Large Mammals from the Plio-Pleistocene of Kashmir Intermontane Basin , India , with Reference to their Status in Magnetic Polarity Time Scale

The larger mammalian fossils (Equus, Elephas, Cervus and Canis) being described from the Plio-Pleistocene Karewa deposits of Kashmir Valley, India, were recovered from the sections already dated by paleomagnetism. Based on these data, the occurrence of Equus and Cervus is dated to about 2.0 Ma. Elephas and Canis are documented only from the Brunhes epoch. The stratigraphic occurrence of these genera is correlated with other known biochronological sequences of subcontinent, and discussed in connection with the Plio-Pleistocene boundary in North India. [Die Großsäuger aus dem Plio-Pleistozän des intermontanen Kashmir-Beckens, Indien, unter Berücksichtigung ihrer Stellung in der paläomagnetischen Zeitskala] Kurzfassung: Reste von Gtoßsäugern (Equus, Elephas, Cer­ vus und Canis) aus der plio-pleistozänen Karewa-Formation des Kashmir-Beckens in NW-Indien werden beschrieben. Da die Fossilien aus paläomagnetisch datierten Schichten stammen, kann hiet erstmals eine zeitliche Korrelation für das Auftreten der Großsäuger für das Kashmir-Becken ge­ geben werden. Equus und Cervus erscheinen erstmals vor etwa 2.0 Mio. Jahren, während Elephas und Canis erst in der Brunhes-Epoche belegt sind. Das Auftreten dieser Gat tun­ gen wird mit andeten bekannten biochronologischen Sequenzen des Subkontinentes verglichen u n d kurz auf die Grenzziehung zwischen Pliound Pleistozän in Nordindien eingegangen. 1. I n t roduc t ion The Kashmir Valley, developed during the Late Cenozoic Era under different tectonic regimes, occupies the depression formed by the bifurcation of the Great *) Authors address: Dr. B. S. K O T L I A , Institut füt Paläon­ tologie, Nussallee-8, D — 5300 Bonn-1 , West Germany. Himalayan Range whose southern arm is known as the Pir Panjal Range and the northern arm as the Main Himalayan Range. In between the complexly folded and faulted mountain ramparts lies the Karewa intermontane basin which has accumulated above 1000 m of the Plio-Pleistocene synorogenic sediments compri­ sing mudstone-sandstone-gravel succession (BURBANK & JOHNSON 1982; KOTLIA 1985a; KUSUMGAR et al. 1986). The rise of the Pit Panjal Range was responsible for impounding the primaeval drainage and giving rise to a vast Karewa lake (SINGH 1982, BHATT 1983). As Pir Panjal continued to rise, the Karewa lake shrank and shifted towards the Himalayan Range. The heterogenous assemblage of lacustrine and glaciofluvial sediments (about 1000 m thick Lower Karewas), the lacustrine deposits (about 100 m thick Upper Karewas) and aeolian deposits (about 25 m thick loess deposits) has been accumulated in fotm of plateau-like terraces, called the "Karewas". The Kare­ wa sedimentation had commenced by about 4.0 Ma ago (BURBANK & JOHNSON 1982, 1983; KUSUMGAR et al. 1985, 1986; A G R A W A L et al. 1987). The first reference of mammalian fossils was given by D E TERRA & PATERSON (1939) and followed by BAD AM (1968), TEWARI & KACHROO (1977), KOTLIA et al. (1982) and SAHNI & KOTLIA (1985). The genera col­ lected or listed by them are Bos, Sus, Felis, Elephas, Equus, Rhinoceros and Sivatherium. While initiating and then carrying out in detail the palaeontological studies in the Karewas, KOTLIA (1985a, 1987a, 1989) made collection of mega and microvertebrate fossils with their stratigraphic occurrence in a composite litho-stratigraphic column. In addition to taxonomy, this paper also discusses the stratigraphic occurrence of various fossils in terms of absolute chronology. Large Mammals from the Plio-Pleistocene of Kashmir Intermontane Basin 39

Himalayan Range whose southern arm is known as the Pir Panjal Range and the northern arm as the Main Himalayan Range.In between the complexly folded and faulted mountain ramparts lies the Karewa intermontane basin which has accumulated above 1000 m of the Plio-Pleistocene synorogenic sediments compri sing mudstone-sandstone-gravel succession (BURBANK & JOHNSON 1982;KOTLIA 1985a;KUSUMGAR et al. 1986).The rise of the Pit Panjal Range was responsible for impounding the primaeval drainage and giving rise to a vast Karewa lake (SINGH 1982, BHATT 1983).As Pir Panjal continued to rise, the Karewa lake shrank and shifted towards the Himalayan Range.The heterogenous assemblage of lacustrine and glaciofluvial sediments (about 1000 m thick Lower Karewas), the lacustrine deposits (about 100 m thick Upper Karewas) and aeolian deposits (about 25 m thick loess deposits) has been accumulated in fotm of plateau-like terraces, called the "Karewas".The Kare wa sedimentation had commenced by about 4.0 Ma ago (BURBANK & JOHNSON 1982, 1983;KUSUMGAR et al. 1985KUSUMGAR et al. , 1986;;AGRAWAL et al. 1987).
While initiating and then carrying out in detail the palaeontological studies in the Karewas, KOTLIA (1985aKOTLIA ( , 1987aKOTLIA ( , 1989) ) made collection of mega and microvertebrate fossils with their stratigraphic occurrence in a composite litho-stratigraphic column.In addition to taxonomy, this paper also discusses the stratigraphic occurrence of various fossils in terms of absolute chronology.

Geologie
The Plio-Pleistocene Karewa deposits are commonly divided into two lithologic units, the Lower Karewa Formation (Lower Karewas) and Upper Karewa For mation (Upper Karewas).Lower Karewa deposits, in general, are gently inclined and better exposed on the Pir Panjal flank and have an unconformable contact with overlying nearly horizontal Upper Karewas which are better exposed on the Himalayan flank.The geo logical map of the area is shown in Fig. 1.The best exposed Lower Karewa sections are along River Rembiara between Dubjan and Krachipatra, and along River Romushi between Ichhagoz and Romu.Well ex posed Upper Karewa sections are at Sombur and Burzahom.The Lower Karewas are characterised by mudstones, unconsolidated sandstones, lignite layers, and conglomerate horizons.Upper Karewas are laminated claystones, sandstones and some conglomeratic layers but without lignites.The Karewas, as a whole, are capped by loess deposits associated with dark humic layers (palaeosols).Abb.1: Geologische Karte des Kashmir-Beckens (Indien), mit den Fundorten sowie der Verteilung der unteren und oberen Karewa-Formation.Measured sections are shown in Fig. 2. Sediments ex posed at Dubjan, Hirpur and Krachipatra localities are exposed along River Rembiara near Hirpur village, about 67 km SW of Srinagar.JOHNSON (1982JOHNSON ( , 1983) ) and KUSUMGAR et al. (1986) developed a new chronology of the Karewas through the use of magnetic polarity stratigraphy and fission track dating of volcanic ashes and correlated various Karea sub-sections.The dates are shown in Fig. 2. With more available data, AGRAWAL et al. (1987) dated the lower part of the Karewas (Hirpur section) to 3.4-2.0Ma, the middle part (Romushi section) to 2.4-0.2Ma, the upper part to < 0.73 Ma, and the loess sequence to < 200 Ka.In this work, the magne tic stratigraphy is taken from KUSUMGAR et al. (1986).(1986).Base of Karewas extends to about 3.8 Ma.Various sub-sections are correlated as described in text.In addition two layers of volcanic ash have been dated absolutely one of 2.4 ± 0.3 Ma given by BURBANK &JOHNSON (1983) Abb.2: Kotrelation zwischen den Fundhorizonten der Wirbeltiere und den paläomagnetischen Daten in der Karawa-Formation (vetändert nach KUSUMGAR et al. 1986) collected from the sediments showing normal polarity, probably belonging to Cochiti magnetic event.The 2nd horizon (dated to Reunion magnetic event), yielding the remains of Cervus is located near village Ichhagoz, about 30 m above the basal conglomerate of Romushi section (see Fig. 2).Based on few antler fragments, this material was originally identified by KOT-LIA et al. (1982) as Cervus sivalensis.However, with recovery of complete and well preserved antlers, it is well identified as Cervus punjabiensis.
About 70 m above this sand bed, is located 3rd horizon which has yielded the molars of Equus sivalensis alongwith numerous bone fragments.This horizon is located both in Romushi and Shaliganga (Husgund) sections.The 4th horizon with an age of about 1.6 Ma is represented by upper part of the Kilar section and has yielded Equus sivalensis.Antlers of Cervus sivalensis are collected from the second conglomerate horizon of the Romushi section, considered here as 5th horizon and dated to about 1.6 Ma.The records of Elephas hysudricus are from the Brunhes chron (younger than 730 Ka), represented by 6th horizon.Among Upper Karewa localities, Sombur section has yielded Elephas hysudricus.This is described as 7th horizon, dated to < 730 Ka.The 8th horizon, located at Burzahom has yielded Canis vitastensis remains.
Several Karewa lithologies were observed to be richly fossiliferous.The main are: dark mudstones interbedded with medium to coarse grained sandstones with gastropod shells as first and 4th horizons; lensoid, loose and current bedded sand beds as 2nd and 3rd horizons; conglomerates with sandstone and mudstone lenses as 5th horizon; medium grained unconsolidated huge sandstones interstratified with siltstones and greyish to yellowish patches of claystones as 6th and 7th hoirzons.

Systematic palaeontology
The large mammals, described by the earlier workers have been referred from villages without reference to measured sections.Thus, there is no information available of the exact location within the sttatigtaphic succession.However, one explanation for earlier finds could be that most of the fossils mentioned by DE TERRA & PATERSON (1939) and BADAM (1968, 1972) are from the Upper Karewas which are, in general, younger than 730 Ka in age.
The fossils described and illustrated herein were obtained in situ and keyed to a mastet litholog.The specimens are housed in the Museum of vertebrate palaeontological Labotatory (VPL) of the Centre of Advanced Study in Geology, Chandigarh.Several specimens collected from the surface lag deposits are not taken into considetation as their exact hotizons could not be ruled out.The main emphasis is given to the motphological description of Equus, Elephas and Cervus.New species of Canis have earlier been de scribed by KOTLIA (1987 b).Since more work on Hexaprotodon material is under progress, description is not given here.While describing equid matetial, dental terminology of AZZAROLI (1979) is used.Dental termi nology of cervids and morphological analysis of antlers are used as given by BROWN (1926)  There is an antetiot tine (btow tine) projecting forward then slightly inward and finally towards the posterior parts of palmate portion of antler.Posteiiotly, the palmate pottion of antler is continued back in beam with a prominent median axis and a tear tine.The rear tine of tetminal fork which forms the conti nuation of beam is long and wider.The posterior tine (beam) seems to have a small ridge along median axis above which the present specimen is broken.The brow tine is circular in cross section with a pointed tip at end.The antler is divergent, the brow tine making an obtuse angle (147 degree) with the beam.
The longitudinal grooves run parallel to median axis of the brow tine.The grooves are more prominent near palmate central portion, less prominent in middle part and are almost absent near the pointed termina tion of brow tine, thus giving rise a smooth surface to the tine.The beam consists of a few longitudinal grooves running parallel to median axis and they are nearly absent on latetal sides and thetefore the beam looks smoother than brow tine.The beam is triangular to circular in outline.Locality and horizon: Romushi section, 5th horizon in Fig. 2.

Description of new material
Left antler with pedicle (VPL/B 3004): The antler is heavily built, rough and supported on a rather long and nearly circular outlined pedicle.The anterior tine projects forward and slightly outward.Posteriorly, the palmate central portion of the antler is continued in beam with median axis and finally it terminates in two points or tines, at which the present is broken.Thus the antler might be said to have a long anteroposterior axis with a single outwardly directed brow tine at the front and a broad double pointed posterior termina tion.The brow tine of this specimen is broken at its free end at a small distance from the central portion of antler.It curves slightly outward and forward away from the beam and is roughly oval in cross section.The brow tine arises immediately above the burr and makes an obtuse angle (122 degree) with the beam which is slender, nearly oval in cross section, curved laterally along median axis showing the longitudinal grooves on its surface.The prominent longitudinal grooves run parallel to the median axis of beam.
The cervid antler has a branched structure consisting of anterior and posterior prongs which are designated by POCOCK (1933) as "a" (anterior) and "p" (poste rior) prongs.According to POCOCK (1933), as the "p" prong branches more, progressively more com plex antlers are found.The anterior prong "a" rarely branches as it projects out over the eye.COLBERT (1936) proposed to call the single anterior tine "a" supposing it to be homologous with "a" tine in other cervids.The backward projecting broad portion (beam) is designated by COLBERT (1936) as "p" and when this part of antlet shows two terminal prongs, these are designated as p e and p m (external and median posterior tines).Following POCOCK (1933), the present specimen can be said to have a base with a 1 and p 1 in the primary fork and only one p 2 in secondary fork.
Remarks: C. sivalensis antler is shaped very much like of C. duvauceli, that is, there is a brow tine directly above burr, while further up the beam divides into two branches, a smaller inner one and larger outer one.C. sivalensis antler is larger than that of Rucervus simplicidens.
It is further distinct from C. punja biensis by the development of brow tine immediately above burr and by having the comparatively more rough surface.Siwalik Hills have yielded a number of species of deer (LYDEKKER 1876;PILGRIM 1913;BROWN 1926;MATTHEW 1929;COLBERT 1935;AZZAROLI 1954;GAUR 1981) Maglio,Trans. Amer. Phil. Soc. 63,pi. 13, Lectotype: Maxillary molar, Fauna Antiqua Sivalensis, pi. 8 (3a).VPL/B 2061: The molar possesses at least eleven plates and is relatively the best preserved.Three antetiotmost plates are joined together and the first four plates from anterior show the sign of abrasion.The surface of wear of all plates which are embedded in cement is concave.The ridges are unworn showing characteristic folded and plicated enamel.The enamel is smooth, thin and plicated and thickness ranges from 2.5 mm to 2.8 mm.The length and height of preserved portion ate 140.5 and 147.0 mm respec tively.
VPL/B 2062: This broken tooth has six to seven ridge plates.Anteriormost plate is much smaller than second and is broken posteriorly.The second plate is slightly bent downwards and therefore leaves much place between the second and third.All plates are un worn and embedded in cement.Maximum enamel thickness is 2.8 mm.Molar shows slight expansion to wards posterior side.In lateral view, the unworn ridge plates are extended straight on the occlusal surface except second and third plates which are bent upward in the opposite directions.
VPL/B 2063: The molar, broken posteriorly, consists of six ridge plates, the anterior most being smaller than second.The plates are unworn and show folded enamel.The maximum thickness of enamel teaches up to 2.7 mm and in general is highly variable from anterior to posterior portion of teeth.
Remarks: The two most important and approximately similar aged species of Indian Elephas, E. hysudricus and E. (Archidiskodon) planifrons can be distin guished by: thickness of enamel (comparatively more in E. planifrons); shape of molars (very broad in E. planifrons); and the ratio of crown height to width of molar (greater in E. hysudricus).E. hysudricus is also distinct from E. planifrons in lacking the median expansion in the plates and in marking by strong pli cations in the enamel.The third species, E. platycephalus, reported by OSBORN (1929), however has a questionable relationship and may tepresent one stage of E. planifrons (MAGLIO 1973).The fourth species, E. namadicus, is younger than first two species, and is distinct from them in having low enamel thickness in teeth as well as a very high crown height, upto 150 percent greater than the width (MAGLIO 1973).

Brief description of material
The ramus is slender exhibiting a prominent sigmoid curve on the lingual side.The paraconid on M 2 is developed and metaconid is the largest cusp on the tooth.Protoconid is high, hypoconid poorly devel oped, entoconid lower than hypoconid, talonid trian gular and shallow basined, trigonid smaller than talo nid, and the external and internal cingula are well developed.The ratio of independent lengths of Ml, M 2 and M 3 of this specimen (in case of M1 and M 3, diameters of alveoli have been taken) is far from mem bers of subfamily Amphicyoninae.The specimen has been easily distinguished from othet species of Canis, i.e., C. curvipalatus and C. cautleyi, reported by PILGRIM (1932) from Upper Siwaliks, in having a better anterior cingulum in form of paraconid and a rounded entoconid which is higher than ridge shaped hypoconid (KOTLIA 1987b).Further the talonid on lower second molar of C. cautleyi consists of a tren chant hypoconid and a small entoconid whereas the talonid on this tooth of C. vitastensis is triangular in outline.
According to KOTLIA (1987 b), the ratio of length and width of the lower molars in fossil and living species of Indian Canis varies from 0.417 to 0.6 (Ml); from 0.667 to 0.706 (M2) and from 0.7 to 0.737 (M3), which fall within the limit of C. vitastensis with ratio of 0.46 (Ml); of 0.667 (M2) and of 0.7 (M3).

Stratigraphic occurrence of the fossils
While discussing the placement of the fossil mammals with respect to magnetic polarity time scale fot Katewas, it is noted that the appearances of Equus, Cervus and Elephas in Kashmir are latet than in the Siwaliks of Indian subcontinent.E q u i d a e : The oldest occutrence of Equus in North America is dated to 3.5 Ma (EVERNDEN et al. 1964).BOUT (1970) proposed an age of 2.5 Ma for the oldest appearance of Equus in the Old World (in Perrier faunas of Auvergne-Velay areas).KELLER et al. (1977) initially based their palaeomagnetic data on the supposed oldest occurrence of Equus at about 1.8 Ma in Pakistan.Subsequently, OPDYKE et al. (1979) recorded earliest appearance of Equus at 2.48 Ma in northern Pakistan.This was confirmed as the oldest appearance of Equus in southern Asia by BARRY et al. (1982).AZZAROLI & NEPOLEONE (1982) recorded Equus in the Indian Siwaliks from the Matuyama magnetic epoch whereas, TANDON et al. (1984) discussed the appearance of Equus at about Gauss/ Matuyama boundary (2.48 Ma).More recently, RANGA RAO et al. (1988) described Equus remains in Jammu Siwaliks from a level, 61 m below this boundary.The present study records the oldest appearance of Equus in Romushi section from Matuyama magnetic epoch (roughly at Reunion event) and in Kilar section from the top of Olduvai event (1.6 Ma), matking 3td and 4th horizons respectively (Fig. 2).The horizon expo sed along River Shaliganga at site Husgund (Raithan) and yielding Equus remains is correlated with the 3rd horizon of Romushi section (KOTLIA et al. 1982;KUSUMGAR et al. 1986 The Lower Karewa deposits have yielded a greater number of specimens than the Upper Karewa depo sits.This disparity in the proportions could be attri buted to greater thickness and better exposed nature of former.Among megavertebrates, Equus, Elephas and Cervus are better represented.As compared with Tatror and Pinjor Formations of nearby Siwaliks where one third of mammalian assemblage consists of bovids, the present assemblage is devoid of bovids.Even in literature, only a couple of bovid specimens (DE TERRA & PATERSON 1939, BADAM 1972) have been documented in the past.Also, the absence of crocodilids and chelonids which are so abundant in the Tatrot and Pinjor Faunas probably indicates that the climate during Karewa deposition was probably colder than that of Upper Siwaliks.It is well known that the rep tiles attain profusion and great size only in the warmer climates.Besides, KOTLIA (1989) described 15 fossiliferous horizons of the cyprinid fishes, mainly from the Lower Karewa localities.Among the micromammals, a number of specimens belonging to arvicolids, murids and soricids from three levels (2.5, 1.6 and < 730 Ka in age) have been collected.Some of the tesults on arvicolids, aged 1.6 Ma, are published (KOTLIA 1985a), and the detailed studies are in pro gress.In addition to vertebrates, numerous specimens of gastropodes and ostracodes have also been sampled.
The greater frequency of gtazing forms such as horses and elephants probably suggests the presence of enough grass covered areas to support large herds of these animals.The occurrence of enough browse alongwith grasses is indicated by cervids.The high proportional percentage of antlers compared to teeth is probably artificial as a single antler can break down into a number of fragments.The habitat spectra of Karewa mammals suggest a mosaic of dominant grass land and wooded grassland with subordinate wood land and some bushland (KOTLIA 1985 b).It appears that the areas close to the ancient lake were more open and supported grassy vegetation with some tree and bush cover.

Pliocene-Pleistocene boundary
The delineation of Plio-Pleistocene boundary has drawn the attention of workers for well over a century.The boundary has been placed in the light of various interpretations comparing the various data available on Siwalik vertebrate palaeontology and glacial sequences, at the base of Tatrot (MATTHEW 1929;LEWIS 1937;DE TERRA & PATERSON 1939;HOOIJER 1955), at the base of Pinjor (SAHNI & KHAN 1964;NANDA 1976) and the base of Boulder Conglomerate (PILGRIM 1944;WADIA 1951;SATSANGI & DUTTA 1971).However, with the advent of magnetic polarity time scale in the Siwaliks, various workers (OPDYKE et al. 1979;AZZAROLI & NAPOLEONE 1982;TANDON et al. 1984;RANGA RAO et al. 1988) came to an agree ment that the Tatrot/Pinjor boundary be equated with the Gauss/Matuyama boundary (2.48 Ma).Most of these workers also agreed that the earliest appearance of Equus coincides with this boundary.The author is in agreement with the view that Tattot / -Pinjot boundary coincides with the Gauss/Matuyama boundary and that the Equus arrived in the subconti nent approximately at this time.The Plio-Pleistocene boundary in Siwaliks has been approximated at Oldu vai event (KELLER et al. 1977;FLYNN & JACOBS 1982;KUSUMGAR et al. 1986), and either at top of Olduvai event or at Gauss/Matuyama boundaty (RANGA RAO 1988;RANGA RAO et al. 1988). OPDYKE et al. (1979) listed a number of fauna: Canis, Mustela, Hyaena, Felis and Machairodus etc., all appearing immediately above Olduvai event.Approx imately at the Olduvai event, the appearance of some micromammals (Mus, Golunda, cf.Rattus) is shown by FLYNN & JACOBS (1982) who consider this level as biostratigraphically very important.KOTLIA (1985a) described from the Olduvai event the arvicolids which are considered to be useful biochronological markers for Plio-Pleistocene. KOTLIA (1989) surveyed the Kare was also for fish remains and found that just a single horizon yielded fossil fishes after Olduvai event com pared to fourteen horizons below this level.The com bined data on mega and microvertebrate palaeontolo gy therefore suggest that a major global fauna change occurred at about Olduvai event.
How did the climate play a tole in the Kashmir Valley and elsewhere at Gauss /Matuyama boundary and at Olduvai event?The oldest appearance of arvicolids Kashmir Valley, developed during the Late Cenozoic Era under different tectonic regimes, occupies the depression formed by the bifurcation of the Great *) Authors address: Dr. B. S. KOTLIA, Institut füt Paläon tologie, Nussallee-8, D -5300 Bonn-1, West Germany.

Fig. 1 :
Fig. 1: Geological map of Kashmir basin showing distribution of Lower and Upper Karewas as well as the localities.

Fig. 2 Fig. 2 :
Fig.2shows the biochronology of megavertebrate fos sils.Only in-situ specimens are used for taxonomic description as well as for biochronology.The base of the Karewas extends to the Gilbert magnetic epoch, about 3.8 Ma.The Romushi and Hirpur sections are correlated based on the earlier results mentioned above.Kilar section is lateral extension of a part of the Romushi section and both are correlated by a promi nent conglomeratic bed which has been physically traced from Kilar to Aglar (Romushi).Sombur section falls well within the Brunhes chron (KUSUMGAR 1980), therefore, is younger than 730 Ka.The horizon loca ted at Burzahom is stratigraphically 20 m below the loess deposit.Since the loess had started depositing in Kashmir Valley at about 200 Ka ago(AGRAWAL et al. 1987), the underlying Karea sediments at Burzahom are savely considered to be slightly older than 200 Ka.
upper part of Lower Karewas are measured along River Romushi and described as Romushi section.Kilar section, about 60 km SW of Srinagar is exposed along Birnai Nala, a tributary of River Romushi.All above mentioned sections are measured on the Pit Panjal flank.Sombur section, about 18 km SE of Srinagar, and Burzahom section, about 12 km NE of Srinagar are represented by the Upper Karewa Formation and measured on the Hima layan flank.
They are described to gether in Hirpur section, the type section for lower part of the Lower Karewas.Ichhagoz and Romu deposits, forming the Late Pliocene to much of the Pleistocene (MIDDLE-Miss 1924; WADIA 1951; BHATT 1975), and of Late Miocene to Pleistocene age (ROY 1975).Based on the preliminary magnetostratigraphic data, KUSUMGAR (1980) indicated that the basal part of the Karewas en compassed the Gilbert magnetic epoch.BURBANK &

Table 1 :
Measurements (in mm) of Equus sivalensis specimens So far, four species of Equus have been described from India, viz., E. sivalensis, E. namadicus, E. palaeonus and E. cautleyi.E. namadicus is of large size whereas E. sivalensis is relatively small.In E. namadicus, protocones of the premolars and molars are much longer than in E. sivalensis.Futthet, E. siva lensis is older in age than E. namadicus.E. cautleyi has larger size of mandibular rami than that of E. is placed in the valley considerably further in than the base of protocone and hypocone.Paracone shows two costae, one median and other at anterior border, the latter produced into an angle on anterolabial side.The costae on both paracone and metacone run perpendicularly to the base of crown but the outer surface of metacone becomes slightly wider as it approaches the base of crown.The maxi mum length and width of the specimen are 22.0 and 23.0 mm respectively.
1935 Equus namadicus FALCONER & CAUTLEY, Col bert, Trans.Amei.Phil.Soc.26,160-161.1982EquussivalensisFALCONER& CAUTLEY, Azzaroli, Paleont.Ital.72,1-79.Lectotype: British Museum No. 16160, a cranium.Cotypes: British Museum Nos.17227, posterior por tion of ctanium; M 2626, left maxilla; M 2698, premaxilla und symphysis; 22107, portion of right ramus; 22108, portion of mandible.4.1.1.1.Description of new material Lower ptemolar (VPL/B 2074).Protoconid is smaller than hypoconid and the outet wall of hypoconid is situated slightly anteriorly compared to protoconid.The metaconid is rounded and slightly bigger than metastylid which is comptessed lingually.The entoconid btoadly opens into a triangular hypoconulid.The metaflexid is wider with its antetiot portion com pressed labiolingually.The entoflexid is compressed along the anteropostetiot length of tooth.Both meta flexid and entoflexid are filled with cement and open lingually through narrow openings.The ectoflexid is deep and pointed; it is separated from the anterior margin of entoflexid and posterior margin of meta flexid.The metaconid and metastylid form the chatacteristic double knot of equids.Lower molar (VPL/B 2072): Protoconid and hypo conid ate well developed, hypoconid being latget than protoconid.The anterior extremity of protoconid runs along lingual side of tooth, giving rise the anterior margin of metaflexid.The hypoconulid is triangular in shape.Metaconid is rounded cusp and narrowly opens into metastylid.The metaflexid is slightly curved and compressed anterolabially.Entoflexid is wider than metaflexid and is having mild crenulations on the outer wall.Linguaflexid is exactly V shaped with lingual sinus open and deep.Ectoflexid is somewhat angulat, deep and narrowly separated from postetiot side.Lower molar (VPL/B 2073): The anterior part of pro toconid, the hypoconid, and a pan of entoconid are broken in the specimen.The anterior exttemity of protoconid runs along anterior side of the tooth and has a thin enamel.The metaconid is latget than metastylid.The posterior wall of metaflexid is directed downward and outer wall of entoconid is crenulated.metastylid.Ectoflexid is narrow, V shaped and set apart from posterior margin of metaflexid.The posterior end of metaflexid and anterior end of entoflexid lie close to each other and the dentine in between them is fairly well developed.The entoflexid is more expanded posteriorly having crenulations on outer wall.extenduptomesostyle.There are two re-entrants, one on posterolingual margin of prefossette and other on postetolingual border of postfossette.The outer mar gin of postfossette reaches upto mesostyle, thus pro ducing a weakly developed metacone.The Plicaballin is slightly curved.Protocone is elongated and oriented anteroposteriorly.sivalensis which also differs with present day E. caballus in possessing smaller protocones.Clearly distin guishable E. sivalensis and E. cautleyi remains are from Pinjor Formation and E. namadicus ftomNarmada alluvium (KHAN 1972).Upper molar (VPL/B 3002): Protocone has a btoad and rounded inner surface with an outer wall exactly parallel to the anteroposterior axis of crown.On lingual side, protocone projects more lingually than hypocone, inner surface of which is slightly angular, sloping from lingual to labial side.Between protocone and hypocone, there is small and slender accessory tubercule which Right antler with pedicle (VPL/B 3003): The light brownish coloured antlet supported on a short and circular pedicle is moderately smooth in outline.The palmate central portion is rounded consisting of a heavily fossilised burr with an irregular margin.

Table 3 :
). C e i v i d a e : OPDYKE et al. (1979) recognized lowest occurrence of cervid with antlers in uppet Gauss epoch.AZZAROLI & NAPOLEONE (1982) repott ed Cervus antlets from Matuyama epoch.RANGA RAO et al. (1988) listed Cervus antlers in Jammu fauna from slightly below Gauss /Matuyama boundary.C. punjabiensis antlers, in this work, are recorded from Matuyama magnetic epoch, roughly at Reunion event (2nd horizon), and C. sivalensis is reported from upper part of Olduvai event (5th horizon).Variation in the comparative measutements of M 1, M2 and M 3 of Indian canids (After KOTLIA 1987 b) Archidiskodon at about Gauss/ Matuyama boundary.E. hysudricus in Karewas is recorded only after 0.73 Ma from Romushi section (6th horizon) and Sombur (7th horizon).Canis from Matuyama epoch.In this work, it is de scribed from the Burzahom section (Upper Karewas) which is younger than 730 Ka and older than 200 Ka (AGRAWAL et al. 1987.The locality is marked as 8th horizon.