Peabody Museum

of Natural History

Yale University

New Haven, CT 06511

Postilla Number 183

Part 1
30 November 1980

A Geologic and Biostratigraphic
Framework for Miocene Sediments
near Khaur Village,

Northern Pakistan

John C. Barry
Anna K. Behrensmeyer
Marc Monaghan

Received 3 March 1980
Abstract

A litho- and chronostratigraphic framework for
approximately 3500 m of late Miocene
sedimentary deposits with lateral exposure of
30-40 km is now well established for Siwalik
deposits in the vicinity of Khaur Village, north-
ern Pakistan. Based on paleomagnetic rever-
sal stratigraphy and the constraints of long
range faunal correlations, the time spanned is
between 15-12 and 6 my BP. Boundaries of
the major lithostratigraphic units are time-
transgressive with respect to a paleomag-
netic isochron traced laterally for 20-30 km.
With reference to this datum, the zone of inter-
fingering of the Dhok Pathan and Nagri For-
mations moves upward through approxi-
mately 1.5 my from east to west in the Khaur
area. The major rock units represent adjacent
alluvial deposits of two or more major fluvial
systems with different characteristics of fossil
preservation. Nearly 350 fossil-collecting
localities can be placed in this litho- and
Chronostratigraphic framework. These
localities span the entire 6-9 my interval, al-
though the lower half of the section is only
poorly fossiliferous. The sequence of faunas
encompasses Pilgrim's Kamlial through Dhok
Pathan faunal zones, but with this new
Chronostratigraphic framework it is now ap-

© Copyright 1980 by the Peabody Museum of
Natural History, Yale University. All rights re-
served. No part of this publication, except
brief quotations for scholarly purposes, may
be reproduced without the written permission
of the Director, Peabody Museum of Natural
History.

parent that the classic faunal zones must
be superseded by a biostratigraphic zonal
scheme.

Key words: Siwaliks, stratigraphy,,Potwar
Plateau, upper Miocene, Hominoidae.

~ 4 2 #NOA
ag Ee a
we ea 7

Introduction

The geology and stratigraphy of fossiliferous
sediments on the northern limb of the Soan
Synclinorium have been summarized in pre-
vious publications dealing with the geology
and paleontology of the Siwalik formations of
the Potwar Plateau as a whole (e.g., Pilbeam
etal., 1977; Pilbeam etal., 1979). In this paper
we present a comprehensive and updated
litho- and biostratigraphic framework for Mid-
dle Siwalik deposits found in the vicinity of
Khaur Village (Fig. 1), where field research of
the joint Yale Peabody Museum-Geologic
Survey of Pakistan (Yale-GSP) has concen-
trated since 1973. The Khaur area provides a
continuous stratigraphic section spanning
over 5000 m and representing the time period
between 12+ and 6 my BP. The section can
be traced for tens of kilometers laterally and is
fossiliferous throughout, with a diverse mam-
malian fauna which includes hominoid pri-
mates. This combination of features provides
a unique opportunity to study the evolution of
terrestrial environments and faunas during a
critical period of Cenozoic history. The most
basic requirement for such goals is an under-
standing of the litho- and chronostratigraphy
of the sedimentary rocks. Through the efforts
of many members of the Yale-GSP team, the
litho- and chronostratigraphy of a significant
portion of the Khaur sequence are now well

Siwalik Stratigraphic Framework

Postilla 183

INDIA

Fig. 1
Drainage map of the Khaur area, from Dhok Pathan
(west) to Kaulial Kas (east), showing the major
canyons or kas’ and the positions of long strati-
graphic sections (straight solid lines). The outcrop
trace of the U sandstone 's indicated by a heavy
black line. Inset shows position of the Potwar
Plateau study area in northern Pakistan. Abbrevia-
tions for the stratigraphic sections’ locations, from
west to east: DP=Dhok Pathan, CH=Choutriwali
Kas, DM=Dhok Mila Kas, HL=Hasal Kas,
DK=Dinga Kas, JB=Jabbi, MK=Malhuwala Kas,
GK =Gandakas, BW=Bhagwa Kas, RK=Ratha
Kas, KL=Kaulial Kas.

Postilla 183

Siwalik Stratigraphic Framework

5 km.

4 Siwalik Stratigraphic Framework

Postilla 183

documented and are presented here as a
basis for further research.

We present the following information in two
parts; the first summarizes physical geologi-
cal and chronological evidence and the sec-
ond discusses faunal evidence in relation to
the overall stratigraphic framework.

The Geological and Chronostratigraphic
Evidence
The physiography of the Khaur area is charac-
terized by canyons or “kas” (plural, kas’)
which cut through dipping strata perpendicu-
lar to strike and are spaced at intervals of
about 3 km along strike. The measuring of
long stratigraphic columns in these kas’ is
thus a relatively straightforward initial ap-
proach to the geology. Sections have been
measured using both tape-and-compass and
Jacob's staff techniques. Frequent changes
in dip and poor exposure, particularly in the
lower portions of the sections, limit the accu-
racy of these methods. Thicknesses pre-
sented here are thus subject to possible error
which is probably not of great significance
with respect to the total sedimentary column.
Since the Khaur area sediments are ex-
tremely variable laterally, correlations be-
tween sections, measured at approximately 3
to 6 km intervals, cannot be done by simple
matching of vertical patterns in the stratig-
raphy. Lateral tracing of anumber of marker
horizons is necessary to establish the litho-
stratigraphic correlations between sections.
In the lower parts of many sections, divides
between the kas’ are mantled with late Pleis-
tocene to subrecent Potwar Silts, and marker
horizons cannot be traced with certainty. To-
ward the tops of the sections, the Soan River
meanders across strata, preventing correla-
tion of marker units with areas to the south and
west (Fig. 1). Inthe upper middle portion of the
Khaur sequence, however, there are nearly
continuous exposures that can be walked out
laterally for some 40 km, and these have pro-
vided critical horizontal links between all
measured sections. Within this belt, a marker
unit referred to as the ‘‘U" sandstone is the
most useful traceable datum.

The large-scale stratigraphic evidence
consists of 6 long measured sections and
five shorter ones correlated using the U
sandstone and two less extensive units.
These provide the lithostratigraphic
framework for the Khaur area which is pre-
sented graphically in Figure 2. We regard the
longest measured section, 3240 m in Kaulial
Kas, as the best reference section for facies
and faunal history of the Lower and Middle
Siwalik deposits of the Khaur area. This sec-
tion has also provided the most complete
paleomagnetic sampling because it is gener-
ally finer-grained than sections to the west.

The lithologically correlated sections can
provide only rough estimates of the relative
time relationships between facies changes or
faunal events in the different kas’ of the Khaur
area. Correlations to more distant Siwalik de-
posits using lithofacies corresponding to for-
mations such as the Chinji or Nagri are far
more subject to possible time differences. The
strong tendency of previous workers to think
of the lithostratigraphic entities (Table 1) as
time-specific has led to problems of interpre-
tation and nomenclature which have yet to be
completely unraveled (Pilbeam et al., 1979).

Until the development and application of
paleomagnetic reversal stratigraphy there
was no way to measure time vertically or trace
it laterally through the Siwalik sediments of the
Potwar, other than through use of the verte-
brate faunas. Volcanic ash horizons are very
rare and when they occur are laterally discon-
tinuous. Vertebrate evidence has been suffi-
cient to indicate that some lithostratigraphic
units were possibly time-transgressive, but it
has not provided any firm biostratigraphic
marker horizons that could be judged as
time-specific events rather than ecological or
sampling phenomena.

The obvious need for a chronostratigraphy
which is independent of faunas and litho-
units, at least in a local sense, has led to
intensive sampling and resampling of the
paleomagnetic reversal patterns in the Siwalik
deposits, particularly in parts of the section
yielding large faunal collections. The work of
J. Barndt (1977), and more recently L. Tauxe
(1978), in conjunction with Dartmouth Col-

Table 1. Siwalik Formations represented in the Khaur area, with local characteristics of the Middle
Siwalik formations. Thicknesses vary as indicated in Figure 2; those given here are based primarily on

the measured section in Kaulial Kas. Further details concerning lithofacies are given in Pilbeam et al.,
WE
Ww
=
©
os
ep)
Age Group Formation Thickness and Lithology in Khaur Area S
&
®
xe)
=
oS
Pliocene to Upper Soan Not yet studied; composed of gravelly sands mn
early Pleistocene Siwalik and brown silts according to Shah (1977). o
a
5
Dhok Pathan 1600 m (5200'); predominantly red-brown =
silts and clays with thinner sand units
(tabular and lenticular): lower 340 m (1100')
include both blue-gray and buff sands, middle
400 m (1300') buff to gray sands and upper &
Upper Miocene Middle 360 m (2800') gray sands with coarse gravel. a
Siwalik =
ra)
Nagri 1300 m (4200'); defined locally as having o
units of blue-gray sand more than 33 m (100')
in thickness, interbedded with thinner units
of variegated silts and clays with occasional
lenses of buff sand.
Chinji Dominantly red mudstones interbedded with
lenticular blue-gray sands; not yet studied
in detail.
Middle to late Lower
Miocene Siwalik
Kamlial Sands and gravels with interbedded red and
purple mudstones; not yet studied.

6 Siwalik Stratigraphic Framework

Postilla 183

lege, the Lamont-Doherty Geological Obser-
vatory and the University of Arizona, has es-
tablished a paleomagnetic reversal stratig-
raphy for the upper 3000 m of the Siwalik
section in the Khaur area. Portions of the re-
versal pattern are duplicated consistently in
seven laterally correlated sections of different
kas’, helping to confirm the reliability of the
data (Tauxe, 1979). In addition, Tauxe has
traced one normal to reversed transition
below the U sandstone laterally over a
distance of about 30 km and the resulting
isochron (referred to as the sub-U isochron)
closely parallels the upper boundary of

the sandstone (Behrensmeyer and Tauxe, in
preparation). Thus we have a horizontal time
link through the stratigraphic sections as well
as an overall vertical reversal pattern. The
long, composite column of Barnat and Tauxe
can be correlated with the current global re-
versal pattern for the late Miocene as shown in
Figure 2. The approximate age of the sedi-
ments based on faunal correlations with
radiometrically dated assemblages in Europe
was previously stated as between 13 and 8
my, with the bulk of the Khaur area primates
and associated fauna estimated at around 9
my (Pilbeam et al., 1977). The revised in-
terpretations of age by Tauxe (1979) show
that a more probable date for this fauna is
about 8 my. Interestingly, this also indicates
an error of at least 1 my in previous attempts at
absolute dating based on long-range faunal
correlations.

We feel confident that the stratigraphic and
chronological framework shown in Figure 2 is
now well established and can serve as a reli-
able basis for analysis of facies and faunas in
the upper half of the Khaur area deposits and
for correlation with other Siwalik sequences in
Pakistan and India. Further work in the Khaur
area is planned to extend the paleomag-
netic stratigraphy through the lower part of the
section.

Major Sedimentary Units

Stratigraphic columns measured in the Khaur
area typically begin in sediments charac-
terized by 1) interbedded red mudstones with

discontinuous blue-gray sands and continue
upward through 2) thick, massive sand-
stones; they are followed by 3) interbedded
sandstones and variegated silts and clays,
and (if there are sufficient exposures) end in 4)
thick red-orange silts with cobble-bearing
sandstone lenses. For anyone familiar with
Siwalik formations, as defined by Fatmi
(1973), there is little difficulty in assigning 1) to
the Chinji Formation, 2) to the Nagri Forma-
tion, 3) and possibly 4) to the Dhok Pathan
Formation, where the distinguishing lithologi
cal features of the units are best expressed
(Table 1). However, difficulties are encoun-
tered in recognizing boundaries between the
units because of their interfingering relation-
ships. Ithas proved impossible to map recog-
nizable boundaries in the Khaur area based
on Fatmi's definitions of the formations

Originally it was thought important to estab-
lish lithological boundaries as reference
points for the fossil collections, especially
when these boundaries were supposed to be
more or less isochronous. Now, however, the
paleomagnetic reversal patterns and particu-
larly the tracing of the sub-U isochron show
that these lithological boundaries are not only
complex in detail but also strongly time-
transgressive over distances of 30-40 km,
with time differences on the order of 1.0-1.5
my (Fig. 2). The major units of the Khaur area

Fig. 2

Lithostratigraphic framework for Lower and Middle
Siwalik deposits in the Khaur area. Vertical lines
represent the measured sections, with dotted ex-
tensions showing where strata continue beyond
what has been measured. The U sandstone, which
closely parallels a paleomagnetic isochron, is used
as a horizontal datum to align the sections, along
with two less extensive horizontal marker units. The
boundaries of the Nagri Formation are defined as
the points where individual sandstone units exceed
33 m in thickness in each long section. The com-
posite paleomagnetic reversal stratigraphy (Tauxe.
personal communication) for the sequence is given
to the right of the KL column, with normal zones
shown in black, epoch numbers to the left of the
column. Abbreviations for sections given in caption
to Figure 1. See also Figure 2 in Gill (1951).

DP CH HL MK RK KE

10.0

Primarily fine-grained with:

500 [

km. “PATHAN

on |

GRAY BUFF SANDS WITH
GRAVELS

BUFF SANDS

| INTERBEDDED BLUE-GRAY
AND BUFF SANDS

yOMEWe! DYydesHi}eNS yes

cC8l BIINSOd

8 Siwalik Stratigraphic Framework

Postilla 183

actually appear to represent differing broad-
scale depositional regimes, some of which
were contemporaneous with one another.

The major sedimentary units can be divided
into anumber of distinctive smaller scale
lithofacies which bear a specific relationship
to particular depositional regimes (e.g.,
channels, floodplains). Three important
small-scale lithofacies, the blue-gray sand,
buff sand and silt/clay facies, have been pre-
sented by Pilbeam et al., 1979. Further work
on these and additional lithofacies is under-
way, but details are beyond the scope of this
paper. The three lithofacies mentioned above
have proven to be consistently recognizable
in the field, and distinguishing features of the
Nagri and Dhok Pathan Formations in the
Khaur area can be described using the rela-
tive proportions of the component facies in
each (Table 1). These small-scale lithofacies
clearly represent depositional environments
and conditions that were not specific to any
particular time period.

The blue-gray sand and the buff sand
facies are strikingly different in composition
and in the shape of their sedimentary units.
Preliminary petrographic results show a
higher proportion of rock fragments in the
blue-gray sands. This supports other evi-
dence (Pilbeam et al., 1979) for the less ma-
ture nature of the blue-gray sands. We cur-
rently favor the hypothesis that the two facies
represent two separate fluvial systems with
differing drainage basins. Blue-gray sands
bear evidence of the dominance of physical
erosion while buff sands indicate greater in-
fluence of chemical weathering. The former
are probably derived from a river system
draining tectonically elevated regions within
the Himalayas, while the latter represent rivers
draining the foothills of the mountain front. The
buff sands are more intimately related to the
silt/clay facies, in terms of intergradational
contacts and lateral interfingering, than are
the blue-gray sands. Thus it appears that the
buff-sand system deposited more of the silt/
clay facies in the Khaur area than did the
blue-gray system.

“Cycles” that have been noted in previous
publications (Pilbeam et al. 1977, 1979) con-

sist of alternations of one or the other of the
sandstone facies with thicker units of the silt
Clay facies. The cycles change laterally along
with overall trends in the major lithofacies, with
blue-gray sand units thickening westward
and silt/clay units thickening eastward. The
zone of the most characteristic cycles, with
2-10 m thick blue-gray sand units separated
by 30-50 m of silt/clay and buff sand facies, is
about 500 m thick and is time-transgressive
toward the west (Fig. 2). In this zone the alter-
nating dominance of the blue-gray and silt
clay plus buff sand facies appears to reflect
cyclic processes intrinsic to one or both fluvial
regimes although there may be broader, ex-
trinsic tectonic or climatic influences on these
systems which we do not yet comprehend. It
may be possible to separate the lithological
record of extrinsic processes from intrinsic
ones as the latter become better understood,
and work is continuing with this goal in mind

Paleogeography and Sedimentary History

Three lines of evidence may be used to re-
construct the large-scale paleogeography of
the Khaur area: 1) the lateral variation of major
rock units parallel to the sub-U isochron,
along the generally east-west strike of the out-
crops, 2) sedimentological characteristics of
small-scale lithofacies and larger-scale units
which indicate source area and depositional
environment, and 3) current directions of
channel sandstone bodies as determined
from axes of trough cross-beds and linear
directions of the sand bodies themselves

At present we can discuss in detail only the
paleogeography of the Nagri and Dhok
Pathan Formations in the Khaur area. How-
ever, the relationships of these formations !n-
dicate what might be expected for the lower
part of the section as wellas for other areas Of
the Potwar Plateau.

Along the sub-U isochron blue-gray sands
predominate in the western part of the Khaur
area and interfinger with the silt/clay facies
and buff sands which predominate in the east.
Lenses of buff sand occasionally are in direct
erosional contact with underlying blue-gray
sands. The thickness and lateral extent of the

) Siwalik Stratigraphic Framework

Postilla 183

buff sand units do not change along strike, in
contrast to the blue-gray sand units which
generally thicken westward and gradually
predominate over the buff sand and silt/clay
facies until they form stacks of “multistoried”
Sandstones in the vicinity of Dhok Pathan (at
the U level).

Paleogeographically, these relationships
indicate the presence of two major low gra-
dient distal alluvial fans interfingering in the
Khaur area during the late Miocene (Fig? 2)
Through time the zone of interfingering moved
westward as the fan of the buff-sand system
“displaced” that of the blue-gray sand sys-
tem, resulting in upward time-transgression to
the west of the Nagri and Dhok Pathan rock
units. The Nagri Formation corresponds to the
part of the Khaur section where blue-gray
Sands are dominant and multistoried, i.e., the
alluvial fan to the west. The Dhok Pathan For-
mation corresponds to the strata with a high
proportion of the silt/clay facies and thinner
units of either buff or blue-gray sandstones,
.€., the alluvial fan to the east. Along the
Sub-U isochron the two formations are thus
laterally time equivalent. Determinations of
Current direction show that flow in the buff
Sand channels trended south to southeast
while the blue-gray sand channels flowed
more consistently eastward (Behrensmeyer
and Tauxe, in preparation). Bedding char-
acteristics of the blue-gray sands imply
sedimentation during periods of high flow ina
low-sinuosity braided system dominated by
Sand, while those of the buff sand and silt/clay
facies indicate deposition in a more continu-
Ously variable regime including both mean-
dering and braided channel belts associated
with fine-grained floodplains. Further details
concerning the fluvial paleoenvironments are
given in Pilbeam et al., 1979 and Badgley and
Behrensmeyer, in press.

Modern analogues for the proposed pale-
Ogeographic reconstruction of two interfin-
gering fluvial systems are present along the
Southern edge of the Himalayas, the clearest
being that of the Kosi River and adjacent riv-
€rs of India (Gole and Chitale, 1966). The Kosi
River is known for its rapid lateral shifts, total-
ling 112 km westward in 200 years, during

which it has left behind a persistent unit of
sand on the order of 2-3 m thick. The Kosi
drains an area of the Himalayas where there
are few temporary storage areas (i.e., valleys)
so that a large load of sand-grade material is
carried to the mountain front and deposited in
a piedmont fan. The Kosi also bears a large silt
component, but because of the relatively
steep gradient on the upper part of the fan (95

“cm/km) this is carried farther downstream be-

fore being deposited. It seems reasonable to
suppose that such circumstances may be
analogous to the depositional regime of sand
units in the blue-gray sand system. The latter
may have been larger in scale than the Kosi
River fan, and possibly was deposited by the
proto-Indus River.

In the case of the Miocene blue-gray and
buff-sand systems, it is interesting to note that
deposits of the two fluvial regimes accumu-
lated at nearly the same rate through time,
with only gradual lateral displacement of the
blue-gray system. An apparently balanced
situation occurred in spite of the fact that the
buff system was dominated by fine-grained
sediments while the blue-gray system was
dominated by sand. This implies tectonic con-
trol of sediment accumulation rates and over-
all thickness in the basin of deposition. Fine-
grained sédimentation was probably slow
relative to sand deposition. In order for equiva-
lent thicknesses to accumulate, east and
west, sand deposition must have been con-
siderably more sporadic or punctuated than
the build-up of finer sediments. There is no
evidence at present to support the possible
alternative hypothesis of increased subsi-
dence toward the western end of the basin.

In the upper 860 m of section in the Khaur
area (Kaulial Kas), coarse gravels are as-
sociated with the buff sands and interfinger
with the silt/clay facies. The sands are less
Clearly of the typical buff-sand facies and
often appear to be a mixture of the buff and
blue-gray compositions. The zone of contact
between the gravelly sands and other litho-
facies rises in the section westward (Fig. 2). It
appears that a source of coarse clastic mate-
rial contributed to the later phases of Khaur
area sedimentation. This may have been due

10 Siwalik Stratigraphic Framework

Postilla 183

to contributions from yet another river system
draining a relatively local area of tectonic uplift
such as the moder Kala Chitta or Khair-i-
Murat ranges. Units of blue-gray sand reap-
pear at the top of the Khaur section in Kaulial
Kas, indicating the return of this fluvial regime
or the influence of a new source area on the
already established systems.

In Gill's (1951) study of the Siwalik se-
quence of the northern Potwar, he notes
coarsening and thickening of the Nagri For-
mation westward, thickening of the Dhok
Pathan Formation eastward, and the appear-
ance of coarse Clastics in the upper part of the
section east of Kaulial Kas. Our study con-
firms his overall broad-scale picture of lateral
variation in the Middle Siwalik formations, and
the sub-U isochron provides new information
on the relationship of both major and minor
lateral facies changes to time.

Paleoecological Implications

Much of the fossil vertebrate material fromthe |
Khaur area occurs in the zone of interfingering —

of Nagriand Dhok Pathan lithofacies, in strata
dominated by silt/clay facies with thinner beds
and lenses of blue-gray or buff sands. Follow-
ing strictly lithostratigraphic definitions of the
formations, this means that the bulk of the
vertebrate collection is derived from the Dhok
Pathan Formation. This formation is over 1400
m thick and spans more than 3 my in Kaulial
Kas while in its type area 40 km to the west it is
about 300 m thick and spans less than 1 my
(Fig. 2). The fauna from the lower part of the
Dhok Pathan Formation between Ratha and
Kaulial Kas’ bears many similarities to the
fauna from the type section of the Nagri For-
mation, on the southern limb of the Soan
Synclinorium, some 60 km south of Khaur. Re-
ferring to the latter fauna as the ‘Nagri fauna,”
as has been frequently done in the past, leads
to confusion when the name is applied to a
similar fauna in the Khaur area which is clearly
from the Dhok Pathan Formation. New ter-
minology is needed for referring to the verte-
brate faunas as biostratigraphic entities, and
an informal preliminary zonation has been

suggested in Pilbeam et al. (1979) as a first
step in this direction. As previously noted, we
are restricting our use of the terms Chinji,
Nagri and Dhok Pathan to the major lithofacies
units which fit the descriptions of their typical
lithologies in the type areas, without reference
to faunas or time.

There is a difference in vertebrate fossil
abundance between the Dhok Pathan and
Nagri lithofacies, which may be due to both
taphonomic and paleoecologic factors. The
fluvial regimes of the two formations were
clearly very different, with the multistoried
Nagrisands reflecting extensive reworking by
relatively high energy flow, whereas the Dhok
Pathan sedimentary units represent variable
conditions of deposition with slow vertical
accretion of the silt/clay facies and sporadic
episodes of channel cut and fill. A priori, it
seems likely that such differences in the over-
all fluvial regime would have an effect on the
preservation of vertebrate remains. However,
exactly why the Dhok Pathan depositional
processes should have been conducive to
bone burial and preservation, especially in the
zone of interfingering, is an intriguing problem
which is currently under investigation.

The Biostratigraphic Evidence

Figure 3 shows schematically the geographic
and stratigraphic positions of most of the
Yale-GSP collecting localities in the area near
Khaur. On five of the columns a series of
letters or numbers marks the stratigraphic
positions of topographic landmarks

chosen as reference points and on four col-
umns there is a summary of the paleomag-
netic data, details of which can be found in
Tauxe (1979) and Behrensmeyer and Tauxe
(in preparation).

As discussed above, all 11 sections have
been correlated by tracing one or more hori
zons laterally between adjacent columns.
Some of these marker horizons, the principal
one of which is the U sandstone, are shown on
the figure as horizontal, dashed or solid lines:
These marker units are mostly thick sand-
stones which in some cases can be traced

i] Siwalik Stratigraphic Framework

Postilla 183

tens of kilometers laterally before they pinch
Out. In the case of the U sandstone it has been
shown by Tauxe (1978) that the unit is essen-
tially an isochron over the whole extent of its
Outcrop. The paleomagnetic and faunal data
Suggest that the age of the top of the Kaulial
Kas section is probably slightly more than 6
my, the U sandstone horizon 8.1 my, and the
Locality 259 level about 9.5 my (Tauxe, 1979,
and personal communication). Interpolating
downwards then, the oldest localities on Fig-
ure 3 might be about 12 to 15 my, and thus
the whole sequence might span 6 to 9 my.

Localities and Survey Blocks

The stratigraphic positions of the localities
and survey blocks are indicated by letters and
numbers at the side of each stratigraphic col-
umn in Figure 3. A collecting locality, as we
Use the term, is a very restricted area of out-
Crop surface on which one or more fossils
Nave been found. In the Siwalik exposures
fossils may occur either as isolated finds or in
discrete concentrations which may have only
& few specimens or as many as several
hundred. Both isolated finds and concentra-
tions occur nearly continuously throughout
the stratigraphic sequence ina variety of
Sedimentary contexts, although some sedi-
ment types and some parts of the sequence
are more fossiliferous than others. Because
few of the isolated occurrences, which are
typically unidentifiable bone and tooth frag-
Ments, are of particular interest, we have
largely focused our efforts on the concentra-
tions of fossils. Such concentrations are usu-
ally scattered over a surface area of less than
1000 sq m and are derived from only one
Sedimentary body. A few of the very largest
localities, however, outcrop over much
larger areas and may be derived from acom-
plex of sedimentary facies and stratigraphic
horizons.

All the Yale-GSP localities have been la-
beled with sequential numbers and their posi-
tions have been markedon topographic maps
and, when available, overlays on aerial
Photographs. The Universal Transverse Mer-
Cator grid coordinates and a brief lithological

description of each locality have also been
recorded in the Yale-GSP field catalogues
and, for recently discovered localities, on
a card file system which also incorporates
sketch maps, small-scale cross-sections,
polaroid ground photographs, and related
information.

All collected fossils have been catalogued
with a field number and notation of the locality
from which the specimen comes. This infor-
mation as well as the taxonomic identification
and a description of the specimen are now
tied into an electronic data processing system
(Pilbeam et al., 1979) so that it is possible to
create lists of the taxa and skeletal elements
found at each of our over 490 collecting
localities.

The fossil collections made by B. Brown in
1922 and G.E. Lewis in 1932 have sufficiently
accurate locality data to justify attempts at
relocating their exact positions. In both these
cases the concept of locality used by the orig-
inal collectors is the same as that used by the
Yale-GSP group and some of these localities
are shown on Figure 3, preceded by a B for
the American Museum collection and an L for
the Yale Peabody Museum collection.

On Figure 3 a series of survey blocks are
marked to the left of five of the sections. These
survey blocks provide an alternative to assign-
ing individual locality numbers to isolated fos-
sils of biostratigraphic interest, such as single
equid or bovid molars, and are useful in
areas where fossil concentrations are not
common. Each survey block is typically be-
tween 30 and 120 m thick and is defined within
the local section by prominent upper and
lower lithological marker horizons. Most ofthe
survey blocks extend laterally 1 or 2 km. The
most complete series is in the Kaulial Kas
section where the upper third of the section
has 14 separate levels. As may be seen from
Figure 3, the upper and lower boundaries of
stratigraphically equivalent blocks in sepa-
rate kas’ do not necessarily coincide (for
example, KL16 and MLOS5).

The stratigraphic horizon of each locality on
Figure 3 has been determined in the field by
tracing its level laterally into one of the nearby
measured sections. Because we have also

Siwalik Stratigraphic Framework

Postilla 183

been able to correlate the sections, it is then
possible, within limits, to assess the relative

stratigraphic positions of all these localities.

There are, however, certain problems stem-

ming from the assumptions we are forced to
use and the difficulties of precisely locating

the localities.

The major assumption we have made when
correlating widely separated localities is that
the stratigraphic thickness between a locality
and the isochronous U horizon is equivalent to
time, so that, for example, localities 500 m
below the U horizon in Kaulial Kas are approx-
imately the same age as localities 500 m
below the U horizon in Hasal Kas. Aside from
the sensitivity to error in measuring the sec-
tions, this approach depends on the assump-
tion that in the local Khaur area there is not
significant thinning or thickening of the sec-
tion in any direction.

There is also a related question of the fine-
ness of resolution possible in a fluvial system
where nondeposition or subsequent downcut-
ting by streams might remove several meters
of section, so that two nearby localities at the
same stratigraphic level might be of signifi-
cantly different ages. Study of this problem
suggests that the lower limit of resolution, at
least in the Dhok Pathan facies, is somewhere
near 10 m, although when very detailed
stratigraphic work is done it is often pos-
sible to refine the resolution considerably
(Behrensmeyer and Tauxe, in preparation).
On Figure 3the accuracy with which a locality
can be placed is limited by the scale to about
20m.

A different set of problems results from the
fact that many of the Yale-GSP localities were
discovered and collected before the strati-
graphic sections were completed. Thus we
frequently have had to relocate Yale-GSP
sites found in the earlier years of work and all
of B. Brown's and G. E. Lewis’ Dhok Pathan
localities. Experience has shown that itis often
difficult to relocate a locality unless individuals
working with the field party are already familiar
with the location of the site. The available to-
pographic maps are hachure maps on which
it is difficult to accurately mark locations. Un-
certainty about the exact geographic location

compounds the difficulty of tracing a locality
laterally and there is therefore frequently
some question about the exact position of in-
dividual localities on the stratigraphic col-
umns, with different localities having different
degrees of uncertainty. We have therefore
developed a two-part classification scheme to
express the degrees of confidence we have in
the geographic and stratigraphic placement
of our fossil localities (Table 2).

With some exceptions the localities on
Figure 3 are allofat least Stratigraphic Class 4
quality. That is, the stratigraphic positions of
these localities are accurate to within 50 m
and the majority are more accurately placed.
The exceptions are all Stratigraphic Class 5
localities, which are of special interest either
because of the richness or uniqueness of the
fossil assemblage or because the locality lies
at a stratigraphic level not otherwise well rep-
resented in the Khaur area. The ranges of
possible stratigraphic positions for all the
Class 5 and some of the marginal Class 3 and
4 localities are indicated on Figure 3 by
dashed vertical lines.

Biostratigraphy

A brief review of the history of the stratigraphic
nomenclature of the Siwaliks is given in Pil-
beam et al. (1977). Ashort discussion is pre
sented here of various critical biostratigraphic
problems and how we have attempted to re
solve them.

The definition and use by Pilgrim (1913,
1934) and others (Cotter, 1933; Colbert, 1935;
Lewis, 1937) of five successive faunal zones
(Kamlial, Chinji, Nagri, Dhok Pathan, and
Tatrot) marked an important advance in reC-
ognizing that there was a series of successive
faunas in the Siwaliks and in attempting to
devise a local nomenclature for discussing
these important faunas. The terms, as first
used by Pilgrim (1913), were conceptually
most similar to the stages of the current
American Code of Stratigraphic Nomencl@-
ture (1961), but subsequent usage has
tended to employ them as both lithostrati-
graphic formations and bio- or chronostratl-
graphic zones or stages with the distinction not

13 Siwalik Stratigraphic Framework Postilla 183

Table 2.7 A classification of the quality of geographic and stratigraphic placements.

A. Geographic
Class Criteria

1 Location known to within 10 m. Location marked on aerial photographs and
maps and sketches and ground photographs have been made. The exact position
of these localities can usually be confirmed by the presence of bone scrap.

2 Location known to within 100 or 200 m, but either because of the absence or
extra abundance of bone scrap it is now impossible to find the exact loca-
tion. In some cases these are localities which have a very wide collection
area, encompassing what are really several distinct localities. Locations are
marked on topographic maps and sometimes aerial photographs. Most Class 2
localities probably cannot be upgraded to Class 1.

3 Location is also known to within 100 or 200 m, but not yet confirmed by
actually revisiting the site. Locations are marked on topographic maps. In
most cases these can be upgraded to Class 1.

4 Location not known to within 200 m. Locations marked on topographic maps, but
we have not yet attempted to relocate them. Some, but not all, could be
upgraded to either Class 1 or 2.

5 Location not known to within 200 m. Locations marked on topographic maps,
but it is not possible to fix the locality's position more precisely.

6 Location essentially unknown other than as to region (i.e. "Chinji,"
"Hasnot"). Probably cannot be ungraded.

B. Stratigraphic
Class Criteria

u Stratigraphic position marked on one or more local measured sections and
accurate to within 3 m. It should be possible to unambiguously order
stratigraphically all the Class 1 localities tied to each local section.

2 Stratigraphic position marked on a local section with a range of precision
between 3 and 15 m. Some placements could be upgraded to Class 1.

3 Stratigraphic position marked on a local section with a range of precision
between 15 and 50 m. Composed mainly of Geographic Class 2 localities
where the uncertainty about the geographic location makes it difficult to
fix the stratigraphic horizon more accurately. Cannot be upgraded.

4 Stratigraphic position marked on a local section with a range of precision
between 15 and 50 m. Composed mostly of Geographic Class 3 localities.
Could probably be upgraded to either Class 1 or 2.

5) Stratigraphic position tentatively marked on a local section with a range of
precision greater than 50 m. May only be possible to state that a locality
is older or younger than some datum. Can be upgraded.

6 Stratigraphic position known only in terms of Pilgrim's faunal zones (i.e.
"Chinji") or completely unknown. Not marked on measured sections.

Siwalik Stratigraphic Framework

Postilla 183

always being clear. Used in a very broad and
loose sense as stages Pilgrim's faunal zones
have been and still are useful, although it is
apparent on close scrutiny that there are three
fundamental problems associated with their
continued use apart from the confusion result-
ing from lithostratigraphic and bio- or
chronostratigraphic terminology.

The first problem results from the lack of
demonstrable superposition for four of the
original faunal zones. As lithological forma-
tions Pilgrim was able to demonstrate by field
evidence the superposition of his five units
and the sections in his 1913 paper show the
same sequences of these rock types
throughout the Potwar Plateau. As faunal
zones, however, only the Nagri and Chinji
faunal zones have any direct evidence of
superposition. The other type areas are all
widely separated geographically and no one
section is fossiliferous throughout. Thus none
of the type faunas can be related to the others
except by lithological correlation.

The second problemis, like the first, closely
related to the confusion between kinds of
stratigraphic terms. Since some of his type
areas were only poorly fossiliferous, in order to
include enough species to make each fauna
distinctive Pilgrim used species from distant
areas that had more fossils. He was therefore
forced to presume that his lithological correla-
tions were also chronological correlations.
The Kamlial zone is thus characterized
paleontologically by species from Dera Bugti
and the Manchars about 500 km to the south,
while the Nagri fauna includes species from
Haritalyangar which is 1000 km to the east.
The Dhok Pathan is a special case. The type
area is very fossiliferous, but Pilgrim added
distinctive species from the lithologically simi-
lar part of the section at Hasnot, which we now
believe (Pilbeam et al., 1977; Lindsay, per-
sonal communication) may be significantly
younger. The Dhok Pathan fauna as a result is
mostly characterized by faunal elements
younger than the type area.

Finally Pilgrim’s zonal concept has no
sense of the length of time during which the
faunas lasted nor of the length of time, if any,
between them. This is the result of not knowing

the exact stratigraphic relationships of the
faunal zones. We now know from our field
research that Pilgrim's Chinji fauna was col-
lected from strata which probably span a very
long period of time. We also now know that
although the Nagri Formation is thick and ob-
viously includes aconsiderable length of time,
the fossils Pilgrim had available from the type
section all came from a single locality which
probably spans only a very short period of
time. The Dhok Pathan fauna, if restricted to
the fossils from near Dhok Pathan Village, !S
from a 300-meter-thick section with the bulk of
the localities being concentrated in the narrow
100-meter middle part of the section, which
surely represents only a‘short period of time.
We have no reason to believe that Pilgrim's
faunal zones are other than a series of se-
quential units. However, we favor discontinu-
ing their use as either biostratigraphic or
chronostratigraphic units because, for the
reasons just stated, they are poorly defined
and not amenable to either more precise
definition or further refinement. In particular,
because the relationships between faunal
zones are not easily demonstrated and their
duration is uncertain, we cannot sharply de-
fine the boundaries between them nor sub-
divide them into smaller units. In place of Pil
grim’s five faunal zones we have suggested 4
biostratigraphic zonal scheme (in the sense Of
the American Code of Stratigraphic Nomen-
Clature, 1961) which is, however, still very
preliminary and likely to be changed as
study of the faunas proceeds (Pilbeam et al.
1979). Such a scheme can be based on the
sections presented in Figure 3 or similar S€C-
tions from the Hasnot region, which can be
correlated independently of the faunas
using paleomagnetic reversal stratigraphy.

Faunal Events and Comments

From a preliminary analysis of the Khaur area
Yale-GSP collections we have established the
approximate stratigraphic levels at which var
ious large mammal species either first appe a”
or finally disappear from the faunal succeS-

sion. These events are noted on the right mar
gin of Figure 3. Three of the most dramatic

faunal events in the Khaur sequence are the

15 Siwalik Stratigraphic Framework

Postilla 183

Successive appearances of equids, large
Giraffes and alarge, hypsodont bovid. Each of
these forms first occurs in several localities at
approximately the same stratigraphic level,
becomes very common, and apparently per-
Sists to the top of the local section. Other
noteworthy faunal events include the succes-
Sive disappearances of Conohyus sindiensis,
Giraffokeryx punjabiensis, Listriodon pen-
tapotamiae, and a species of Deinotherium, as
well as the appearance of Hippopotamodon
Sivalense, Propotamochoerus hysudricus, a
very small Dorcatherium species, the small
hyaenid Palhyaena sivalense, the burrowing
rhizomyid Protachyoryctes tatroti, and col-
Obine monkeys. In addition, we have estab-
lished the presence at certain levels of other
important but rare taxa and these and the
localities at which they are found are listed in
Table 3. The difficulties of making identifica-
tions on fragmentary material and the uneven
distribution of localities throughout the section
limit our certainty about the exact level at
which each event took place. With more col-
lecting and further analysis we expect to see
changes in the level of at least some of these
faunal events.

Genera in our Khaur area collections from
levels at or below Localities 126 and 259
include Hyainailouros, Herpestes, Percro-
Cuta, Deinotherium, Listriodon, Conohyus,
Merycopotamus, Dorcabune, Dorcatherium,
Giraffokeryx, Elachistocerus, and either
Ramapithecus or Sivapithecus. Ator between
Localities 259 and 395 we have the highest
Stratigraphic occurrences of Listriodon, Cono-
hyus, and Giraffokeryx and the lowest oc-
Currences of hipparion equids and large gi-
raffes (both co-occurring at Locality 395 with
Listriodon). This short stratigraphic interval
(100 m) may prove to be a time of extensive
faunal replacement, since we have indica-
tions of change of some of the other faunal
elements as well. However, the faunal turn-
Over is by no means complete and taxa
as different as hominoids, amphicyonids,
and anthracotheres survive on into
younger ages apparently unchanged.
Faunal events above the Locality 395
level seem to be more evenly spaced

throughout the stratigraphic sequence.

The oldest Khaur area hominoid is from
Locality 259 in Kaulial Kas, but many other
specimens from what are undoubtedly older
levels have been found near Chinji Village, on
the southern side of the Soan Synclinorium.
The youngest hominoid we have found in the
Khaur area is from Locality 442, which is also
in Kaulial Kas. Both of these localities are
separated from the other hominoid occur-
rences by long stratigraphic intervals. We do
not yet know whether this pattern is only an
artifact of collecting a rare group or if there
were actually periods during which hominoids
were not members of the region’s fauna. Nor
do we know whether hominoids later became
locally extinct, after their last occurrence in the
Kaulial section.

The youngest localities we have found in the
area near Khaur are thought to be slightly
older than 6 my. None of these localities have
any fossils of the Hippopotamus Hexaproto-
don, which is acommon fossil in the Tatrot
beds near Hasnot. The absence of Hexa-
protodon at Khaur is in accordance with its
appearance in the region after 5.5 my (Op-
dyke et al., 1979). The oldest localities shown
on Figure 3 are 233 and 234. We have, how-
ever, one undoubtedly much older locality
near the small village of Gali Jagir, about 15
km south of Fatehjang. This locality is prob-
ably in the lower part of the Murree Formation,
but it is not possible to correlate it as yet to any
ofour measured sections. Its fauna includes a
shark.

Summary and Conclusions

The combination of litho-, chrono- and bio-
stratigraphic information presented in Figures
2 and 3 establishes a framework for the ongo-
ing study of Siwalik faunas and sedimentary
history. The detailed information resulting
from seven years of work in the Khaur area will
allow us to calibrate local variation in faunal
occurrences, lithofacies and the record of
paleomagnetic reversals and to discover the
consistent patterns that are of more than local
significance. These broader patterns, includ-
ing the time-transgressive nature of the major
lithofacies and the appearance and disap-

Siwalik Stratigraphic Framework Postilla 183

Table 3. Rare Lower and Middle Siwalik mammal taxa found at Yale-GSP localities
in the Khaur area. The stratigraphic horizon of each locality is indicated on

Figure 3.

Taxon Localities

Lorisidae, gen. et sp. indet. Age apy 272'5)]

?Presbytis sivalensis 370

Ramapithecus punjabicus S222 by 224 226s 222k 2005 aUos
310),<317., 23505, .409,414,. 41652463

Sivapithecus indicus S75 eee Ole 20/5 2h 224, 0227 soe ous
2605 2261 OSES Ae Si S2eyeso OU. aLO,
414, 416

Hominoidae, gen. et sp. indet. 259, 3275442

Hyainailouros sulzeri 3035330

Herpestes spp. UO US 2 Zoos to i.

Manis sp. 34

Orycteropus sp. 260, 365

Deinotherium sp. oP lG Oy A Seal Ae O25 21 253 aeods
2434) Zool, Say S295. 3625445

Schizochoerus gandakasensis 1 7/Sy2 09

Sivahyus punjabiensis S75

Tetraconodon magnus 251, 325, 493

Chalicotherium salinum 158. 82," 2loy) 2274 245, 269, S002 O05

S35) 2084 45

324,

17 Siwalik Stratigraphic Framework

Postilla 183

pearance of important vertebrate taxa, can
then be used as a firm basis of compari-
son with other regions. Eventually this strati-
graphic framework should lead to overall cor-
relations of other major fossil-producing areas
of the Soan Synclinorium and Siwalik se-
quences in India.

At the present stage of analysis, we can
offer some preliminary statements concerning
faunal and sedimentary history in the Khaur
region, based on the newly established
Chronostratigraphic framework.

The succession of fluvial sediments ap-
pears to be continuous throughout the 3250 m
documented so far, in the sense that there is
no lithostratigraphic evidence for major time
gaps (€.g., erosion surfaces, abrupt major
Changes in lithology). The observed micro-
and macropatterns of lateral interfingering of
the blue-gray sands with buff sands and
finer-grained facies indicates long-term
deposition in two adjacent fluvial regimes on
the piedmont belt of the Himalayas. Within the
upper half ofthe section, the local dominance
of one system to the other in the Khaur area
shifted slowly throughtime. Basin subsidence
apparently kept pace with the input of sedi-
ment. Within this overall depositional system,
there are differences in the two fluvial re-
gimes, particularly exemplified by the ap-
parently isochronous blue-gray sheet sands
(e.g., the U sandstone) which may reflect re-
peated tectonic or climatic events in the
source area. The sedimentary record on afine
scale is composed of channel cut and fill
episodes, and vertical floodbasin aggrada-
tion with local hiatuses during which paleosols
formed. The preserved sedimentary and
faunal record is thus subject to short-term time
gaps in any one section due to such local
vagaries of the depositional system. However,
for the Khaur area as a whole, the geologic
record can be considered essentially con-
tinuous, at least over the 6 to 9 my time span
indicated by the magnetostratigraphy.

The faunal record is subject to taphonomic
processes resulting in a patchy distribution
within the sedimentary units, both laterally and
through time. This limits the fine-scale resolu-
tion of the evolutionary record. Faunal remains

from a similar level relative to the sub-U iso-
chron generally must be combined inorder to
have a large enough sample for comparison
with other levels in the section. The net result is
that such combined faunal samples may rep-
resent time spans on the order of 10,000 to
100,000 years, or more for the lower part of the
section. For the upper part of the section, the
degree of time-averaging can be made com-
parable from level to level by combining fos-
sils from equal stratigraphic thicknesses, re-
sulting in a faunal record which is continuous
over a4 million year period, given the stated
limits of time resolution. The nature of this
record contrasts with that of the well-known,
and comparably thick, Clark Fork Eocene
sequence of northern Wyoming (Ginger-

ich, 1974; Bown, 1979) in that the latter has
more distinct levels of fossil concentration
which probably represent shorter periods of
time-averaging.

Within the limits imposed by the sedimen-
tary and taphonomic characters of the Khaur
Siwalik sequence, we are able to document
13 large mammal faunal events. These in-
clude three local extinctions and two appear-
ances which are clustered together about 500
m below the U horizon, and a sequence of
events which are more regularly spaced up
through the succeeding stratigraphic levels.
The large mammals thus suggest that there
was a major turnover of the fauna at about
9.5 my BP, at which time such forms as Cono-
hyus, Listriodon, and Giraffokeryx became
locallly extinct and hipparions and large gi-
raffes made their local appearance. This
episode was then apparently followed by a
long period during which single local extinc-
tions and appearances slowly removed or
added species to the fauna. With our present
collections from the Khaur area we are not
able to determine the nature of faunal change
at the stratigraphic levels below the Locality
126/259 level, if there was in fact any change.
The paleomagnetic evidence indicates, how-
ever, that the underlying sediments were de-
posited over a considerable length of time
and, as discussed in Pilbeam et al. (1979),
preliminary analysis of the fossil collections
from the Chinji type section suggests that

Siwalik Stratigraphic Framework

Postilla 183

there was some faunal change during the
period in which the Kamlial and Chinji forma-
tions were deposited.

Aside from the lithofacies changes noted in
the above discussions, there is no obvious
indication of environmental change in the
sedimentary record through the Khaur area
sequence. The lithofacies changes bear little
relationship to faunal appearances and dis-
appearances, as Currently understood. In-
stead it seems that the biological record is
simply Superimposed on the succession of
facies marking the build-up of the sub-
Himalayan alluvial plain. These lithofacies
preserve different densities of fossil remains
due to variability in taphonomic processes,
and possibly also due to differing ecological
conditions in the different fluvial systems.
Based on our Current understanding of the
relationship of the faunas to the sediments
through time, we do not see any indications

that processes affecting the fluvial deposits
were linked, in an observable way, to proces-
ses affecting faunal change.

Acknowledgments

This study was completed under NSF Grant
No. BNS 772 5984 and SFC Grant No. FC
80254100. We would like to thank the many
members of the Yale-GSP project for both
ideas and criticisms, and particularly Dr. S.
Ibrahim Shah of the Geological Survey of
Pakistan and Dr. David R. Pilbeam of the Yale
Peabody Museum whose support has made it
possible to complete the field and laboratory
research. We thank Lisa Tauxe of the
Lamont-Doherty Laboratory and Dr. Everett
Lindsay, Larry Flynn, Richard Haskin,
Yukimitsu Tomida, and Louis Taylor of the
University of Arizona for providing the critical
paleomagnetic work referred to in this paper.

Literature Cited

American Commission on Stratigraphic Nomenclature. 1961. Code of Stratigraphic Nomenclature

Am. Assoc. Pet. Geol. Bull. 45:645-665.

Badgley, C. E. and A. K. Behrensmeyer. In press. Paleoecology of Middle Siwalik sediments and faunas,
northern Pakistan. Paleogeogr. Paleoclimatol. Paleoecol.

Barndt, J. 1977. The magnetic polarity stratigraphy of the type locality of the Dhok Pathan faunal stage,
Potwar Plateau, Pakistan. Dartmouth College, M.A. thesis.

Behrensmeyer, A. K and L. Tauxe. In preparation. Isochronous sedimentary environments in Siwalik

deposits, northern Pakistan.

Bown, T. M. 1979. Geology and mammalian paleontology of the Sand Creek facies, lower Willwood
Formation (lower Eocene), Washakie County, Wyoming. Geol. Surv. Wyoming Mem. 2. 151 pp.
Colbert, E. H. 1935. Siwalik mammals in the Americal Museum of Natural History. Trans. Am. Phil. Soc., N.S.

26;)1—401.

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. 10:1-80.

Gill, W. D. 1951. The stratigraphy of Siwalik Series in the northern Potwar, Punjab, Pakistan. Quart. J. Geol.

Soc. (London) 107:375-394.

Gingerich, P.D. 1974. Stratigraphic record of early Eocene Hyopsodus and the geometry of mammalian

phylogeny. Nature (London) 248:107-109.

Gole, C. V. and S. V. Chitale. 1966. Inland delta building activity of the Kosi River. J. Hyd. Div., Proc. Am.

Soc. Civ. Eng. 92:111-117.

iy Siwalik Stratigraphic Framework Postilla 183

Lewis, G. E. 1937. A new Siwalik correlation (India). Am. J. Sci., Ser. 5, 33:191-204.

Opdyke, N. D., E. Lindsay, G. D. Johnson, N. Johnson, R. A. K. Tahirkheli, and M. A. Mirza. 1979.
Magnetic polarity stratigraphy and vertebrate paleontology of the Upper Siwalik subgroup of northern
Pakistan. Paleogeogr. Paleoclimatol. Paleoecol. 27:1-34.

Pilbeam, D.R., J. C. Barry, G. E. Meyer, S. M. I. Shah, M. H. L. Pickford, W. W. Bishop, H. Thomas, and
L.L. Jacobs. 1977. Geology and palaeontology of Neogene strata of Pakistan. Nature (London) 270:684—
689.

Pilbeam, D. R., A. K. Behrensmeyer, J. C. Barry, and S. M. I. Shah, eds. 1979. Miocene sediments and
faunas of Pakistan. Postilla (Peabody Mus. Nat. Hist., Yale Univ.), No. 179:1—45.

Pilgrim, G.E. 1913. The correlation of theSiwaliks with mammal horizons of Europe. Geol. Surv. India Rec.
43:264-326.

——— 1934. Correlation of Ossiferous Sections in the Upper Cenozoic of India. Am. Mus. Novitates, No.
704:1-5

Shah, S. M. I, ed. 1977. Stratigraphy of Pakistan. Geol. Surv. Pakistan Mem. 12:1-138.

Tauxe, L. 1978. A comparison of isochrons defined by short-term paleomagnetic events and lithostrati-
Qraphic units in Siwalik rocks of the Potwar Plateau, Punjab Province, Pakistan. Yale College Scholar of the
House thesis.

——— 1979. Anew date for Ramapithecus. Nature (London) 282:399-401.

The Authors
John C. Barry and Anna K. Behrensmeyer, Marc Monaghan, Department of Geology
Peabody Museum of Natural History and De- —_ and Geophysics, Yale University, New Haven,
partments of Anthropology, and Geology and CT 06511.

Geophysics, Yale University, New Haven, CT
06511

Kaulial
Jabbi

KL15
se Fi 370,386,377.374,372 *LSD ?Presbytis sivalensis
1
aes 437 458 %
i]
BBS KLO7 | ina i=
376378 369) A
« Malh : i |
alhuwala !
KL101
Nila SS
KLo9 !
OSS Oversoil Sand H
KLO8 T pe st LSD Protachyoryctes tatroti
4
os 472,399
392 171
176,214
Dhok Pathan i ae

*“HSD Deinotherium sp.
174

394,155 473 RPY
31, 27, £15, L56, £19, 113, £11, L14, £12, L16 3

3

1

H

HN KL12

97, 98, 99, 129, 130, 131, 116, 133 1 167.192 ;

7 ce 1

} he _
107, 108, 110, 90, 13, 25, 19, 20, 33. B46 !

i‘: Gandak
SORE Gee ep ee ee Lenco wee een ene nn as ~ 156,389
Oe a Bau meme ieee nae TTT : Dhok Mila A : andakas \

3 T
21,15,B104 J
1 11, 112, 29, 109, 16

4a
y BE
1 99, 23, 92, 93, 118, 111, 132, B43 +

t bovid
poorpeeny LSD large, hypsodon

477
MLO5T 387,407,445 1 408 <
T 201 169 360
ines 7, 2, L6 15, B44,B105,838 1 1 I MLO4I 388,158 469,468,470
Hilltop a OO Nala 1113, 119, 32, 28, B42 1 Kh K ! ~ K 421,146 TOS 164,165, 166,178
Se rata | a Mercere Te aur Ras jee Dinga Kas = “y= -------28.------- ~ 7 .
uu - Ie $ m li ®@- Hi 355 MLO6 415 Gray Tracer H haonbioniae 406,449
Choutriwali a ee same 9)
4 DMo3i,, | @- w
319 Hi LSD Palhyaena sivalense
196,197,193,474,480 ' 320,205 = --- 181,198 44
Le - : DMo2T 366,967.397,402,484 1 oe = DKo1 a, MLO7, 420 : 446 4,
” Kot Mal 1 c- - H : ite
1 Ll = 137
1 oO allaran ee Mco8 ot :
Ss DMO1T . | G9 4 182,208 260,409,465,466,412,414,471,485
243
T
| “ 231,229 1 240,239,385 944 o x
314 T ' 1 213,255 1351 8 237 416
122 ia 317 310 1 316,318,308,307,313 - 334,226,359 ae Gace 163, 154.245 L T 191 H KLO3; yi
A i thc L 232 225 224,227,356 160,159,200 = wr i i 0
“ 121, 89 T 269,350,365 T
7] 123 1 140 270,482 !
i | 235 N-
I T 273
Marc1 919217 .>3——<_$——_——_____M- L
218 228 KLO5 £ 262,481
@4 330,362,329 , LSD small tragulid
= 426
250 | 337,338,340,268 eis 761428 4g
O+?P T H M4 424,276
EJ 425
7 H ee a 427
T vr z
H pee H KLO4
1120, 127 Baas ©4 N : + 246,249 K- pt ?LSD Propotamochoerus hysudricus
i L 284,285,286
Ja
101, 102
17 | 289 LSD Hippopotamodon sivalense
251,248,493
357
H H
125 ad @- Ko | 252 Go
i 258,368
@- 364,334
124, 128 = ee
LSD large giraffe
1 : B+C eed HSD Listriodon pentapotamiae
T
H LSD “Hipparion” sp.
GtH H E4 mn HSD Giraffokeryx punjabiensis
1 253 % 7 | 254 ?HSD Conohyus sindiensis
1
I 1 1 is
126 —— ; KL17)
!
@ ~ iy H re
7 I
T
1
1
1 302, 303, 304, 305, 103
: @-7 D-
1
i
rs
ul
a
Legend Fig. 3
a
C- Stratigraphic positions for Lower and Middle Siwalik
fossil localities in the Khaur area. The vertical scale
T D4 Ratha Sarton i is 1 cm = 60 meters; the horizontal scale is
1 > en henene approximately 1 cm = 0.5 km. Locality numbers are
1 332,333 to the right of each local section and survey blocks
4 are to the left. Localities prefixed with an “L” are
@> 335,336 N 1 Polarit those found by G.E. Lewis; those prefixed with a“B”
C4 a / ne are those found by B. Brown. The collections made
r by these men are in the Yale Peabody Museum and
BS 1000. 399 Survey : the American Museum of Natural History. The
Block )RKO11 — Reversed Polarity horizontal line linking all the sections is the sub-U
= = Hi isochron. The right margin shows the level of
a | - é — Age in Millions of Years 13 biostratigraphic events. First appearances
| a (LSD) are shown above the line marking the
Gg Ny 200 appearance. Last appearances (HSD) are shown
- below the line marking the local extinction. Special
@ = ' 500 4 symbols are noted in the legend.
1
Hl 7 T Range of Possible
1 360,381 Sing Marc2 | ._ Stratigraphic Placements ‘
! H 1 247 Postilla 183
1 489 I 7 Marker TOS :
= Begs == = L
4 !
! Feet oo Meters 251,249 — Fossil Localities B4
Map
Reference @
Points AS
*LSD: Lowest Stratigraphic Datum
*HSD: Highest Stratigraphic Datum
3
1
I
1
1 233,234
1
1
I
I
x
A Drafted by Akefeh K. Nowroozi
0
rs
Zé
=
gs |$% ;
pss = @
. Bo
8 a2
5 3/38
= 8/88 § A y
+4 Q@Fqs 5
oe z 6 =3 £ c
foo s=> 4 >s =
Se x» =
2.358 562
~~ >S&e goes
~# os55 ee
oexM“ Fe OS
O~ ®E+s c So
a. = os s ° = eS
ao @q@uc2 3542
8B
ci. © a te a
2S > ~~ X
3230 = aes
=I3S tg
Se
Tb5c 8
Gal
oO oO
(OZ Dz
a) & re) FZ

Om
/

t
G