ripen Peabody Museum of Natural History Yale University New Haven, CT 06520 Posti j la Number 179 30 November 1979 Miocene Sediments and Faunas of Pakistan Edited by David R. Pilbeam A.K. Behrensmeyer John C. Barry S.M. Ibrahim Shah Table of Contents pean PAD SIKAG easter cbieie te renee emote eh been Uh er ch ped eden y eine eet eRe eet eR eS 3 FOIReN Mo] fo) x renee \ Gk anor naia Metre me Coty a lee, A a: Mano N tines Wee a iactren art, peniheec are 4 ITROGUCTION es. cree te ee ba chee kee cr aN nl nS yua oe nmnTe ceca a vane ania ae 5 [© [=£o] Foye | (im Narre abner NA Aatey shares coal aa peu ea eo ao ORE aes AOE ar We Se ea Pe Ae ae 8 a. PotwanGeology-and:Siratigraphy-of Khaur Region: 7 0).-7s.).05 2s 2... ee 8 Db PaleoMmagnellc SttalGra@Vamieette vn emit pecitcaG comets arn cran tet cicrme oe 11 c. Lateral Lithofacies Relationships and Paleogeography ...................005. 7 ok [MoI ORTIEt(ofrelol aherelnlob rele) alolalolnew/emnchgs ice cee ne txt tt Coenosans a yoann 23 PES LUNIA LN decree ee mtenra tec aE hee oO mnths Ine Utne Mey toeeie le ahea! cattins its 25 PaleOntOlog Varah OM i City el car ecco anne ame Scar ence oxenut ige i eee LO CIUOTIO blair lathe Sey. te eter Rv nike Ms iM nmr, Mien ch sias We nah cree ets 27 Die ElSCOnieN Walia RrOCESSInCuer nun cain site Wesmierc niu tinates ct )cten st, aware totus 27 lege G {elorelalcalalo|(ofo\ifan patent an eine oie Wena unban er iedncis Nrah «canal ayaa coart Maen Neath ge een! 30 Cla Ginicll Seen arts ee ow cee oer cone eater ee ee OMe Ma cats at deat encectiencstrar eter: Mente nate Cy2 fee FIN [ola tantelnninatclitG roll ols eaten an ae astro tema in ceay Cuneta Mey Jeri os Sr camer rate ates. 38 fae CON] aulomzancalnLercOnbmental COMMECIOMS vine aac comme wey kaso sts cireelees 38 GadealCOCCU OG Vp teint aie atz. ats wr Pati hatoa cee ean itd ccna ean en ae ou SummMahy and CONnClUSION eis aera) tre tienes peters sen etre sean 41 (GioAnUi Al eLUWOL ESE ak Pvc lee nce Remedi en nT ce re ieee acne ees Pe cee Re My trae 42 [SILOTALURGSOLUG Clare ee mewn ts earee oe emia ee Gn ree, erates ec EG vs a easele ne erias serene E 43 Unpublished Manuscripts, Reports, and Other Documents ................ 000s eee 45 HSER GIG Gcokee SA aia te ca ca nrcren Ey heen acme sermyit Frm a here aeRO frase telee vee 45 (Received 8 March 1979) Abstract The results of five field seasons of work on the Siwalik sediments of Northern Pakistan have greatly expanded our knowledge of these Miocene sediments and their vertebrate faunas. We have measured six long strati- graphic sections on the north limb of the Soan synclinorium near the town of Khaur. These columns, the longest of which is over 3,000 meters, provide the stratigraphic framework for our paleontological studies and give a detailed description of the lithological sequences in the Khaur region. We have con- Cluded that the formational units of previous workers are poorly defined and of little prac- tical value for biostratigraphy or chronostrati- graphy. We recognize three major lithological facies: a blue-gray sand facies; a buff sand facies; and a silt/clay facies. The results of Intensive paleomagnetic sampling allow a Provisional correlation to the La Brecque Magnetic time scale. The paleomagnetic Sampling has also defined a series of iso- Chrons, one of which we have followed later- ally along a 30 km-long belt of outcrop. Cer- tain lithological horizons may also be reliable chronostratigraphic markers. We hypothesize that the two sand facies correspond to two separate river systems which co-existed for millions of years. The Characteristics of the blue-gray sandstones Suggest a large braided river carrying sedi- ment derived from a freshly weathered terrain. Miocene Sediments and Faunas of Pakistan Edited by David R. Pilbeam A.K. Behrensmeyer John C. Barry S.M. Ibrahim Shah The buff sandstones display characteristics of both meandering and braided channels. The sediments of this river system were derived from an area of intense weathering. The silt/ clay facies represent levee and floodplain deposition. The pattern of interfingering of the two sandstone facies, with broad overlap ona scale of at least 30-40 km, indicates periodic fluctuations in the dominance of one or the other river system. These fluctuations are seen as the result of periodic, extensive in- fluxes of the blue-gray system. Increased pro- duction of sand in the source area might have been the result of climatic or tectonic changes. Fossils are usually found only in the buff sands or their laterally equivalent fine-grained flood- plain and levee deposits. We recognize three types of fossil localities based on the charac- ters of the fossil assemblages and the sedi- ments. Localities in channel-related deposits were formed as composite events averaged over time and space, and therefore provide information suitable for paleoecological re- constructions. On the basis of appearances of key species we are provisionally defining a series of eleven biostratigraphic zones. These span the sequence from the Lower Siwaliks to the base of the Upper Siwaliks. The faunas of the three lowest zones show similarities to the Asteracian faunas of Europe and to the East African middle Miocene faunas. Zones 4 through 8 appear to be a period of faunal en- demism although there are some resem- blances to European and Asian faunas. Cor- relations of these middle zones are to Eurasian 4 Miocene Sediments and Faunas of Pakistan Postilla 179 localities dated between 10 and 8 million years (m.y.). Beginning in Zone 9 the faunal endemism is disturbed by aseries of immigra- tions and emigrations. Most of the inter- changes seem to be with Africa. Correlations suggest ages of 8to6m. y. for Zones 9 and 10. Paleomagnetic evidence places the base of Zone 11 at 5.1 m.y. Particularly important among the many thou- sands of fossils we have found are over one hundred new specimens of the hominoids Ramapithecus, Gigantopithecus, and Sivapithecus. Among the new finds are post- cranial elements which can be attributed to these three hominoids. The bulk of the homin- oid collection comes from a stratigraphic level provisionally dated at 9 m.y. From the geological evidence we infer the large river systems were not stable through time. River floodplains were well drained with few lakes or ponds except in cut-off channels. Most of the time the shifting, braided channels of the buff river system were dominant, creating a mosaic of habitats by constant destruction and renewal of plant successions resulting in avegetational mosaic of grassland, bush, and woodland. There is little detectable change in the trophic structure of the herbivores from Zone 1 through Zone 8. The faunal change in Zone 9 does suggest an underlying habitat change. Foreword Since 1973 many people have been involved in the Yale Peabody Museum-Geological Survey of Pakistan Siwalik Project. Simply listing them cannot possibly convey my debt to them. 1973-74 party: Glenn Conroy, Tony Gaston, Phillip Gingerich, Margaret Egan Gingerich, Grant Meyer, Mahmood Raza 1974—75 party: Grant Meyer, Tony Gaston, Martin Pickford, Jeff Barndt, Fred Sibley, John Barry, Wendy Barry, Mohammed Iqbal, Mahmood Raza, Catherine Badgley, Horace French; 1975-76 party: Grant Meyer, John Barry, Wendy Barry, Herbert Thomas, Mahmood Raza, Martin Pickford, Horace French, Mohammed Iqbal, Jane Conroy, Glenn Conroy, Tauqir Shuja, Louis Jacobs, Robert J. Poreda, Jeff Barndt, Holt Ardrey, W. William Bishop; 1976-77 party: John Barry, W. William Bishop, Andrew Hill, John Damuth, Martin Pickford, Grant Meyer, Louis Jacobs, Everett Lindsay, Iqbal Cheema, Lisa Tauxe, Mahmood Raza, Bruce MacFadden, Kay Behrensmeyer, Catherine Badgley, Horace French, Herbert Thomas, Pascal Tassy; 1977-78 party: Horace French, Iqbal Cheema, Louis Jacobs, Martin Pickford, John Barry, Kay Behrensmeyer, Pat Shayler, Marc Monaghan, Andrew Hill, Catherine Badgley, Lisa Tauxe, Mahmood Hassan, Bruce MacFadden, Bonnie Lipschutz. | am especially indebted to Dr. A.N. Fatmi, then Director, Geological Survey of Pakistan, with whom | began this collaborative project in 1973, and to Dr. S.M. Ibrahim Shah, Director, Geological Survey of Pakistan, my current collaborator and Codirector with whom | have worked so amicably since 1974. The then Director General, Geological Survey of Pakis- tan, Mr. A.N. Khan, now Joint Secretary, Min- istry of Fuel, Power, and Natural Resources, and the subsequent Directors General, Mr. S. Tayyab Ali and Dr. Asrakullah have been of tremendous assistance. The Geological Survey and the Ministry have given us their fullest cooperation at all stages. Various U.S. Government agencies have also been of great help. | feel it necessary to single out a few people for special thanks. Mahmood Raza and Grant Meyer made it possible to begin and sustain the project. Martin Pickford performed pro- digious feats of collecting and mapping. Horace French was always dependable. John 5. Miocene Sediments and Faunas of Pakistan Postilla 179 Barry was quietly efficient and creative throughout and now forms, with Kay Behrensmeyer, a productive team here at Yale as well as in the field. To these, and many, many others, | am eternally grateful. The report is edited by me, Ibrahim Shah, Kay Behrensmeyer, and John Barry; substantial portions were written by Drs. Behrensmeyer and Barry. It benefited greatly from the atten- tion of the ‘“‘Khaur Collective,” especially Drs. Lindsay and Jacobs, and Marc Monaghan. Names in parentheses following section titles indicate authors, or those with primary responsibility. The work has been supported throughout by the National Science Foundation and the Smithsonian Foreign Currency Program: Current grants are NSF BNS 772 5984 and SFCP FC80254100. This report was completed in December 1978, after five seasons. David R. Pilbeam 1. Introduction (Pilbeam) Ithas been known for more than a century that the Neogene Siwalik Group rocks of Indo- Pakistan contained abundant faunal remains, including hominoid primates. In the past two decades particular attention has focussed on one of the Siwalik hominoids, Ramapithecus Punjabicus, considered by many to be alikely early hominid. Whether hominids or not, the Siwalik primates were sampled from a time period during which hominids evidently dif- ferentiated (5 to 15 m.y.), and the demon- Strable completeness of the sedimentary Sequence is therefore particularly important. Initial and quite detailed Siwalik faunal analyses were completed before the 1940s, and although there have been a few studies since then there is a considerable need for modern comparative and functional work on the fauna. In addition, earlier reports and syntheses concentrated on integrated faunal studies, little attempt being made to elucidate precise biostratigraphic ranges or to recon- struct species associations. Further, the geological surveys completed prior to 1973 were generally of extremely broad scope. We began work in the Potwar Plateau of Pakistan (Figs. 1 and 2) in 1973 and have now completed five seasons. The research has involved collaborative field work by groups from the Geological Survey of Pakistan (GSP) led by Dr. S.M. Ibrahim Shah and the Peabody Museum of Natural History, Yale University (YPM). We have concentrated our efforts on Lower and Middle Siwalik rocks (roughly, those spanning middle and late Miocene and early Pliocene time), and have collaborated with a group from Howard University, also working with GSP and led by Dr. S. Taseer Hussain, who have worked on similar age rocks outside the Potwar Plateau. In addition, we have worked closely and exchanged personnel with a team from Peshawar— Dartmouth—Lamont—Arizona that has con- centrated its efforts on Upper Siwalik rocks (late Pliocene and early Pleistocene). The advantages of working in Siwalik Group rocks are considerable. They offer a virtually continuous record of the last 14 or 15 million years of Neogene time and make possible the potential definition of faunal community struc- tures and habitats at a particular time, and of changes in communities and habitats through time, within a sequence that can be calibrated independently from the faunas. The disadvan- tages are that the predominantly fluvial sedi- ments preserve relatively incomplete verte- brate remains, and that the tremendous scope of the project involves such a long term com- mitment of time, effort, and money. The Potwar Plateau (Fig. 2) is a folded and faulted block of Neogene molassic sediment, roughly 20,000 km? in area. Much is covered by late Pleistocene alluvium, but middle Miocene through early Pleistocene sediments are widely exposed both on the Plateau and in river and stream channels. Sedimentation is repetitive, consisting in any one section of Miocene Sediments and Faunas of Pakistan Postilla 179 Potwar PAKISTAN 4 5 yk *Ramnagar aoee oe" iHaritalyangar INDIA Fig. 1 Map of Indian subcontinent showing the location of the Potwar Plateau. alternating silts/clays and sandstones. The sequence was divided by earlier workers into a series of formations (from oldest to young- est: Kamlial, Chinji, Nagri, Dhok Pathan, Tatrot, Pinjor; see, for example, Cotter, 1933) and into a set of successive faunas and time zones with the same sequence of names. With the recognition that the formational bounda- ries are frequently highly time-transgressive, and that a series of four or five discrete faunal units may underestimate the modality or com- plexity of faunal changes, we have tried to make amore realistic assessment of the situa- tion by adopting an informal faunal zone ter- minology, tied to accurate and detailed strati- graphic mapping; we have also proposed new informal chronostratigraphic terms. The classic faunas from the Potwar Plateau come mainly from three areas: Chinji-Nagri, Hasnot-Tatrot, and Dhok Pathan (Fig. 2). We have greatly expanded collecting in areas to the north and east of Dhok Pathan, around Khaur and Kaulial. Unfortunately, until recent- ly collections from the three areas have been lumped together in faunal analyses; when they are separated it becomes clear that in each area somewhat different time periods are probably represented by the bulk of par- ticular local samples. Thus collections from Chinji-Nagri largely antedate those from Dhok Pathan-Khaur and those in turn are mostly older than those from the Hasnot-Tatrot area. Our first field season in 1973-74 involved reconnaissance around Chinji and Dhok Pathan, this being extended to the Khaur area in 1974-75. In 1975-76 and 1976-77 more systematic collecting and geological efforts were concentrated around Khaur and at Nagri. In 1977-78 preliminary work began at i Miocene Sediments and Faunas of Pakistan Postilla 179 KALA CHITTA HILLS Rawalpindi e@ A Kaulial ° seule SoSGandakas ©Dhok Mila ben 4 4km ie) 10 20 30 40 50 Fig. 2 Map of the Potwar Plateau showing the location of the three major research areas. A) Dhok Pathan-Khaur-Kaulial; B) Chinji-Nagri; C) Tatrot-Hasnot. Hasnot-Tatrot. Our aims have been as far as Possible to locate previously known sites, to fit these and any new localities to local and re- gional stratigraphic columns, to date the rocks and their contained faunas, to reconstruct Past habitats, and to elucidate changing habitat and faunal patterns through time; all this we see as absolutely essential to a clearer understanding of hominoid evolution. We believe that we have been relatively Suc- cessful in achieving many of these aims. Results of the first four seasons were sum- Marized in Pilbeam et al. (1977a, b) anda More expanded account of all five seasons is Included here. We have a fairly good under- Standing now of the stratigraphic relationships of many earlier fossil discoveries, of Potwar lithostratigraphy and magnetostratigraphy, and an expanded knowledge of paleogeog- raphy and paleohabitats. In addition, we have collected many new fossils and are beginnig to understand better the patterns of faunal evolution and migration. Also, we have doubled the previously known collections of hominoid primates, expanding our knowledge of those species by perhaps an order of mag- nitude because many of the new specimens are significant in their completeness or in adding information about new body parts (postcranials, for example). The first part of this paper surveys the geolog- ical work and is followed by a paleontological summary. Miocene Sediments and Faunas of Pakistan Postilla 179 2. Geology a. Potwar Geology and Stratigraphy of Khaur Region (Monaghan) The Potwar Plateau of the Punjab Province, Pakistan (72° 30’ E, 33° 00’ N) is an elevated area of some 20,000 km2, bounded to the north by the Kala Chitta and Margala Hills, south by the Salt Range, east by the Jnelum River and west by the Indus River. The plateau is underlain by Neogene molasse which was deposited in subsiding basins on the southern flanks of the rising Himalayas. After deposition the area was folded, eroded, and covered by Pleistocene alluvium. Substantial amounts of the Neogene rocks are exposed in canyons cut through the plateau. In the study area Neogene sediments form the Soan synclinor- ium, the axis of which runs roughly east-west, and are also exposed in anticlinal and mono- clinal belts along its northern and southern margins. Sediment thickness in the study area exceeds 5,000 m. Neogene sediments in the Potwar Plateau have been described and subdivided by a number of authors (Pilgrim, 1910 and 1913; Anderson, 1927; Cotter, 1933; Colbert, 1935; Lewis, 1937; Gill, 1952; Fatmi, 1973; and Pilbeam et al., 1977a and b). The work by Cotter, as represented in a geological map of the Khaur area (Fig. 3), reflects the results of work done by many of these authors and the framework within which many more recent workers discuss these sediments. To con- struct this map, particular lithological boun- daries (e.g., Nagri-Dhok Pathan) assumed to be isochronous were strike- mapped out of an area of clear vertical lithofacies contrast into other areas where the vertical contrast was not always so clear (see Fig. 4). In 1973 the Stratigraphic Committee of Pakis- tan formally defined purely lithostratigraphic nomenclature for these Neogene sediments: Age Formation Lithology early Pleistocene Soan (includes Conglomerate, to Tatrot and Sandstone, late Pliocene Pinjor of earlier Siltstone and workers) Clay middle Pliocene DhokPathan Sandstone and to Clay middle Miocene Nagri Sandstone Chinji Claystone and Sandstone middle Miocene Kamiial Sandstone to Murree Sandstone and early Miocene Claystone The lithostratigraphic nomenclature defined by the Stratigraphic Committee of Pakistan is virtually the same as that which was often used by earlier workers when describing bio- or chronostratigraphic units. This has resulted in some confusion. The Khaur Area The Khaur area lies on the northern flank of the Soan synclinorium. The major structures of the area are the Khaur and Dhulian anticlines (Fig. 3). Bedding plane dips vary between vertical and horizontal. In the fossiliferous part of the section, dips vary between 0 and 25° (to the southeast). Reverse faults of minor throw (1 to 20 feet) are uncommon but do occur through- out the study area. Six long stratigraphic sections (2,000 meters) and many more short sections were measured by G. Meyer, M. Raza, A.K. Behrensmeyer, J. Barry, M. Monaghan and P. Shayler. Details are available on request from the archives of the Yale Peabody Museum. A summary of por- tions of the measured sections is given in Figure 5. On the basis of studies done in 1978, sedi- ments can be grouped into three lithofacies associations: g) Miocene Sediments and Faunas of Pakistan Postilla 179 GEOLOGICAL MAP OF THE KHAUR AREA us STRATIGRAPHIC UNITS (FROM COTTER, 1933) DP = DHOK PATHAN N = NAGRI C = CHINUI K = — KAMLIAL M — MURREE ; UNIT BOUNDARY ———— q RIVER AA s STREAM -———- : TOWN 6 MEASURED SECTION | DHULIAN © 4 s \ if Fig. 3 Map of Dhok Pathan-Khaur-Kaulial area (A of Fig. 2) showing lithological units according to Cotter (1933) and location of measured sections. 1) The Blue-Gray Sand Facies, consisting of Quartz, feldspar, hornblende, garnet, mica, and schist-clast rich, arenitic, angular, me- dium to fine grain, blue-gray sandstone and gray-green to blue-gray silts. The sandstones extend as sheets which laterally grade into Silts over distances of several kilometers. In- dividual sandstones are up to 20 m in thick- ness. In some areas multistoried sandstones are formed by individual sandstones laid one Upon another. These multistoried sandstones May be over 125 m in thickness. The sand- Stones are often cross and planar bedded, Overwise massive. Conglomeratic layers are rare. The sheetlike nature of the sandstones and the continuity over great areas of paleosol horizons lying beneath and within a few feet of the base of these sheets indicate that the sandstones were laid down quickly and may be used as isochronous units. 2) The Buff Sand Facies, consisting of quartz, feldspar and garnet-rich, silty, subrounded, medium to fine grain, buff to yellow-brown sandstone. The sandstones are lenticular and complexly layered. Individual sandstones are rarely greater than 15 m in thickness. The sandstones are cross-bedded, and carbonate nodule conglomerate interbeds are common. The complex juxtaposition of buff sandstones and adjacent silt/clay layers indicates that 10 Miocene Sediments and Faunas of Pakistan Postilla 179 STRATIGRAPHIC UNITS (as mapped by Cofter, 1933) E W ee — DHOK PATHAN SS oom NAGRI CHINUI Fig. 4 Schematic diagrams showing the chronostrati- graphic units mapped by Cotter (1933) and others. Major lithostratigraphic boundaries interfinger and ifadhered to in mapping would not follow the chron- ostratigraphic boundaries (i.e., Nagri-Dhok Pathan boundary). Sand is stippled, silt-clay blank. these sediments cannot be used as isochron- ous units. 3) The Silt/Clay Facies, consisting of indis- tinctly bedded, poorly sorted, red, brown and brown-orange silts and clays. Paleosol hori- zons are delineated by hematite and car- bonate concretions, textural and color varia- tions. This lithofacies is volumetrically dominant in the eastern part of the study area. The Blue-Gray Sand Facies and the Buff Sand Facies interfinger and are interbedded. In general, the Buff Sand Facies and the Silt/ Clay Facies are predominant lower in the section and eastward across the mapped area. Previous work by Gill (1951) and Cotter (1933) indicates that there is a west to east fining trend in the region. Thick, cobble conglomerates are found to the west of the town of Pindi Gheb in the unit mapped as Nagri by Cotter (Fig. 3). At the present time our group uses lithofacies in a descriptive and environmental context and notin achronostratigraphic context. More field mapping and section measuring is needed before we can describe and map boundaries of lithostratigraphic units on a regional scale. Further discussion on litho- facies is presented in a following section. wi Miocene Sediments and Postilla 179 Faunas of Pakistan 0 STRATIGRAPHIC COLUMNS vertical We scale (meters) 200 300 Fs sandstones > 12m. thick an 400 260 locality GK 260 _} reversed interval ” Cig Dae 261-4 251 258 259 b c a ae 0.8 | Ain 1 kilometers kilometers Fig. 5 Simplified measured sections corresponding to locations given in Figure 3. The position of the “U” Sandstone, Ganda Kas reversed intervals, and representative fossil localities are shown. Solid lines Connecting the columns are lithologically con- tinuous horizons walked out between sections. Only Sands thicker than 12 m are shown; thinner sands Occur throughout the columns (i.e., “U"). b. Paleomagnetic Stratigraphy (Tauxe, Behrensmeyer) Work is still in progress, but much has already €en accomplished in identifying reversals of the Earth's magnetic field as preserved in the Siwalik rocks. The reversal pattern in strati- Yraphic sections, or paleomagnetic strati- graphy, is useful both in local and regional Correlation, ue Paleomagnetic stratigraphy of the Siwalik roup in the Khaur area (Figs. 6, 7, 8) has been studied by J. Barndt of Dartmouth Col- lege (1977) and L. Tauxe of Yale. Reconnais- sance sampling by Barndt was done at widely spaced intervals through 3500 m of section to establish the broad pattern of polarity changes (Fig. 6). Ageneral pattern consisting of a long period of dominantly normal polarity punctuated by several short reversals was identified on the northern rim of the Soan Syncline in the area mapped by Cotter as Nagri (Fig. 3). During 1976-77 and 1977-78, Tauxe sampled a second long column and also traced laterally several of the reversed 2 Miocene Sediments and Postilla 179 Faunas of Pakistan 3600-4 UN ie 32004 } GK ‘ S a DM HL tie. ® 2800 cy 5 aeomei er or 4 tl | i i shy ; Game ais = 24004 7 at J gO") l < pal © 2000-44 ; ae i (ht eel lie i Geese orient ea Kode ienla:Shgos 160044 cae ateec| @ io ce 2007 65 490 a +s 800- qq z 2) 400-J uc Z p ieee oA peo E 3 a oO an Fig.6 @& a or Magnetic polarity sections from the Dhok Pathan- 8 S bd 2 ‘ ts ‘ ' =z WwW WwW Khaur-Kaulial area. Positive values indicate normal = & 4 polarity, negative values reversed (revised from < < - fz Barnat et al., 1978). < Ke re) zx a UN) Utran Kas; KM) Khot Maliaran Kas; DM) Dhok Eee: o Ff Mila Kas; HL) Hasal Kas; MKM) Malhuwala Kas; r GK) Ganda Kas. “U" and “'T” are major blue-gray 5 sand units. & 2800-4 o cA % 2400 Ww rs x 2 F 20004 Q PS a C4 6 600 1 z & te ” 12004 Jy 800 4 g Fig. 7 » eee ’ Dhok Pathan-Khaur-Kaulial area composite section V (Hasal Kas and Malhuwala Kas sections) com- (oped Oo «(6 EPOCH II pared with three published reversal scales (modi- fied from Barndt et al., 1978). 13 Miocene Sediments and Postilla 179 Faunas of Pakistan A B Cc D EF G 2 a = ou" Fi SANDST. & a 3 $3 ee St pats r A 4 = fo} o Fig. 8 Middle Siwalik magnetostratigraphy and probable Correlation to standard magnetic polarity sequence, noting different levels of resolution. A) composite of detailed sections (Tauxe, B.S. thesis, also Fig. 9); B) composite of regional mag- netic stratigraphy (Tauxe, RK and KUL sections); C) magnetic subzones: D) magnetic zone; E) Chronostratigraphic units; F) composite of regional Magnetic stratigraphy (Barndt et al., 1978, HL and MKM sectional); G) magnetic polarity time scale (see Barndt et al., 1978). events within the long normal zone of Barndt, Using closely spaced vertical sampling to €stablish the details of the paleomagnetic record. The magnetic memory, or remanence, of these rocks is carried by hematite, a highly Stable iron-oxide. A complete study of the Magnetic mineralogy is underway in order to determine definitely the time of acquisition of the remanence, but we presently believe that itwas acquired either during or soon after deposition (Tauxe, in preparation). POLARITY NORMAL REVERSED Sampling of the Siwalik sediments for paleo- magnetic purposes is constrained by anum- ber of factors which ultimately affect the detail of local columns and therefore the reliability of magnetostratigraphic correlations. In the Khaur area, fine-grained lithologies suitable for paleomagnetic sampling make up less than 50% of the stratigraphic column. Sand- stones up to or more than 125 m thick occur throughout the section, and sampling is restricted to the fine-grained sequences between them. Thus, due to the nature of the sedimentary record, it would be relatively 14 Miocene Sediments and Faunas of Pakistan Postilla 179 easy to miss short-term polarity changes if sandstones represent significant segments of time. The working assumption has been that the dominant long-term reversal pattern should still be apparent in long stratigraphic sections, even at low sampling densities, and correlations can be made using “broad” patterns. However, lateral correlation of spe- cific polarity changes can easily be compli- cated by the combined sampling problems, even over distances of a few kilometers. Barndt (1977) sampled in five river channels (kas) which cut through the south-dipping sediments, from Utran Kas at Dhok Pathan to Ganda Kas (Figs. 6, 11). His longest paleo- magnetic record is from Hasal Kas. This column includes a stratigraphically thick segment of normal polarity, with three reversed events, overlying a considerable thickness of alternately normal and reversed sediments. Using faunal correlations which suggest that the normal polarized section should be in the range of 8-11 m.y. in age, Barndt, et al. (1978) correlate the “long nor- mal” in Hasal Kas to Epoch 9 and marine magnetic Anomaly 5 of the magnetic time scale (LaBrecque et al., 1977). Epoch 9 is thought to span some 1.5 m.y. and to fall between about 10.0 m.y. and 8.5 m.y. Asystem of nomenclature for the magnetic stratigraphy is presented in Figure 8. We pro- pose the name “Khaur Normal” for the long normal unit in the Khaur area, together with a local chronostratigraphic unit, which contains the Khaur Normal, called the “Dhurnal unit.” Tauxe concentrated her sampling further to the east from Ganda Kas, where fine-grained sediments make up more of the stratigraphic column. In Ratha Kas, 1300 m of section were sampled at an average density of 1 sample per 15-20 m. The upper and lower boun- daries of the Knaur Normal have been identi- fied in Ratha Kas and the Kaulial section has extended the paleomagnetic stratigraphy into younger rocks than Barndt’s Malhuwala sec- tion (Fig. 8). Eight other short sections were sampled at higher densities (up to 1 per 2-3 m) along approximately 25 km of outcrops spanning the so-called “U” sandstone (Fig. 9). This sandstone has been used by us as a marker horizon for placing widely separated fossil localities in a stratigraphic framework. The “U” sandstone occurs in the sixth reversed “event” (from the bottom) in the composite section representing the Dhurnal unit (Fig. 9). Samples spaced at intervals of 2-3 m spanning the “U” sandstone event in Ganda Kas show that the “event” is actually com- posed of two reversed and one short normal periods (Figs. 8, 9). These have been referred to informally as the Ganda Kas (GK) reversed interval. At least seven major reversed inter- vals characterize the Khaur Normal overall, rather than the four revealed by lower sam- pling density (Fig. 8). Tauxe has found that the number of reversals documented in a strati- graphic column increases with sampling density. Thus, there may be many more rever- sals in the Khaur Normal than have been detected so far. This presumably reflects the fact that sedimentation rates for fine-grained Siwalik sediments were often relatively higher than in the marine sediments where the cur- rent “standards” for the Miocene magneto- stratigraphic record have been established. If so, then the reversal record in the Siwalik Group may eventually be more detailed than in marine deposits (though less continuous overall) and this must be taken into account in making correlations from region to region. The paleomagnetic patterns vary laterally for the reversed interval (GK) which spans the “U" sandstone (Fig. 9). However, the initial reversal after a relatively thick sequence of normally polarized sediments can be identi- fied within a few meters of the base of ‘U” in most of the laterally correlated sections. This correlation establishes the closest approxi- mation yet possible for an isochron through the fluvial sediments. This isochron is essen- tially parallel to the base of the ‘“U”, which therefore appears to be isochronous in the study area. Lateral variation in thickness and 15 Miocene Sediments and Faunas of Pakistan Postilla 179 23 KM A METERS 04 | Pa fs | | Fig. 9 Composite of all stratigraphic sections bracketing “U” and “Buff-E” sandstone horizons showing associated virtual geomagnetic pole latitudes. The dots (three specimens in good agreement using Criteria of Irving, 1964) are Class | data; letters (R and N) are Class II data. For section abbreviations See Figure 6. Solid lines indicate paleomagnetic Correlations, dashed lines lithologic correlations. lithology of the “U” sandstone are described In the following section. The Stratigraphic and paleomagnetic records for the “U” level east of Ganda Kas lie ina direction nearly perpendicular (S-N) to the depositional axis (W-E, as indicated by Paleocurrent directions; see Figs. 11, 12). The 'Sochron indicates that lateral spread of'sand away from this axis was not time transgres- Sve, and that “U” (and presumably other simi- lar sandstones) can be used as time markers 'N north-south exposures. However, the sands May be more time transgressive west to east, Parallel to the depositional axis, and therefore '€SS precise as chronostratigraphic markers In these directions. y os l Ae sus —"y" \SOCHRON N [SS ea B @ ok REVERSED EVENTS The sediments included in the reversed inter- val discussed above can be compared later- ally as representing contemporaneous sedimentary environments. Likewise, fossil localities included in these sediments can be compared for faunal differences that are due to ecological or taphonomic factors rather than evolutionary ones. This part of the analy- sis in the Khaur area remains to be done. It would be helpful to know the time represented by the entire reversed event, but at present there is no reliable information on the absolute time spanned by reversed events in the Khaur Normal. However, relative rates of sedimenta- tion in the various laterally equivalent sections can be calculated, and this should provide some information on local areas of rapid ver- 16 Miocene Sediments and Faunas of Pakistan Postilla 179 JEVELOF'U SANDSTONE AND IGANDA KAS REVERSED INTERVAL 5 Fig. 10 Map of Khaur area showing normal (dots) and reversed areas; major normal region is the Khaur Normal zone. Letter abbreviations are as in Figure 6. Circles indicate polarity changes documented in long columns by Barndt (B) and Tauxe (7) The narrow reversed bands within the Khaur Normal zone are Tauxe’s interpretation of the most prob- able correlations between reversed events in the various sections (see also Barndt et al., 1978, fig. 6). sus slow sediment accumulation. This, in turn, should be interesting in relation to fossil-bear- ing horizons. Accurate identification of field reversals in several columns spanning the Khaur Normal in addition to that at the “U” level will even- tually permit other isochrons to be drawn through the section. The faunas and sedi- ments inthe Khaur area will then have a firmer chronostratigraphic framework which it is hoped can be extended to other areas. Pat- terns of evolution, community succession, extinction, and environmental change can be worked out in more detail using this relative time reference, even if its precise correlation to absolute time scales remains uncertain. > WH Miocene Sediments and Faunas of Pakistan Postilla 179 c. Lateral Lithofacies Relationships and Paleogeography (Behrensmeyer) A study of lateral lithofacies variation in the fossil-bearing Siwalik sediments of the Khaur area (Fig. 11) was undertaken in 1978. The Purposes of this study were to: 1) test the lateral continuity of prominent sandstone units used as local stratigraphic markers; 2) coor- dinate lateral stratigraphic sections with Paleomagnetic sampling in order to identify 'sochronous surfaces in relation to the sedi- ments; 3) provide information on the nature and scale of lateral variation in lithofacies in Order to reconstruct paleogeography; 4) €stablish the relationship between ‘“‘Nagri” and “Dhok Pathan” lithofacies in an area where they appear to interfinger (Pilgrim, 1910, 1913; Cotter, 1933; Lewis, 1937; Gill, 1951). Previous work on lateral relationships of facies and sandstone correlations between the can- yons or kas in the Khaur area by M. Pickford Indicated that a number of the “blue-gray Sandstones” could be followed along strike for Many kilometers. Two of these sandstones, U" and “T” originally designated by Pickford, were chosen for the 1978 study. In the type area of Pickford’s sandstone units, Malhuwala Kas, these two units are interbedded ina Sequence of buff-sand and silt-sand facies which include many of the important Miocene fossil localities. The methods used for study Included mapping of the “U” sandstone and associated red beds over a distance of 30 km and Measuring detailed sections through “U” at intervals along strike (Figs. 11, 12). A total of 15 sections spanning 50 to over 100 ™ Of stratigraphic thickness were measured (Fig, 12). Documented rock characteristics Included texture, color, sedimentary struc- tures, bedding thickness, diagenetic features ale as Carbonate nodules, and fossil con- Chit As noted by Monaghan in Section 2a, there are two distinctive types of sandstone in the Khaur area Middle Siwaliks sequence. The blue-gray sandstone is characterized by its distinctive color and is well sorted, clean and immature with angular quartz, chert and feld- spar grains and abundant fresh mafic miner- als and rock fragments. The buff sandstone is typically buff to gray-brown, more mature in composition with fewer mafic minerals and more weathered rock fragments and feld- spars. Rock fragments are commonly per- meated by hematite, and some hematitic staining occurs in the carbonate cement. The blue-gray sandstones occur as extensive sheets which grade laterally into finely bedded gray sands and silts. The buff sand- stones occur as sheets and as lenses that abruptly thicken or thin and occasionally pass laterally into interbedded fine brown to red sands and silts. For the most part the two lithologies are distinct throughout the study area although a few occurences of apparent mixing have been noted. The blue-gray facies occurs more commonly low in the section (relative to the “T—U” level) and in the west toward Chouktriwali Kas (Fig. 11). Buff sand- stones dominate the upper part of the section and are more common east of Ganda Kas. In addition to the sandstones, the sedimentary sequence includes fine-grained deposits. These sediments are highly variable in texture and sorting, but all contain a significant amount of silt or clay. Color is typically red- brown but can vary from gray-brown to red- orange and yellow-brown. Bedding is variable but usually is indistinct in fine-scale, and there is evidence for extensive bioturbation and physical soil-forming processes. In summary, the Siwalik sediments of the Khaur areacan be characterized in terms of at least three major lithofacies: 1) the Blue-gray Sand Facies, 2) the Buff Sand Facies, 3) the Silt Clay Facies. These lithological components are convenient in describing vertical and lateral variations in the overall stratigraphy. At finer levels of resolution they include significant degrees of variability which forms a limitless source of information for fine-scale paleoenvironmental and paleoecological studies. The blue-gray sandstones allow definition of what have been informally called “cycles” in 18 Miocene Sediments and Postilla 179 Miocene Sediments and Postilla 179 Faunas of Pakistan Faunas of Pakistan As vibe ese at CHOUKTRIWALI k, "227 aes ae hs y i ° 1 2 3 Cae ee I 7 "U" SANDSTONE CURRENT } ®~ "u" SANDSTONE DIREC TIONS Q BUFF SANDSTONES IN CYCLE UNDER “U" X SECTIONS Fig. 11 Map of Khaur area showing positions of lateral sections (“Xs”) along the “U" sandstone. Small dots indicate fossil localities. Current directions suggest that the trend for “U”" (blue-gray) system is east- south-east while that for underlying buff sandstones is more southeast (current indicators averaged for each kas). 20 Miocene Sediments and Faunas of Pakistan Postilla 179 "air" a ( SANDSTONE Fig. 12 Fence diagram showing the “U” sandstone and major overlying buff sandstones toward the north- east. Some of the buff sand bodies under “U” are indicated. Dotted line under “‘U” is the paleomag- netic isochron determined by Tauxe for the base of areversed event in the Khaur Normal magnetozone A PRIMATE LOCALITIES a SILT-CLAY FACIES Y BLUE-GRAY SAND FACIES BUFF SAND FACIES CE gi 100430 rh Nee eT KI f 270 20 50 M SCALES as it related to the various lithofacies. Points at the top of “U” marked by “X” correspond to points marked by “X” on the map in Figure 11. Letter abbreviations for sections also correspond to kas or localities indicated in Figures 6 and 11. the Khaur area. In the lower part of the section in Malhuwala Kas, atypical cycle consists of a 10-50 m thick blue-gray sandstone conform- ably overlain by 60-80 m of clay-silts which include buff sandstone lenses, interbedded red to buff silts and sands, thick beds of red, brown and orange silty clays and occasional units of drab gray to green thinly bedded sand and silt. The next blue-gray sandstone over- lies these clay-silts on an erosional contact (generally with less than 3 m of relief). The Middle Siwalik deposits have long been recognized as fluvial, and attempts have been made to designate the fluvial regime(s) as meandering and/or braided. In the Khaur area detailed study has shown that the depositional system was complex and was probably influenced by a number of factors including intrinsic cycling mechanisms of the river sys- tems, tectonics, climate, and source materi- als. The sand deposits primarily represent river channel deposition. The blue-gray sand- stone sheets, with their lateral continuity and complex bedding, suggest a braided river carrying a high proportion of sand. The buff sandstone lenses display characteristics of both meandering and braided channels, with cutbank and oxbow features, variable current directions, abrupt vertical alternation of sand and silt, and some point-bar facies. However, classic upward-fining fluvial bedding sequences are not common. The fine-grained deposits (‘red beds”’) represent levee and floodplain deposition. Evidence for persistent floodplain swamps or lakes is notably scarce. Most of the silt and clays represent overbank deposition on land surfaces where bioturba- tion permitted relatively few primary bedding features to survive. The fine-grained sedi- ments were contributed by overbank flow from both the braided and meandering channels. 21 Miocene Sediments and Faunas of Pakistan Postilla 179 Fig. 13 Preliminary interpretation of the paleogeography of Khaur area Middle Siwalik deposits. Heavy dashed line indicates line of sections in Figure 12. Relative distance to mountain ranges is conjectural. In the Khaur area, the sand deposits toward the west and the finer sediments toward the Northeast apparently kept pace with each other as each system continued to aggrade. This balance probably was controlled by the rate of subsidence and the base level which both regimes shared. The two sandstone types may represent two different paleo river systems in the Khaur area. IN addition to the distinctive lithologies, other €Vvidence supports this hypothesis: 1) current directions for the buff sandstones around the U" sandstone level are consistently more BLUE-GRAY SAND FACIES Fe BUFF SAND FACIES ie} io 20 8% | aeeiere emer Mees | APPROXIMATE SCALE variable and trend south to southeast, whereas those for the blue-gray sandstones are east to southeast (Fig. 11); 2) the tabular shape of the blue-gray sand bodies indicates a different fluvial regime from that of the more lenticular buff sand bodies; 3) the blue-gray sandstones at the ‘‘T—U” level thin and pass laterally into finely bedded sand and silt and eventually pinch out in a direction (N—NE) perpendicular to the east-trending currents (Figs. 11 and 12). Thickening of blue-gray sands south and west suggests that the main axis of blue-gray sand deposition lies in that direction, while that of the buff sand lies toward the north or east. Following the two-river hypothesis, a tentative 22 Miocene Sediments and Faunas of Pakistan Postilla 179 overall paleogeography of the Khaur area at the ‘“T—U” sandstone level is reconstructed in Figure 13. Both river systems were contributing to aggradation in the same subsiding trough. Therefore, their differences are likely to be due to different conditions in the source areas. Braided streams indicate a load larger than the river is competent to transport, meander- ing streams are either at grade or are locally undersupplied in terms of sediment load. It seems that the source of the “blue-gray river” provided abundant fresh sand and little by way of chemical weathering products (clay, etc.), whereas the “buff river” drained an area of soils and sedimentary deposits with more fine-grained and chemically weathered mate- rial. Taking modern Himalayan geography as amodel, the blue-gray river may have drained a high relief area from within the mountain ranges, such as the modern Indus does today, whereas the buff river drained a lower relief region on the southern margin of the mountains, as does the modern Jhelum. Cur- rent directions and depositional strike sug- gest that in the Miocene these rivers trended locally more southeast, perhaps following the axis of a trough formed by the junction of the Indian and Asian continental plates. The pattern of interfingering of the two sand- stone facies, with broad overlap on a scale of at least 30—40 km, indicates periodic fluctua- tions inthe dominance of one or the other river system. Paleomagnetic sampling of the sedi- ments associated with blue-gray sandstone “U” show that its base is essentially isochron- ous in the study area (Fig. 12). Thus it appears that the blue gray river occasionally swamped the region with great influxes of sand. These varied in lateral extent and decreased in importance upward in the middle Siwaliks deposits. The pulses of sand created the boundaries for the “cycles” which character- ize these deposits near Khaur. The buff river seems to have been more persistent in its sedimentation patterns, dominating the Khaur area except when “displaced” by the blue- gray river. Causes of the punctuated, extensive blue- gray sand influxes could be extrinsic (¢.g., climatic, tectonic), intrinsic to the river system, or possibly a combination of both. Extrinsic effects, such as increased sand production in the source area, are attractive hypotheses because such controls on stratigraphic sequence imply a dominance of mega proc- esses, which therefore could be inferred from local stratigraphy. However, more proximal, intrinsic processes must also be considered in explaining the blue-gray sandstones. Such intrinsic controls would include periodic “waves” of headward erosion (as suggested by Schumm, 1977) which might result in increased sand deposition relative to silt and also increased lateral spread of sands from river channels. Continuing work on both field evidence and sedimentation models will it is hoped develop ideas to help discriminate between intrinsic and extrinsic controls on the Siwalik stratigraphic sequences. The lateral lithofacies study provides geologi- cal evidence for the paleoenvironments and paleogeography of the Miocene ecosystems of the Khaur area. There appear to have been at least two major rivers within 50-100 km of each other, the sandy and more braided blue- gray river to the southwest and the more meandering buff river to the northeast. Other smaller streams were present as well, and all drained off the rising Himalayan hills and mountains across what must have been a broad, coalescing piedmont fan (Fig. 13). Floodplains were relatively well drained with few lakes or ponds except in cut-off channels. Periodically sand from the blue-gray river spread across the Khaur area, temporarily altering the soil conditions and thereby affect- ing vegetation and fauna. For most of the time, however, the shifting channels of the buff river were dominant, creating a mosaic of habitats through constant renewal of plant succes- sions near the channel axes. Most of the ver- tebrate fossils in the Khaur area are derived from deposits of the buff river system; thus it appears that either ecologic or taphonomic factors (or both) have favored preservation in this paleoenvironmental setting. From the 23 Miocene Sediments and Faunas of Pakistan Postilla 179 paleogeographical reconstruction of the Khaur area we can infer that the fluvial envi- ronments associated with the Miocene faunas of the Dhurnal chronostratigraphic unit were homogeneous on a large scale and probably also through time. There may have been gradual ecological change toward the blue- Qray river system, and relatively sudden shifts during periods of lateral spread of the blue- gray sands. On a smaller scale, the buff river system provided a mosaic of depositional environments and habitats. Habitats used by vertebrates probably varied over distances of 10-1,000 m, but were recurrent and relatively homogeneous over larger distances across the piedmont fan system. d. Microstratigraphy and Taphonomy (Badgley) In 1976 we began taphonomic field studies of the fossil accumulations and sediments at Selected Potwar localities. This work has Concentrated on analysis of both the sedimen- tary environments in which fossils occur and the characteristics of bone assemblages from different sedimentary environments. The pur- Pose of these studies is to determine the nature of the preservational biases and how they affect the reconstruction of the Miocene Community from fossil assemblages. Most of the studied localities are in the Khaur region, although a few are near Chinji, Sethi Nagri, or Dhok Pathan. All of the Khaur area localities are in the upper half of the Dhurnal Chronostratigraphic unit, as depicted in Fig- Ures 12 and 14. Although the Dhurnal chrono- Stratigraphic unit spans about 1.5 m.y., there IS little faunal change during this time period. Thus the studied fossil assemblages are derived froma single chronofauna. Study of the lateral variation of the representation of this fauna allows us to define taphonomic and €cologic effects on the faunal composition of the fossil localities. We have made detailed field analyses of the Stratigraphic and sedimentary features of 40 localities. Field data includes: composition and grain size, sorting, color, texture, bedding features, geometry, extent of bioturbation, lateral and vertical variation of fossiliferous units, and the relationships of fossiliferous units to adjacent strata. At 14 of the 40 localities we made systematic surface samples of the fossils. Taphonomic information recorded from these samples includes: faunal and skeletal-part composi- tion, variation in weathering stages present, completeness of each specimen, presence or absence of articulated material, signs of bone damage caused by carnivores or rodents, and (where possible) orientations of bones in the sediment. From these data the minimum number of individuals and the number of specimens per individual were calculated. The results of the lateral lithofacies and micro- stratigraphic studies suggest that while isolat- ed fossils are sparsely distributed throughout the stratigraphic interval studied, concentra- tions containing remains of more than one individual are regularly associated with only a few of the many lithologies present. In the Khaur area such concentrations seem to occur most often in the sediments of the buff- sand complex. Few concentrations are known from the extensive blue-gray sandstones or the laterally equivalent thin-bedded gray silts. This disparity may be partly related to the greater volume and areal exposure of the exposure of the buff sands and associated fine-grained deposits. All fossil concentrations generally fall into one of three sedimentary contexts. This pattern is thought to reflect the transport history of each fossil assemblage. Within the sediments of the buff-sand system, the large concentrations of fossils tend to occur in the coarser-grained facies within, or close to, the domain of great- est current activity. The most common deposits containing large bone concentrations are the intraformational conglomerates which usually occur within cross-bedded sandstones. These conglom- erates and their fossils are interpreted as lag 24 Miocene Sediments and Faunas of Pakistan Postilla 179 deposits associated with persistent current activity. The conglomerates are mixtures of clay- and silt-clasts derived from previously deposited floodplain deposits, plus calcare- ous nodules and algal oncolites of local origin, sand pebbles, and occasional pebbles for- eign to the system. The matrix varies from moderately sorted sand to poorly sorted mix- tures of clay, silt, and sand. Bioturbation of the sediments is minimal. Most of the bones are broken and show a wide variety of weathering stages. Animals of a large size range are present; the average number of bones per individual is low; and bones are never articulated. The second most common fossil-bearing deposits are heterogeneous silty-sandy sediments marginal to cross-bedded sand bodies interpreted as channel axes. These channel margin deposits may have been formed in topographic depressions within a large sandstone body which were filled with finer, more heterogeneous sediment. Biotur- bation structures are extensive, occurring as root casts, burrows, mottling, and footprints. Primary bedding features such as mud drapes, silt lenses, and thin sandy-silt layers are usually still in evidence. Most of the bones from localities in this lithology are broken and show a mixture of weathering states. Animals from a large size range are present and the average number of bones per individual is low, but slightly greater than in the conglom- erate localities. Articulation of bones is rare. While the intraformational conglomerate local- ities show predominantly the influence of fluv- ial processes and little or no evidence of exposure as land-surfaces, the channel mar- gin environments show the results of the oper- ation of both fluvial and land-surface proc- esses. This suggests an alternation between periods of fluvial deposition and periods of dessication which may have been accompa- nied by occupation of the site by plants and animals. The third type of fossil-bearing deposit is the fine-grained floodplain sediments of the buff- sand system. The predominant lithologies are silt and clay with variable amounts of said, so that the deposits vary from slightly sandy silt to silty clay and clayey silt. Colors include numerous hues of orange-brown, yellow- brown, red-brown, maroon-brown, and dark brown. Gray and gray-green fine-grained units are uncommon and rarely contain fos- sils. These fine-grained beds attain great thicknesses between prominent sandstones but few clues to their primary depositional mode are evident. The contacts between fine- grained units indicate both gradational and sudden changes in depositional mode but are never erosional. Primary bedding features are rare and presumably have been erased by bioturbating agents. The activity of these agents is evidenced by abundant small root casts and scars, burrows, and horizons of blue-gray mottling. All identified paleosol horizons occur within the fine-grained sequences and various soil-forming proc- esses must also have affected the sediments. Taken together the features of the fine-grained units indicate they formed on aggrading land surfaces. Such localities sometimes contain many bones from a single individual. These bones are often intact and sometimes articulated. The preliminary analysis of the microstrati- graphic and surface-sampling data from 14 localities suggests to us several of the proc- esses and events which may have been important in forming the fossil concentrations. For example, unusual breakage of large mammal bones indicates that the fresh bones were trampled before burial, while the pres- ence in a fossil assemblage of specimens with different degrees of weathering indicates that a bone assemblage was formed over a sub- stantial period of time. Similarly, the activities of carnivores and rodents are clearly shown by tooth marks and breakage patterns, and one unusual accumulation with several rela- tively complete small animals may have been a carnivore’s food cache. There are pronounced taphonomic differ- ences between fossil assemblages from 29 Miocene Sediments and Faunas of Pakistan Postilla 179 Channel-related and floodplain-related locali- ties. The effects of fluvial transport in accumu- lating bones from a large geographic area, thereby removing them from the site of death and possibly even reworking them from pre- viously deposited sediments, dominate the intraformational conglomerate localities. In channel margin localities such effects are INComplete and there is a mixture of trans- Ported and untransported elements. In chan- nel deposits skeletons may be widely dis- persed and, while there may be large numbers of bones with many individuals and Species, rarely is an individual represented by More than five bones and equally rarely are Specimens articulated. In the fine-grained deposits, however, fluvial transport was weak Or nonexistent and skeletons were widely dis- Persed only through the activities of predators and scavengers. The sedimentary and taphonomic features Indicate that channel-related localities were for med as composite events averaged over time and space. There is no evidence that any channel-related accumulations resulted from Catastrophic mortality. The most common taxa In Channel-related localities, as expressed by minimum numbers, are medium-sized bovids and equids (40-200 kg weight class), with tragulids, suids, giraffes, elephants, and rhinos well dispersed (present in many locali- ties) but less abundant than bovids or equids. Carnivores are slightly less well dispersed but are often common where they occur. Homi- NOids follow the same pattern as carnivores but are generally rarer. Anthracotheres are UNCommon and chalicotheres are very rare. Large, high diversity localities in fine-grained deposits are uncommon. Most floodplain &ccumulations contain only a few taxa, but Certain rare taxa have only been found in such localities. This distinction may be the result of habitat separation or differences in the pres- €rvation processes associated with fine- 9rained versus coarse-grained facies. The direct record of the biological properties of the Preserved animals, such as habitat prefer- €Nce or behavior of predators is, we believe, best preserved in the floodplain localities, perhaps partially perserved in the channel margin deposits, and poorly preserved in the channel conglomerate deposits. However, since channel-related concentrations are time- and space-averaged, they may best record the species compositions and relative abundances of the Dhurnal chronofauna as a whole. e. Summary The essential descriptive framework for geo- logical interpretation is now complete enough in the Khaur area to permit some preliminary interpretations and to indicate the most impor- tant directions for further research. Long stratigraphic columns in six kas show that a broad pattern of vertical change occurs from lower through Middle Siwalik facies, corresponding to the Chinji-Nagri-Dhok Pathan Formations, as described by Fatmi (1973). Lateral variation in lithology on both a large and small scale makes boundary defini- tion extremely difficult, however, and the three distinctive lithofacies cannot be regarded as simple, successive time-rock units in the Khaur area, or mapped in the sense of “proper” formations. In the Khaur area, large-scale lithostrati- graphic units are most easily defined on the basis of the thickness and lateral extent of the sandstones. Thin tabular sandstones are characteristic of the lower part of the section and are comparable to sandstones of the Chinji Fm. in its type area. These are overlain by a sequence of thick, blue-gray sandstones comparable to those of the Nagri Fm. in its type area. Thinner blue-gray and buff sand- stones characterize the upper part of the Khaur sequence, and these can be traced laterally to the type section of the Dhok Pathan Fm. Over approximately 40 km of laterally continuous exposures in the Khaur area, there is amarked decrease in the proportion of sand in section from west to east, and the transition zone between thick, Nagri-type sandstones 26 Miocene Sediments and Faunas of Pakistan Postilla 179 and the Dhok Pathan lithofacies drops lower in the section in this direction as well. Because of the problems in mapping boundaries between the three dominant lithofacies and in correlating them to the type areas for the Chinji and Nagri Fms., at present we prefer to use “Chinji lithofacies” and “Nagri lithofacies” rather than “formation” in referring to these divisions of the Khaur section. Measured stratigraphic sections document between 1,000 and 2,000 m of continuous deposits bearing Miocene faunas. Work aimed at dividing this long column into bio- and chronostratigraphic units is well under way. Paleomagnetic reversal patterns have established a provisional temporal framework for environmental and faunal studies in the Khaur area. If a relatively long normal period, locally named the Khaur Normal, is equivalent to Epoch 9 of the marine sedimentary paleo- magnetic record and to Anomaly 5 of the marine magnetic anomaly time scale (Barndt et al., 1978), then many of the fossil localities are between about 8 and 10 m.y. in age. Informal biostratigraphic zonation of the sec- tion based on the mammalian faunas will be discussed later. Detailed study of lateral facies near the top of the Dhurnal chronostratigraphic unit (which records the Khaur Normal period) shows that the sedimentary deposits resulted from a complex of fluvial systems rather than a single aggrading river. The rivers responsible for the Khaur area sediments probably drained topographically different areas of the rising Himalayas. Evidence suggests that the Khaur area lay on the distal portions of coalescing, low-angle alluvial fans in a subsiding trough which paralleled the suture line of the Indian and Asian plates. East to southeast current directions indicate that the rivers at this time may have flowed along this trough rather than southward as does the modern Indus River. One of the major river systems in the Khaur area periodically deposited widespread sheets of blue-gray sand and sandy silt. One of these sand sheets, the ‘‘U” sandstone, appears to be essentially isochronous based on paleomagnetic correlations, and others probably are as well. The other major system deposited buff sands and red-brown, fine- grained sediments in a less clearly repetitive pattern than the blue-gray river system. Dif- ferences in the patterns of sedimentation may be related to source area tectonics and cli- mate, or both, or to processes intrinsic to the river systems. Nagrilithofacies in the Khaur area correspond to thick, multistoried sand deposits of the blue- gray river system. The Dhok Pathan Fm. con- sists of interfingering deposits of the buff and blue-gray river systems. Thus, the terms “Nagri” and the “Dhok Pathan,” which have been used not only as formational names but also as bio- and chronostratigraphic entities, actually appear to be rock bodies that repre- sent the fluctuations in space and time of two (or more) contemporaneous river systems draining different parts of the rising Himalayas. Fossil vertebrates are found primarily associ- ated with the buff sand system, in discrete patches or low density scatters in specific microfacies. The densest concentrations occur in carbonate-clast conglomerates in channels, in mixed sand and silt facies at the tops of channels and occasionally in fine- grained overbank deposits. Low-density scatters are found in most channel sands, and isolated bones occur rarely in the fine-grained red beds. The taxa represented tend to be more diverse in the channel-related deposits. Skeletal part ratios indicate that assemblages in channel lag sediments are most altered by fluvial sorting. The localities in the upper part of the Dhurnal unit are usually characterized by numerous remains of Hipparion and medium-sized bovids; suids, elephants, rhinos, tragulids, and giraffes are also pres- ent. Anthracotheres, primates, and chalico- theres occur less commonly. The patterns of occurrence of these groups may reflect char acteristics of the original mammal community: 2 Miocene Sediments and Faunas of Pakistan Postilla 179 The geological evidence for paleoenviron- ments indicates broadly homogeneous, well- drained fluvial plains with smaller-scale envi- ronmental mosaics associated with the Channel axes. Sands at or near the ground Surface were probably important as aquifers and their areal distribution may have exerted Considerable control on vegetational patterns, Particularly if the climate was strongly Seasonal. 3. Paleontology a. Introduction A general outline of biostratigraphy and bio- Chronogy has been given in Pilbeam et al., 1977a. Since that report was compiled we have collected more specimens, expanded Our stratigraphic knowledge, and become aware of several new problems and issues. What follows is a brief account of our knowI- Edge of the fauna, first assessed by taxonomic group: almost all of the groups are undergoing revision and reanalysis and we can expect Significant advances in our knowledge soon. At the moment, these brief accounts should be regarded as provisional. Second, we dis- Cuss biostratigraphic and biochronologic Implications before analyzing the possible Significance of the fauna in assessing inter- Continental migrations and past habitats. However, before the faunal analyses we start with sections describing our progress in elec- tronic data processing, and follow this with a general discussion of our approach to geo- Chronology. b. Electronic Data Processing (Barry) I. Introduction With nearly 15,000 specimens from over 350 localities and an ever increasing complexity of relevant data even the most basic analysisis a formidable task. Therefore, one objective of the YPM-GSP Siwalik Project has been the development of a computer-based data retrieval system which would serve not only as a collection management tool but also as an aid to research. Data retrieval, as distinct from other aspects of electronic data processing, is the searching of a file of data records by a computer for those records that contain some particular item of information — for example, records of speci- mens found at one particular locality. A data retrieval system uses logical expressions to select a record that meets the stated condi- tions; it may then print the entire record or abstract some part of it. A computer thus can be used to organize the data, and because of the speed at which it operates, a computer- based data retrieval system can efficiently sift through a file containing thousands of individ- ual records. To date the following has been accomplished: 1) We have determined the categories of information needed in the specimen file and developed a workable coding system. 2) We have collected data on approximately 4,000 fossils including all those found during the 1977 and 1978 field seasons plus all of the primates, carnivores, giraffes and anthraco- theres. In addition, records are now being compiled for all of the bovids, suids, tragulids, rodents, rhinos, proboscideans, birds, anda large fraction of the equids. When completed, this will bring the specimen file up to approx- imately 10,000 records. Data have also been collected on part of the 1932 Lewis (YPM) collection. 3) Over 4,000 of these records are now written on magnetic disks and are available for use. 4) We have written and proofed a series of system programs (Barry, n.d.). These include file creation and maintenance programs as well as sorting and retrieval programs. The Society of Vertebrate Paleontologists has published guidelines (S.V.P. News Bulletin No. 97) concerning data conventions in verte- brate paleontology. Those guidelines recom- mend several categories of data which we find are not relevant to our proposed research 28 Miocene Sediments and Faunas of Pakistan Postilla 179 although some are implicit in our locality category (i.e., all levels of geographic and stratigraphic data). Furthermore, we have made major modifications in recording other categories, principally the skeletal element descriptions. Three of our categories (Taphonomic Indicators, Size, and Hydraulic Sorting Factors) are entirely new. Each record is a fixed length, alphanumeric character string 84 characters long. Data categories are organized as fields having a fixed position within the record, whereas the file is a matrix with individual records as rows. Thus any one value can be accessed by indexing both row and columns. Indexing for row alone accesses an entire record and indexing for column accesses all values for any category. Information is generally coded either directly with a numeric code, or with a binary present/ absent code. The third data category is some- thing of an exception and because of its com- plexity needs special comment. In this cate- gory, information is coded with a series of abbreviations in seven separate fields, four of which are independently searchable (Pri- mary, Side, Condition, and Wear). The other three subcategories modify the Primary field and can only be used in conjunction with it. They may be used either separately or in combinations. Currently we have four specimen files: 1) PRIMATE, which contains only the records of the Primates; 2) CARNIVORE, which contains only the records of carnivores; 3) MAMMAL, which contains the records of all the mam- mals; and 4) MASTER, which contains the records of all catalogued specimens. We also contemplate using locality files in the future but the data standards for them are still being developed. Data files are stored in on-line disk storage and are accessible from the APL (A Programming Language) language through the APL auxiliary processors. Small data files may also be stored in on-line storage as part of named APL workspaces which also contain processing programs. With files of more than 1,000 records, however, this is not a cost- effective technique. We also maintain backup copies of the files and programs on magnetic tape. These could be used for jobs run in batch. In the field, data are recorded on computer- printed worksheets. Records are then entered into disk storage through a remote terminal using an APL program. Other APL programs will display, correct, add, or delete data from individual records. ll. Data Standards and File Organization Data standards refer to the type of information collected and the manner in which it is coded. In our system all specimens have their own record, even if they are articulated bones of one individual, and within each record there are seven basic categories of information, six of which have individually searchable sub- categories. These categories and searchable subcategories are: 1) Specimen identification: A. Collection identification letter B. Specimen identification number 2) Taxonomic identification: As stored in the data file these are expressed as numbers but on either input or output scientific names are used; the system automatically converts from names to numbers and back again. Taxonom- ic categories are hierarchical and allow a specimen to be identified at various levels within the hierarchy. A. “Highest Taxon,” generally class B. Order C. Family D. Subfamily E. Tribe F. Genus G. Species 3) Element description: A. Anatomical description 1. Primary 2. Side 3. Secondary: In three separate fields which can be linked to the 29 Miocene Sediments and Faunas of Pakistan Postilla 179 primary field either individually or in combinations. These hold two types of data: a. Completeness b. Further anatomical description B. Condition of specimen C. Wear (for teeth only) 4) Locality number 5) Taphonomic indicators A. Indicators of predepositional transport (graded sequence): 1. Abrasion 2. Rounding 3. Polishing B. Indicators of predepositional weather- ing (graded sequence): 1. Cracking 2. Checking 3. Exfoliation C. Toothmarks: 1. Parallel (caused by rodents) 2. Nonparallel (caused by carnivores?) 3. Feathering (caused by carnivores?) 4. Puncture D. Predepositional fractures: 1. Spiral (green bone) 2. Local crushing 3. Fibrous (green bone?) 4. Regular [= step] (dry bone?) 5. Irregular E. Specimen articulated with another specimen F. Evidence of a pathological condition G. Evidence of etching by stomach acids 8) Hydraulic Sorting Factors: A. Density (scaled from 0-10) B. “Volume” C. Shape indices 7) Size: A. Size estimate based on specimen B. Size estimate based on average for taxon to which it belongs Ill. Command Language We have written a series of computer pro- grams that will accept a set of logical instruc- tions, check their syntax, and then do a search through the data file and print the requested information. The command language in which the logical instructions are written is rigorous but English-like in syntax and vocabulary. To begin, the user writes a sentence indicating which subprogram he wants to use, what item of information he wants compiled, and what logical conditions he wants to impose on the search. Each instruction begins with the name of the subprogram, followed by the requested item and a conditional statement. Conditional statements are constructed from three clauses (the IF, FOR, and WITH clauses). These clauses are written with parentheses, logical connectors, and certain keywords. An example of a valid instruction is: “LIST GENERA IF LOCALITIES ARE YOITORVAS2 These instructions would cause the subpro- gram LIST to search the data file for all records of specimens from localities Y311 and Y182. From these records LIST would compile (retrieve) a list of genera found at either local- ity Y311 or locality Y182 or both. This list would then be printed showing the number of spec- imens at each locality for each genus. In this instance “IF,” “LOCALITIES,” and “ARE” are keywords and “OR” is a logical connector. Our programs are written in APL and stored as named APL workspaces on accessible disk memory. Currently there is one master pro- gram which coordinates one or more subpro- grams (LIST, SORT, and PLOT) plus a set of programs to create and maintain the data files. These programs are accessible through either TSO (Time-Sharing Option) or batch. From TSO, jobs may be submitted to batch and printed on the system printer, or run in TSO and printed at the terminal. By using var- ious APL auxiliary processors, the results of any job or the data file itself may be passed to any other program in the operating system for secondary analysis. 30 Miocene Sediments and Faunas of Pakistan Postilla 179 IV Subprogram Functions and Organization The three current subprograms and their func- tions are: 1) LIST — compiles and prints lists of requested data items plus accumu- lated counts; 2) SORT — sorts data records by stated crite- ria; and 3) PLOT — determines stratigraphic ranges of taxa or sets of taxa as controlled by the conditional statement; prints as a graphic display. In addition, a fourth subprogram (CALCULATE) is contemplated. This will perform standard or user-specified calculations on the results of LIST and SORT. c. Geochronology (Barry, Lindsay) The key to understanding many of the pale- ontological problems of the Siwaliks lies in the development of a reliable chronostratigraphic framework, which in turn depends on the resolution of the lithostratigraphic and bio- stratigraphic problems. Many of these problems are discussed in Pilbeam et al. (1977a), but no final solutions have yet been presented as our studies in these areas are in an early stage of development. Good pro- gress has been made on the lithostratigraphy and we now have a collection adequate for generally characterizing the faunal sequences, although this collection is still sparse at the bottom and top of the sequence. Until all the faunal revisions are completed, it is difficult to define and characterize ade- quately biostratigraphic and chronostrati- graphic units; however, some preliminary attempts are in order. Because of the nature of fossil occurrences in the Potwar Plateau, it may be possible to define formal biostratigraphic and chrono- stratigraphic units such as is frequently done with invertebrate assemblages but rarely with vertebrates (James, 1963; Lindsay, 1972). In the Khaur region alone there is a continuous rock sequence spanning nearly ten million years and, although not uniformly fossiliferous throughout, this sequence does preserve mammalian fossils at nearly every level. By strike mapping it is possible to determine the stratigraphic relationships of nearly all collect- ing localities so that we are able to establish the true stratigraphic ranges of many of the mammalian taxa. On this basis then, we will be able to erect a series of biostratigraphic zones, defining and characterizing them prin- cipally on the successive appearances (generally immigration events) of faunal ele- ments (Murphy, 1977; Woodburne, 1977). These biostratigraphic zones can be broad- ened and correlated by physical stratigraphy Fig. 14D Correlations of the magneto-, litho-, and biostrati- graphic zones of the Chinji, Hasnot, and Khaur regions. Two vertical scales are used. The scaie for the magnetic epochs is in millions of years; the scale for the rest of the figure is in meters and is based on the thicknesses of the formations in the Khaur region. Correlation between these scales is based on the correlation of the top and bottom of the Khaur Normal Magnetozone to the top and bottom of Magnetic Epoch 9. The dates for these boun- daries, as marked by broken lines, are generally put at around 8.5 and 10.0 million years. Each biostrat- igraphic zone is numbered and labeled with the name of the taxon whose local appearance marks the lower boundary. Large-face numbers indicate localities at which the various taxa appear and in effect define the lower boundaries of the zones. Small-face numbers are additional localities we have tentatively placed in the stratigraphic sequence. Letters prefixing the numbers refer to localities found by G.E. Lewis (L), Barnum Brown (B), the Yale-GSP group (Y), and the Dartmouth- Peshawar-Arizona group (DP). Those without letter prefixes are also Yale-GSP localities. Zone 11 is placed in this scheme on the basis of the first appearance of Hexaprotodon relative to the paleo- magnetic sequence. All other zones are placed 07 the basis of the position of the large faced localities in the lithostratigraphic sequence. Miocene Sediments and Faunas of Pakistan Postilla 179 Magnetic Chrono- | Magneto- Forma- Biostrat Chinji Hasnot Khour & epochs” zones zones tions Zones region region region + Gilbert GF af Z | | Hexaprotodon pp2i 5 Y) =e VY) 6 Y 7 Z Z 2370 369 IO Indarctos L97 86-8 BI35}_ Jabi press ps S LO ee 97 9924 B38 Hasal Dhok {9 Lycyaena 8 Pathan = v8 fm 158 166 ae IQ smal! suid 153 7 Palhyaena YI98 El Percrocuta 182 260 a 6 grandis 227 Y3I7 Khaur 390 ao 9 Dhurnal | Normal Nagri 362 235 330 on fm 3 337 261 Cormo- i 5 hipparion 82 BUS Teka small hipparion 4 complex ___ _ _]94 BI7-23 254 i128 ieee 259 ——— B108 om: 859-60 76 333-305 10 336 335 a8 Chinji Mer ycopotamus a fm dissimlis 38 Y54 B25 (38/41 Bora — r— Conohyus 189 BI44 sindiense Base BI43. lames LListriodon 7m 29 Kamlial l2 Pa fm VSN Ei Stee Hipparion Datum 62 Miocene Sediments and Faunas of Pakistan Postilla 179 and paleomagnetic zonation to establish a chronostratigraphic framework reliable for the Khaur area and other areas of Siwalik deposi- tion. The resulting chronostratigraphic frame- work should be useful beyond the limits of the Siwalik deposits because of the duration and completeness of those deposits and the appearances of mammalian immigrants. The informal units discussed in Pilbeam et al. (1977a) (Dhok Mila, Utran, Gandakas, and Kundvaii ‘‘units”’) represent an early attempt at defining chronostratigraphic units. Although these units were discussed in terms of litho- logy, they were meant to have a strictly time value. (In fact, the defining time-event bound- aries were sandstone bodies which were, and still are, thought to be minimally time- transgressive.) We have revised this earlier lithologic frame of reference to one based on more generally applicable time-events such as magnetic reversals and faunal changes. A preliminary biostratigraphic zonation is shown in Figure 14. In this scheme the bottom of each interval-zone (Hedberg, 1976) is defined solely by the appearance of some particular taxon (listed on Fig. 14), whereas the top is defined by the bottom of the suc- ceeding zone. As ameans of emphasizing that this biostratigraphic scheme is highly provisional we have chosen to refer to the zones by numbers rather than formal names. The ranges of the zone-delimiting taxa are not necessarily restricted to one zone and, in fact, most of these taxa range considerably higher than just one zone. Our faunal studies are still in progress and our analyses are incomplete; therefore we have not attempted fully to char- acterize these zones (Murphy, 1977), nor have we established complete reference sec- tions, although the Khaur region may be taken as the primary reference area for all but the oldest and youngest zones. In time we expect to refine and improve this scheme by the addi- tion of more taxa and events, and will formally define and characterize the zones and desig- nate stratotypes. These zones will then serve as the basis for a series of chronostratigraphic zones. For now, the main purpose of the zones is to provide an internal biostratigraphic frame- work for current research and later chrono- stratigraphic reference. The framework is meant to be time-significant, for although these proposed zones are biostratigraphic we have based the boundaries on what we believe are time-dependent events. We recognize, however, that when all the faunal studies are completed, these biostratigraphic units may be equally significant for paleoeco- logical studies and geochronology. In addition to the biostratigraphic zonation we are also proposing an informal and prelimin- ary chronostratigraphic zonation for the Khaur area which is based on the local paleomagne- tic sequences. As shown in Figure 14 the boundaries between the Bora and Dhurnal chronostratigraphic units and the Dhurnal and Hasal chronostratigraphic units are marked by the top and bottom of the Khaur normal magnetozone. d. Mammals As noted above, these brief summaries are quite preliminary and will be considerably expanded as various specialists complete their reports. 1. Primates (Pilbeam, Rose) A very brief report on new primate material from Pakistan was published by Pilbeam et al. (1977b). Other discussions have also appeared (Pilbeam, 1978a, b) or are in press (1979); adetailed description is in preparation (Pilbeam et al). Previously, four species of hominoid primate have been described from the Indo-Pakistan Miocene: Ramapithecus punjabicus, Dryo- pithecus (Sivapithecus) sivalensis, D. (S.) indicus, and Gigantopithecus bilaspurensis. In addition, a prosimian (/ndraloris lulli), a colobine (?Presbytis sivalensis), and a cer- copithecine (Macaca-paleindicus) are also known. 33 Miocene Sediments and Faunas of Pakistan Postilla 179 We have recovered more prosimian material, including a possible prosimian specimen from locality 259 in ?zone 3, a small lorisid lower molar from locality 182a in zone 6, and asso- Ciated jaws, cranial, and postcranieal parts of a small lorisid from locality 363 in ?zone 9. Associated lower teeth of a colobine were also found at locality 370 in ?zone 10 or 11. Remains of at least three hominoid primate Species, totaling over 100 specimens from More than 50 individuals, have been found during the past four seasons. They represent Ramapithecus, Gigantopithecus, and Siva- bithecus indicus (previously D. (S.) indicus), although other species may be sampled. Of particular interest are postcranial remains which seem also to represent at least three Species. They are the first recovered from Indo-Pakistan and some of the first middle Miocene hominoid postcranials ever found. Although these were previously divided into two groups, Hominidae and Pongidae, Pil- beam etal. (1977b) proposed classification in a single family, Ramapithecidae. Ramapithe- Cids are characterized by thick-enameled Cheek teeth, some or all have enlarged cheek teeth, some have relatively reduced canines; Postcranially some were probably similar to the African apes. Stratigraphically, almost all our hominoids Come from zones 5 and 6. Previous collections Probably also sample zones 72, 3, and per- haps 7 or 8. 2. Small Mammals: Insectivora, Tupaiidae, Chiroptera, and Rodentia (Jacobs) Small mammals are conspicuously rare in Previous Siwalik collections (see Black, 1972). However, we have been successful in recov- €ring numerous small mammal taxa from Many horizons, many of which are repre- Sented by large samples recovered by screening from a few key localities. Many of these taxa have been or are being described by Jacobs (1977a, b; 1978; 1979a, b, c). Our localities Y41 and Y 39 in the Chinji Forma- tion at Chinji are possibly in zone 2 and have yielded Antemus chinjiensis, Kanisamys sp., Rhizomyoides sp., Copemys sp., Megacri- cetodonsp., Sayimys sp., anda soricid. Black (1972) described Paraulacodus indicus, Sivacanthion complicatus, Rhizomyoides punyabiensis, and Kanisamys indicus, all of which probably come from zone 2 or 3. From locality Y259, which is probably in zone 3, we have a ?tupaiid, an erinaceid, a glirid, a flying squirrel, a tree squirrel, a ctenodactylid, a cricetid, Kanisamys sp., and the murid cf. Progonomys sp. Locality Y450 in zone 4 has yielded the anterior portion of a tupaiid skull (Jacobs, 1979a) and actenodactylid molar. Our locality Y311, which we tentatively place in zone 5, yielded a microchiropteran, a soricid, a sciurid, a ctenodactylid, a new species of cricetid, a rhizomyid, a murid, and a grooved rodent incisor similar to that reported in Paraulocodus (but see Black, 1972). Locality Y182a in zone 6 has yielded a tupaiid, a soricid, a glirid, a sciurid, Kanisamys sivalensis, Progonomys debruijni, Karnimata darwini, Parapodemus sp., and Rhizomyoides cf. nagrii. At localities Y24 and Y34 in zone 9 we have recorded Kanisamys sp., Progonomys sp., and Karnimata sp. Locality Y369, which is in either zone 10 or 11, produced one Protach- yoryctes tatroti, while locality DP13, probably in zone 10, has a soricid, a leporid, Rhizomy- oides cf. nagrii, Mus auctor, Karnimata huxleyi, and Parapelomys robertsi. A Pleis- tocene locality in the Pabbi Hills, DP24, yielded a soricid, a modern Mus, cf. Rattus sp., and Golunda kelleri. We plan extensive sampling for small mam- mals and hope eventually to record small mammals from all zones in the Khaur and Chinji areas. Present collections suggest that the rodent faunas change radically at the 34 Miocene Sediments and Faunas of Pakistan Postilla 179 levels of zones 2, 6, and 10. Although the tempo and mode of change in the Siwalik small mammal faunas is still poorly resolved, the basically ctenodactylid-cricetid-rhyzomy- id faunas of zones 2 and 3 are clearly different from the murid-dominated faunas of younger zones. At least two groups of murids can be distinguished in these later faunas, and are also usually found with rhizomyids. Changes in the two major groups of murids apparently do not occur simultaneously nor do changes in the murids necessarily coincide with changes in the rhizomyids. 3. Carnivora and Creodonta (Barry) For the purposes of this discussion we have divided the Siwalik carnivores into three size groups: 1) the large carnivores being those over approximately 40 kg; 2) the small car- nivores those much less than 40 kg; and 3) an intermediate-sized group. This divi- sion is based on the preservational biasing effect of size (Behrensmeyer and Dechant, 1979) which is, we believe, reflected in the nature and detail of our knowledge of these forms. Large carnivores are fairly commonly found and we have relatively complete knowledge of them. For the stratigraphic levels most intensely collected we have probably recorded most of the large carnivore taxa, know something of their stratigraphic ranges, and may even be able to determine their rela- tive abundances. The large carnivores are, as a result, of some use in stratigraphic correla- tions and paleoecological reconstructions. The preservation and discovery of a small carnivore, on the other hand, has amuch greater element of chance and, although we have found a number of very interesting small species, our knowledge of this part of the fauna is very incomplete. Small carnivores are valuable ecological indicators, however. The third, intermediate-sized group has charac- teristics of both the others. The following lists of carnivores are based mostly on the GSP, YPM, and American Muse- um of Natural History collections. Additional taxa are known but are not listed because they either are of questionable status or our knowl- edge of their stratigraphic ranges is too imperfect. Large carnivores Hyaenodontidae Dissopsalis | zones2and3 carnifex Hyainailouros zones 1 and3 sulzeri Amphicyonidae Amphicyon sp. zones 2, 3, 6, 8 and 9 Vishnucyon zone2 chinyjiensis Agnotherium zone 2 or 3? n.sp.? Hyaenidae Percrocuta zones 2, 3, 4,5 carnifex and 6 Percrocuta zones 6, 7, 8 and grandis 9? Lycyaena zones 9 and 10? macrostoma Felidae Sansanosmilus zone 2 or 3 sp. Paramachaer- zone 6 odus sp. Megantereon zone9 sp. Other large carnivores are: Vinayakia sp., which is probably a hyaenodontid and not a felid; Arctamphicyon lydekkeri; Agriotherium palaeindicum; and Indarctos punjabiensis. Intermediate-sized carnivores These are estimated to be slightly smaller than 40 kg but are commonly enough found that we can probably determine their stratigraphic ranges. Collecting has not been intense enough at most levels to ensure that we have sampled adequately. 35 Miocene Sediments and Faunas of Pakistan Postilla 179 Hyaenodontidae Metapterodon? zones 2 and 5 n. sp. Mustelidae Plesiogulo crassa Eomellivora sp. zones 2 or 3 as well as 5 and 6 zone 10 Enhydriodon zone 11 sp. Hyaenidae Progenetta, zones 3 through 7 3 species Palhyaena zones 7, 8 and 9. sivalense May be suc- ceeded by Palhyaena indicum in zone 10? Felidae “Sivaelurus” zone3 chinjiensis Sivasmilus zone 2 or3 copei Small carnivores. At this stage it is difficult to say anything about the small carnivores other than to make a list and a few comments. Mustelidae Lutrinae: There are several otters present, the most common being Sivaonyx bathygna- thus in zone 6. Vishnuonyx chinjiensis: not an otter but rather a ferret badger, zone 3. “Martes” lydekkeri: more similar to Enhydricitis, zone 2. cf. /schyrictis: three species; one in zone 3? plus small species in zone 6 anda large species at Hari Talyangar. Viverridae Pilgrim’s “Lutrinae genus indet. hasnoti’ is a large, probab- ly new viverrid genus. We have excellent material from zone 5. “Viverra” chinjiensis: zone 3 and perhaps zone 5 ?Paradoxurinae: zone 5 ?Hemigalinae: zone 9 Herpestinae: two species from zones 3 and 6 Felidae small felid from Hari Talyangar Felis sp.: zone 10 Comments The carnivore fauna as a whole has anumber of interesting peculiarities. Among these are the absence of canids from the pre-Pinjor faunas and the apparent absence of anumber of mustelids character- istic of the Vallesian/Turolian faunas of Europe, Turkey, and China (i.e., Proputorius, Parataxidae, Melodon, Trocharion, Trochictis, Promeles, and Promephities). Combined with these intriguing absenses is the relatively late survival of anumber of archaic forms, princi- pally the amphicyonids and creodonts. The cooccurrence of hipparions and acreodont at about 10 m.y. is of particular interest. Finally we have found no evidence of any procyonids and are in agreement with Tattersall (1968) that the Siwalik species of Sivanasua are referable to the Primates (cf. /ndraloris). 4. Pholidota (Pickford) A single specimen of a small pholidote is known from zone 9. This specimen, a medial phalanx, is referred to Manis sp. (Pickford, 1976a) and is the earliest record for the pan- golin in the Indian subcontinent. 5. Tubulidentata (Pickford) Aardvarks also comprise a very rare element of the Siwalik fauna. Pickford (1978) now recognizes only one large species instead of two (Colbert, 1933, 1935a, and Lewis, 1938). Remains of this species have been found in zones 2, 3, 5, and 79. In addition we have found remains of a smaller, undescribed species from zones 1 and 2 (Pickford, 1978). 6. Proboscidea (Tassy) Trilophodont Gomphotheriinae (two taxa) and Amebelodontinae (at least one taxon) are present in the Chinji Fm. Choerolophodon- tinae (one genus: Choerolophodon = Syncono- 36 Miocene Sediments and Faunas of Pakistan Postilla 179 lophus) specimens, which are rare, are somewhat comparable to the Fort Ternan/ Ngorora species Ch. ngorora from East Africa. A small Choerolophodon species, which is similar to the Siwalik Ch. palaeindicus and the East African Ch. kisumuensis, may be repre- sented by a molar fragment from the Kamiial Fm. (?zone 1). In contrast to its occurrences in the Chinji Fm., Choerolophodon is common in the Nagri and Dhok Pathan Fms. with one species (Ch. corrugatus Pilgrim) being close to Ch. pentelici of the Vallesian/Turolian beds of the eastern Mediterranean basin. This species is known from zones 5 through 9. A part of amolar from the Chinji Fm. (zone 3) isa possible early Stegolophodon and could be related to a Stegolophodon sp. known from zones 6 and 7. This latter form is clearly a gomphotheriid close to the European Tetra/lo- phodon. A lower tusk with tubular dentine from zone 8 is referred to cf. Platybelodon. The cross-section of this tusk is different from PI. grangeri from Tung Gur and it could be related to the tusk of “Tetralophodon” grandincisivus from Maragha. The Maragha tusk is the type of the species and it has tubular dentine. 7. Equidae (MacFadden) As is true of most of the Siwalik taxa, increas- ingly detailed knowledge of the stratigraphic occurrence of these forms has led to an increasingly complex story of their phylogeny and paleoecological associations. We now believe that the Siwalik hipparions represent at least two genera and three species and their first appearance may well involve a dif- ferent taxon than is found in Europe (MacFadden and Bakr, 1979). The only cer- tainly identified species is Cormohipparion theobaldi which defines the base of zone 5 and may persist into.zone 11. The small species ““Hipparion”’ antilopinum is of uncer- tain generic assignment (contra Skinner and MacFadden, 1977), whereas a second small species appears to be a true Hipparion (MacFadden and Bakr, 1979). These last two are not easily separated on dental criteriaand it is not Clear which is present at any strati- graphic level. Small hipparions define the base of zone 4 (which may be equivalent to the Hipparion datum of Berggren and Van Couvering, 1974) and are abundantly present throughout the section to about the level of at least zones 9 and 10. Small and large hippar- ions are often associated at the same locali- ties. Equus has not been found in any of our sections. Once hipparions appear they rapidly become an abundant part of the large mammal fauna. 8. Chalicotheriidae (Pickford) Pickford (ms.) notes that the chalicothere material from the Siwaliks can be considered as belonging to a single, quite variable species, Chalicotherium salinum, from zones 2 through 9 or 10. It is possible that two species are being sampled. 9. Rhinocerotidae (Guérin) Rhinocerotids have been provisionally assigned by Guerin (in preparation) to four groups, with some specimens remaining indeterminate. The groups are: Gandaithe- rium browni and G. vidali (zones 3 through 8); Aceratherium sp. cf. A. simorrense (zone 6); Chilotherium intermedium (ones 5 through 9); and Brachypotherium perimense (zones 6 through 8). Relative abundances of the vari- ous species differ between zones, and this probably has some paleoecological signifi- cance. Zone 5 is noteworthy since rhinocero- tids are both abundant and diverse. 10. Suidae and Tayassuidae (Pickford) The eleven genera of Siwalik suids are readily divisible into three age groups (Pickford, ms.): an older group with Listriodon pentapotamiae, Conohyus sindiensis, and Hyotherium pilgrim! ranges from zones 1 through 3 (but not all occur in all zones). A younger group includes Tetraconodon magnus, Hippopotamodon sivalense, Lophochoerus nagrii, and Propota- mochoerus hysudricus from zones 4 through 9. Hippohyus sivalensis and H. lydekkeri, Sivachoerus prior, ?Sus, and Sivahyus pun- jabiensis appear or become abundant only in zone 10 and above. There is avery small suid of uncertain affinities from zone 8. 37 Miocene Sediments and Faunas of Pakistan Postilla 179 The peccaries Pecarichoerus orientalis and Schizochoerus gandakasensis are known from zone 2 and zones 5 and 8 respectively (Pickford, 1976b and in press). The last- Named species may also occur in zone 6. It remains to be seen how abruptly the transi- tions between suid assemblages occur, iden- tifications of specimens are now virtually Complete, making stratigraphic range data for this important group possible to compile. 11. Anthracotheriidae (Barry) Anthracotheres are fairly common but the fragmentary nature of our material makes Specific identifications difficult. Several Species are found from zones 2 through 11. 12. Hippopotamidae, Camelidae and Cervidae (Barry) The sole hippopotamus, Hexaprotodon sival- ENSis, is limited to the uppermost of our defined zones (zone 11). We have found remains of Hexaprotodon sivalensis at Tatrot and they are common in the younger parts of the Siwalik section. We have not found remains of either the Camelidae or the Cervi- dae (contra Brown, 1927 and Colbert, 1935a) 'N Our areas. 13. Tragulidae (Hamilton) he tragulids of the Siwaliks are represented by at least two and possibly three genera. Of the primitive Dorcabune, we have in our col- lections material of two species: the very large D. anthracotheroides, found so far only in Zones 2 and 3, and the somewhat smaller D. Nagrii from zone 6. Dorcatherium has four SPecies. Two of these are the very large D. Majus which ranges from zone 2 into zone 5, and the smaller D. minus which appears to be 'estricted to zones 2 and 3. A third and as yet Unnamed species is much smaller than D. Minus and appears to have a long geological history, while a fourth very small species, also unnamed, has been found in rocks from zone 8. This last species may be close to Tragulus Sivalensis. These tragulids, particularly the &rger ones, may be much more common than generally thought since fragmentary speci- mens are easily confused with both bovids and anthracotheres. 14. Giraffidae (Barry) The giraffe family holds considerable promise for biostratigraphy in the Siwalik but before this potential can be realized it will be neces- sary to develop taxonomic diagnoses based on the teeth and postcranials as well as on horncores. As the matter stands now, it is only possible to state a few generalities on the basis of the GSP and YPM collections and the published information in Colbert (1935a). The relatively small Giraffokeryx punjabiensis is restricted to zones 2 and 3 where it is a com- mon element in the fauna. All the larger giraffes are absent from these levels, first appearing (Bramatherium megacephalum) in zone 5. (The range of Giraffokeryx may extend upward and that of the larger taxon downward into zone 4. There may also be more than one species of Giraffokeryx present (Hussain et al., 1977).) The Sivatherium giganteum specimens collected by Barnum Brown came from the “upper Siwaliks” (Colbert, 1935a). It would appear that this species is restricted to our zone 11 or even younger sediments. Two other taxa, the “Giraffa” punjabiensis and Giraffa sivalensis of Colbert (1935a) are also known; the former from zone 10? and the latter from the Upper Siwaliks (zone 11 or young- er?). 15. Bovidae (Thomas) Lower and Middle Siwalik bovids can be sub- divided into three groups, with little strati- graphic overlap. The oldest group consists of Protragocerus gluten, Miotragocerus grad- iens, Kubanotragus sokolovi, Pseudotragus potwaricus, Sivoreas eremita, and a Gazella sp., and ranges through zones 1 through 3. A second group consists of Miotragocerus punjabicus, Selenoportax vexillarius, ?Pseudotragus sp., Elachistoceras khauris- tanensis, and a second Gazella sp., and is found in zones 4 through 8. The third and youngest of the lower and mid- dle Siwalik groups contains, among other forms, Kobus sp. and cf. Prostrepsiceros 38 Miocene Sediments and Faunas of Pakistan Postilla 179 houtumschindleri, and apparently comes from zone 9. Other bovids remain to be identified. Earlier studies by Pilgrim are currently under revision by Thomas (1977 and in preparation); when identification of bovid material is com- plete it will be possible to compile more com- plete stratigraphic range information. e. Nonmammal Groups 1. Aves Birds found by this expedition are the first to be reported from the lower and middle Siwaliks. Seven species have been identified from material collected prior to 1978 (Harrison and Walker, in preparation). These include a pelican Pelecanus cf. cautleyi from either zone 3 or 4, a stork Leptoptilos siwalikensis from zone 6, and a vulture Torgos cf. trachelio- tus from zone 3. Four new species are also being described by Harrison and Walker (in preparation), for the stork genus Xenorhyn- chus (from zone 9), the pheasant genus Lophura (zones 3, 5, and 6) and the rail genera Urmornis (zones 3 and 5) and Por- phyrio (zone 6). 2. Amphibia and Reptilia J.C. Rage has identified the following amphib- ia, lizards, and snakes from the Yale-GSP col- lections, expanding on previously known material (Hoffstetter, 1964): Anura, genus and species indeterminate Varanus sp. Acrochordus sp. Python sp. Colubridae, genus and species indeterminate The following turtles have been identified by F. de Broin: Geochelone (s.|.) sp. zone 5? Testudo sp. zones 5 and 6 Emydidae, genera indeterminate, zones, 5, 6, and 8. These terrapins represent three generic groups: Kachuga, Geomyda, and either the Batagur or Hardella groups. Trionyx (s.|.) Sp., Zones 2, 3, 5,6 and 8 ?Chitra sp., zones 2,3 and 6 Lissemys cf. punctata, zone 8 cf. Lissemys sp. 1, zones 5 and 6 cf. Lissemys sp. 2, zones 2, 3, 5and6 The many crocodilian teeth and postcranial fragments recovered to date and studied by E. Buffetaut establish the presence of croco- dilians at almost all stratigraphic levels in the Siwaliks but rarely allow the identification of species. However, sufficiently complete remains have been assigned to four species, two crocodiles and two gavials. These are Crocodylus palaeindicus from zone 6, ?Croc- odylus sivalensis from zone 1, Gavialis gan- geticus from either zone 6 or 7 and zone 9, and finally, Gavialis hysudricus from zone 5. f. Correlation and Intercontinental Connections Pilbeam et al. (1977a) briefly discussed cor- relation of the faunas from the Potwar Plateau with those elsewhere in Eurasia and Africa. Since then some further progress in our understanding has been made, although it should be clearly stated that until more identi- fications are completed and until we have a better understanding of stratigraphic ranges, reliable correlations will not be possible. At present several important groups (bovids, suids, rodents) give the impression that there are afew rather distinct faunal units separated by periods of rapid change, although the pic- ture from other groups (carnivores) suggests a series of more gradual and less clear-cut changes. Presumably the issue of tempo and mode of faunal change will be clarified in the future. Faunas of zones 1, 2, 3 show anumber of similarities to Astaracian faunas in Europe, Barstovian faunas in North America, and East African middle Miocene faunas. Thus, rodent faunas with cricetids are similar to those in Europe (Anwil in Switzerland, for example) and North America, dated at 11 to 15 m.y. Three or four species of bovid are identical to 39 Miocene Sediments and Faunas of Pakistan Postilla 179 those from Fort Ternan and Ngorora in Kenya, with an age range of 14 to around 10 m,y., while the suids resemble Astaracian forms. Similar ties are to be seen in some other groups (i.e., proboscideans) and it seems likely therefore that zones 1 through 3 fall with inthe 11 to 15m.y. span. At that time, faunal resemblances of Pakistan and other south Asian faunas were fairly general, showing ties to North America, west and central Eurasia, and Africa. Possibly we are witnessing a peri- Od of moderate endemism following the initial Mixing of African and northern faunas but before significant barriers to migration developed. Zone 4 is poorly sampled, as is zone 5, although zones 6, 7, and 8 are becoming well-documented. The south Asian faunas at this Stage show few similarities to those in Africa or North America (except for hipparions which originate in the New World and are first abundant in zone 4). There are some resem- blances to European and west Asian faunas, Particularly among rodents and suids. Corre- lations are to localities in Eurasia dated between 10 and 8m.y. The first appearance of Hipparion species in Europe is now thought to be at a little under 11 my., in East Africa at around 10 m.y.; if correlative with zone 4 this Matches the approximate age estimates for Our first hipparions. This relative faunal endemism is disturbed by a faunal change, involving both emigration and immigration, recorded in zone 9. Several Groups document changes, among them Primates, carnivores, suids, bovids, and rodents. Thus hominoids are apparently greatly diminished in diversity or disappear Entirely, to be replaced (apparently) by colo- bine monkeys; relatively more modern hye- Nids turn up, as do pigs of more modern aspect; a reduncine species similar to that from Mpesida and Lukeino in Kenya appears, @S does one found at Samos and Maragha; N€w rodents may indicate ties with Africa. here is evidence of interchange with North America, although at present most of the exchanges (emigrants as well as immigrants) seem to be with Africa. Correlations to dated localities suggest ages of perhaps 8 to6 m.y. for zone 9 and younger zones. As noted above, this very simple scheme may well need modifying in the direction of increased complexity, although we believe that the age estimates are unlikely to be signif- icantly in error. g- Paleoecology (Badgley, Behrensmeyer) Ideally, we would like to understand the eco- logical factors involved in both the succession of faunas through time and the composition of the fauna at any single level in the Siwalik section. Such questions concerning pale- oecology must be designed with the fossil record and its limitations in mind, since infor- mation about past communities of plants and animals differs substantially from information available for living communities. Recent studies in vertebrate paleoecology and taphonomy have attempted to define ques- tions that can be answered using the small and often highly biased subset of data avail- able in the fossil record. With regard to the 12 to 10 m.y. of faunal succession represented in the Siwalik section of the Potwar Plateau, we have concentrated our research on both the “vertical” (succes- sion through time) and the “horizontal” (single level) paleoecology. In the first case, our goal is to interpret broad faunal changes through time from an ecological perspective, consid- ering possible interactions of the faunas, their habitats, and climatic change. In the second case, the goal in studying horizontal pale- oecology is to define limits of variation (e.g., in the faunal composition of fossil assemblages) so that this can be distinguished from variation due to evolutionary change through the sec- tion. At present, the sources of data for these two goals are rather different. We shall summarize 40 Miocene Sediments and Faunas of Pakistan Postilla 179 the available information and preliminary con- clusions regarding them separately in the following paragraphs. Broad interpretations of paleoecology over the entire Siwalik section in the Potwar Plateau must be based on what we know of the tax- onomic composition of the fauna and the overall paleoenvironments since detailed lith- ofacies and taphonomic studies are presently available for only one level. Biostratigraphic reference sections for the vertebrate faunas are in apreliminary state for the Potwar Pla- teau as a whole, but superposition of fossil localities within each of the three major regions (Khaur, Chinji-Nagri, Hasnot) is clear. The interpretations of faunal patterns given below are based on this superpositional evi- dence plus inferred feeding and locomotor behaviors, the array of herbivore body sizes, and known habitats of modern relatives for mammals represented in the Middle Siwalik fossil assemblages. The Siwalik vertebrate fauna included few medium or large mammals that might have been exclusively forest species, although some of the small species probably were forest forms. There were also few species that appear to represent adaptations to open grassland. Rather, we seem to be dealing with an assemblage of forms sampled from a “mixed” woodland-grassland habitat. If this hypothesis is to be further justified, we must address several potential interpretive prob- lems. First, there are few analogues today of such mixed habitats that have been left undis- turbed yet still contain a good sample of their natural faunas. Second, many members of mixed habitat faunas of the middle and late Miocene may have been ecologically different from modern descendants or presumed anal- ogues. Third, mixed habitats are, because of their mosaic nature of “graininess,” likely to support a similarly mosaic fauna with more forest-adapted species living quite close to more open country forms. Fossil samples thus may easily represent many different micro- habitats. However, if particular depositional environments tend to sample particular micro- habitats, then we may be able to distinguish taxonomic groups representing different por- tions of the overall ecological mosaic through time. Until we have detailed lateral sampling at a number of levels, we shall not be able to resolve this problem. There is little detectable change in taxa regarded as paleoecological faunal indica- tors from zone 1 through zone 8. Prehipparion faunas contain more “primitive” elements, but these are not necessarily “forest” species. Leinders has argued (1976) that Listriodon, for example, was likely to have been an open woodland form. The appearance of hipparions is almost certainly due to their rapid spread throughout the Old World following immigra- tion from North America of already evolved lineages. Although not itself necessarily indi- cating habitat change, the introduction of the equids, as important new grazing elements, may have spurred some habitat shifts or may have initiated a “grazing succession” among the large herbivores. The faunal changes marking zone 9 do suggest an influx of some- what more open country species, perhaps recording a habitat change. Our second, horizontal approach to paleoe- cology has both geological and taphonomic components; most of the work aimed at detailed reconstruction of paleoenvironments and communities has focused on the sedi- ments and fossils of zones 5—6 (defined in this paper), in part because of the high density of fossil localities at this level and in part because most of the hominoids recovered by this expedition come from this level. Here, there has been an attempt to integrate the study of lateral lithofacies variation with the microstratigraphic and taphonomic work on individual fossil localities. The resulting pale- oecologic reconstructions can be done at various levels of resolution. At the scale of 30 km, the lateral lithofacies work establishes the overall physiographic character (an ‘‘aerial photograph view” of the Siwalik environment) with the inferred presence of two large river 41 Miocene Sediments and Faunas of Pakistan Postilla 179 Systems whose courses were not stable through time, well-drained floodplains and laterally shifting sedimentary environments which probably caused a constant renewal of vegetational habitats in these areas. At the other end of the scale, with resolution on the Order of 10's of meters, the microstratigraphic Study indicates fluvial and postdepositional Processes including bioturbation by plants and animals that may have acted over as little as a year’s time. These geological and tapho- nomical analyses are not completed, but we advance preliminary interpretations below (Badgley and Behrensmeyer, ms.). Fossil localities in zones 5—6 are dominated by medium-sized mammals within the size range of 20-200 kg. The lack of abundant Smaller mammals is probably due to tapho- nomic size bias at every stage of preserva- tional process. The lower representation of the larger mammals is probably due both to eco- logical (lower density and lower turnover rates) and preservational factors. For species within the “common‘ size range, it may be possible to estimate relative abundances; analysis awaits completion of taxonomic identifica- tions. The fossil assemblages do not appear to indicate differences in fine-scale habitats and habitat associations per se: that is, spe- Cies with more open-country adaptations, Such as the hipparions, cooccur with brows- INg herbivores. Perhaps these cooccurrences reflect an original patchiness of canopied and Open areas. Certain bone concentrations appear to be predator accumulations or sites Of repeated predation, including one of the Spectacular hominoid localities (Locality 260). The size structure of the herbivore trophic level, with the predominance of small to med- 'um-sized ungulates, indicates an abundance of browse and suggests a vegetational mosa- \C of different successional stages, including Qrassland, bush, and woodland; this view based on the herbivores is harmonious with Paleoenvironmental reconstruction based on the fluvial sedimentary regime. In Summary, our present view of Siwalik pale- 0€cology is broader than many previous inter- pretations in that we see evidence in both faunas and geology for a habitat mosaic including everything from forest and wood- land to open grassland. The scale or “graini- ness” of this mosaic was such that faunal elements were consistently sampled from dif- ferent parts of it to make up the fossil assem- blages found in typical channel (i.e., trans- ported) contexts. Thus, the habitat “grain” was probably on the order of tens to hundreds of meters rather than kilometers. The degree of mixing of different faunal ““subcommunities” may have changed through time but in order to see this, lateral analyses of several levels are needed throughout the section. For zones 5 and 6, our present evidence allows us to define a consistently cooccurring group of taxa that made up the overall large mammal “paleocommunity” of the fluvial system. This was dominated by Hipparion and Miotrago- ceras and included other bovids, probos- cideans, suids, giraffes, and rhinos as the most common major taxa. Less common taxa, including anthracotheres, tragulids, carni- vores, primates, and chalicotheres, were either absolutely less abundant or had differ- ent, more restricted patterns of occurrence indicating that they represent partially distinct subcommunities. This interesting possibility can only be resolved through further sampling and taphonomical analysis. 4. Summary and Conclusion (Pilbeam) The classic Siwalik faunas from the Potwar Plateau of Pakistan come from fluvial sedi- ments in three major areas, separated from each other by significant distances. These areas are around Chinji and Sethi Nagri, type areas of the Chinji and Nagri Fms.; Dhok Pathan to the north, where the Dhok Pathan Fm. lies above Nagri equivalent rocks; and to the east around Tatrot and Hasnot where the bulk of the Middle Siwalik faunas come from the Bhandar Beds which are probably younger than those of the Dhok Pathan Fm. 42 Miocene Sediments and Faunas of Pakistan Postilla 179 This project began in 1973 with work at Chinji and at Dhok Pathan. Since then we have greatly expanded our geographical range. Our initial aims were to get a broad overview of regional geography and stratigraphy; locate and, where possible, recollect earlier locali- ties; and place them stratigraphically as best as possible. This we have achieved reason- ably well in the Chinji and Dhok Pathan areas, and with less success around Hasnot. New collections, now yielding well over 10,000 specimens from more than 400 localities have been made in all major areas, but particularly around Khaur. However, it is around Khaur that we have con- centrated and will continue to concentrate our efforts, in the belief that a detailed sequence of rocks and faunas is best defined in one area, and then extended to others. The progress we have made towards defining litho-, magneto-, bio-, and chronostrati- graphic units in the Khaur area, and extending them elsewhere, is summarized in sections 2 and 3 and in Figure 14. We still have a great deal of work to do, but a clearer picture is beginning to emerge of an evolving series of interlocking fluvial systems close to the emergent Himalayas, covered with a mosaic of mostly woodland-type habitats. Faunas included a wide variety of forms, and among the browsing element there were probably several species of hominoid primates. These primates were of diverse ecological charac- ter; some at least were relatively terrestrial forms, perhaps spending a good deal of their time feeding on the ground. They document a significant shift of hominoids from a mainly forest, arboreal style of life to one in more mixed open and forest conditions. Included among them may well be the earliest recog- nizable human ancestor, or a species closely related to that ancestor. The Siwalik rocks of Pakistan are one of the few places where these earliest traces of the process of “becoming human” can be documented. Contributors David R. Pilbeam, Department of Anthropology and Department of Geology and Geophysics and Peabody Museum of Natural History, Yale University, New Haven, Connecticut. A. Kay Behrensmeyer, as above John C. Barry, as above S.M. Ibrahim Shah, Geological Survey of Pakistan, Quetta Marc Monaghan, Department of Geology and Geophysics, Yale University Lisa Tauxe, Lamont-Doherty Geological Observatories, Palisades, New Jersey Catharine Badgley, Department of Biology, Yale University Everett H. Lindsay, Department of Geosciences, University of Arizona, Tucson Michael D. Rose, Section of Gross Anatomy, Department of Surgery, Yale University School of Medicine Louis L. Jacobs, Museum of Northern Arizona, Flagstaff, Arizona Martin Pickford, TILLMIAP, Nairobi, Kenya Pascal Tassy, Laboratoire de Paléontologie de Vertébrés et de Paléontologie Humaine, Université Paris VI Bruce MacFadden, Florida State Museum, Gainesville, Florida Claude Guerin, Departement des Sciences de la Terre, Université Claude Bernard, Lyon Roger Hamilton, British Museum (Natural History), London Cyril Walker, as above Colin Harrison, British Museum (Natural History), Tring, Hertfordshire Herbert Thomas, Institut de Paléontologie, Musée d'Histoire Naturelle, Paris J.C. Rage, as above F. de Broin, as above E. Buffetaut, as above 43 Miocene Sediments and Postilla 179 Faunas of Pakistan Literature Cited olay R.V.V. 1927. Tertiary stratigraphy and orogeny of the northern Punjab. Geol. Soc. Am. Bull., 38: 65-725. Barndt, J. 1977.A magnetic polarity stratigraphy of the type locality of the Dhok Pathan Faunal Stage, Potwar Plateau, Pakistan. M.A. thesis, Dartmouth College. 105 pp. Barndt, J., N.M. Johnson, G.D. Johnson, N.D. Opdyke, E.H. Lindsay, D.R. Pilbeam, and R.A.H. Tahirkheli. 1978. The magnetic polarity stratigraphy and age of the Siwalik Group near Dhok Pathan Village, Potwar Plateau, Pakistan. Earth and Planet Sci. Lett., 41: 355-364. Behrensmeyer, A.K. and D.E. Dechant. 1979. The recent bones of Amboseli National Park, Kenya, in relation to East African paleoecology. in A.K. Behrensmeyer and A. P. Hill [eds.] Fossils in the making. Univ. Chicago, Chicago. Berggren, W.A. and J. Van Couvering. 1974. The late Neogene: biostratigraphy, geochronology and Paleoclimatology of the last 15 million years in marine and continental sequences. Paleogeogr., Paleocli- Matol., Paleoecol., 16: 1-216. Black, C.C. 1972. Review of fossil rodents from the Neogene Siwalik beds of India and Pakistan. Palaeon- tology, 15: 238-266. Brown, B. 1927. A new deer from the Siwaliks. Am. Mus. Novitates, No. 242, 6 pp. Colbert, E.H. 1933. The presence of Tubulidentates in the Middle Siwalik Beds of Northern India, Am. Mus. Novitates, No. 604, 10 pp. ~ 1935a. Siwalik mammals in the American Museum of Natural History. Trans. Am. Phil. Soc. N.S., 26: 401. ~~ 1935b. Distributional and phylogenetic studies on Indian fossil mammals, |, American Museum Collecting localities in Northern India, Am. Mus. Novitates, No. 796, 20 pp. Cotter, G. de P. 1933. The geology of the part of the Attock District west of longitude 72° 45’ E. Geol. Surv. India Mem., 55: 63-161. Fatmi, A.N. 1973. Lithostratigraphic units of the Kohat-Potwar Province, Indus Basin, Pakistan. Geol. Surv Pakistan Mem. 10: 80 pp. Gill, W. D. 1951. The stratigraphy of Siwalik series in the northern Potwar, Punjab, Pakistan. Quart. J. Geol. Soc. (London) 107: 375-394. Hedberg, H. D. (ed.) 1976. International stratigraphic guide. John Wiley and Sons, New York, 200 pp. Hoffstetter, R. 1964. Les serpents du Néogéne du Pakistan (couches des Siwaliks). Bull. Soc. Géol. France, 7em série, 6, 467-474. Hussain, S.T., R.M. West, J. Munthe, and J.R. Lukacs 1977. The Daud Khel Local Fauna: a preliminary report on a Neogene vertebrate assemblage from the Trans-Indus Siwaliks, Pakistan. Contrib. Biol. Geol. Milwaukee Public Mus. No. 16, 17 pp. Irving, E. 1964. Paleomagnetism and its application to geological and geophysical problems. John Wiley and Sons, New York Jacobs, L.L. 1977a. Small mammal fossils from Neogene Siwalik deposits, Pakistan. Ph.D. dissertation, University of Arizona. ~—1977b. A new genus of murid rodent from the Miocene of Pakistan and comments on the origin of the Muridae. Paleo Bios, No. 25. ~~1979a. Tooth cusp homology of murid rodents based on Miocene fossils of Pakistan. Cas. Miner. Geol., Praha, 24. ~——— 1979b. Siwalik fossil tree shrews. /n W. Luckett [ed.] Comparative biology and evolutionary relation- Ships of tupaiids. Plenum, New York. -1979c. Fossil rodents (Rhizomyidae and Muridae) from Neogene Siwalik deposits, Pakistan. Bull. Mus. North Ariz. James, G.T. 1963. Paleontology and nonmarine stratigraphy of the Cuyama Valley Badlands, California. Calif. Univ. Publ. Geol. Sci. 45: 1-145. LaBrecque, J.L., D.V. Kent, and S.C. Conde 1977. Revised magnetic polarity time scale for the late Cretaceous and Cenozoic. Geology 5: 330-335. Leinders, J.J.M. 1976. A functional interpretation of the dental morphology of Listriodon (Suidae, Artio- dactyla). Proc. Kon. Ned. Akad. Weten. Ser. B., 81: 61-69. Miocene Sediments and Postilla 179 Faunas of Pakistan Lewis, G.E. 1937. A new Siwalik correlation (India). Am. J. Sci. Ser. 5, 33: 191-204. ——— 1938. A skull of Orycteropus pilgrimi. Am. J. Sci. 236: 401-405. Lindsay, E.H. 1972. Small mammals from the Barstow Formation, California. Calif. Univ. Publ. Geol. Sci. 93: 1-104. MacFadden, B. J. and A. Bakr. 1979. The horse Cormohipparion theobaldi from the Neogene of Pakistan, with comments on Siwalik hipparions. Palaeontology 22: 439-447. Murphy, M.A. 1977. On time-stratigraphic units. J. Paleontol. 51, 213-219. Pickford, M.H.L. 1976a. An Upper Miocene Pholidote from Pakistan. Pakistan J. Zool. 8: 21-24. ———-1976b. Anew species of Taucanamo (Mammalia: Artiodactyla: Tayassuidae) from the Siwaliks of the Potwar Plateau, Pakistan. Pakistan J. Zool. 8: 13-20. ——— 1978. New evidence concerning the fossil Aardvarks (Mammalia, Tubulidenta) of Pakistan. Tertiary Res. 2: 39-44. Pilbeam, D.R. 1978a. Rearranging our family tree. Human Nature, 1 (6): 38-45. ———1978b. Rethinking human origins. Discovery (Peabody Mus. Nat. Hist., Yale Univ.) 13: 2-9. — 1979 (in press). Major trends in human evolution. /n L.K. Konigsson [ed.] Current argument on early man. Pergamon, Elmsford, New York. Pilbeam, D., J. Barry, G.E. Meyer, S.M.1. Shah, M.H.L. Pickford, W.W. Bishop, H. Thomas, and L.L. Jacobs. 1977a. Geology and paleontology of Neogene strata of Pakistan. Nature 270: 684-689. Pilbeam, D., G.E. Meyer, C. Badgley, M.D. Rose, M.H.L. Pickford, A.K. Behrensmeyer, and S.M. 1. Shah. 1977b. New hominoid primates from the Siwaliks of Pakistan and their bearing on hominoid evolution. Nature 270: 689-695. Pilgrim, G.E. 1910. Preliminary notes on a revised classification of the Tertiary fresh-water deposits of India. Geol. Surv. India Rec. 40: 185-205. 1913.The correlation of the Siwaliks with mammal horizons of Europe. Geol. Surv. India Rec. 43: 264-326. Schumn,, S.A. 1977. The fluvial system. John Wiley and Sons, New York, 338 pp. Skinner, M.F. and B.J. MacFadden. 1977. Cormohipparion n. gen. (Mammalia, Equidae) from the North American Miocene (Barstovian-Clarendonian). J. Paleontol. 51: 912-926. Tattersall, |. 1968. A mandible of Indraloris (Primates, Lorisidae) from the Miocene of India. Postilla (Peabody Mus. Nat. Hist., Yale Univ.) No. 123, 10 pp. Tauxe, L. 1978. A comparison of isochrons defined by short-term paleomagnetic events and lithostrati- graphic units in Siwalik rocks of the Potwar Plateau, Punjab Province, Pakistan. Yale College Scholar of the House thesis. Tedford, R.H. 1970. Principles and practices of mammalian geochronology in North America. Proc. N. Am. Paleontol. Conv. Part F, 666-703. Thomas, H. 1977. Un nouveau Bovideé dans les couches a Hominoidea du Nagri (Siwalik moyens, Miocene superieur), Plateau du Potwar, Pakistan: Elachistoceras khauristanensis gen. et sp. nov. (Bovidae, Artio- dactyla, Mammalia). Bull. Soc. Geol. France Ser. 7, 19: 375-383. Woodburne, M. O. 1977. Definition and characterization in mammalian chronostratigraphy. J. Paleontol. 51: 220-234. 45 Miocene Sediments and Postilla 179 Faunas of Pakistan Unpublished Manuscripts, Reports and Other Documents Badgley, C. and A.K. Behrensmeyer, ms., Paleoecology of Middle Siwalik Sediments and Faunas, Northern Pakistan. Barry, J.C. n.d., Electronic data processing programs and guides to their use and recording data are available from the authors of this report. Guerin, C., ms., Etude préliminaire des restes de rhinocéras recueillis au Pakistan dans la région de Khaur (Plateau du Potwar, Séries des Siwaliks). ot C.J.O. and C.A. Walker, ms., Fossil Birds from the Middle and Upper Miocene of Northern akistan. Pickford, M., ms. Miocene Chalicotheriidae of the Potwar Plateau, Pakistan. ~~—ms., The Miocene Suidae of the Indian Subcontinent. Thomas, H. ms., Le Probleme du Groupe Se/enoportax-Pachyportax (Boselaphini, Bovidae). ~——ms., A Propos de la Présence de Cervidae et d'une Antilope Nouvelle dans la Formation de Chinji. TE diea Etude du Groupe Miotragocerus (= Tragocerus)-Tragoportax (Boselaphini, Bovidae), dans les Iwaliks. The Editors David R. Pilbeam, A.K. Behrensmeyer, and John C. Barry. Peabody Museum of Natural History and Departments of Anthropology, and Geology and Geophysics, Yale University, New Haven, Connecticut 06520. S.M. Ibrahim Shah. Geological Survey of Pakistan, Quetta, Pakistan.