Early Pliocene anuran fossils from Kanapoi, Kenya, and the first fossil record for the African burrowing frog Hemisus (Neobatrachia: Hemisotidae)

Abstract

Isolated amphibian bones from the early Pliocene of Kanapoi (West Turkana, Kenya) help to improve the scarce fossil record of the late Neogene and Quaternary amphibians from East Africa. All currently available 579 bones are referable exclusively to the Anura (frogs and toads). More than half of the remains (366) are identified as Hemisus cf. Hemisus marmoratus, an extant species that still inhabits Kenya, but apparently not the northwest of the country and the Turkana area in particular. The rest of the remains are identified simply as Anura indet. because of poor preservation or non congruence with the relatively few African extant taxa whose osteology is known in detail. The Hemisus material represents the first fossil record for Hemisotidae, an endemic African family of peculiar, head-first burrowing frogs, whose sister taxon relationships indicate a divergence from brevicipitids in the Late Cretaceous or early Paleocene. The ecological requirements of extant H. marmoratus suggest that the Kanapoi area surrounding the fluvial and deltaic settings, from where the fossil remains of vertebrates were buried, was likely a grassland or relatively dry, open low tree-shrub savanna.

Introduction

Gardner and Rage (2016) recently summarised the fossil record of anurans and other lissamphibians from Africa, updating and significantly expanding earlier attempts by Van Dijk, 1995a, Van Dijk, 1995b and Roček and Rage (2000). The African record is biased towards Neogene and Quaternary occurrences (about two-thirds of the total number of localities covered in Gardner and Rage's (2016) review were from those intervals). Most of those post-Paleogene localities are geographically restricted to the northern and southern portions of the continent and consist entirely of extant anuran taxa, many of which still live in the same areas. Tihen's (1972:12) statement “Africa remains terra incognita in the fossil record” written nearly 50 years ago in reference to African toads, remains broadly applicable to the fossil record of the entire continent for amphibians, mostly because the lack of extensive use of the appropriate collecting techniques represents a limiting factor to the retrieval of their remains (others likely include: difficult access to potential localities; limited and scattered exposures; limited interest in finding, collecting, and studying small vertebrate fossils). At the same time, the knowledge of the osteology of extant African anurans is astonishingly limited if we consider the number of species inhabiting the region: Gardner and Rage (2016) calculated that about 1100 extant species inhabit Africa, its associated islands, and the Arabian Peninsula, but detailed comparative osteological treatments are available for only a few species. This lack of knowledge significantly hinders the identification of fossil material belonging to extant taxa and highlights the need for comprehensive comparative collections, which so far are not available (Delfino et al., 2004). Among the relatively few papers dealing at least partially with the skeletal morphology of extant African anurans, and potentially useful for identifying fossil remains are the following: De Villiers, 1930, De Villiers, 1931, De Villiers, 1933 reported on the cranial osteology of Breviceps, Phrynomantis, and Probreviceps; Laurent, 1940, Laurent, 1941a, Laurent, 1941b, Laurent, 1942a, Laurent, 1942b, Laurent, 1942c, Laurent, 1943, Laurent, 1946, Laurent, 1951, Laurent, 1978 published a number of papers dealing with the osteology of Breviceps, Phrynomantis, and Trichobatrachus, plus several rachophorids and ranids; Emerson (1976) documented digging behavior and associated musculoskeletal features in Amietia and Hemisus; Clarke, 1979, Clarke, 1981 investigated the osteology of Raninae (now Raninae s.l.); Channing (1995) and Van Dijk (2001) listed a few characters differentiating Breviceps from Hemisus; Delfino et al. (2009) described the osteology of Barbarophryne brongersmai; Barej et al. (2014) described the osteology of Odontobatrachus and compared a few characters with those of Petropedetes and Arthroleptides; Matthews et al., 2015, Matthews et al., 2016 provided information on selected skeletal elements of a few extant taxa (Amietophrynus, Heleophryne, Hyperolius, Kassina, Ptychadena, Tomopterna, and Xenopus) used by them to identify fossil material from South Africa (moreover, see also Cope, 1867, Noble, 1922, Vergnaud-Grazzini, 1966, Blommers-Schlösser, 1993, Blackburn et al., 2015).

Due to the lack of osteological information for most of the extant African species, it is paradoxically easier to identify and contextualize Mesozoic and Paleogene remains that belong to extinct taxa, rather than Neogene and Quaternary remains that likely represent extant taxa. For post-Paleogene anurans, pipids are the best represented group, but alytids, brevicipitids, bufonids, heleophrynids, hylids, hyperoliids, pelobatids, ptychadenids, pyxicephalids, and ranids also have been recognized (Gardner and Rage, 2016, and literature therein). However, these occurrences are largely in northwestern and southern Africa. The East African record is limited to undetermined anurans from the middle Miocene Kenyan sites of Maboko, Ombo, and Kulu-Kosala, as well as from the Early and Late Pliocene of Kanapoi and the Laetoli area, respectively, and undetermined bufonids and the pipid Xenopus sp. from the Early Pleistocene of Olduvai Gorge in Tanzania (Leakey, 1967, Andrews et al., 1981, Harris et al., 2003, Manthi, 2008, Rage and Bailon, 2011, Gardner and Rage, 2016). This extremely sparse fossil record does not even minimally reflect the extraordinary diversity of East African extant amphibians, which according to Channing and Howell (2006) totals over 200 species referable to 40 genera, 10 families, and two orders, with nearly all the taxa being anurans.

A revision of the anurans from the Kenyan site of Kanapoi is the subject of this contribution. The age of the fossiliferous layers at Kanapoi is constrained by McDougall and Brown (2008) between 4.195 and 4.108 Ma, while the stratigraphy and depositional setting of Kanapoi are discussed by Feibel (2003; see also the literature cited by Manthi, 2008).

The exceptional number of anuran remains from Kanapoi (previously mentioned by Harris et al., 2003, Manthi, 2008, who reported abundant but unidentified anurans) provides an opportunity to improve our knowledge about the evolution of anurans in East Africa by opening a window on an early Pliocene assemblage. Moreover, the Kanapoi anuran assemblage may help to better characterize the local paleoenvironment (a topic previously explored by Schoeninger et al., 2003, on the basis of stable isotopes) of a site that has yielded a rich vertebrate fauna including the type material of Australopithecus anamensis (e.g., Leakey et al., 2002, Harris and Leakey, 2003, Manthi, 2008, Werdelin and Manthi, 2012, Geraads et al., 2013, Ward et al., 2013).

Anatomical nomenclature is based on Sanchiz (1998); taxonomic nomenclature follows Frost (2016). Comparative material of Hemisus marmoratus is housed in the collections of the Department of Earth Sciences of the University of Turin (Italy) and of the National Museum of Kenya (Nairobi) under the accession numbers, respectively, MDHC 411 and KNM-OR 457 and KNM-OR 479. The National Museum of Kenya is also the repository for all the anurans from Kanapoi (KNM-KP). The KNM-KP fossil collection includes a mixture of individually numbered and bulk numbered anuran bones. In the lists of specimens in the following descriptive accounts, counts for bones catalogued under each number are reported using a colon and Arabic number at the end of each catalogue number. For example, “KNM-KP 68623: 1” contains one bone, whereas “KNM-KP 68636: 51” contains 51 bones.