Important new records of Pelomedusa species for South Africa and Ethiopia

Because of a recent taxonomic revision, the species identity of helmeted terrapins (Pelomedusa) became unclear in many regions of their wide distribution range. Based on mtDNA sequence data, here we present the first record of Pelomedusa subrufa sensu stricto for the South African province of Mpumalanga. In South Africa, this species was previously known only from a single record in the province of Limpopo. In addition, we provide evidence for the occurrence of at least two distinct Pelomedusa species in Ethiopia. A sample from southern Ethiopia (Omo Region) turned out as P. neumanni, while another sample from Koka Lake (Oromia Region, central Ethiopia) represents P. somalica. Also a historical museum specimen from Ethiopia, most likely collected south of the Shebelle River (Oromia Region), belongs to P. somalica. However, these two Ethiopian specimens of P. somalica represent highly distinct genetic lineages, which may actually correspond to two different species.


Introduction
Several papers provided evidence that the widely distri buted subSaharan helmeted terrapins of the genus Pelo medusa represent a species complex (Vargas-ramírez et al. 2010;Wong et al. 2010;Fritz et al. 2011Fritz et al. , 2014;;Petzold et al. 2014;nagy et al. 2015) and not a sin gle species as thought before (e.g.Wermuth & mertens 1977;Fritz & haVaš 2007;Van dijk et al. 2014).The recent revision by Petzold et al. (2014) formally recog nized ten distinct Pelomedusa species.However, five other genetically distinct lineages have been identified, which may correspond to additional species (Petzold et al. 2014;nagy et al. 2015).While some of these spe cies and genetic lineages are morphologically distinctive, others are difficult to tell apart using external morphol ogy alone (Fritz et al. 2014(Fritz et al. , 2015;;Petzold et al. 2014).Thus, in many cases, helmeted terrapins need to be stud ied genetically for reliable species determination.This leads to the uncomfortable situation that many of the old records across Africa (cf.iVerson 1992) cannot be as signed to a certain Pelomedusa species.
In Ethiopia, the species identity of all Pelomedusa records has become unclear.Helmeted terrapins are widely distributed in this country, with records along the southern border, in the Great Rift Valley, in the Blue Nile drainage system, and in the east of Ethiopia in the Ogaden Region (largen & sPaWls 2010).However, two genetically verified museum specimens collected in 'Abyssinia' raised the possibility that two distinct species of Pelomedusa, P. gehafie and P. somalica, might occur in Ethiopia (Petzold et al. 2014).These museum speci mens from the late 19 th and early 20 th centuries lacked exact locality data, and when the specimens were col lected, Abyssinia was often identified with present-day Ethiopia and Eritrea.Hence, the two terrapins could also have originated from the latter country, where P. gehafie occurs (Fritz et al. 2014;Petzold et al. 2014).Besides Eritrea, there exists only one somewhat questionable re cord for P. gehafie, from the region of Khartoum, Sudan (Petzold et al. 2014).Genetically verified records of P. somalica are only known from the Awdal Region, Somalia.There, two highly distinct lineages have been recorded approximately 40 km apart (Fig. 1), suggestive of two parapatric or even sympatric species (Vargas ramírez at al. 2010; Petzold et al. 2014).
For South Africa, the situation is different.There are many genetically verified records, providing evidence for the occurrence of at least two distinct Pelomedusa spe cies, P. galeata and P. subrufa sensu stricto.Like P. so malica, P. galeata is comprised of two deeply divergent genetic lineages, which may actually correspond to two distinct species (Petzold et al. 2014).One of these lin eages is widely distributed over most of South Africa, while there are only two sites known for the other one, both in the westernmost range of P. galeata (Northern Cape and Western Cape provinces; Fig. 1).Pelomedusa subrufa sensu stricto, widely distributed from southern Angola and Namibia to East Africa and introduced in Madagascar, has been recorded until now only from the South African province of Limpopo (Hoedspruit, western border region of Kruger Park; Petzold et al. 2014).
In the present paper, we report the second record of P. subrufa sensu stricto for South Africa and the first unambiguous, genetically verified identifications of hel meted terrapins for Ethiopia.

Materials and Methods
We sampled four helmeted terrapins from South Africa (one terrapin each from Eastern Cape, Northern Cape, KwaZuluNatal, and Mpumalanga) and one from south ern Ethiopia (Omo Region; see Fig. 1 and Table 1 for exact locality data).Small pieces of skin or tissue were clipped off and preserved in pure ethanol.Using these samples and the laboratory procedures described for fresh material in Fritz et al. (2014), we sequenced 343344 bp of the mitochondrial 12S rRNA gene (12S), 671674 bp of the cytochrome b gene (cyt b), and 810815 bp of the ND4 gene plus adjacent DNA coding for tRNAs.In addi tion, using the laboratory procedures for museum mate rial of Fritz et al. (2014), we generated homologous se quence data for a dry specimen from the collection of the Hungarian Natural History Museum, Budapest (HNHM 2004.84.1).This terrapin was collected in the early 1960s in central Ethiopia (Oromia Region; Fig. 1, Table 1), and is figured and described in detail in dely (1971).For this specimen, we obtained 252 bp of the 12S gene, 319 bp of the cyt b gene, and 435 bp of the ND4 gene.The DNA se quences of each individual were concatenated and merged with the alignment used by Petzold et al. (2014) andnagy et al. (2015).Then, Maximum Likelihood analyses were calculated as described in Petzold et al. (2014) and the phylogenetic placement of the studied terrapins was used for species identification.GenBank accession num bers of new DNA sequences are LN824005 -LN824022.

Results and Discussion
The concatenated sequences of the helmeted terrapins from KwaZuluNatal, Eastern and Northern Cape, South Africa, clustered together with the more widely distrib uted clade of Pelomedusa galeata, while the terrapin from Mpumalanga, South Africa, corresponded to P. sub rufa sensu stricto (Fig. 2).Also the small size and colour pattern of the Mpumalanga terrapin argue for this spe cies (Fig. 3).Moreover, its triangular pectoral scutes do not meet at the midline, a character sometimes found in P. subrufa sensu stricto, but never in P. galeata (Petzold et al. 2014).This represents the second record of P. sub rufa sensu stricto for South Africa and the first record for the province of Mpumalanga.The collection site is approximately 90 km away from the previous record in Limpopo province (Petzold et al. 2014).Both records are in the western border region of the Kruger Park, sug gesting a wider distribution of P. subrufa sensu stricto in the Kruger Park area.
The museum specimen from central Ethiopia (HNHM 2004.84.1;Fig. 4) was assigned to one of the two dis tinct lineages of P. somalica (Fig. 2).Consequently, it represents the first unambiguous record for P. somalica in Ethiopia and outside Somalia.The localities of P. so ma lica in Somalia (Petzold et al. 2014) are approxi mately 480 km away from the site in central Ethiopia.In contrast to this specimen, the helmeted terrapin from southern Ethiopia was revealed as P. neumanni (Figs 2 and 5), a species previously only known from Kenya and Tanzania (Petzold et al. 2014).The nearest genetically verified locality of P. neumanni lies in Kenya (South Horr, Marsabit;Petzold et al. 2014), approximately 320 km distant from the record in southern Ethiopia.It seems possible that the ranges of P. neumanni and P. somalica abut or overlap in the region of the Great Rift Valley lakes of Ethiopia, even though our records of the two species are approximately 470 km apart.
We cannot exclude the possibility that more species of Pelomedusa occur in Ethiopia.As outlined above, Petzold et al. (2014) (Pagenstecher 1902;Von erlanger 1904;sPrigade 1904), it is likely that NMW 24449 (P.somalica) was collected in presentday Ethiopia, whereas the itinera ry of William Jesse (Finsch 1869: plate 23) suggests for ZMB 15693 (P.gehafie) an origin in Eritrea.Never theless, P. gehafie could occur in Ethiopia because hel meted terrapins are known from the region of the Blue Nile (largen & sPaWls 2010; Fig. 1).This river con  1905).NMW 24449 has on the right side of the cara pace an additional fifth costal scute (Fig. 6), exactly as described by tornier (1905) as an abnormality for one of the seven specimens from von Erlanger's and Neumann's   expedition.NMW 24449 represents the same genetic lin eage of P. somalica as previously identified from a sin gle fresh Somali sample (Rugi, Awdal Region, N9.96980 E43.43250;Petzold et al. 2014), whereas the specimen from Koka Lake (HNHM 2004.84.1) belongs to the other highly distinct lineage of P. somalica.Until now, the lat ter lineage was only known from three terrapins from the vicinity of Borama (Awdal Region, Somalia, N9.97050 E43.14600; Petzold et al. 2014).If both lineages should turn out as distinct species, Ethiopia would harbour at least three species of Pelomedusa.

Fig. 2 .
Fig. 2. Maximum Likelihood tree for 189 Pelomedusa terrapins, rooted with Pelusios sinuatus.The tree is based on the 1848bplong alignment of mtDNA used by Petzold et al. (2014) and nagy et al. (2015) plus sequences from the six terrapins studied for the present paper.Chimerical sequence data from a helmeted terrapin from Swellendam, Western Cape, South Africa, were excluded (see Petzold et al. 2014).Numbers along branches indicate bootstrap values greater than 50 (not shown for terminal clades with short branch lengths).Terrapins discussed in the present paper are highlighted in orange.For explanation of other sample codes, see Petzold et al. (2014) and nagy et al. (2015).Inset: Pelomedusa neumanni, Omo Region, Ethiopia.
reported two genetically verified historical museum specimens from 'Abyssinia', one rep resenting P. somalica (voucher in the Naturhistorisches Museum Wien, NMW 24449, coll.Carlo von Erlanger 18991901) and the other P. gehafie (voucher in the Museum für Naturkunde Berlin, ZMB 15693, coll.William Jesse 1868).Abyssinia was identified by many 19 th and early 20 th century authors with what is now Ethiopia and Eritrea (e.g.rüPPell 1835; Finsch 1869).When the itinerary of Carlo von Erlanger is considered

Table 1 .
Helmeted terrapins examined for the present study.Abbreviations: HNHM -Hungarian Natural History Museum, Budapest; MTD -Museum of Zoology, Senckenberg Dresden (Tissue Collection).It is most likely that NMW 24449 is one of the specimens collected by Carlo von Erlanger and Oscar Neumann on 20 June 1900 south of the Shebelle River (= Wabbi or Webbi River, Oromia Region, Ethiopia, ap prox.N7.80000 E41.00000; neumann 1905; tornier (Petzold et al. 2014)he region of Khartoum, Sudan, from where a questionable record of P. gehafie exists(Petzold et al. 2014).