus of primary freshwater crab (Yeo et al. 2014) in South Africa, with 16 described species occurring in the country, two having been described in recent years (Daniels and Bayliss 2012;Phiri and Daniels 2014).
The iSimangaliso Wetland Park forms the southernmost region of the Maputaland centre of endemism and constitutes the focus of biodiversity conservation in the region (Smith et al. 2008). Recent ecological and biodiversity surveys of Lake St Lucia have been conducted in an attempt to update local taxonomic records, identify undescribed species, highlight the change in diversity over time, and provide illustrated and annotated checklists to use as identification tools (Nel et al. 2012;Peer et al. 2014;Perissinotto et al. 2014). Surveys have revealed the existence of an undescribed species of Potamonautes along the western shores of False Bay (Fig. 1) that most closely resembles P. lividus Gouws, Stewart & Reavell, 2001 in morphological appearance, exhibiting a rounded vaulted carapace and the potential ability to spend a large amount of time out of water. However, genetic analysis showing a 7.4-7.8% difference across the combined 16S rDNA and COI gene fragments in comparison to P. lividus (G. Gouws unpubl.) and morphological analysis (present study) indicate that the two species are indeed distinct.
In this paper we describe Potamonautes isimangaliso sp. n. from the sand forests of the iSimangaliso Wetland Park. NP and GG wrote the taxonomic part of this study, including the description of the new species, while the contribution of the other authors dealt with natural history and ecological observations.

Collection of crabs
Crabs were collected from four localities (Fig. 1) using a sweep net in pans or by active hand capture. The unidentified species was found in 2012-2013 during routine surveys in the area as part of an ongoing project on the biodiversity of Lake St Lucia, supported by the iSimangaliso Park Authority and the provincial conservation authority, Ezemvelo KZN Wildlife, and subsequently during a dedicated survey undertaken in February 2015. Specimens were preserved in 10% formalin or 70% ethanol, once photographs were taken using a Canon Powershot G12.

Morphological and morphometric analyses
In the laboratory, 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.
Abbreviations for depositories and provinces: For the morphometric analyses, eight variables (CL, PFCD, CWP, ED, CH, AW6, CRPL and CRPW) were log transformed and used to run a stepwise discriminant function analysis in STATISTICA v 12.5 (Statsoft 2004). Data for Potamonautes lividus were obtained from G. Gouws and represent the specimens used for the de-scription of P. lividus (Gouws et al. 2001). Classification functions were calculated and individuals were then reassigned to groups based on a priori probabilities. Canonical scores were plotted for both species on a frequency histogram to support distinction between the two forms. Lastly, a linear regression analysis was used to examine variation for specific variables.

Genetic analysis
DNA was extracted from each specimen and amplification of the mitochondrial cytochrome c oxidase subunit I (COI) and 16S ribosomal DNA genes were carried out following protocols outlined by G. Gouws (unpubl.). Amplifications were confirmed by electrophoresis in 1% agarose gels with an ethidium bromide stain. The product was then viewed on an ultraviolet transilluminator. Sequences were generated from a representative of the new species, using approaches described elsewhere (Daniels et al. 2002;Phiri and Daniels 2014;G. Gouws (unpubl.)).

Morphometric analysis
The new species (Potamonautes isimangaliso sp. n.) was distinguished from P. lividus by its larger size, flatter carapace and more rounded posterior. The carapace variables CL, CH and CWP contributed the most to distinguishing between the two forms in the discriminant analysis. Fig. 2 highlights the morphometric distinction between the two species.
The classification function was calculated for both species as follows: Individuals were reassigned to groups based on a priori probabilities using these classification functions. 100% of both forms were correctly classified with no individuals being reassigned. Three variables (CWP, PFCD and CH) were regressed over CL and a significant difference was seen between the two species as follows: CWP/CL -SS = 0.1, df = 2, F = 2.29, p < 0.001; PFCD/CL -SS = 0.1, df = 2, F = 2.29, p < 0.001; CH/CL -SS = 0.1, df = 2, F = 5.99, p < 0.001 (Fig. 3a, b, c)  Diagnosis. Main distinguishing features of P. isimangaliso from P. lividus Gouws, Stewart & Reavell, 2001 as follows: slightly granulated, horizontal anterolateral margin more rounded than in P. lividus; downward projection of postfrontal crest at exorbital edges; uniform colouration of dark purplish brown with lighter or orange coloured joints, cheliped tips and pereopods tips. Potamonautes isimangaliso is larger than P. lividus, with a maximum size of 37 mm CL recorded in males.
Third maxillipeds (Figs 4c, 5e). Filling entire buccal frame except transversely oval respiratory openings at top lateral corners. Ischium slightly scabrous, with pronounced groove running vertically. Flagellum on exopod of third maxilliped fairly long, curving upward at distal ends.
Mandibular palp (Fig. 5c, d). Consisting of two segments; terminal segment undivided and smooth, with dense tuft of setae protruding from base; hirsute margins; light covering of setae on posterior surface; subterminal segment enlarged distally where it joins with terminal segment.
Pereopods (Figs 4a,b,5a,b). General right-handedness and prominent inequality of chelae where CRDL/CLDL = 1.32. Dactyl of major chela moderately arched; large interspace formed in major cheliped when fingers are closed, long slim interspace formed by closing of fingers in minor cheliped. Twenty four cutting teeth present on the dactyl of major cheliped and 29 on dactyl of minor cheliped; 3 larger and more prominent than the rest. Propodus fairly inflated; right propodus larger (CRPL/ CLPL = 1.41) and wider (CRPW/CLPW = 1.75) than left. Left pollex with 25 cutting teeth and right propodus with 18. Carpus on either side containing one prominent tooth followed by one smaller tooth. Meri strongly granulated around margins; slender pereopods (pereopod 2: ML/MW = 3.28; pereopods 5: ML/MW = 2.17), pereopod 3 is longest and pereopod 5 shortest; ventral margins of meri smooth; ventral margins of propodi slightly serrated; dorsal margins of meri and propodi bearing fine sharp bristles; dactyli serrated and ending in sharp points. Pleon (Fig. 4b). First five segments broad and short, with segments 6 and 7 longer; segments 1-6 four sided, with triangular distally-rounded terminal segment (telson). Pleopods (Fig. 5f, g, h, i). Gonopod 1 bearing short terminal segment only 0.23 times the length of the subterminal segment. Terminal segment curving slightly away from midline when viewed posteriorly. Gonopod widest at base, with both subterminal and terminal segments tapering and ending in sharp point. Inner lateral margin of subterminal segment irregular; outer lateral margin curving in a concave manner towards middle of gonopod; both margins hirsute. Groove extending almost through entire length of gonopod, visible on posterior surface, lined with setae. Distal margin of subterminal segment irregularly curved. Gonopod 2 consisting of two segments; terminal segment relatively long (0.47 times length of subterminal segment), very slim; subterminal segment wide at base, sharply becoming very narrow around 0.4 of length at which point narrow process forms, leading up to terminal segment. Small tuft of setae present on outer margin of base of subterminal segment. Gonopod 2 curved, moving outwards away from medial line of gonopod proximally, curving back towards medial line distally.
Variation. The major cheliped does not always display a pronounced interspace when fingers are closed. In juveniles and in the female allotype this was less prominent. The arching of chelipeds varies too, with some (particularly the minor chelipeds) bearing straight dactyli while others are fairly rounded. All collected specimens display a pronounced heterochely and all appear to be right-handed.
Live colouration. Colouration of carapace may vary between light brown, maroon, purplish-brown and almost black. Similarly, tips of dactyli may be either orange, bright yellow or a dull yellow.
Distribution. Currently only known from the False Bay region of the iSimangaliso Wetland Park on the north-east coast of South Africa. Etymology. The species is named after the iSimangaliso Wetland Park, located in northern KwaZulu-Natal, where it is currently thought to be micro-endemic. This is significant as the iSimangaliso Wetland Park falls within the Maputaland centre of endemism (Smith et al. 2008), highlighting the importance of this park as a global biodiversity hotspot. The Park is a UNESCO World Heritage Site and contains three Ramsar Wetlands of International Importance.
Potamonautes perlatus (H. Milne Edwards, 1837), P. granularis Daniels, Stewart & Gibbons, 1998, P. sidneyi Rathbun, 1904, P. barbarai Phiri & Daniels, 2014and P. barnardi Phiri & Daniels, 2014 all have sharply-defined scabrous or granular epibranchial corners and prominent postfrontal crest, while P. isimangaliso has a heavily rounded smooth epibranchial corner and poorly-defined postfrontal crest. Potamonautes parvicorpus Daniels, Stewart & Burmeister, 2001 also displays a finely granulated anterolateral margin and rounded epibranchial corners, but the resemblance to P. isimangaliso is superficial, as it differs in the indentation of its anterior margin where P. parvicorpus bears a slightly indented anterior margin while that of P. isimangaliso lies straight. Further differences are seen in locality as the habitat of P. parvicorpus is restricted to high mountain streams in the Western Cape (Daniels et al. 2001).
Potamonautes clarus Gouws, Stewart & Coke, 2000, P. depressus (Krauss, 1843, P. brincki (Bott, 1960), P. flavusjo Daniels, Phiri & Bayliss, 2014 and P. lividus Gouws, Stewart & Reavell, 2001 all have smooth anterolateral margins and rounded smooth epibranchial corners but bear differences compared to P. isimangaliso. One of the diagnostic characters of P. depressus is the dorsally flattened carapace, where CL/CH = 2.3-2.6. Potamonautes isimangaliso has a more vaulted carapace with a CL/CH ratio of 1.6-1.8. Potamonautes brincki and P. clarus are smaller crabs (max CL = 27 mm in males for both species), preferring fast-flowing mountain stream habitats. Potamonautes flavusjo is ecologically distinct from P. isimangaliso and can be found in the Mpumalanga Highveld. In addition to this, the species is smaller and has flattened chelipeds, not adapted for burrowing . Light yellow chelipeds and ventral surfaces of pereopods characterise P. flavusjo.
Potamonautes lividus shares a similar distribution, outward appearance and preference for air-breathing with P. isimangaliso. However various differences exist between the two species.The level and angle of anterolateral margin differ, where P. isimangaliso bears an anterolateral margin lying on the same horizontal plane as the anterior margin. Conversely, P. lividus has an anterolateral margin which angles downward to join the anterior margin and thus sits higher than the anterior margin. The downward angle of postfrontal crest at exorbital edges is seen in P. isimangaliso but not in P. lividus. Carapace flatness is indicated by the CL/CH ratio which equates to 1.5 for P. lividus and 1.8 for P. isimangaliso holotypes. The maximum size (37 mm CL in P. isimangaliso and 25.5 mm CL in P. lividus), colouration (dark blue carapace with bright orange chelipeds in P. lividus and dark brown/purple carapace with brown or dull yellow cheliped in P. isimangaliso), inflation of chela with gap between propodus and dactyl (dactyl of P. lividus is more arched than that of P. isimangaliso) and the number of poorly-developed teeth on carpus (P. lividus containing one prominent and three rudimentary teeth; P. isimangaliso containing one prominent and one rudimentary tooth) further distinguish the two species. Gonopods of both species are very similar with the only difference being the tuft of setae found at the base of pleopod 2 in P. isimangaliso. Specimens resembling P. lividus were found in the Dwesa Forest, Eastern Cape and appear to be genetically nearly identical to P. lividus . This further substantiates the genetic distinctiveness of P. isimangaliso. The smallest mature male of P. isimangaliso recorded had a CL of 13.2 mm whilst all females recorded were mature (min CL = 18 mm).
Habitat and ecology. Potamonautes isimangaliso sp. n. inhabits freshwater ephemeral pans (maximum salinity recorded: 0.75) which fill up with fresh, oxygen-deprived water after rainfall events, mainly during the summer wet season. These pans are located along the western shores of False Bay, Lake St Lucia (Fig. 6a-d) in clearings of the sand forest biome of False Bay and are generally partially shaded. Potamonautes isimangaliso and P. lividus are found in close proximity although P. lividus has not been found in False Bay Park and P. isimangaliso has not been found outside of the park. Furthermore, a difference in habitat type between P. isimangaliso and P. lividus Gouws, Stewart & Reavell, 2001 is seen, where the latter is known to inhabit burrows well above the surface water line in Ficus and Barringtonia swamps (Gouws et al. 2001), while the new species was found in close association with ephemeral pans in sand forest habitat with burrows extending below the surface waterline. Vegetation types include the dominant canopy species Cleisthantus schlechteri, Hymenocardia ulmoides, Psydrax fragrantissima, Croton pseudopulchellus and Drypetes arguta (Kirkwood & Midgley, 1999), as well as various Acacia spp. (Moll 1980). Grass species such as Paspalum vaginatum and Eleochoris sp. are also closely associated with this environment (Moll 1980). Aquatic plants associated with the ephemeral pans include the reed Typha capensis, the sedge Juncus kraussii, the mangrove fern Acrostichum aureum and the duckweed Lemna sp. (Howard-Williams 1980).
Potamonautes isimangaliso adults form burrows on the banks of these pans (Fig. 7a), while juveniles are found either in burrows or free-crawling in shallow water (2-50 cm). The species lives sympatrically with P. sidneyi but is separated by habitat type, with P. sidneyi inhabiting flowing streams and able to withstand higher salinities of up to 9 (18 May 2013, Mpophomeni Stream, 27°57'7.17"S, 32°22'37.21"E). Oxygen levels in the pans inhabited by P. isimangaliso are quite low compared to flowing streams (Table 2).
Although the species appears to be more closely associated with water than its morphologically closest congener, P. lividus (Gouws et al. 2001), the low levels of oxygen characteristic of the pans along with the ephemeral nature of the waterbodies indicate a greater affinity for a terrestrial lifestyle, as it may be more effective to obtain oxygen through air-breathing. This has been recorded previously in various African freshwater brachyuran genera and a high-vaulted carapace may be indicative of this change, where periods of dryness favour the evolution of burrowing semi-terrestrial, air-breathing tendencies (Cumberlidge 1999;Cumberlidge 2009). Specimens of P. isimangaliso have been observed in deep burrows (30-50 cm) around desiccated pools. Because the rainy season in this area is generally restricted to the period November-April (late Austral spring to early Autumn), much of the population hibernates deep in the mud, where traces of moisture persist throughout the dry season. Crabs return to the surface only after major rainfall events have filled up the ephemeral pools. The summer of 2014-2015 had been particularly dry in the area, with substantial rain falling in the False Bay area starting only in the middle of January (69 mm during 15-17 Jan, 54 mm during 28-30 Jan 2015; False Bay Park Meteo Station). Numerous adult and sub-adult crabs were observed from 31 Jan to 3 Feb in the newly filled ephemeral pools but hardly any young juvenile, indicating that the previous drought conditions had not allowed spawning to happen yet.
The feeding ecology of the species is largely unknown, although Potamonautes crabs are generally thought to shift from a diet of aquatic invertebrates to a more herbivorous or opportunistic diet with age (Hill and O'Keeffe 1992). The chelar dentition is serrate and the larger crusher chela lacks rounded or molariform occlusive geometry in the proximal region, probably due to wearing down over time. The dentition of the chela is indicative of an opportunistic omnivorous diet (Yamada and Boulding 1998). 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.