karroica bogerti and Pachydactylus scutatus angolensis ( Sauria , Gekkonidae ) : new distribution records , comments on type localities and taxonomic status

In 1944 Loveridge described two new geckos from Angola, Afroedura karroica bogerti and Pachydactylus scutatus angolensis. The descriptions of both species have vague and confusing type localities and refinements are suggested based on early expedition reports, historical accounts from the region, and a review of cartographic material. Numerous new distribution records are reported for both species from expeditions undertaken from 1956– 2016 by the authors or their colleagues. The taxonomic status of both species has changed, but new material from diverse habitats, altitudes and geological substrates indicates that further taxonomic adjustments are likely in order to reflect additional cryptic diversity.


Introduction
Studies on the herpetofauna of Angola have entered a new phase following increased collaboration on regional and national biodiversity surveys, multi-authored modern taxonomic reviews, and the emergence of young Angolan scientists (Branch 2016).Almost all recent surveys (e.g.Huntley 2009;Brooks 2012Brooks , 2013;;Ceríaco et al. 2014Ceríaco et al. , 2016;;Ernst et al. 2014;Huntley and Francisco 2015;Conradie et al. 2016), have uncovered new national and regional distribution records (Branch and McCartney 1992;Branch and Conradie 2013;Conradie and Bourquin 2013;Ernst et al. 2014), as well as new species of amphibians (Conradie et al. 2012a(Conradie et al. , 2013) ) and reptiles (Conradie et al. 2012b;Stanley et al. 2016).However, problems arise when attempting to integrate these new findings with early literature and original species descriptions.Comparison is complicated by numerous factors, particularly the generally vague locality details for type material in early descriptions.It is not always obvious which localities or regions in these descriptions refer to modern towns in different parts of the country that have similar or identical names, and may be affected by the variant spellings of explorers and researchers.Exacerbating this is the loss of almost all of the material studied by José Vicente Barbosa du Bocage, the father of Angolan herpetology, in a fire that destroyed the Museu Bocage collections in 1978.As a consequence, many type specimens are no longer available and replacement 'topotypic' material requires confidence in the documentation of type localities.To illustrate these problems we consider two recent examples, Afroedura karroica bogerti Loveridge, 1944, and Pachydactylus scutatus angolensis Loveridge, 1944, where knowledge of the vague type localities given in the original description have become more defined following recent investigation and the integration of new and early observations.Both taxa are subsequently been elevated to specific status (see below).

Angolan Flat Gecko (Afroedura bogerti)
History.The Angolan Flat Gecko (Afroedura bogerti) is a beautiful and zoogeographically interesting species.For many years knowledge of the species was based on a single specimen collected during the Vernay Angola Expedition (VAE) in 1925.Charles Bogert, after whom the gecko is named, was the Curator of Herpetology at the American Museum of Natural History (AMNH), and prepared a detailed report on the snakes collected by the VAE (Bogert 1940).He never published on the extensive lizard collections, but did draw Loveridge's attention to a specimen that he thought might be a new species, and for which a superb drawing had been prepared (fig. 1 in Loveridge 1944).Loveridge confirmed this assessment and described the species in Bogert's honour.He considered it closely related to a South African species, Oedura karroica Hewitt, 1925, but did not consider the few African representatives of Oedura to be congeneric with Australian Oedura Gray, 1842, which had been based on the Australian species Oedura marmorata Gray, 1842.He therefore placed all African representatives of Oedura in a new genus, Afroedura, with his new subspecies, Afroedura karroica bogerti as the genotype.
Taxonomic status of Afroedura bogerti.Onderstall (1984) investigated relationships between Afroedura species, and on morphological grounds recognized three species groups.He examined material collected by WDH in 1971 and1974, and based on its possession of two pairs of scansors per digit and a verticillate tail placed it in his 'transvaalica ' group, along with A. transvaalica (Hewitt, 1925) andA. loveridgei Broadley, 1963.His 'africana' group, which included A. karroica, was characterized by having three pairs of scansors per digit and a verticillate tail, and he therefore elevated A. bogerti to a full species (Onderstall 1984).Based on material collected in 2009 from Tambor in the Iona National Park (see below), significant genetic divergence of A. bogerti from all other Afroedura (Jacobsen et al. 2014;Makhubo et al. 2015) confirmed its specific status.However, there remains confusion over its phylogenetic relationships.Jacobsen et al. (2014) and Makhubo et al. (2015) found that A. bogerti was part of the A. transvaalica group, although they differed in the latter group's relationships to other species groups.This association remains of zoogeographic interest as the range of A. bogerti is separated by nearly 2000 km from other members of the A. transvaalica group (sensu Jacobsen et al. 2014).However, among Angolan geckos it is easily recognized by its dorso-ventrally flattened body, smooth homogenous dorsal scalation, possessing 1-2 pairs of scansors under the fourth toe, and the verticillate tail.
Type locality of Afroedura karroica bogerti.In his description of Afroedura k. bogerti, Loveridge (1944) noted that the type specimen (AMNH 47841) was an adult male from Namba (Mombolo), Cuanza Sul Province, Angola, collected by Harry and Allan Chapman, between September and November 1925.Although the locality was not defined in more detail by Loveridge, Bogert (1940) had previously given general geographical co-ordinates for Mombolo (12°10'S, 14°50'E), but did not mention Namba.Crawford-Cabral and Mesquitela (1989) mapped the vertebrate collections recorded from Angola but overlooked Loveridge's (1944) two gecko descriptions.They did record Bogert's (1940) Mombolo (as 'Mombola') using the details given in his paper, and also listed Namba (Missão da Namba, 11°55' S, 14°51'E), approximately 27 km north of Bogert's Mombolo locality.It should be noted that the geographical co-ordinates listed for both these localities are relatively vague and provide only degrees and minutes.

A B
There is no detailed history or route for the VAE, but a brief summary of the expedition is given in the Mammals of Angola (Hill and Carter 1941, p: 3).It records (salient comments abstracted): Messrs. Herbert Lang and Rudyerd Boulton, collectors, went to Angola in April, 1925, and  The geographical locations of sites detailed in Hill and Carter (1941) are again relatively vague: i.e.Chipepe, near Cassongue (12°S, 15°E); Monte Victoria Verdun (12° 05'S, 15°E); and Namba (Mombolo) (11°35 S, 14°25' E).
Although Loveridge (1944) recorded that the type specimen of A. k. bogerti was collected during the VAE, he specifically states that it was "collected by Harry and Allan Chapman, between September and November, 1925."This was after the American members of the VAE had departed.This discrepancy is probably explained by a comment in Hill and Carter (1941, p: 5) where they note that their mammal material included "….. fifty mammals … purchased from Mr. C. P. Chapman, collected at Chipepe, Namba, and Monte Victoria Verdun.", and it is thus possible that Loveridge's type specimen was obtained in the same way.
Many of these locations are discussed in the recently published 'Reminiscences' of William Chapman, the father of Charles, Harry and Alan Chapman (Stassen 2010).The latter two sons (note corrected spelling of Alan, not Allan) are the collectors of the type specimen of A. k. bogerti (Loveridge 1944).William Chapman was the son of the famous South African explorer James Chapman (Chapman 1868), and one of the original 'Trek Boers', who moved away from British rule in the Cape region, and eventually settled  in southern Angola.
Chapman notes "Ernst Meyer had selected a farm in the Mombolo country at the base of a mountain range called Namba" (Stassen 2010, p: 295).He later comments "The next day I crossed the Etala stream [a tributary of the Cuchem] with my wagon and went up to the small plain below the koppie on which I subsequently built my house" (Stassen 2010, p: 297).He called it Monte Victoria-Verdun (Stassen 2010, p: 25 notes it was originally called Sandula, but see below).
It was noted (Stassen 2010, p: 13) that William Chapman: ….. liked the company of scientific explorers and quite a number of them were invited to stay for some time on his farm.As a result the type specimens of quite a few species of fauna and flora were recorded at Sandula, Namba or Chipepe, the farms of the Chapman clan near Mombolo.
A thorough examination of official detailed cartographic material supported by the historical accounts, allowed us to identify the possible location of Chapman's farm on topographic maps as "Fazenda Monte Verde", located on the northeastern slopes of the Lupangue mountains, close to the headwaters of the Etala stream, which drains between rock outcrops from the mountain.We were able to confirm this following a site visit (5-6 November 2016), and from inspection of the original "Titulo de Concessão", dated 28 January 1930 (Fig. 1A), that confirms the farm's ownership by William Chapman.The farm was purchased by the Kath-Brock family, who renamed it as Monte Verde.The original farm map attached to the title deed reveals that the "koppie" (hill) noted by Chapman was originally called Sandula (Fig. 1B).The brick farm buildings of the Brock family (12°10'09"S, 15°01'42"E, 1798m a.s.l.) were built on, and incorporate some of, the original Chapman house, but both are now derelict, whilst Sandula Hill (12°11'04"S, 15°01'41"E, 2242m a.s.l) is locally now called Monte Verde.This prominence is capped in granite outcrops, with intrusive rock outcrops on its lower slopes and is close to the farm site.These outcrops form suitable habitat for flat geckos (Fig. 2).
Stassen (2010, p25) notes "Mombolo is a small village on the Benguela highlands in the Cuanza Sul district and is also called Omambolo, Nakamombolo or Makamombolo".Although two small quimbos (villages) to the southwest of Chapman´s farm have been identified in the cartographic material referenced as Mombolo, we consider that the name is more correctly associated with the general region to the south of Missão da Namba, as is shown in the Carta de Angola at the scale of 1/100 000, where the area immediately to the south of Missão da Namba is named Mombolo.This is an extensive grassy area named as the Anhara do Mombolo, and at the southern edge of which lies the present town of Maka-Mombolo (12°11'54"S, 14°52'06"E, 1760m asl) mentioned by Stassen (2010, p25).Local people refer to extensive flat natural grasslands as 'mombolos', areas suitable for planting staple crops.Missão da Namba is a Seventh-day Adventist church established by James Delmes Baker, who settled in the area in November 1928 (Anon, 2017).

New distribution records for
Currently Angolan Afroedura bogerti is known from approximately 15 localities that appear to fall into disjunct populations occurring above and below the Great Escarpment, and in diverse habitats and on different geological substrates.Colouration and scalation varies between the known populations of A. bogerti (Fig. 4) and preliminary morphological and genetic analysis corroborates the existence of cryptic diversity within the species (Branch et al. 2017).Molecular sequences for A. bogerti incorporated into recent phylogenies of the genus were based on tissues collected from lowland populations within the Angolan Namib region (Jacobsen et al. 2014, Tambor CAS 248780-81;Makhubo et al. 2015, Omauha Lodge, PEM R17936-37).A flat gecko, referred to as Afroedura cf.bogerti, was discovered in a rock crevice on the summit of the Otjihipa Mountains, northern Opuwo District, Namibia (approx.17°18'18"S, 12°36'35"E, 1170m) in 1992 (Branch 1998;Griffin 2003).However, no voucher material is available in the State Museum, Windhoek (A.M.Bauer, pers. comm. May 2016), and the taxonomic status of the Namibian population cannot be assessed until further material becomes available.

Angolan Thick-toed Gecko (Pachydactylus angolensis)
History.In the same paper in which he described Afroedura k. bogerti, Loveridge also described the small terrestrial gecko Pachydactylus scutatus angolensis Loveridge, 1944. Loveridge's (1944) description was based on three specimens.The holotype (AMNH 47874) was "an adult male from Hanha, Benguela Province, Angola, collected by Arthur Vernay, Herbert Lang and Rudyerd Boulton, May 17, 1925".One of the two paratypes was also collected at the type locality but on 13 May 1925, and the other collected earlier "from Lobito Bay, Angola … by Herbert Lang, April 24, 1925".These dates, unlike those for A. bogerti, indicate that the geckos were actually collected during the VAE.
Taxonomic status of Pachydactylus angolensis.Bauer et al. (2002), in a review of Namibian P. scutatus Hewitt, 1927 (including the description of P. parascutatus Bauer, Lamb and Branch, 2002), elevated P. angolensis to a full species, noting the consistent nasal arrangement.A recent molecular phylogeny of the genus Pachydactylus (Heinicke et al. 2017) confirms the specific status of P. angolensis, but reveals its sister relationship to P. cariculicus not P. scutatus.With the elevation of P. angolensis to specific status, P. scutatus was considered to be restricted to Namibia north of the Brandberg.However, WDH (unpubl.) had already collected the first Angolan P. scutatus from Iona National Park (TM 40751), and others from Espinheira (TM 40615-18), when he visited the region in April 1971.Recent material from the latter locality has also been reported by Ceríaco et  2016).The TM geckos from Iona and Espinheira have the characteristic narrow and dark-edged white nape band illustrated in Boone and Barts (2006) which is not evident in P. angolensis or mentioned in its description (Loveridge 1944).Thick-toed geckos (Pachydactylus) are poorly represented in Angola, with only seven currently recognized species recorded; P. cariculicus, P. scutatus, P. angolensis, P. punctatus, P. oreophilus, P rangei and P. vanzyli.This contrasts dramatically with Pachydactylus diversity in Namibia (38 species) and South Africa (30 species).Pachydactylus angolensis can be distinguished from most Angolan Pachydactylus (e.g.P. cariculicus, P. punctatus, P. rangei and P. vanzyli) by its keeled dorsal scalation, from P. oreophilus by its smaller size, and from P. scutatus by lacking a darkedged white nape band and by the exclusion of the first upper labial and rostral from the nostril (P.scutatus has a well developed nape band, and both the first upper labial and rostral enter the nostril).

Type locality of Pachydactylus scutatus angolensis.
Ango lan localities referenced under the toponym "Hanha" may cause confusion.Two main places are of concern, and both are situated in Benguela Province: Hanha do Norte (also Hanha Estate) and Hanha do Cubal.Bauer et al. (2002) noted that the holotype was ".. collected at Hanha (probably Hanha do Norte in quarter degree square 1213Bc; another locality, Hanha do Cubal, 1314Aa is also in the same province)".The above assumption is here accepted, insomuch as it avoided confusion between the two localities, and furthermore is corroborated by the fact that Hill and Carter (1941) had earlier noted that the VAE did not visit Hanha do Cubal, but rather the Hanha Estate oil palm plantation.Unfortunately, although the geographic description of Hill and Carter (1941) is basically correct, the co-ordinates they give (13°30' S, 14°30' E) are incorrect and refer to another Hanha toponym that lies southeast of the estate.It is indicated on a 1934 map of the region, and is now known as Cambondongolo (13°21'0''S, 14°24'0''E).To avoid further confusion we recommend that the type locality for Pachydactylus scutatus angolensis Loveridge, 1944 be restricted to -Hanha do Norte (12°14'42"S, 13°42'27"E), approximately 20 km northeast of Lobito, Benguela Province, Angola.This site is illustrated in fig. 1 of Hill and Carter (1941).Recent collections of geckos currently attributed to P. angolensis are grouped into coastal and inland populations (Fig. 5), with coastal populations (<300m a.s.l.) usually associated with consolidated marine deposits, while inland populations (>600m a.s.l.) are found in granite or intrusive outcrops.Preliminary morphological investigation indicates variation in dorsal colouration and habitus (Fig. 6) between these populations that are indicative of further cryptic diversity in Angolan geckos, and will be investigated with integrative morphological and genetic studies (Branch in prep.).

Figure 1 .
Figure 1.A: (left) Photograph of the original "Titulo de Concessao" (28 January 1930) confirming William Chapman's ownership of the farm that was purchased by the Kath-Brock family.B: (right) the farm map from the "Titulo de Concessao" showing 'Alengo Sandula' at middle of lower boundary.

Figure 2 .
Figure 2. The ruins of farm Monte Verde, surrounded by eucalyptus trees (middle distance), overlay those of William Chapman's farm Monte Victoria-Verdun.Viewed from an intrusive rock outcrop on the lower northern slopes of Sandula Hill.

Figure 3 .
Figure 3. Geographical distribution of Afroedura bogerti.Insert (left) shows the polygon that encompasses the area from which the type specimen was collected.The problematic Namibian specimen is included.