PALEVO 1631-0683 Elsevier S1631-0683(09)00170-5 10.1016/j.crpv.2009.12.003 Research article Systematic palaeontology (Vertebrate palaeontology) Paléontologie systématique/Systematic palaeontology (Paléontologie des vertrébrés/Vertebrate palaeontology) First mammal evidence from the Late Cretaceous of India for biotic dispersal between India and Africa at the KT transition Premier témoignage par des mammifères du Crétacé supérieur indien, de la dispersion biotique entre Inde et Afrique à la transition KT Prasad Guntupalli V.R. guntupalli.vrprasad@gmail.com a Verma Omkar b Gheerbrant Emmanuel c Goswami Anjali d Khosla Ashu e Parmar Varun f Sahni Ashok e Department of Geology, University of Delhi, Delhi – 110 007, India Centre for Biodiversity Studies, School of Biosciences and Biotechnology, Baba Ghulam Shah Badshah University, Rajouri 185 131, Jammu and Kashmir, India UMR 7207 du CNRS, CR2P, CP 38, département histoire de la Terre, centre de recherches sur la paléobiodiversité et les paléoenvironnements, Muséum national d’histoire naturelle, 8, rue Buffon, 75005 Paris, France Department of Genetics, Evolution, and Environment, University College London, Wolfson House 408, 4, Stephenson Way, London NW1 2HE, UK CAS in Geology, Panjab University, Chandigarh 160 014, India Department of Geology, University of Jammu, Jammu 180 006, India 9 1-2 S1631-0683(10)X0002-1 63 71 © 2009 Académie des sciences. Published by Elsevier B.V. All rights reserved. 2009 Académie des sciences Full (PDF)

The Late Cretaceous record of mammals from India assumes great significance in view of the fact that it is the only Gondwanan landmass that has yielded definitive eutherian mammals. These mammals have variously been assigned to palaeoryctids, archontans or Eutheria incertae sedis. Well preserved lower molars recovered from a new mammal-yielding Deccan intertrappean site near Kisalpuri village, Dindori District, Madhya Pradesh (state), India, are described here under a new species Deccanolestes narmadensis sp. nov. The new fossil material indicates close phylogenetic relationship between Deccanolestes from India and Afrodon (Adapisoriculidae) from the Late Palaeocene of Africa and Europe. In view of older age and more primitive state of Deccanolestes teeth, it is inferred that Deccanolestes represents an ancestral morphotype from which the African/European adapisoriculid Afrodon may have been derived. This is the first compelling terrestrial fossil evidence for an early dispersal between India and Africa. Such a dispersal possibly involved an East African contact with India at the KT transition.

Les gisements à mammifères du Crétacé supérieur indien, d’intérêt majeur, ont livré les seuls euthériens connus dans le Gondwana à cette époque. Les espèces indiennes, de parentés longtemps incertaines, ont été identifiées comme des paléoryctidés, des archontes ou des euthériens incertae sedis. Un nouveau matériel dentaire bien conservé, découvert dans le nouveau site de Kisalpuri (District de Dindori, Madhya Pradesh), dans les inter-trapp du Deccan, est décrit et rapporté à la nouvelle espèce Deccanolestes narmadensis sp. nov. Celle-ci indique des relations phylogénétiques étroites entre le genre indien Deccanolestes et l’adapisoriculidé Afrodon du Paléocène supérieur d’Europe et d’Afrique. En accord avec son âge ancien et sa morphologie primitive, Deccanolestes est donc reconnu ici comme représentant le morphotype ancestral dont sont issus les adapisoriculidés cénozoïques tels qu’Afrodon. Il s’agit de la première indication fossile précise à l’appui d’une dispersion ancienne entre l’Afrique et l’Inde. Celle-ci a probablement impliqué une connexion Est-Africaine à la transition KT.

 Cretaceous, Mammals, Deccanolestes, Afrodon, Adapisoriculids, Dispersal, India, Africa Crétacé, Mammifères, Deccanolestes, Afrodon, Adapisoriculidés, Dispersal, Inde, Afrique presented Presented by Philippe Taquet Introduction

The fossil record of Cretaceous mammals is very scanty from the Gondwanan continents. Although known from the Early Cretaceous of Australia and Africa, until now, no definitive mammals have been recorded from the Turonian–Maastrichtian interval (ca. 94–65 m.y.) of these continents. Late Cretaceous mammals are known only from South America, Madagascar and India. The Late Cretaceous mammal fauna of South America is represented by gondwanatheres (a specialized Gondwanan group), dryolestoids, and triconodontids (Bonaparte, 1990). Likewise, Late Cretaceous Madagascan mammal fauna is known by gondwanatheres (Krause et al., 1997), a possible marsupial (Krause, 2001) or zhelestid eutherian (Averianov et al., 2003) and some unidentified forms (Krause et al., 1994). A probable gondwanathere mammal was also documented from the Cretaceous rocks of Tanzania, the precise age of which is yet to be confirmed (Krause et al., 2003). As compared to this, India is the only Gondwanan continent that yielded undoubted Late Cretaceous eutherian mammals, and thus it is very important for understanding the evolution of eutherian mammals in the Gondwanan continents. These mammals were first reported from the sedimentary beds sandwiched between the Deccan volcanic flows 2 km north-east of Naskal, Rangareddi Distrit, Andhra Pradesh (Fig. 1) and were assigned to Deccanolestes hislopi with possible affinities to the Laurasian insectivore family Palaeoryctidae (Prasad and Sahni, 1988) (cimolestids of Kielan-Jaworowska et al., 2004). Following this, additional fossils belonging to D. hislopi and a second species D. robustus were described from the same site (Prasad et al., 1994). The dental morphology of Deccanolestes was regarded closely comparable to that of North American cimolestids (Cimolestes and Procerberus) and basal eutherians from Asia (Kennalestes and Otlestes) (Prasad et al., 1994). In addition to the dental remains of D. hislopi and D. robustus, ankle bones falling within the size range of these two species have also been described (Godinot and Prasad, 1994 and Prasad and Godinot, 1994). Based on close similarity of tarsal morphology to that of archontans, it was inferred that Deccanolestes adapted for an arboreal way of life and archontans had possibly originated on the Indian subcontinent from Deccanolestes-like animal (Godinot and Prasad, 1994 and Prasad and Godinot, 1994). Rana and Wilson (2003) documented additional fossils represented by D. hislopi, D. robustus and a new taxon Sahnitherium rangapurensis from the intertrappean beds of Rangapur that occur 4 km southeast of Naskal mammal-yielding site along its strike extension (Fig. 1). Rana and Wilson (2003), though accepted the boreoeutherian affinity of these taxa, felt that morphological similarities with Cimolestes and Procerberus are mostly plesiomorphic and hence placed Deccanolestes and Sahnitherium in Eutheria incertae sedis. In addition to eutherian mammals, gondwanatheres were documented from the intertrappean beds of Naskal and Gokak (Fig. 1) (Das Sarma et al., 1995, Krause et al., 1997 and Wilson et al., 2007) and haramiyids from the intertrappean beds in the vicinity of Naskal (Avashishta bacharamensis) (Anantharaman et al., 2006). One stumbling block in the interpretation of true phylogenetic relationship of the Cretaceous eutherians from India is the poor preservation of specimens from both Naskal and Rangapur sites. The teeth from these sites, particularly from Naskal, are heavily pitted supposedly due to groundwater action (Khajuria and Prasad, 1998).

More recently, a new Late Cretaceous mammal yielding site was discovered in the eastern Deccan volcanic province about 1.5 km south-west of Kisalpuri village (Fig. 1), Dindori District, Madhya Pradesh situated about 700 km north of Naskal site (Khosla et al., 2004). This site has yielded a new gondwanathere mammal (Bharattherium bonapartei) (Prasad et al., 2007a) and an eutherian mammal (Kharmerungulatum vanvaleni) (Prasad et al., 2007b) with affinities to archaic ungulates. In addition to these two taxa, several upper and lower teeth that fall within the morphological framework of Deccanolestes have also been recovered from the Kisalpuri site. The newly recovered teeth from Kisalpuri are fairly well preserved and provide us with an opportunity to decipher the crown morphology more precisely. In the present article, the new fossil material, especially lower molars, are described and the taxonomic relationship of Deccanolestes is reassessed in light of latest comparisons with eutherian mammals of Late Cretaceous age from North America and Asia, Late Palaeocene of Morocco and Germany.

Institutional abbreviations

VPL/JU/IM: Vertebrate Palaeontology Laboratory, University of Jammu, intertrappean mammal catalogue numbers, VPL/JU/NKIM: Vertebrate Palaeontology Laboratory, University of Jammu, Naskal intertrappean mammal catalogue numbers, ITV/R/Mm: Intertrappean vertebrates, Rangapur mammal catalogue numbers, Garhwal University, India, THR: Adrar Mgorn1, collections of the Université des sciences et techniques du Languedoc, Montpellier II catalogue numbers, CR, CY; Cernay, Bn: Braillon, Ph: Phelzion collections in Muséum national d’histoire naturelle (MNHN), Paris, UCMP: University of California Museum of Paleontology, Berkeley.

 Systematic palaeontology

Class Mammalia Linnaeus, 1758 Infraclass Eutheria Gill, 1872 Order ? Euarchonta Waddell et al., 1999 Family Adapisoriculidae Van Valen, 1967 Genus Deccanolestes Prasad and Sahni, 1988 Species Deccanolestes   narmadensis sp. nov.

Holotype

VPL/JU/IM/8 (Rm1 or Rm2).

 Referred Material

VPL/JU/IM/7 (Rm1 or Rm2), VPL/JU/IM/6 (Rm3), ITV/R/Mm-8 (Rm1 or Rm2).

 Horizon, type locality and age

Deccan intertrappean beds exposed 1.5 km southwest of Kisalpuri village, Dindori District, Madhya Pradesh, Central India, Upper Cretaceous (Maastrichtian).

 Stratigraphic range

Upper Cretaceous (Maastrichtian) intertrappean beds of Rangapur, Rangareddi District, Andhra Pradesh, South India.

 Specific Diagnosis

Differs from D. hislopi, D. robustus and Deccanolestes sp. (ITV/R/Mm-9) in the following traits: anteriorly projecting and lingually placed, large individualized paraconid placed in line with the metaconid rather than crest-like as in D. hislopi; trigonid cusps forming an open triangle; paraconid, metaconid and entoconid in line with each other; talonid as wide as and slightly shorter than or as long as the trigonid; and in presence of a postmetacristid.

 Etymology

Species named after Narmada river which flows through the study area.

 Description

The species is known only by lower teeth. VPL/JU/IM/8 has well-preserved cusps except for the abraded metaconid and hypoconid tips. The trigonid cusps are arranged in an open triangle. The protoconid is the highest and largest cusp of the trigonid. The metaconid is slightly smaller than the protoconid and its posterior base is slightly posterior to that of the protoconid which makes the posterior trigonid wall and protocristid slightly obliquely oriented to the transverse axis. The paraconid is the smallest cusp of the trigonid occurring at a lower level than the protoconid and metaconid; it is placed at the lingual margin in line with the metaconid and the entoconid (Fig. 2.2A). The paraconid is a relatively large and well individualized cusp slanting anteriorly and separated from the metaconid by a broad ‘V’ shaped groove (Fig. 2.2D). The paraconid is broken to its base on VPL/JU/IM/7, but in its girth it appears to be substantially large (only slightly smaller than the metaconid) and is placed on the lingual margin (Fig. 2.1A). The paracristid has a ‘V’ shaped notch. The precingulid is prominent on both the teeth, stoutly built and subvertical in orientation with a small cuspule at its dorsal end (Fig. 2.1A, 2.2A–C,). The trigonid is nearly as long and as wide as the talonid (Table 1). It is nearly one and half times higher than the talonid (Fig. 2.1B, 2.2B). The hypoconid is the largest and highest cusp of the talonid. The hypoconulid is slightly smaller than the hypoconid. The hypoconulid and entoconid are subequal in size on VPL/JU/IM/8 and the hypoconulid is slightly larger than the entoconid on VPL/JU/IM/7. The hypoconulid is closer to the entoconid than to the hypoconid (Fig. 2.1A, C, 2.2.A, E) The entocristid is short and low. The cristid obliqua joins the trigonid wall below the protocristid notch on VPL/JU/IM/8 and slightly lingual to it on VPL/JU/IM/7. There is a thin distal metacristid at the posterolingual base of the metaconid. It is more prominent on VPL/JU/IM/7 than on VPL/JU/IM/8. Posteriorly, a shallow notch is present between the hypoconid and the hypoconulid. The talonid basin is short, deep, and bowl-like. The hypoflexid is deep.

In m3 (VPL/JU/IM/6), the trigonid is acute-angled (Fig. 2.3A) as compared to that of VPL/JU/IM/7 and 8 and the metaconid is as small as the paraconid. The paraconid is a small, well individualized conical cusp placed lingually in line with the metaconid (Fig. 2.3A). As compared to the metaconid and the paraconid, the protoconid is very high (Fig. 2.3D). The trigonid and talonid junction is lingually twisted. The trigonid basin is open through a wide groove between the paraconid and metaconid. The paraconid tip is broken (Fig. 2.3D–E). Though the paraconid and the metaconid appear subequal in height, in girth the metaconid is slightly smaller than the paraconid (Fig. 2.3D). The distal metacristid descends posterolingually from the protocristid notch in continuation with the lingual crest of protoconid and from the posterior mid-height of metaconid it turns lingually and then descends to the lingual base of metaconid to merge with the entocristid (Fig. 2.3C). The trigonid is shorter and wider than the talonid. The long talonid and posteriorly projecting hypoconulid imply that the tooth represents an ultimate molar (Fig. 2.3A, B, D). Slightly labially shifted hypoconid is the largest cusp of the talonid which is separated from the hypoconulid by a broad notch. The hypoconulid is smaller than the hypoconid, the entoconid and entocristid are completely worn and thus the talonid opens lingually. The cristid obliqua contacts the posterior trigonid wall below the protocristid notch and extends dorsally. The talonid is shallow and slopes anterolingually (Fig. 2.3A).

 Remarks

D. narmadensis sp. nov. differs from the lower molars of D. hislopi and D. robustus in having a relatively large paraconid placed at the lingual margin of the trigonid and projecting anteriorly, the talonid nearly subequal in width to the trigonid and slightly shorter or as long as the trigonid, and the hypoconulid slightly closer to the entoconid than to the hypoconid. The paraconid is at a lower level than the protoconid and metaconid and the trigonid lacks the flat triangular area formed by trigonid cusps in D. hislopi. In these features and having a paracristid with ‘V’ shaped notch, a deep and broad ‘V’ shaped notch separating the paraconid from the metaconid, deep hypoflexid, small precingulid cuspule, and cristid obliqua contacting the trigonid wall slightly lingual to protocristid notch, VPL/JU/IM/7 and VPL/JU/IM/8 compare well with ITV/R/Mm-8 described as Deccanolestes? sp. (Rana and Wilson, 2003). Because of this close similarity in crown morphology, ITV/R/Mm-8 is also assigned to the new species. However, in ITV/R/Mm-8, the hypoconulid is equidistant from the hypoconid and entoconid while it is closer to entoconid in the new teeth from Kisalpuri. But this feature is variably present in teeth referred to D. hislopi. No comparison is possible with Sahnitherium, another eutherian taxon known from the Late Cretaceous of India, as it is based on an upper molar. As the crown morphology of VPL/JU/IM/6-8 and ITV/R/Mm-8 is distinct from that of D. hislopi and D. robustus, naming of a new species, D. narmadensis sp. nov., for these specimens is justified. The mammal collection from the intertrappean beds of Kisalpuri is overwhelmingly represented by lower molars and includes only a few incompletely preserved upper molars. Hopefully, well preserved upper molars referable to D. narmadensis sp. nov. may be recovered in future when large scale screen-washing is undertaken at this site.

 Comparisons

Comparison of Deccanolestes, including the new specimens from Kisalpuri, with the casts and original specimens of Late Cretaceous cimolestids (previously described as palaeoryctids) from North America (Cimolestes, Procerberus Batodon) and some basal eutherians from Asia (Kennalestes, Asioryctes) at UCMP, Berkeley, and Late Palaeocene cimolestids (Cimolestes, Aboletylestes) at MNHN, Paris by GVRP and EG indicated that the similarities in the crown morphology between Deccanolestes and these taxa are primarily plesiomorphic. Deccanolestes is more derived than Cimolestes, Procerberus, Batodon, Aboletylestes, Kennalestes, Asioryctes, and Otlestes in having upper molars with subequal parastylar and metastylar lobes, reduced parastylar hook, labial cusps well separated to their bases (predilambdodonty), reduced transverse width of the crown, and protocone moderately expanded mesiodistally and lingually, and reduced height difference between trigonid and talonid. Therefore, any close relationship with the above mentioned taxa is ruled out for Deccanolestes.

Among all known eutherian taxa, the upper and lower tooth morphology of two adapisoriculid species Afrodon germanicus known from the Late Palaeocene of Walbeck, Germany and Cernay and Berru localities (France) (Russell, 1964) and Afrodon chleuhi documented from the Late Palaeocene (Thanetian) Adrar Mgorn site in Ouarzazate basin, Morocco (Gheerbrant, 1988 and Gheerbrant and Russell, 1989) is very similar to that of Deccanolestes. The upper molars of D. hislopi and D. robustus compare well to those of A. chleuhi and A. germanicus in many derived characters such as the differences listed above from cimolestids (more or less subequal development of parastylar and metastylar lobes, reduced parastylar hook, labial cusps separated as far as their bases (Fig. 3.4C, 3.6C–D) reduced transverse width (Fig. 3.1, 3.2, 3.3, 3.4A, 3.5A, 3.6A), moderate mesiodistal and lingual expansion of the protocone base (Fig. 3.1, 3.2, 3.3, 3.4A, 3.6A), and the paraconule placed between the paracone and the protocone, the metaconule small and closer to the metacone (Fig. 3.1, 3.4A, 3.5A, 3.6A), and moderately labially shifted protocone apex (Fig. 3.4B, D, 3.5B–D, 3.6B). In fact, THR-168, originally described as a didelphodontid (Gheerbrant, 1995), but subsequently identified as a m1 of A. chleuhi (Gheerbrant et al., 1998) is remarkably similar to the holotypes of D. hislopi (VPL/JU/NKIM/10) and D. robustus (VPL/JU/NKIM/13) in it's crown morphology.

Likewise, the new lower molars assigned to D. narmadensis sp. nov. are very similar to those of A. chleuhi and A. germanicus in having trigonid and talonid subequal in width (Fig. 2.1A, 2.2A, 2.4A, 2.5), reduced height difference between the talonid and trigonid (trigonid nearly one and half times higher than the talonid) (Fig. 2.1B, 2.2B, 2.4B, 2.7C), well individualized and substantially large paraconid slanting anteriorly and placed lingually in line with the metaconid (Fig. 2.2A, D, 2.4A, C, 2.5, 2.7A), the paracristid and protocristid making an angle of more than 50°, metaconid slightly smaller than the protoconid, a large and salient hypoconid, the hypoconulid closer to the entoconid than to the hypoconid (Fig. 2.1A, 2.2A, 2.4A, 2.5, 2.6, 2.7B, C), the distal metacristid along with the cristid obliqua delimiting a deep depression in the linguobasal region of the posterior wall of the trigonid, and a deep bowl-like talonid basin.

Of the two known species of Afrodon, A. germanicus appears to be more derived than A. chleuhi and Deccanolestes in its narrow stylar shelf and in the greater relief and extension of cristid obliqua on to the posterior trigonid wall, sharp cusps on lower molars, proportionately wider talonid with a more labially shifted hypoconid (Fig. 2.6) (linked to a dilambdodont trend) which makes the talonid slightly wider than the trigonid. A. germanicus and A. chleuhi also differ from Deccanolestes in the nature of connection that cristid obliqua makes with the posterior wall of the trigonid. In Deccanolestes, the cristid obliqua joins the posterior trigonid wall slightly lingual to the protocristid notch (Fig. 2.1A, 2.2A), while in A. germanicus and A. chleuhi it meets the posterior trigonid wall in the middle of posterior base of the metaconid (Fig. 2.4A, 2.5, 2.7B). Deccanolestes is surprisingly more derived than Afrodon in the position of centrocrista on the upper molars. In Afrodon, the centrocrista is rectodont (predilambdodonty), whereas in Deccanolestes it is labially placed (dilambdodonty). However, in the development of conules and metacingulum Afrodon is more derived than Deccanolestes. The paraconule and metaconule are well developed in Afrodon, while the paraconule is well developed in all upper molars of Deccanolestes, the metaconule is either well developed as in VPL/JU/IM/14 (Fig. 3.5A) or weakly developed as in VPL/JU/IM/17 (Fig. 3.4A). Moreover, the metacingulum is absent in Deccanolestes, whereas it extends slightly on to the posterior face of the metacone in Afrodon. Besides, the postparaconule crest of Afrodon continues on to the lingual flank of paracone giving it an angular aspect. On the whole, Deccanolestes is more similar to A. chleuhi than to A. germanicus in the small size of molars, in the presence of robust cusps on lower molars, and in having a talonid nearly subequal in width or slightly narrower than the trigonid and less clearly marked cristid obliqua with a low relief. In these traits, the European species A. germanicus is more derived than Deccanolestes and A. chleuhi.

The m3 VPL/JU/IM/6 referred to D. narmadensis sp. nov. is strikingly similar to a left lower third molar (THR-51) assigned to A. chleuhi (Gheerbrant, 1995). In both these teeth, the paraconid is a comparatively large, well individualized cusp nearly equal to the metaconid and anterior to it in position; the metaconid is substantially smaller than the protoconid; the trigonid is slightly raised in lingual view as compared to the talonid; the hypoconid is the largest cusp of the talonid; the hypoconulid and entoconid are smaller than the hypoconid and subequal in size; the hypoconulid is median in position; the hypoconid and hypoconulid are separated by a broad notch posteriorly; the talonid basin is shallow and anterolingually sloping and partially open lingually. The trigonid is acute-angled in both THR-51 and VPL/JU/IM/6. The only difference between the two teeth is in the position of cristid obliqua with respect to the posterior trigonid wall. In THR-51, it joins the posterolingual base of the metaconid at its middle and extends up the trigonid wall as compared to its position slightly lingual to the protocristid notch on VPL/JU/IM/6.

 Discussion Affinities of Deccanolestes

The marked similarity between the upper and lower teeth of Deccanolestes and Afrodon germanicus and A. chleuhi (Adapisoriculidae) implies close phylogenetic relationship between the two genera. Decannolestes is consequently referred here to the Adapisoriculidae family, aside to Afrodon. The Indian and African forms appear especially more closely related with respect to the more derived European species Afrodon germanicus. Other European adapisoriculids such as Adapisoriculus and most remarkably Remiculus, are even much more specialized.

A. germanicus and A. chleuhi have been assigned to the family Adapisoriculidae (Gheerbrant, 1995 and Gheerbrant and Russell, 1989). The phylogenetic relationship of adapisoriculids has remained a vexed problem in palaeontology. Van Valen (1967) created Adapisoriculinae for Adapisoriculus (Lemoine, 1883) known from the Late Palaeocene of Cernay, France. Adapisoriculus? germanicus first described from the Late Palaeocene rocks of Walbeck, Germany (Russell, 1964) was later found in coeval deposits in Cernay and Berru in France. The taxonomic affinity of Adapisoriculus has been debated for long and didelphid marsupial (Teilhard de Chardin, 1922), tupaiid (Van Valen, 1965) or mixodectid insectivore (Butler, 1972) relationships have been suggested. Van Valen (1967) placed it in the subfamily Adapisoriculinae (Family Tupaiidae), while Butler (1980) expressed doubts on this familial attribution. Gheerbrant and Russell (1989) placed it in its own family Adapisoriculidae and included predilambdodontids (Afrodon and Bustylus) and dilambdodontids (Adapisoriculus, Garatherium, and Remiculus). Gheerbrant (1995) placed adapisoriculids with some reservation in lipotyphlan insectivores. However, in a latest paper (Storch, 2008), small postcranial bones associated with A. germanicus (Gheerbrant and Russell, 1989) and Bustylus cf. cernaysi (Gheerbrant and Russell, 1991) of Late Palaeocene Walbeck fauna of Germany were assigned to Euarchonta and plesiadapiform primates. This hypothesis is interestingly congruent with the tupaiid relationship suggested by Van Valen for Adapisoriculus (1967). The close similarity in dental morphology of Deccanolestes to Afrodon and an arboreal habitat inferred from the postcranial bones morphology of Deccanolestes (Boyer et al., 2009, Godinot and Prasad, 1994 and Prasad and Godinot, 1994) lends some credence to this interpretation. The currently best available working hypothesis that Deccanolestes and the adapisoriculids are euarchontans needs to be further tested in a formal cladistic framework.

A phylogenetic tree of Adapisoriculidae showing a proteutherian with zalambdodont structure (otlestid?) as a possible ancestral morphotype for adapisoriculid insectivores was proposed by Gheerbrant (1995, text fig. 2). Gheerbrant (1995) expressed the opinion that Prokennalestes is the most directly related form to the hypothetical ancestry of adapisoriculids as far as molar structure is concerned. But at the same time he argued that it is a very old ancestral form going back to more than 45 m.y. before adapisoriculds during which many things could have happened in the early mammalian evolution. Deccanolestes with its close resemblance in molar morphology to A. germanicus and A. chleuhi fills in partially this time gap. In view of the close dental similarity of Deccanolestes to Afrodon germanicus and A. chleuhi, and because of slightly primitive state of Deccanolestes molars, we propose here a working hypothesis that the latter represents the ancestral morphotype from which euarchontan adapisoriculids such as Afrodon were derived.

 Palaeobiogeographic significance

The Late Cretaceous vertebrate fauna of India exhibits biogeographic affinities with both Gondwanan (Krause et al., 1997, Prasad, 2005 and Sahni and Prasad, 2009) and Laurasian fauna (Prasad and Rage, 1995, Prasad and Sahni, 1988 and Sahni and Prasad, 2009) during the northward drift phase of India. Based on gondwanathere mammals, abelisaurid dinosaurs and baurusuchid crocodiles, a Late Cretaceous terrestrial connection was envisaged between South America and India–Madagascar via Antarctica and Kerguelen Plateau (Krause et al., 1997 and Sahni and Prasad, 2009) (but see Ali and Aitchison (2008) for a contrary view). On the other hand, a northern sweepstakes mode of dispersal between India and Eurasia across Kohistan-Dras island-arc system was suggested for the presence of eutherian mammals, discoglossid frogs and anguimorph lizards of Laurasiatic origin in the Late Cretaceous of India (Prasad, 2005, Prasad and Sahni, 1999 and Sahni and Prasad, 2009).

The paleobiogeographic relationships between India and Africa during the Cretaceous and Early Tertiary are enigmatic, especially because of the African fossil gaps (Gheerbrant and Rage, 2006). Some authors suggested that Africa was the main terrestrial link between India and both Laurasian and Gondwanan continents when Greater India was close to the northeastern Africa (Somalia) (Briggs, 2003 and Chatterjee and Scotese, 1999). This hypothesis could not be tested rigorously as Late Cretaceous (Turonian–Maastrichtian) fossil yielding sites are not known from Africa and the fossil record from India is mainly from the Upper Cretaceous (Maastrichtian) deposits. The present study espouses a close adapisoriculid taxonomic relationship of Deccanolestes, especially with A. chleuhi known from the Late Palaeocene rocks of Morocco. Because of the close similarity, but primitive construction and old age with respect to Afrodon, we suggest here that Afrodon may have evolved from a Deccanolestes-like ancestral morphotype. When and how exactly such a morphotype reached Africa/Europe before giving rise to Afrodon remains to be understood as there is no documented fossil record of Late Cretaceous mammals from Africa. However, the relationship of Deccanolestes and A. chleuhi is the most compelling and first direct evidence for an Indo-African dispersal at the beginning of the Tertiary (before the Late Paleocene). The northern dispersal hypothesis that favours sweepstakes mode of dispersal across the Tethys sea via island-arc systems to explain the presence of discoglossid and pelobatid frogs, anguimorph lizards, and eutherian mammals of Laurasiatic origin in the Late Cretaceous of India (Prasad and Sahni, 1999 and Sahni and Prasad, 2009) may also hold equally good for the adapisoriculids. The close relationship between A. chleuhi and Deccanolestes however, weighs in favour of an East-African connection. But this will become clear only when the phylogeny of adapisoriuculids is better understood.

Ali and Aitchison (2008) while refuting the Kerguelen Plateau terrestrial connection between South-America and India–Madagascar suggested that the South American forms may have crossed over the proto-central Atlantic and Mozambique channel to India–Madagascar through West and Central Africa. This is in consonance with the early Late Cretaceous (95 Ma) biotic link suggested by Sereno et al. (2004) between South America and Africa based on abelisaurid dinosaur fossils from the Cenomanian of Niger, Northwest Africa. Another biogeographic link that demands close examination is the Late Cretaceous East African link envisaged for the Indian mobile raft (Briggs, 2003 and Chatterjee and Scotese, 1999). Crucial to testing the latter biogeographic model is the Late Cretaceous fossil record of Africa which is poorly documented at present. Further, the East African connection for India needs to be firmly supported by marine geological and geophysical data from the basins west of India and a rigorous phylogenetic analysis of dental and postcranial remains of Deccanolestes and Afrodon in a cladistic framework (work in progress).

Acknowledgements

GVRP acknowledges Samuel P. & Doris Welles Research Grant, UCMP, Berkeley and the MNHN “Maître de Conférence Invité” grant, which enabled him to undertake comparative study of the Late Cretaceous eutherian mammals of India with those of North America, Asia, Europe and Morocco. Thanks are also due to William A. Clemens and Patricia Holroyd of UCMP and P. Tassy and C. Sagne of MNHN for allowing access to holotypes and casts of Late Cretaceous and Palaeocene eutherian taxa from North America, Europe, Asia, and Africa SEM photomicrography of the mammalian teeth by S. Shome of the Geological Survey of India. Kolkata is duly acknowledged. GVRP acknowledges financial support from the Department of Science & Technology (DST), Govt. of India (Grant No.SR/S4/ES/24/2002) for this work. OV(Grant No.SR/FTP/ES-33/2008) and AK (SR/S4/Es-382/2008) are also thankful to DST for research grants.

 Ali J.R. Aitchison J.C. Gondwana to Asia: plate tectonics, paleogeography and the biological connectivity of the Indian subcontinent from the Middle Jurassic through Latest Eocene (166-35 Ma) Earth Sci. Rev. 88 2008 145–166 Anantharaman S. Wilson G.P. Das Sarma D.C. Clemens W.A. A possible Late Cretaceous haramiyidan from India J. Vert. Paleontol. 26 2 2006 488–490 Averianov A. Archibald J.D. Martin T. Placental nature of the alleged marsupial from the Cretaceous of Madagascar Acta Palaeontol. Pol. 48 2003 149–151 Bonaparte J.F. New Late Cretaceous mammals from the Los Alamitos Formation, northern Patagonia Nat. Geog. Res. 6 1990 63–93 Boyer, D.M., Prasad, G.V.R., Goswami, A., Krause, D.W., Godinot, M., 2009, New material of Deccanolestes (Mammalia, Eutheria) from the Cretaceous of India: evidence for habitus and relationships. J. Vert. Paleontol. 29, Supplement to Number 3, Abstract 68A. Briggs J.C. The historical biogeography of India: isolation or contact Syst. Zool. 38 2003 322–332 Butler, P.M., 1972. The problem of insectivore classification. In: Joysey, K.A., Kemp, T.S. (Eds), Studies in Vertebrate Evolution. Oliver and Boyd, Edinburgh, pp. 253–265. Butler P.M. The tupaiid dentition Luckett W.P. Comparative biology and evolutionary relationships of tree shrews 1980 Plenum Press New York 170–204 Chatterjee S. Scotese C.R. The break-up of Gondwana and the evolution and biogeography of the Indian plate Proc. Ind. Nat. Sci. Acad. 65A 1999 397–425 Das Sarma D.C. Anantharaman S. Vijayasarathi G. Nath T.T. Rao Ch.V. Palaeontological studies for the search of micromammals in the infra- and intertrappean horizons of Andhra Pradesh Rec. Geol. Surv. Ind. 128 1995 223 Gheerbrant E. Afrodon chleuhi nov.gen., nov. sp., « insectivore » (Mammalia, Eutheria) lipotyphlé (?) du Paléocène marocain : données préliminaires C. R. Acad. Sci. Paris, Ser. II 307 1988 1303–1309 Gheerbrant E. Les mammifères paléocènes du Bassin d’Ouarzazate (Maroc). III. Adapisoriculidae et autres mammifères (Carnivora;?Creodonta, Condylarthra? Ungulata et incertae sedis) Palaeon tographica A 237 1995 39–132 Gheerbrant E. Rage J.-C. Paleobiogeography of Africa: How distinct from Gondwana and Laurasia? Palaeogeogr. Palaeoclimatol. Palaeoecol. 241 2006 224–246 Gheerbrant E. Russell D.E. Presence of the genus Afrodon [Mammalia, Lipotyphla(?) Adapisoriculidae] in Europe; new data for the problem of trans-Tethyan relations between Africa and Europe around the K/T boundary Palaeogeogr. Palaeoclimatol. Palaeoecol. 76 1989 1–15 Gheerbrant E. Russell D.E. Bustylus cernaysi, nouvel adapisoriculidé (Mammalia, Eutheria) paléocène d’Europe Geobios 24 4 1991 467–481 Gheerbrant E. Sudre J. Sen S. Abrial C. Marandat B. Sigé B. Vianey-Liaud M. Nouvelles données sur les mammifères du Thanétien et de l’Yprésien du bassin d’Ouarzazate (Maroc) et leur contexte stratigraphique Palaeovertebrata 27 3–4 1998 155–202 Godinot M. Prasad G.V.R. Discovery of Cretaceous arboreal eutherians Naturwissenschaften 81 1994 79–81 Khajuria C.K. Prasad G.V.R. Taphonomy of a Late Cretaceous mammal-bearing microvertebrate assemblage from the Deccan intertrappean beds of Naskal, peninsular India Palaeogeogr. Palaeoclimatol. Palaeoecol. 137 1998 153–172 Khosla A. Prasad G.V.R. Verma O. Jain A.K. Sahni A. Discovery of a micromammal-yielding site near Kisalpuri, Dindori District, Madhya Pradesh Curr. Sci. 87 3 2004 380–383 Kielan-Jaworowska Z. Cifelli R.L. Luo Z.-X. Mammals from the age of dinosaurs: origins, evolution and structure 2004 Columbia University Press New York, USA p. 630 Krause D.W. Fossil molar from a Madagascan marsupial Nature 412 2001 497–498 Krause D.W. Hartman J.H. Wells N.A. Buckley G.A. Lockwood C.A. Wall C.E. Wunderlich R.E. Rabarison J.A. Randriamiaramanana L.L. Late Cretaceous mammals Nature 368 1994 298 Krause D.W. Prasad G.V.R. Koenigswald W.v. Sahni A. Grine F.E. Cosmopolitanism among Gondwanan Late Cretaceous mammals Nature 390 1997 504–507 Krause D.W. Gottfried M.D. O’Connor P.M. Roberts E.M. A Cretaceous mammal from Tanzania Acta Palaeontol. Pol. 48 2003 321–330 Lemoine V. Sur l’Adapisorex, nouveau genre de Mammifère de la faune cernaysienne des environs de Reims C. R. Acad. Sci. Paris 97 1883 1325–1327 Prasad G.V.R. Mammalian perspective of Late Cretaceous palaeobiogeography of the Indian subcontinent Gond. Geol. Magz. 8 2005 111–122 Prasad G.V.R. Godinot G. Eutherian tarsal bones from the Late Cretaceous of India J. Paleontol. 68 1994 892–902 Prasad G.V.R. Rage J.-C. Amphibians and squamates from the Maastrichtian of Naskal, India. Cret. Res. 16 1995 95–107 Prasad G.V.R. Sahni A. First Cretaceous mammal from India Nature 332 6164 1988 638–640 Prasad G.V.R. Sahni A. Were there size constraints on biotic exchanges during the northward drift of the Indian plate Proc. Ind. Nat. Sci. Acad. 65A 3 1999 377–396 Prasad G.V.R. Jaeger J.-J. Sahni A. Gheerbrant E. Khajuria C.K. Eutherian mammals from the Upper Cretaceous (Maastrichtian) intertrappean beds of Naskal, Andhra Pradesh India. J. Vert. Paleontol. 14 1994 260–277 Prasad G.V.R. Verma O. Sahni A. Krause D.W. Khosla A. Parmar V. A new Late Cretaceous gondwanatherian mammal from Central India Proc. Ind. Nat. Sci. Acad. 73 1 2007 17–24 Prasad G.V.R. Verma O. Sahni A. Parmar V. Khosla A. A Cretaceous hoofed mammal from India Science 318 2007 937 Rana R.S. Wilson G.P. New Late Cretaceous mammals from the intertrappean beds of Rangapur, India and palaeobiogeographic framework Acta Palaeontol. Pol. 48 3 2003 331–348 Russell D.E. Les mammifères paléocènes d’Europe Mem. Mus. Nat. Hist. Natur., Ser. C 13 1964 1–324 Sahni A. Prasad G.V.R. Geodynamic evolution of the Indian plate: consequences for dispersal and distribution of biota Mem. Geol. Soc. Ind. 66 2009 203–225 Sereno P.C. Wilson J.A. Conrad J.L. New dinosaurs link southern landmasses in the Mid-Cretaceous Proc. R. Soc. Lond. 271B 2004 1325–1330 Storch G. Skeletal remains of diminutive primate from the Paleocene of Germany Naturwissenschaften 95 10 2008 927–930 Teilhard de Chardin P. Les Mammifères de l’Éocène inférieur français et leurs gisements Ann. Paleontol. 11 1922 9–116 Van Valen L. Tree shrews, primates, and fossils Evolution 19 2 1965 137–151 Van Valen L. New Paleocene insectivores and insectivore classification Bull. Am. Mus. Nat. Hist. 135 1967 221–284 Wilson G.P. Das Sarma D.C. Anantharaman S. Late Cretaceous sudamericid gondwanatherians from India with palaeobiogeographic considerations of Gondwanan mammals J. Vert. Paleontol. 27 2 2007 521–531

Map of India showing the location of the Upper Cretaceous mammal-yielding intertrappean sites (denoted by stars). 1. Naskal, 2. Rangapur, 3. Kisalpuri, 4. Gokak.

Carte de l’Inde montrant la localisation des sites inter-traps pourvoyeurs de mammifères du Crétacé supérieur (indiqués par des étoiles).

1–3. Deccanolestes narmadensis sp. nov. from the Upper Cretaceous Kisalpuri intertrappean beds, central India. 1. Right m1 or m2 (VPL/JU/IM/7) in occlusal (1A), labial (1B), posterior (1C) and lingual (1D) views; 2. Right m1 or m2 (VPL/JU/IM/8, holotype) in occlusal (2A), labial (2B), anterior (2C), lingual (2D) and posterior (2E) views; 3. Right m3 (VPL/JU/IM/6) in occlusal (3A), labial (3B), posterior (3C), lingual (3D) and anterior (3E) views. 4. Left m1 of Afrodon chleuhi (THR 214) from the Late Palaeocene of Morocco in occlusal (4A), labial (4B) and lingual (4C) views. 5–7. Afrodon germanicus from the Late Palaeocene of Germany and France. 5. Left m2 of A. germanicus (CY 590) in occlusal view; 6. Right m2 (CR-1-Bn) in occlusal view; 7. Right m1 (Wa/16118B) in lingual (7A), posterior (7B), and labial (7C) views. Figs. 4 and 5 are redrawn from Gheerbrant (1995) and Figs. 6–7 are after Gheerbrant and Russell (1989, fig. 4a–c, 5). Scale bar is same for Figs. 1–3.

1–3. Deccanolestes narmadensis sp. nov. provenant des lits inter-traps du Crétacé supérieur de Kisalpuri, Inde centrale. 1. m1 ou m2 droite (VPL/JU/IM/7) en vues occlusale (1A), labiale (1B), postérieure (1C) et linguale (1D) ; 2. m1 ou m2 droite (VPL/JU/IM/8, holotype) en vues occlusale (2A), labiale (2B), antérieure (2C), linguale (2D) et postérieure (2E) ; 3. m3 droite (VPL/JU/IM/6) en vues occlusale (3A), labiale (3B), postérieure (3C), linguale (3D) et antérieure (3E). 4. m1 gauche d’Afrodon chleuhi (THR 214) du Paléocène tardif du Maroc en vues occlusale (4A), labiale (4B) et linguale (4C). 5–7. Afrodon germanicus du Paléocène tardif d’Allemagne et de France. 5. m2 gauche de A. germanicus (CY 590) en vue occlusale ; 6. m2 droite (CR-1-Bn) en vue occlusale. 7. m1 droite (Wa/16118B) en vues linguale (7A), postérieure (7B) et labiale (7C). Les Figures 4 et 5 sont redessinées selon Gheerbrant (1995) et les Figures 6–7 selon Gheerbrant et Russel (1989, figs. 4a–c, 5). La barre d’échelle est la même pour les Figures 1 à 3.

1. Left M2 of Afrodon germanicus (CR-181-Ph) in occlusal view. 2–3. Deccanolestes from the Upper Cretaceous intertrappean beds of Naskal, South India. 2. Right M2 of Deccanolestes cf. D. hislopi (VPL/JU/NKIM/15) in occlusal view; 3. Right M1 of D. hislopi (VPL/JU/NKIM/10, holotype) in occlusal view. 4–5. cf. Deccanolestes hislopi from the Upper Cretaceous Kisalpuri intertrappean beds, central India. 4. Left M1 (VPL/JU/IM/17) in occlusal (4A), anterior (4B), labial (4C) and posterior (4D) views; 5. Right M2 (VPL/JU/IM/14) in occlusal (5A), posterior (5B), anterior (5C) and lingual (5D) views. 6. Right M2 of A. germanicus (CR-19-Bn) in occlusal (6A), posterior (6B), labial (6C) and lingual (6D) views. Figures 1 and 6 are after Gheerbrant and Russell (1989, figs. 1a–d, 5) and Figures 2 and 3 are after Prasad et al. (1994, figs. 6A, C). Scale bar is same for Figures 4 and 5.

1 M2 gauche d’Afrodon germanicus (CR-181-Ph) en vue occlusale ; 2–3 Deccanolestes des lits inter-traps du Crétacé supérieur de Naskal, Inde méridionale. 2. M2 droite de Deccanolestes cf. D. hislopi (VPL/JU/NKIM/15) en vue occlusale. 3. M1 droite de D. hislopi (VPI/JU/NKIM/10, holotype) en vue occlusale. 4–5. cf. Deccanolestes hislopi des lits inter-traps du Crétacé supérieur de Kisalpuri, Inde centrale). 4. M1 gauche (VPL/JU/IM/17) en vues occlusale (4A), antérieure (4B), labiale (4C) et postérieure (4D) ; 5. M2 droite (VPL/JU/IM/14) en vues occlusale (5A), postérieure (5B), antérieure (5C) et linguale (5D). 6. M2 droite de A. germanicus (CR-19-Bn) en vues occlusale (6A), postérieure (6B), labiale (6C) et linguale (6D). Les Figures 1 et 6 sont dessinées selon Gheerbrant et Russell (1989, figs. 1a–d, 5) et les Figures 2 et 3 selon Prasad et al. (1994, figs. 6A, C). La barre d’échelle est la même pour les Figures 4 et 5.

Measurements (millimeters). L: length; TRW: trigonid width; TAW: talonid width; PDH: protoconid height; TAH: talonid height; TAL: talonid length; AML: premetaconid length.

Mesures en millimètres. L : longueur ; TRW : largeur du trigonide ; TAW : largeur du talonide PDH : hauteur du protoconide ; TAH : hauteur du talonide ; TAL : longueur du talonide ; AML : longueur du prémétaconide.

Specimen number L TRW TAW PDH TAH TAL AML VPL/JU/IM/6(m3) 1.23 0.60 0.60 1.33 0.83 0.56 0.56 VPL/JU/IM/7(m1 or m2) 1.32 0.67 0.66 1.34 0.80 0.59 0.56 VPL/JU/IM/8(m1 or m2) 1.25 0.65 0.64 0.97 0.55 0.57 0.58