LIBRARY OF THE

UNIVERSITY OF ILLINOIS

AT URBANA-CHAMPAICN

K

F4S£

,. V4

GEOLOGY

1*w

sponsfbTeS°for i?saS?g th,s material ,s re
LotestDotestamAdb:^0" or before the

UNIVERSITY -u: „ , „. ° 'V Univ""V

-*V IUINOIS UBRJdP »t

APR lo

OCT^

DEC o 7

Mb© • ^

OLIGOCENE SEDIMENTATION, STRATIGRAPHY, PALEOECOLOGY

AND PALEOCLIMATOLOGY

IN THE BIG BADLANDS OF SOUTH DAKOTA

JOHN CLARK

JAMES R. BEERBOWER

KENNETH K. KIETZKE

3

en

FIELDIANA: GEOLOGY MEMOIRS
0 g WSfc* VOLUME 5

o f\ $*$ PubltsJied by

J FIELD MUSEUM OF NATURAL HISTORY

U^nr- DECEMBER 29, 1967

GEOLOGY LIBRARY

FIELDIANA: GEOLOGY MEMOIRS

VOLUME 5

\y o/ *^ .

,'N NATURAL °V
'"t HISTORY £"

JSLw

[>■ .1^ "|." ;■

FOUNDED Br I

^/cAgO

FIELD MUSEUM OF NATURAL HISTORY

CHICAGO, U.S.A.

1967

OLIGOCENE SEDIMENTATION, STRATIGRAPHY, PALEOECOLOGY

AND PALEOCLIMATOLOGY

IN THE BIG BADLANDS OF SOUTH DAKOTA

OLIGOCENE SEDIMENTATION, STRATIGRAPHY, PALEOECOLOGY

AND PALEOCLIMATOLOGY

IN THE BIG BADLANDS OF SOUTH DAKOTA

JOHN CLARK

Associate Curator, Sedimentary Petrology, Field Museum

JAMES R. BEERBOWER
Department of Geology, McMaster University, Hamilton, Ontario

KENNETH K. KIETZKE

FIELDIANA: GEOLOGY MEMOIRS

VOLUME 5

Published by

FIELD MUSEUM OF NATURAL HISTORY

DECEMBER 29, 1967

Edited by Edward G. Nash

Patricia M. Williams

Library of Congress Catalog Card Number: 67-31598

PRINTED IN THE UNITED STATES OF AMERICA
BY FIELD MUSEUM PRESS

CONTENTS

I. Introduction 5

II. General Geography 8

III. General Geology 9

Stratigraphic Column 9

Regional Structural Setting 13

IV. The ? Slim Buttes Formation 16

V. Geology, Paleoecology, and Paleoclimatology of the Chadron Formation 21

Introduction 21

Acknowledgements 21

Structural Relationships 21

Lithology 22

Systematic Paleontology 25

Stratigraphic Paleontology 55

Paleogeography 59

Interpretation of Chadron Sedimentation 60

Paleoecology 67

Interpretative Summary 72

Conclusions 74

VI. Paleogeography of the Scenic Member of the Brule Formation 75

Introduction 75

Topography and General Stratigraphy 75

Lithology 77

Stratigraphic Relations of Sedimentary Lithotopes 85

Paleogeographic Interpretation of the Scenic Member 92

Relationship to Orellan Stratigfaphy and Sedimentation in Nebraska 102

VII. Paleoecology of the Lower Nodular Zone, Brule Formation, in the Big Badlands of

South Dakota Ill

Introduction Ill

Acknowledgements Ill

General Philosophy of Field Work Ill

Pertinent Field Data for Individual Collections 112

Curating and Identification 114

Variables and Biases Affecting Interpretation of the Population Statistics 114

Analysis of Faunas 120

Ecologic Relationships of Particular Genera 123

VIII. Interpretative Summary 138

IX. Conclusions 141

References 144

Appendices 147

Index 156

VII

LIST OF ILLUSTRATIONS

TEXT FIGURES

1. Oligocene rocks of western South Dakota 6-7

2. Columnar section for the Big Badlands 9

3. Upper Chadron channel directions 10

4. The Slim Buttes Formation and the Ahearn Member 11

5A. View of Locality V 12

5B. Slim Buttes Formation 12

5C. Inherited Slim Buttes sediments 12

6. Structural axes in the Big Badlands 13

7. Cross section of Figure 6 14

8. Slim Buttes-Chadron contact zone 18

9. Rate of Deposition of the Chadron Formation 25

10. The genus Parictis 28

11. The genus Parictis 29

12. Measurements of Parictis (Campylocynodon) parvus 29

13. Posterior accessory cuspule on premolars of Daphoenus 30

14. The genus Daphoenocyon: mandibles 31

15. The genus Daphoenocyon: paratype 31

16. Comparative measurements of Daphoenocyon, Daphoenus, and Parictis 32

17. Measurements of Eusmilus sp 33

18. Measurements of Chadron Mesohippi 34^46

19. Type specimens of Daphoenocyon minor, mesohippus grandis and unnamed species of Merycoidodon .... 49

20. Measurements of Mesohippus grandis 49

21. Phylogeny of Chadron Mesohippi 49

22A. Type specimen of Merycoidodon lewisi 53

22B. Type specimen of Merycoidodon lewisi 53

23. Measurements of Merycoidodon lewisi 54

24. Suggested correlations of certain Oligocene formations 57

25. Ecology of the known Chadron fauna 58

26. Tertiary Paleogeographic data — sedimentary 62-63

27. Tertiary Paleogeographic data — vertebrate fossils 70-71

28. Chadronian paleoecology by genera 71

29. Chadronian paleoecology index 72

30. Location of numbered sections and places 76

31. Photograph of sample G 4077 78

32. Photograph of G 3743 78

33. Paleogeography of the Scenic Member 79

34. Riosome #3, near Cottonwood Pass 80

35. Riosome #3, near Cottonwood Pass 81

36. Drenajesome #4 82

37. Viscous flow marks on a skull of Archaeotherium 84

38. Layered concretions in a silty mudstone 85

39. Columnar sections, Scenic Member 87-90

40. Scenic member mudstones 91

41. NE - SW cross-section of the Scnic Member 93

42. Cross-sectional diagram of a stream in a depositional regimen 94

43. Components of an environment of fluvial sedimentation 95

44. Perthotaxy on temperate steppes 100

45. Archaeotherium skull in situ 101

46. Cross-section of Toadstool Park area 103

47. Photograph of "the Bench" 104

48. Laminated sediments, near Toadstool Park 105

49. View westward along fault, Toadstool Park 106

50. Fault-face cutting channel-fill sandstone 107

51. Diagrammatic interpretation of Figure 49 108

52. Probable miscorrelations of lithologic units 109

53. The relationship of a life population to a fossil collection 117

54. Graphs I - XIII, population statistics 130-136

55. Internal consistency of Open Plains and Near Stream collections 137

56. Total fauna represented in collections used in Chapter VII 137

VIII

Chapter I

INTRODUCTION

Successive faunas preserved in the classic Oligocene
sequence of western South Dakota offer clear evidence
that the Oligocene Period was a time of climatic change
in western North America. The strata themselves have
long been recognized as "typical floodplain deposits."
This study attempts to analyse critically these two gen-
eralities by applying modern techniques and theory to
both the fossils and the sediments. The data resulting
from this analysis are used to reconstruct the Oligocene
paleogeography of western South Dakota in detail. An
attempt is then made to reconstruct the successive
paleoclimates of the area. Finally, the interaction of
the physical environments upon the faunas is con-
sidered.

Study of the fossils has been complicated by the fact
that the philosophy of collection, as well as the docu-
mentation, of the classic collections makes them unsuit-
able for analytical use. It has been necessary to make
collections specifically for this study, and modern col-
lections in the area are subject to certain unpredictable
variables which need not have operated at the time the
older collections were made. The limited statistical treat-
ment justified has, nevertheless, revealed sufficient in-
formation to merit its presentation here.

Study of the sediments has naturally devolved about
several basic problems:

1. Did the position of South Dakota relative to
the Earth's rotational poles differ materially from its
present position, or did it change significantly during
Oligocene time?

2. Which characteristics of the sediments are at-
tributable respectively to their ancestry, to their paleo-
graphic position, to the paleoclimate during deposition,
and to epigenetic changes?

3. What geomorphic controls determined that
Oligocene streams would deposit thick floodplain sedi-
ments within 30 miles of their headwaters?

4. What is the length of time represented by the
total sedimentary pile; what was the thickness of indi-
vidual increments; and what time is represented by the
increments as opposed to the intervening episodes of
non-deposition?

5. What is the geomorphic pattern established
by a group of streams whose primary regimen is deposi-
tion rather than erosion?

6. Do fossil faunules reflect the local paleogeog-
raphy evidenced by the sediments?

The first question can be answered in advance with
a probable negative. Even the most positive advocates
of various hypotheses of continental drift seem to feel
that post-Oligocene movement of North America rela-
tive to the poles of rotation has been slight. Abundant
paleobotanical and marine-paleontologic evidence tends
to confirm this. The study has, therefore, been conducted
with the primary assumption that South Dakota occu-
pied essentially the same position relative to the North
Pole, and to the continents of North America and Eu-
rasia, which it occupies today.

Chapters V, VI, and VII, which form the core of this
report, were originally written as separate papers, in-
tended for separate publication. The research was done
at somewhat different times, under quite different aus-
pices. Each chapter retains its own acknowledgments in
order to make clear the extent of assistance received.
The senior author wishes to acknowledge here the un-
stinted assistance of Dr. B. G. Woodland, who checked
most of the author's identifications of mineral species of
sand grains for various chapters. He also wishes to ex-
press his appreciation of Dr. T. Perenyi's patient and
skillful art work on the charts, maps, and diagrams.

This study raises approximately as many new prob-
lems, both local and fundamental, as it solves. I have
attempted to indicate these in all cases where the prob-
lem seems either to cast some doubt upon my favored
hypothesis, or to be fundamental enough to merit gen-
eral interest.

gocene Rocks of Wester

Fig. 1. Oligocene rocks of western South Dakota. The overlap of Oligocene sediments upon the Paleozoics is clearly indicated. Note
also that elastics derived from the Laramide intrusive of the northern Black Hills are available to Box Elder and Elk Creeks at present.

+ +
-t ± t

Laramide intrusives
Precambri an

iouth Dakota

Chapter II

GENERAL GEOGRAPHY

The area under consideration is roughly 60 miles
long, E-W, by 15-30 miles wide. It lies in western South
Dakota, east and southeast of the Black Hills (Fig. 1).
It includes parts of Pennington, Shannon (including old
Washington and Washabaugh), and Jackson Counties.
Rapid City to the northwest, Kadoka to the east, and
Wall to the north constitute the principal towns. South
Dakota route 40, and U.S. 16, 16A, and 18 provide ac-
cess. A fair network of county roads and an excellent
series of trails make almost every spot easily accessible
by light truck. About half of the area under study lies
within the Badlands National Monument; much of the
remainder is scheduled to be included in the future.

Physically, this district comprises the classic fossil-
collecting areas of the Big Badlands, plus the finest
exposures of Oligocene sediments in South Dakota. It
consists of short-grass prairie uplands and table-top
buttes, separated by badlands topography from short-
grass prairie lowland flats and brushy floodplains. Local
relief is usually 40-300 ft., with an extreme of 485 ft.
at the south rim of Sheep Mountain, and of 600 ft.
southeast of the Pinnacles (Sec. 17, T. 2 S., R. 16 E.).
Elevations vary from about 2450 ft., along Cheyenne
River and southeast of Interior, to 3282 feet, at the
southwest edge of Sheep Mountain (Sec. 32, 43 N., R.
44 W.). Cheyenne River drains the Badlands north-
westward, White River southward, and Bad River
northeastward.

The present climate is continental-temperate semi-
arid. The average annual rainfall of 17 in. has fluctuated
between 5 and 29 in.; the average annual temperature

is 47° F., with recorded summer highs of 115° and winter
lows of —35°. Local, diurnal, and annual fluctuations
make the averages almost meaningless for any particu-
lar year and district.

The vegetational cover at any one spot is necessarily
adjusted to the climatic extremes for that spot, which
means that it is usually sparse, small, and consisting of a
limited number of species. The bulk of plant material is
so small, and oxidation so active, that vegetational
decay rarely, if ever, contributes significantly to chemi-
cal weathering processes.

The native fauna is of no geologic significance. Graz-
ing, mostly by cattle, usually keeps the grass cover down
to the minimum necessary to maintain itself, and un-
doubtedly does somewhat aid erosion. The ungrazed
areas within the National Monument .have developed a
heavier grass cover, but I have seen no evidence that
erosion has been significantly influenced thereby.

Human activity has altered the geologic situation in
two ways. First, construction of small dams across
draws and gullies has progressed to the point where
run-off and the regimen of intermittent streams have
been significantly altered. Second, local rockhounds and
commercial collectors have so actively worked over most
of the classic collecting areas that proper collection of
faunas for statistical purposes is possible only in the
most difficultly accessible spots. This is a serious im-
pediment to detailed paleoecologic work: since there is
no correlation between paleogeography and ease of
access, several interesting paleoecologic environments
have been ruined for study, while certain others remain
untouched.

Chapter III

GENERAL GEOLOGY

A. Stratigraphic column.

The stratigraphic column within the area studied is
shown on the chart, Figure 2. Dips in the Pierre shale
are usually indeterminable at any one outcrop, due to
the fact that unloading expansion, followed by weath-
ering, alters bedding structures to a depth of about 30
ft. Alteration of disseminated pyrite to gypsum has
usually progressed to completion down to about the
same depth, which further disrupts bedding.

The 6-70 ft. of Pierre shale directly underlying the
Cretaceous-Tertiary unconformity is in most places
colored yellow, brown, purple, green, or red. This color
change accompanies a slight increase in Fe203 and
Mn02, and decrease in Ca and Mg, which suggests that
the colored zone is a pre-Chadron soil. Distribution of
the colored zone, thickest where the overlying Chadron
is thinnest, and thinnest underlying the old valley which
the Ahearn member of the Chadron fills, further sug-
gests soil, best developed in the gently-rolling hills of
the old upland. X-ray diffraction analyses of the colored
clays have not been run in sufficient quantity to reveal
generalities, but four analyses known to me show a
consistent increase in kaolin over the unaltered Pierre.
Southwest of the town of Pine Ridge, again in T. 36 N.,
R. 46-48 W., and in several places between Fairburn
and Buffalo Gap, the same colored zone is developed on
the Niobrara formation.

The usual color succession of the thicker sections is:

Yellow to pale blue

Brown

Purple

Gray

Unweathered gray or black.

Three to five feet of red form the very top, in the area
of Sage Ridge but not elsewhere (see Fig. 3). Limy
nodules up to 6 in. in diameter, with rough surfaces
or actual zones of blending into the surrounding sedi-
ments, often lie at the contact between the red zone and
the overlying greenish Chadron. Since equally thick
sections of colored Pierre in the eastern and southern
portions of its area of outcrop do not display this upper-
most red zone, it is possible that the red zone represents
topsoil weathering of a Pierre stratum originally some-
what more calcareous than those exposed elsewhere.
The Sage Ridge (Clark 1937, p. 289) apparently existed
as a structural uplift subject to pre-Chadron erosion,
which might well have exposed a stratum covered else-
where.

COLUMNAR SECTION

I lOO'
ROSEBUD

CHADRON
SLIM BUTTES

POLESLIDE-

Fig. 2. Columnar section for the Big Badlands.

The entire colored zone was named the "Interior
Formation"by Ward in 1921, and recognized as a soil
zone rather than a formation by Wanless in 1923. The
latter interpretation is now generally accepted; all of
the pertinent data observed since Wanless' original
interpretation support it.

Slim Buttes Formation: This formation was first
recognized, and a standard section established, in the
Slim Buttes of Harding County, northwestern South
Dakota (Malhotra and Tegland, 1959). The name is
applied here to sediments of similar gross lithology and
appearance, roughly similar mineralogy, and similar
stratigraphic situation. The fauna of the Slim Buttes
formation in its standard locality has not yet been
adequately studied, and the outcrops here referred to it
have yielded no identifiable fossils except wood, so
correlation based upon fossils is impossible.

The formation consists characteristically of white to
very pale greenish and olive-white sandstones, generally
quite well-sorted, bimodal, comprising a clay and a

T

o

z
o

a

<

Fig. 3. Upper Chadron channel directions.
10

E . E .

«$

z
o

oc
D
<
I
U

X
<

z
o

s

^ = -

C

3

«

3

E
o

T3

3

is

c
<v

E

t3

<v

43

as

T3

E

J2
5

C

B
o

o

fa

3

J3

11

12

FIELDIANA: GEOLOGY MEMOIRS, VOLUME 5

Fig. 5A. View of locality V, looking west at the SW \i of Section 2, T 3S, R. 20E, Jackson Co. Shows the Slim Buttes Formation
(light colored with a dark band near the top) resting upon the Interior Zone soil (dark grassy gullies in foreground). The overlying Chadron
cuts out the Slim Buttes in the middle of the picture. Slopes of both Chadronian channel-banks are visible, also Chadron resting on Interior
in the middle of the picture.

Fig. 5B. Slim Buttes Formation resting upon Interior Zone soil and overlain by Chadron, which has penetrated cracks and joints in
the Slim Buttes. In locality IV, Section 12, T 4S, R 17E, Jackson Co.

Fig. 5C. Inherited Slim Buttes sediments reconstituted as an Ahearn sandstone, in locality VI, Big Corral-Little Corral draw divide.

fine to medium-grained sand. The unit frequently in-
cludes a basal conglomerate.

The formation occurs as discontinuous lenses (Figs.
4, 5) sometimes several miles in extent, up to 40 ft.
thick, resting disconformably upon the Interior zone.
Most of these lenses wedge out with no visible indica-
tion of erosion by Chadron streams. However, a Chad-
ron channel-fill clearly bisects a lens of Slim Buttes
sandstone and even cuts somewhat into the underlying
Interior zone, north of Weta (SW y2, Sec. 2, T. 3S.,
R. 20 E., Jackson Co.).

The sediments referred to the Slim Buttes formation
are those previously described by Clark (1937, pp. 277-
278) as "restricted white channel fills" of the Ahearn
Member of the Chadron Formation.

Chadron Formation: Since the Chadron Formation
will be the subject of Chapter V, only the briefest sum-
mary need be given here.

The lowest, or Ahearn Member, consists of 0-80 ft.
of greenish and red-mottled sandstones with subordinate
gray, green, and tan mudstones, occupying an east- west
trough. There is usually a basal conglomerate, and con-

glomeratic channel-fills are common, especially in the
lower two-thirds of the member.

The middle, or Crazy Johnson Member, is made up
of greenish to gray conglomeratic channel-fills asso-
ciated with gray, greenish, and bluish mudstones. It is
20-40 ft. thick.

A predominance of greenish, gray, and tan to orange
mudstones characterizes the upper, or Peanut Peak
Member, 20 30 ft. thick. Conglomeratic channel-fills
are just as well-marked but smaller and much less
numerous than those of the members below.

The Chadron-Brule contact consists of a zone 6 in. to
3 ft. thick, of limey layers intercalated with greenish and
tan mudstones. Occasionally, the contact comprises a
single limestone up to one foot thick. Characteristically,
the limestones are partially to completely opalized.

Brule Formation : The Scenic Member of the Brule
comprises 89 130 ft. of tan, yellow, gray, red, and green-
ish mudstones with some siltstones, thin beds of fine-
grained sandstone, and channel fills some of which in-
clude coarse conglomerate and others which consist only
of fine sand. Chapter VI will describe the Scenic Mem-
ber in detail.

CLARK: GENERAL GEOLOGY

13

The Poleslide Member, 270 ft. thick, conformably
overlies the Scenic Member. It consists of yellow-to-tan
mudstones, with a considerable admixture of volcanic
ash. Sandstone channel-fills occur approximately over-
lying similar channel-fills in the Scenic Member, but are
much narrower, thinner, and usually finer-grained than
their underlying predecessors.

Rosebud Formation: Use of this name for the basal
Miocene formation of the Big Badlands has been dis-
puted by MacDonald (1958). However, the type locality
designated by Matthew and Gidley is almost in lithologic
continuity with the basal Miocene of the area under dis-
cussion ; deposition of sediments was generally outward
from the Black Hills, the same source: Macdonald con-
fuses the type section with beds "along the Niobrara
River to the South in Nebraska," and then invalidates
the name Rosebud by quoting an Oligocene date for the
beds in Nebraska. For these reasons, it seems preferable
to continue use of the name Rosebud for the basal Mio-
cene beds of the Big Badlands.

The Rosebud consists of a basal member, named
Rockyford by Nicknish (1957), up to 55 ft. thick,
composed of water-deposited volcanic ash with minor
clastic mud, which appears white in the field. The base of
the ash has been used for decades as an arbitrary marker
for the beginning of the Miocene. There is no evidence of
disconformity at this horizon, merely a sudden increase
in the amount of ash. Overlying the Rockyford Member
is a series of bedded, gray to tan siltstones with a high
ash content, the top of which is cut across by the Mis-
souri Plateau surface.

B. Regional Structural Setting

The regional structural setting of the Big Badlands
is much more complex and orderly than one might sup-
pose from the outcrop map, or from a hasty field inspec-
tion. Figures 6 and 7 make clear that the structures con-
sist of a series of parallel linear faults diminishing along
their strike into low folds and domes, at intervals of a
few miles, trending roughly N 70°W. The downthrown
side is generally south except for the Pine Ridge Struc-
ture, which is downthrown some 1200 ft. on the north
side. Regional dips between the Sage Fault on the
north and the Pine Ridge Structure on the south trend
south-southeast. The overall picture is an asymmetrical
structural trough, hinged along its northeastern margin
and downthrown to the southwest. Small structures
within the trough parallel the bounding faults. Local
sags and swells occur along all of the structures.

Since the Cretaceous shales and Tertiary mud-
stones involved are completely incompetent to transmit
stresses a distance of tens of miles, these linear move-
ments presumably constitute drape structures over dis-
placements in the Precambrian-Paleozoic basement.
They are part of a system extending south into Ne-
braska, where the direction changes to more nearly
east-west.

Geomorphologic evidence suggests that the Bad-
lands structures extend westward into the Paleozoics of
the Black Hills. The drainage of the east flank of the
Hills consist of a series of sub-parallel streams, tending
ESE, each of which, from Elk Creek south, passes
through the Paleozoic-Mesozoic strata via a structure-
controlled gap.

Fig. 6. Structural axes in the Big Badlands.

14

FIELDIANA: GEOLOGY MEMOIRS, VOLUME 5

Furthermore, the five divides between Bear Butte
Creek and Spring Creek are strongly asymmetrical,
with short, steep north-facing slopes and long, gentle
southern slopes. Plumley (1948, pp. 538-539) suggested
that southward tilting of the whole area after the present
drainages were established could have caused the asym-
metrical drainages. This seems improbable for the fol-
lowing reasons:

1. Table-top gravels occurring east of the present
upper course of Cheyenne River indicate that streams

CROSS -SECTION

tures along the east flank of the Hills, without regional
tilting.

6. Regional tilting can cause lateral migration of
streams only in the presence of alternately harder and
softer beds. These asymmetrical divides are composed
esentially of homogeneous Pierre Shale; some crests
include remnants of Oligocene mudstones for part of
their length, and frequently the crests consist of
?Pleistocene gravels a few feet thick. None of these
materials shows measurable south dips, nor are there

ALONG

LINE

IN

FIGURE 6

KEY
TM VARIOUS MIOCENE FORMATIONS

TOB BRULE
TOC CHADRON

KP PIERRE

KN NIOBRARA

Fig. 7. Cross-section along line A-A, Figure 6.

flowed eastward from the Black Hills at about the Mis-
souri Plateau level during early Pleistocene time. Cap-
ture of the Black Hills drainage by Cheyenne River was
a middle to late Pleistocene event.

2. The table-top remnants of the Missouri Plateau,
scattered throughout the Big Badlands, show no sign of
a southward tilt. Rather, they slope northwestward and
southeastward from an apparent old divide trending
ENE-WSW from the south end of Sheep Mountain to
the Pinnacles.

3. Missouri Plateau remnants preserved on top of
the Railroad Buttes (Sec. 6, 7, 8, T. 2 S., R. 11 E., Pen-
nington Co.), west of Cheyenne River between Spring
Creek and Rapid Creek, slope northward, not south. If
the entire area had undergone a southward tilt, the
plateau remnant on this asymmetrical divide should
slope southward.

4. A southward tilt of even one degree would, in the
40 mile distance from Bear Butte Creek to Spring Creek,
result in a vertical uplift of 3680 ft. north relative to the
south. Actual elevations in the north average about 100
ft. lower than southern elevations of corresponding
topographic situation and distance from the Hills.

5. Such an uplift would have caused notable dis-
placement of the Cretaceous formations involved. No
such displacements occur. Rather, the attitude and
position of all beds exposed are geometrically explicable
only in relation to the known major and minor struc-

any layers significantly harder than the mass as a whole.
Regional tilting of these sediments, even if it occurred,
would not cause drainage migration.

7. The northward course of Cheyenne River above
its junction with the Belle Fourche is known to be post-
middle Pleistocene. It seems almost impossible that a
northward-flowing stream could develop and capture
the drainage of an area which was being actively tilted
southward.

For these reasons, a Pleistocene southward tilting of
the area east of the Black Hills seems most improbable.

Two alternative explanations remain for the asym-
metrical divides between the parallel streams east of
the Black Hills. First, they may be a response to cli-
matic-vegetational controls. North-facing slopes at
present are steeper, and develop more trees but poorer
grass cover, than south-facing slopes. A combination of
wind direction, snow accumulation, and insulation could
cause vegetational differences which would favor more
rapid erosion on north or south slopes, as the case might
be. Many isolated buttes on the Tavaputs Plateau of
Utah demonstrate this situation excellently, although
under the Utah climatic regimen erosion is faster on
south slopes.

However, this hypothesis for asymmetrical divides
does not explain the general parallelism of the master
streams with each other and with the known trend of
regional structures in the Badlands. Also, it does not

CLARK: GENERAL GEOLOGY

15

explain the fact that these streams flow east-southeast-
ward rather than radially outward, normal to the moun-
tain front and the regional strike. It seems reasonable,
therefore, to propose as a structural hypothesis that the
linear structures observed in the Badlands continue
northwestward to the Black Hills. The major streams
emerge through structurally-controlled gaps in the Hills
and follow these old structural lines to the points of
their relatively recent capture by Cheyenne River. Since
the structures are draped normal faults and folds, gen-
erally downthrown a few feet or tens of feet south, the
fault planes dip south. The streams may have migrated
down the dip of the fault zones, producing asymmetrical
divides, or recent tilting of the individual blocks be-
tween faults might have occurred, or climatic-vegeta-
tional control of tributary gullies might produce asym-
metrical divides between structurally-controlled streams.
The second of these hypotheses seems least probable,
but I have not yet found evidence disproving it.

The question of whether the structural system evi-
dent in the Badlands extends westward to the Black
Hills is of subordinate interest to the present study. The
time or times of movement are of much greater signifi-
cance. Clark (1937, p. 289) suggested that a pre-Chadron
topographic ridge paralleling the Sage Fault was evi-

dence of pre-Chadronian movement along the fault;
displacement of early Miocene strata clearly indicates
post-early Miocene movement. The Pine Ridge Fault
brings the Niobrara Formation into contact with pre-
sumably middle Miocene beds. Absence of the entire
Pierre indicates a pre-Chadron uplift of at least several
hundred feet, erosion down to the Niobrara, deep
weathering of the Niobrara, deposition of the Oligocene-
Miocene group, then reactivation of the fault with at
least 300 ft. of post-middle Miocene movement. The
entire fault system has, therefore, undergone at least
two periods of activity, one pre-Oligocene (probably
Laramide) and one post-middle Miocene, probably part
of the general late Tertiary movements farther west
(Eardley, 1962, pp. 493-514).

Chadron and Brule sediments are known to have
overlapped the Oligocene Sage Ridge to the north and
the ridge to the south, which will be referred to as the
Pine Hills, in order to differentiate the Oligocene topo-
graphic feature from the Recent Pine Ridge. The influ-
ence of these two features upon Oligocene sedimenta-
tion suggests, however, that the higher parts of the Sage
Ridge were exposed at least through Orellan (early
Brule) time. Evidence for this will be presented in Chap-
ter VI.

Chapter IV

THE SLIM BUTTES FORMATION

In 1959, Malhotra and Teglund described a new
Tertiary formation, the Slim Buttes Formation, with its
standard section in the NW 34 of the SW % of Section
11, Township 16N, Range 8E, Harding County, South
Dakota. This is in the northwest corner of South Da-
kota, approximately 100-120 miles north of the western
part of the Big Badlands (Fig. 1). Since correlation of
strata in the Badlands with this formation in its stand-
ard area is not certain, and since the original description
of the Slim Buttes Formation was published in a journal
of extremely limited distribution, I shall here para-
phrase such of Malhotra and Teglund's description as
applies to the suggested correlation :

The Slim Buttes Formation occurs as a lens 0 to 40
feet thick, several miles in diameter, which outcrops in
the southern two-thirds of the Slim Buttes and in the
East Short Pine Hills. It rests disconformably upon
Paleocene sediments, and is overlain with apparent
conformity by the Chadron. The formation consists
essentially of white, cross-bedded sandstones. Thin
gravel bands consisting of chert and quartz pebbles up
to 19 mm diameter festoon the outcrops in many places.
The upper contact of cliff-making, white Slim Buttes
sandstone with overlying crumbly slopes of greenish-
gray Chadron mudstones appears sharp at a little
distance, but closer inspection of fresh exposures reveals
that gradational zones a few inches thick are common.
The standard section includes several strata of reddish,
brown, and purple mottled mudstone which contain
vertebrate fossils. Elsewhere, the sandstone is remark-
ably uniform, with only occasional thin laminae of pale
greenish clay.

Microscopically, the light fraction of the sand con-
sists of quartz and feldspar. Much of the latter has been
partially altered, producing pure white grains which,
with the pale greenish to white cementing clay, are re-
sponsible for the overall color of the rock.

The associated heavy minerals can be divided into
several groups, according to source:

I. Autochthonous
Barite
Limonite
Leucoxene
Hematite
II. Reworked from sedimentary rocks
Glauconite
Highly rounded garnet, tourmaline, zircon

III. Ubiquitous

Biotite, generally anhedral, but one specimen
euhedral

IV. Igneous or contact metamorphic
Magnetite, both euhedral and rounded
Apatite

Sphene

Zircon, clear to violet-gray

Tourmaline, black to clear, ranging from
euhedral to subhedral

Rutile, red, resinous, angular
V. Igneous or metamorphic

Hornblende, black to dark green

Garnet
VI. Metamorphic only

Staurolite, brown and resinous

Chlorite, green and platy
The tourmaline and garnet are abundant, hornblende
and biotite uncommon.

Mudstones and concretionary sands at the standard
section and within a mile to the east, south, and west
have yielded a fairly large fauna, which is at present
housed in the South Dakota School of Mines Museum of
Geology. (Due to extremely hard matrix and soft bone,
this collection was not fully prepared by the summer of
1964-J.C). Preliminary observations indicate the pres-
ence of an undescribed tapiroid, a large species of Epi-
hippus, Megalamynodon, Eotrigonias, and a titanothere
the size of Teleodus. Assuming that these tentative
identifications are correct, an Eocene-Oligocene tran-
sitional age is indicated.

This ends the paraphrase of Malhotra and Teglund's
description, which contains much significant matter not
included here.

My own observations indicate that a homogeneous
mixture of fairly well-sorted sand with a white clay mat-
trix characterizes the Slim Buttes formation in its
standard area. Feldspar grains the same size as the
quartz grains are all altered to a white clay, and some of
the white matrix shows fragmentary ghosts of feldspar
cleavages. It seems probable that this texture indicates
syngenetic or epigenetic alteration of an originally well-
sorted, highly arkosic sand. Neither the crossbedding
nor the sorting resembles that of the known Tertiary
fluvial deposits of South Dakota; I am therefore in
agreement with Malhotra and Teglund that these sedi-
ments may well be lacustrine. However, I am not con-
vinced by their evidence that the source of the elastics

16

CLARK: THE SLIM BUTTES FORMATION

17

lay to the west or northwest, and at present consider a
Black Hills source more probable.

Sediments which occupy the same relative strati-
graphic position as the Slim Buttes Formation occur in
five areas in the Big Badlands (see Fig. 4). In all cases
they rest disconformably upon the Interior weathered
zone of the Pierre Shale. In all cases they are overlain
by the Chadron, but at least two localities offer definite
evidence of a time lapse between deposition of the two
formations.

The five localities differ sufficiently to require sepa-
rate description.

Locality I. Western Badlands, Exposed in upper
Battle Creek Draw, in section 12, and in Battle Creek
Canyon, sections 10 and 15, T 42N, R 46W, Shannon
County.

Two lenses of coarse white sandstone up to 20 ft.
thick and 400 yd. in diameter. A very coarse conglomer-
ate of quartz and chert pebbles, up to 6 in. greatest
diameter, lies at the base and crossbedded into the
lower several feet of the thicker sections.

Individual pebbles of rose quartz and "Fairburn"
agate plainly indicate a gravel with its source in the
southern Black Hills ("Fairburn" agates are placer
agates derived apparently from the Hell's Canyon agate
nodules in the Pahasapa Formation, Mississippian, of
the southern Hills.) The rose quartz clearly proves that
the Precambrian pegmatites of the southern Hills had
been exposed by the time these pebbles were eroded
from their parent outcrops. Complete absence of granite
and feldspar pebbles indicates active syngenetic weath-
ering, producing a mature gravel within 30-40 miles of
its source. The contrast with richly arkosic Chadron and
Brule gravels exposed in the same area is striking.

The petrology of the sands in these lenses has not
yet been studied.

No fossils, other than a few obviously transported
pieces of silicified wood, have been collected from this
locality.

The lenses are overlain by greenish Chadron mud-
stones, with a sharp contact.

Locality II. Northeast of the town of Scenic, center
of section 11, T3S, R 13E, Pennington County, (FMNH
specimen G4007).

One lens, up to 8 ft. thick, of greenish-white to
yellowish-white sandstone, with a basal conglomerate of
chert pebbles up to 2 in. in diameter. The lens is exposed
over an area of less than 100 yards; badlands gullies
surrounding it reveal that this is about its maximum
diameter. The overlying greenish Chadron consists of
sandy siltstones, and rests with an erosional contact
upon the fine-grained sandstone. There is no actual
trenching or gullying, but the lens seems to feather out
by erosion of its upper surface rather than by confine-
ment within a small depositional basin.

The sandstone is homogeneous and well-sorted, but
bimodal, with the principal mode at about 0.125 mm
and the secondary mode in the clay sizes. Grains are

generally sharply angular, with a few very well rounded
and slightly frosted ones. These might be presumed to be
inherited if the extreme wear were restricted to quartz
grains, but such is not the case. Possible alternative
hypotheses will be discussed after detailed description
of localities has been completed.

The light fraction of the sand consists of quartz,
angular chips of yellow, orange, gray, and brown chert,
a small percentage of gray, sugary quartzite, and
scattered grains of white, glossy material which is almost
completely weathered feldspar.

Barite both euhedral and interstitial, dominates the
heavy fraction. Next in abundance come magnetite
and a black, metallic mineral which is not sufficiently
magnetic to adhere to a hand magnet. Both of these last
vary from very well-rounded, spherical grains to euhe-
dral double pyramids and sharply angular fragments.
Next in abundance are shards of pink garnet and small,
euhedral crystals of colorless, pale yellow, and pink
zircon. Smoky to olive tourmaline is the only other
common heavy mineral. The rarer minerals include
yellow-brown allanite; dark green, clear, well-rounded
but usually columnar grains of actinolite; and red,
columnar transparent grains, both well-rounded and
angular, of either rutile or cassiterite. A few grains of
blue, transparent indicolite have also been observed.

The quartz grains 0.125 mm and smaller fall into two
sharply differentiated varieties; colorless, and pale to
bright yellow citrine. Both varieties include both highly
angular and well-rounded grains. The large-diameter
fractions of the sample, above 0.125 mm, consist of
limonite nodules, yellowish chert, and a very few angular
quartz grains with slightly etched or abraded surfaces.

Petrologically, the assemblage is notable for imma-
turity, absence of micas, and a high percentage of
minerals from igneous rocks as opposed to those from
metamorphics or from older sediments. It differs from
the Slim Buttes assemblage in the absence of leucoxene,
glauconite, apatite, biotite, sphene, hornblende, and
chlorite. Some of these differences may be due to over-
looking of rare, single grains in the Scenic sample, but
even if so, the relative proportions of the minerals named
must be sufficiently lower to be significant.

Absence of feldspars from the gravel, deep weath-
ering of the few feldspar grains remaining in the sand,
and the homogeneous mixture of greenish-white clay,
presumably derived from in situ weathering of feld-
spars, with otherwise well-sorted sand, are the only
indication of active syngenetic or epigenetic chemical
decay.

No fossils have been found here.

Locality III. Southeast of Wall. A major area of
exposure at the intersection of sections 25, 26, R. 16E.,
and 30 and 31, R. 17E., T. IS., Pennington County.
The total area of exposure is roughly two-thirds of a
mile in diameter. A remnant outlier is exposed on the
southwest flank of a butte, in the NE 1/4 of the SE 1/4
of the SE 1/4, Sec. 10, T. IS., R. 16E. (FMNH specimen
4076).

18

FIELDIANA: GEOLOGY MEMOIRS, VOLUME 5

The larger area comprises horizontally-bedded strata
approximately 35 ft. thick. The lower portion is a white
to yellowish green, fine-grained sandstone. Chert and
some quartz pebbles up to 3 in. diameter make up a
basal conglomerate, near the top of which the pebbles

Fig. 8. Slim Buttes-Chadron contact zone, showing (middle of
picture) darker Chadron clays penetrating joints and cracks in the
light-colored Slim Buttes sandstones. Vertical white streaks are
pick marks, made during removal of weathered surface material.

are not over 2 in. in diameter. This is overlain by a zone
1-2 ft. thick of a peculiar clinkery sediment which varies
from an exceedingly impure limestone to calcareous,
gritty, gray material with white flecks. Fifteen to twenty
feet of pale yellow, pale gray, and brighter reddish-
purple mudstones overly the clinkery zone. A second
clinkery zone caps the formation, overlain directly by
40 ft. of grayish Chadron clays.

The outlier consists of 10 to 15 ft. of medium-
grained, greenish-white sandstone including one stringer
of purplish sand and clay. It also is overlain with ap-
parent conformity by grayish Chadron clays.

The sediments of this locality have not yet been
analyzed petrologically. No fossils have been found in
them.

LOCALITY IV. Southwest of Interior, in Section 12,
T. 4S., R. 17E., Jackson County and in Sections 23 and
24, T. 43N., R. 40W., Washabaugh County. This is the
area described and figured by Clark (1937, p. 277, also
Figs. 5B and 8).

The outcrops north of White River, in section 12,
never exceed 10 ft. in thickness. They consist of a basal
conglomerate of chert and quartz pebbles up to 2 in. in
diameter, overlain by greenish-white, massive, poorly-
cemented sandstone. The sandstone is fractured into
angular blocks several inches in diameter; greenish clay
from the overlying Chadron and purple clay from the
underlying Interior zone have intruded the fractures,
producing a mosaic effect.

The sandstone has a considerable content of white
to greenish-white clay, which is high in kaolinite,

montmorillonite and fine-grained quartz. A few identi-
fiable feldspar grains remain, but many more are weath-
ered almost beyond recognition.

The large outcrop area south of White River includes
a basal conglomerate and cross-bedded to massive lenses
of coarse to fine, greenish-white sandstones. Lenses up
to 3 ft. thick of exceedingly plastic, pale gray clay, also
characterize this locality. The clays consist of fine-
ground quartz, kaolin, and subordinate quantities of
montmorillonite and mixed-layer clay.

No fossils have been found at this locality, and the
detailed petrology has not been adequately studied.
(FMNH specimens G3729, G3730, G3731).

Locality V. North of Weta; sections 1, 2, 3, 11, 12,
14, T 3S, R 20E, Jackson Co. and scattered localities to
the east.

Erosion of the white sandstones previous to deposi-
tion of the overlying, typical Chadron sediments is
clearly demonstrated at this locality. Figure 5, looking
west from the road one-quarter mile north of the south
edge of section 2, shows white sandstone north and south
of a central gap, with sloping shoulders of Chadron cut-
ting it out completely. Outcrops not visible in the
picture have Chadron resting directly upon Interior
Zone clays. The top of the white sandstone is fractured,
and the fractures have been intruded by Chadron clay,
as at Locality IV. Figure 8 shows the fracture zone as
exposed by stripping off the weathered surface on the
shoulder of the northern side of the pre-Chadron eroded
gap in section 2.

The basal conglomerate reaches a thickness of 3 ft.,
with bands of smaller pebbles cross-bedded into the
sandstones as much as 6 ft. above the main gravel mass.
Red and yellow jasper, and "eye agates" of gray and
black chert comprise the bulk of the gravel. The largest
pebbles are about 3 in. in diameter. Less than 1% of
the pebbles over 2 in., and less than 3% of the pebbles
3^-2 in. in diameter, are composed of quartz. Rose
quartz, Fairburn type agates, and black-and-white,
banded, schistose quartzite are all absent here, in
contrast with Locality I where they comprise the bulk
of the gravel.

The white to greenish-white sandstone varies from
coarse to fine, with more greenish color on the finer
material. As elsewhere, the interstitial white clay and
the white clay granules show ghosts of feldspar grains.
Altered feldspar grains form an appreciable proportion
of the larger-sized light sands, but are absent in the
finer sizes.

Microscopically, the light fraction consists of quartz
of two types, colorless and lemon-yellow citrine. Both
types occur usually as sharply angular grains, with a
few highly spherical, well-rounded and frosted ones.
The grains show much less abrasion and frosting, es-
pecially in the sizes greater than 0.25 mm, than do those
at Locality III. Subordinate quantities of yellow,
orange, gray, and brown chert, and a few grains of gray,
sugary quartzite, characterize the +0.25 mm fraction.

CLARK: THE SLIM BUTTES FORMATION

19

The heavy mineral suite differs notably in propor-
tion from that of the outcrops near Scenic. Euhedral to
subhedral barite is the principal constituent; apparently
it is autochthonous and functions as a cement. Of the
clastic heavy minerals, magnetite and a black, metallic,
nonmagnetic mineral resembling it are most abundant;
both include grains varying from euhedral double pyra-
mids and freshly fractured granules to very well-rounded
spheres. Euhedral zircons, colorless, pale yellow, and
pink, are next in abundance. Pale orange to pink garnet
and smoky-olive tourmaline (which would probably be
black if seen in larger pieces) occur in roughly equal
proportions. Clear, dark green, well-rounded prismatic
grains of actinolite, and transparent red columnar
grains of rutile (or cassiterite?) are uncommon. The
rutile includes both angular and well-rounded grains.
Very rare grains of yellow-brown allanite and glassy
blue indicolite complete the list of minerals identified to
date.

The Weta assemblage differs from that near Scenic in
the absence of staurolite, abundance of magnetite and
other metallics, relative decrease in tourmaline, and
considerable decrease in garnet. However, changes of
proportions of heavy minerals cannot be interpreted
arbitrarily in this case, because the sample from Scenic
occurred in a fine-grained sandstone and the sample
from Weta in a much coarser-grained rock.

The abundance of magnetite, the presence of rutile
and indicolite, the absence of biotite and of micas gener-
ally, the presence of smoky tourmaline in some quantity,
and the presence of zircon in some quantity, are all
features in which these rocks differ from the sandstones
of the overlying Chadron. Smoky tourmaline occurs in
abundance in many Chadron sands, but not associated
with abundant magnetite.

The outcrops in section 2 include a bed of red-brown
mudstone (see Fig. 5) near the top of the formation.
This situation resembles that at the type section of the
Slim Buttes formation. A very careful search was made
for fossils and two partial ribs of an animal the size of a
large Trigonias or small titanothere were discovered. No
other fragments of vertebrates have been found.

The basal conglomerate in section 2 includes pieces
of fossil wood up to 2 ft. long. Many of these have,
unfortunately, been removed by local rock collectors,
but a few are in the collection of the South Dakota
School of Mines and Technology, and one, G3729, is in
the Field Museum of Natural History. Since these
specimens are many times larger than the largest clastic
pebbles in the conglomerate, it is obvious that they
were not mechanically transported as boulders from
some other formation. The pieces of unfossilized wood
must have floated in and lodged in their present posi-
tions, but whether they came from a neighboring or a
distant source is not known. The wood has been
tentatively identified as cedar, and of no value in precise
age determination.

A sixth area (marked VI on the map, Fig. 4) consists
of Slim Buttes sands reworked into the basal Chadron,

with so little admixture as to suggest that the Slim
Buttes material was transported only a very short dis-
tance.

General conclusions: Whatever their age, the present
lenses seem to represent remnants of a formerly wide-
spread sheet of elastics. They resemble each other and
the Slim Buttes Formation in the following respects:

1. Distribution as scattered lenses.

2. Stratigraphic position.

3. Relatively good sorting.

4. Grain size generally within the limits of sand.

5. Much better sorting than other Oligocene sedi-
ments.

6. Bimodal size distribution, with maxima in the
medium sand and clay sizes.

7. Material either massive or crossbedded, but only
the rare beds of a really persistent mudstone flat-bedded.

8. White color, due to weathered feldspar grains and
cement of kaolin.

9. Clays with relatively higher content of kaolin and
silica, and lower montmorillonite and mixed-layer, than
other Oligocene sediments.

10. Feldspars highly weathered, in contrast with
other Oligocene sediments which contain predominantly
fresh feldspars.

11. General absence or low percentage of micas.

12. Absence of fresh or weathered pumice shards.

13. General absence of fossils, except in the rare
local, dark-colored mudstones.

14. Presence of both angular and highly rounded
quartz grains.

In view of these points of similarity, and in the ab-
sence of any evidence to the contrary, the white lenses
of the Big Badlands are tentatively correlated with the
Slim Buttes Formation.

The sorting, bedding, and grade of sediment suggest
that this formation is lacustrine. Also, the presence of
both angular grains and well-rounded, spherical grains of
quartz, citrine, and tourmaline at Locality V cannot be
satisfactorily explained by assuming that the worn
grains are inherited. Inheritance of well-rounded grains
of the same varieties of quartz and tourmaline as occur
in first-generation elastics at the same spot, within 75
miles of the source, would be almost impossible. A much
more logical explanation would be that certain grains
have been rounded by wind and wave action along
beaches, while others were introduced immediately prior
to final deposition and suffered no appreciable abrasion.

This does not mean that the Slim Buttes Formation
represents one continuous lake covering most of western
South Dakota. If the whole area were somewhat below
grade, it might be expected that a series of lakes would
form, each receiving sediment from its respective feeder
stream. The similarities noted between the sediments at
different localities are mostly attributable to coeval
deposition in similar lacrustrine environments under a

20

FIELDIANA: GEOLOGY MEMOIRS, VOLUME 5

uniform climatic regimen, with the Black Hills as the
source of elastics. Differences are consistent with differ-
ences in the local headwater districts of the Black Hills,
plus small differences in the province of deposition.

Judging from the paucity of micas and montmoril-
lonite, and the absence of glass shards, little or none of
the Slim Buttes formation is composed of pyroclastics.
This suggests that deposition occurred at a time either
when vulcanism was relatively inactive or when the
prevailing winds carried the major ash falls away from
western South Dakota.

The deep weathering of feldspars, with the develop-
ment of a high percentage of kaolinite relative to mont-
morillonite, suggests a climate more consistently humid
and probably warmer than that of Chadron time.

It must be freely admitted that the presence of

lacustrine deposits over so wide an area poses a serious
paleogeographic problem. The extremely large size of
pebbles in conglomerates many tens of miles from their
source offers a second difficulty. Finally, the fact that
lenses occur on both sides of the Sage Ridge, as well as
on the upland bordering the Red River Valley (which
was old by earliest Chadronian time), is not at present
explicable. In the absence of any general interpretation
which would bring these facts into accordance with
Chadronian history, I have simply interpreted the sedi-
ments as they occur and let the anomalies stand. This
seems a sure way of exposing my interpretations to
searching analysis. Naturally, I intend to study these
rocks further. It is hoped that this chapter will incite
others to investigate them independently from fresh
points of view.

Chapter V

GEOLOGY, PALEOECOLOGY, AND PALEOCLIMATOLOGY
OF THE CHADRON FORMATION

John Clark and J. R. Beerbower

INTRODUCTION

The fauna of the early Oligocene Chadron formation
of South Dakota represents a transition between the
Eocene jungle forest faunas and the later plains faunas
of western North America. Therefore, a study of this
fauna is of particular importance in clarifying the nature
of major faunal changes and in determining some of the
factors controlling these changes. The senior author
began this work with a detailed description of Chadron
stratigraphy (Clark, 1937), but an adequate faunal
characterization was possible only for the upper member
of the formation at that time. In field studies during
1940, 1941, 1946, 1953, and 1954, the senior author ob-
tained sufficient additional information and specimens
to justify an extended interpretation of Chadronian
paleogeography, paleoclimatology, and paleoecology,
and of the relationships of South Dakota faunas to
certain other Oligocene assemblages.

The Chadron Formation as now denned comprises
three members (Clark, 1954) :
Peanut Peak 20-30 ft.
Crazy Johnson 20-40 ft.
Ahearn 0-80 ft.
Within the Big Badlands, the formation ranges in thick-
ness from 8 to 130 ft. The Ahearn is definitely absent
and the two upper members are not separable in the
area of minimum thickness; probably only the Peanut
Peak member occurs there.

The known fauna includes a wide variety of mam-
mals plus poorly-known fish, amphibians, and reptiles. In
general, fossils are more fragmentary and less abundant
than in the overlying Brule Formation; many species
are known only from a few specimens.

This chapter was originally written to be published
as a separate paper; it was completed in 1957. It has
been withheld since that date, pending completion of the
other chapters of this study. The authors have modified
the 1957 manuscript only by bringing the taxonomy up
to date and by modifying certain interpretations in view
of pertinent studies by others. No attempt has been
made to establish detailed correlations with formations

and faunas unknown in 1957, or to include specimens
collected since then. This would entail completely re-
organizing the study, which would unduly delay the
whole project.

ACKNOWLEDGMENTS

The Carnegie Museum of Pittsburgh sponsored field
work in 1940, 1941, and 1946. Princeton University and
the Yellowstone-Bighorn Research Association spon-
sored the field work in 1953 and 1954. The research was
aided by Geological Society of America Grant 658-54.
Messrs. A. D. Lewis, H. O. Woodbury, L. Stagner, H.
Stoll, T. Harrison, and D. T. Taylor served most com-
petently as field assistants.

Studies of Chadron paleogeography and sedimentary
petrology were carried forward during 1963 and 1964,
under the sponsorship of the Field Museum of Natural
History; Mr. Kenneth K. Kietzke served as field as-
sistant.

The authors greatly appreciate the co-operation of
geologists of the staffs of Princeton University, Lafay-
ette College, the Field Museum of Natural History, the
University of Chicago, and Harvard University. These
gentlemen have so freely given of their time and experi-
ence, through numerous informal conferences, that it is
impossible to credit each with the ideas he has con-
tributed. The authors also wish to thank Mr. E. H.
Taylor and Mr. D. T. Taylor, of Scenic, South Dakota,
for their hospitality in permitting the use of their ranch
as field headquarters.

STRUCTURAL RELATIONSHIPS

The Ahearn Member of the Chadron fills an old
valley 70 90 ft. deep and approximately four miles wide,
cut in the Pierre Shale. The valley walls slope not over
2-5°, and consist of the Interior zone weathered to
depths of 30-50 ft. (Figs. 2 and 4).

The Crazy Johnson and Peanut Peak Members can
be distinguished from each other only where they overly
this old valley. Laterally on both sides, they extend as a
25-40 ft. blanket of bentonitic mudstones to the struc-

21

22

FIELDIANA: GEOLOGY MEMOIRS, VOLUME 5

ture-controlled Sage ridge and Pine Hills. The Chadron
thins locally to 8 ft. over the crest of the Sage Ridge, and
in one small area it fails to cover the crest of the Pine
Hills ridge. It thickens again to over 30 ft., north of the
Sage Ridge and south of the Pine Hills.

The Chadron is everywhere conformably overlain
by the Brule.

Occasional sedimentary dikes a few inches in width
and less than one-half mile long transect the Chadron.
They occur singly and are too few to give satisfactory
statistical evidence for or against a preferred orienta-
tion. The fill consists either of identifiable Chadron
material from lower strata cut by the dike, or of ma-
terial probably Chadron but not surely identifiable.
Chalcedony veins extend downward from the lower
Brule into the upper few feet of the Chadron at many
places. Genesis of the fissures occupied by the dikes and
veins is not known, but it probably is not a reflection of
basement structures.

LITHOLOGY

This section is based upon field observations plus
laboratory studies of the following specimens:

G3635
G3637
G3638
G3733
G3734
G3735

G3736
G3740
G3741
G3941
G3942
G3943

Ahearn Member

G3944
G3945
G3946
G4013
G4019
G4102

The basal conglomerate of the Ahearn member oc-
cupies the middle of the old Red River Valley, but thins
to disappearance along the margins. It reaches total
thickness of 3-4 ft., is poorly bedded, and usually very
poorly sorted. Cementation is also usually poor; many
of the sandier patches can be dug with the fingers. The
cement is calcite, occasionally intergrown with a little
silica. At several places the cement directly at the Pierre
Shale contact is pyrite, apparently produced by the
reaction of sulfates in the Pierre with acid Chadron
waters.

The conglomerate comprises pebbles from three
sources: (1) remanie chert and quartz pebbles reworked
from the Slim Buttes basal conglomerate; (2) sand and
gravel directly derived from the Black Hills; (3) oc-
casional pebbles and limonite concretions from the
underlying Pierre.

The pebbles over about 20 mm diameter are all, ap-
parently, secondarily derived from the Slim Buttes con-
glomerate. All pebbles above that size consist of chert,
quartz, and quartzite, while the smaller gravel in the
same outcrops is richly feldspathic, with unweathered
granite pebbles. Since it is impossible to distinguish
primary from inherited chert and quartz grit or sand,
the relative amounts of each in the finer grades cannot
be estimated. Failure to recognize the dual source of
sediment led the senior author in 1937 (p. 279) to believe

that increase in percentage of feldspar upward in the
Chadron section was climatically controlled. He now
believes that this increase in feldspar reflects progres-
sively less dilution with Slim Buttes material, upward
in section.

Above the basal conglomerate is 0-10 ft. of poorly-
cemented olive green sand, overlain by a red, silty clay,
0-12 ft. thick, which includes bands of intraformational,
red shale-pebble conglomerate. This basal sequence
generally totals about 20 ft. thick. It is not uniform in
development and is by no means sharply separable from
the superjacent beds.

The overlying strata consist mostly of cross-bedded,
greenish silty sandstones, streaked with pink and lav-
endar. Cross-bedding dips up to 30° are not rare, but
the general average is 20-25°. Individual cross-bed sets
generally reach dimensions of 100-300 yards laterally
and 5-15 ft. vertically. The cross-beds dip east and
southeast; a few dip northwest or south. Most of these
sands contain considerable muscovite. The bentonite
content varies from large amounts to none. Interbedded
with these are channel fills of arkosic conglomerate,
which project as irregular ledges among the vertically
fluted, organ-pipe columns of cross-bedded silty sand-
stones.

The upper 20 ft. of the Ahearn Member is made up
of light tan to orange siltstones, mostly poorly cemented,
with numerous scattered concretions less than 2 in.
diameter, of mixed calcite and limonite cement. Green-
ish, conglomeratic channel-fill sandstones are much
smaller and less common than in the underlying strata,
and are, in general, finer grained. In general aspect, the
upper Ahearn resembles the Peanut Peak Member.
Fossils are more abundant than in the lower part of the
Ahearn, and some coprolites occur. The top of the
Ahearn was in some places slightly eroded to depths of
2-3 ft. before deposition of the overlying Crazy Johnson
sandstones.

Each of these lithologic subdivisions coarsens west-
ward, without altering the general vertical gradation
from relatively coarser to finer materials. Thus, at
Indian Creek the basal conglomerate of 3 in. pebbles
is overlain by sandstone lenses containing % in. peb-
bles and these by silty claystone. Five miles to the west
in Big Corral Draw, the basal conglomerate of 6 in.
pebbles is overlain by sandstone lenses with 134 in.
pebbles and these by sandy siltstones.

Two lenses, both near the top of the Ahearn Mem-
ber, merit separate description.

The first is a bed of pure ash, approximately 3 ft.
thick and 150 ft. long, N-S, which is exposed beneath a
sandstone ledge in the SE \i of SE %, Sec. 4, T. 4S.,
R. 10E., Custer Co. Bedding is non-evident; planes of
parting develop upon breaking the rock out. Micro-
scopically, the material consists of abundant biotite,
quartz, and devitrified glass, with subordinate oligo-
clase and apatite. The biotite and feldspar are both
perfectly fresh. No isotropic glass has been observed.

CLARK AND BEERBOWER: THE CHADRON FORMATION

23

Rare, very small zircons complete the list of minerals
identified.

The second is a lens 0-3 ft. thick, 20 ft. wide, and at
least 1/2 mile long, outcropping in the NW 1/4 of NW
1/4, Sec. 13, T. 4S., R. 12E., Pennington Co., in the
drainage of Indian Creek, approximately 12 miles east
of the first lens. It consists of pure white bentonite with
abundant biotite. The percentage of oligoclase to quartz
is higher than in the ash lens, and zircon has not been
observed. Apatite occurs about in the same proportion
as in the first lens.

Presumably, these two lenses were formed by ash
falls into temporary ponds or lagoons. The difference in
mineralogy suggests that two falls are represented rather
than one. The absence of elastics and the generally
monolithic character of both deposits indicate that each
was produced by a single fall. This gives a useful meas-
ure of the amount of fine ash deposited by single falls,
as well as of the local relief on a late Ahearnian deposi-
tional surface. The climatological environment neces-
sary for devitrification and bentonitization of the glass,
without noticeable weathering of either biotite or feld-
spar, is not known.

Crazy Johnson Member

In its central area of outcrop, where it overlies the
Ahearn, the base of the Crazy Johnson Member con-
sists of sandstones or arkosic conglomerates. Where it
rests upon the Interior zone, the base is usually either
greenish mudstone or nodules of white calcium carbo-
nate. Greenish to bluish bentonitic siltstones comprise
the bulk of the member.

Cross-bedded sandy siltstones, resembling those of
the Ahearn Member in all respects except that they lack
red color, commonly occur in the central area, associated
with numerous thin, flaggy sandstones and arkosic
conglomerates. The conglomerates and sandstones de-
crease in grade eastward. Maximum diameter of pebbles
in the conglomerates at Quinn Draw (Sec. 14, T. 4S.,
R. HE.) is 2 in.; 5 miles east at Indian Creek (Sec. 2,
10, 11, T. 4S., R. 12E.), it is 3/4 in.; at Cain Creek
(Sec. 21, T. 43N., R. 42W.), 18 miles farther east, about
1/4 in.; and east of Weta (Sec. 14, T. 3S., R. 20E.), 48
miles farther, the coarsest sand is 0.2 mm. This size
diminution does not represent that along the course of a
single stream, since the heavy-mineral content differs
significantly (see below) between some of these localities.
However, it does represent transportation at increasing
distances by contemporaneous streams of comparable
size with a common source, the Black Hills, and common
regimen.

Pseudoconglomerates are a characteristic feature of
the Crazy Johnson Member. They consist of structure-
less lumps of gray, impure calcium carbonate up to 10
in. in diameter, which are jumbled in a matrix of fine,
sandy siltstone. The lumps resemble cobbles of soft,
massive limestone rather than concretions, are rounded,
lack ramifying extensions, and are never intergrown.
Since the Crazy Johnson Member contains no other

pebbles larger than 2 in. diameter, and since the matrix
of the pseudoconglomerates is fine-grained, these lime
cobbles are probably autochthonous rather than stream-
transported boulders of an older limestone. Their
roundness and soft surface suggest that they are chunks
of calcium carbonate ooze, broken up before complete
induration by flood invasion of the shallow swamps or
ponds in which the ooze was accumulating. Unfortu-
nately, no unbroken deposits of this type have been
found : the known pond limestones show extensive algal
structures and contain fossil snails or mussels. The
origin of these cobbles is, therefore, obscure, and the
term "pseudoconglomerate" is used for them in prefer-
ence to the much more definitive "intraformational con-
glomerate." In any case, these pseudoconglomerates,
more than any other sedimentary type, are associated
with the famous "graveyards" of titanothere bone which
occur at the base of the Member.

The senior author described in an earlier paper
(Clark, 1937, p. 301) a humus zone in association with a
titanothere graveyard and a pond limestone. Re-ex-
amination has shown that the supposed humus is a zone
stained with diffused, powdery manganese dioxide.
Study of the other known dark zones within the area
has demonstrated that they all derive their color from
manganese dioxide, and that no humus is preserved in
the Chadron of the Big Badlands.

The Crazy Johnson Member thins westward to less
than 20 ft.; in contrast, the Ahearn thickens to 80 ft.,
but the total thickness of the two remains approximately
100 ft.

Both north and south of its central area of coarser
sediment, overlying the Ahearn Member, the Crazy
Johnson sediments grade rapidly to pale bluish, green-
ish, and gray bentonitic mudstones with occasional im-
pure limestones of small areal extent, a few inches
thick. These sediments blend indistinguishably with
those of the overlying Peanut Peak Member to form a
single unit.

Peanut Peak Member

The Peanut Peak Member rests comformably on the
Crazy Johnson Member; its thickness ranges from 20-30
ft. A pond limestone, flaggy sandstone, or channel fill
conglomerate usually marks the base of the Peanut Peak
in the central area of deposition; north and south of this
district, as described above, the two members are not
separable.

The Peanut Peak comprises light tan to greenish
silty clays, mottled with orange in many places. In the
western part of the central area, the entire member is
orange for distances of more than a mile. Sandstones and
channel fills are notably smaller, less numerous, and
finer-grained than in the immediately underlying Crazy
Johnson, over the entire area from Big Corral Draw
eastward. To the west, a few major channel fills persist
throughout the member, and into the overlying Brule,
carrying pebbles as large as any in the underlying Crazy
Johnson Member.

24

FIELDIANA: GEOLOGY MEMOIRS, VOLUME 5

The upper few feet consist of silty mudstone inter-
bedded with thin, discontinuous limestones. Some of the
latter form definite lenses a few hundred feet in diame-
ter, composed of algal limestone with associated snails
and ostracods. Others are less than an inch thick, un-
fossiliferous, highly irregular, and anastomose from one
local horizon to another. These apparently represent
lime concentrations or caliches formed under semi-arid
conditions, while the fossiliferous limestones are actual
pond deposits. The contact with the overlying Brule is
almost everywhere a pond limestone.

General Observations

The Chadron Formation exhibits foreset types of
cross-bedding throughout most of its central area of
deposition, exclusive of channel fills proper. The local
depositional relief indicated is up to 10 or 12 ft. The
average grade of clastic sediment is notably coarser than
that of the succeeding formations, and finer than that
of the Slim Buttes formation. The cross-bedding some-
what resembles that of the Slim Buttes, but differs from
the horizontal bedding which characterizes the inter-
fluve deposits of the succeeding formations.

Individual channel fills can be differentiated only
rarely and for short distances. When they can, as at the
standard section of the Crazy Johnson Member (SE 1/4,
Sec. 10, T. 4S., R. 12E.), they exhibit cutbanks with a
maximum slope of 60° and 5 to 10 ft. of relief. The great
majority have banks which slope 1-5°, and a relief of
3-5 ft.

The channel fills show division on a scale of indi-
vidual channelways 1-5 ft. wide, and usually 1-3 in.
deep, immediately contiguous to each other. That is, a
cross-section normal to the direction of flow transects
any one stratification surface in a continuous curve
reversing direction, like a series of open S-curves lying
on their sides. The interchannel bars are neither flat-
topped nor larger than the channels. The channel sedi-
ments usually form sand waves with rounded crests, of a
few inches wave length, which V downstream. No de-
formed cross-strata, oversteepened dips, or recumbent
folds have been observed. Figure 6 indicates directions
of channel fills. No true meanders occur and only a few
open bends of more than 45° change in direction have
been noted.

The sands fall into two sharply different mineralogic
groups. All of the Ahearn sands, and all but one of the
Crazy Johnson-Peanut Peak sands in the main area of
outcrop contain abundant black to smoky-olive tourma-
line, pink to gray garnet, and brown staurolite, with
subordinate glauconite, muscovite, and rounded, pearly
chips of fossil bone. Pyrite, limonite, and barite may be
absent to common, and seem to be authochthonous
when they occur. Biotite is rare, although it occurs quite
commonly, as does tourmaline, as tiny inclusions in
quartz grains. This assemblage was determined by
Seefeldt and Glerup (1958) to be typical of sediments
derived from the southern Black Hills, from the Harney
Peak area southward.

The light portion of the large sizes (above 250m)
consist of angular to subangular quartz grains; pink,
yellow, and greenish feldspars; gray, quartzose schist;
occasional yellow to gray chert; and greenish clay pel-
lets.

In the smaller grades, some very well rounded,
frosted quartz grains appear, presumably inherited from
older sandstones in the Black Hills. The pink, cream-
colored, and greenish feldspars show considerable al-
teration to white kaolin, increasing in the smaller sizes
until very few recognizable feldspars remain. Colorless
feldspar grains with a microperthite structure, whose
optical properties place them in the oligoclase-sanidine
range, show no alteration at all. The suggestion is that
the clastic feldspars were altered during erosion and
transport, while the volcanic feldspars underwent quick
burial at the approximate site of their airborne arrival.

A second group of sandstones is represented by two
outcrops: one channel fill 75 ft. wide and 15 ft. thick,
trending S 20° E., at the top of the Crazy Johnson
Member 2y2 miles west of Scenic, NW % of NE \i,
Sec. 19, T. 3S., R. 12E., Pennington Co.; and one chan-
nel fill trending SSE, in the SE H of Sec. 14, T. 3S.,
R. 20 ., Jackson Co., \x/l miles east of Weta, prob-
ably in the Crazy Johnson Member (individual mem-
bers cannot be distinguished here).

The heavy minerals of this group are much more
varied than in the southern-derived group. The list
includes:

Epidote, greenish yellow

Magnetite, euhedral to rounded

Actinolite, greenish, acicular

Sphene, lemon yellow

Garnet, pink

A black, nonmagnetic, isometric mineral

Biotite (not common)

Limonite (not common)

Glauconite (not common).

In addition to these, which are common to both
localities, the Scenic locality contains:
Hornblende
Muscovite
Gypsum.

The Weta locality contains the following which have
not been identified at Scenic:

Staurolite

Barite (autochthonous)

Black tourmaline

Colorless zircon

Rutile

Pyrite (autochthonous)

Waxy, red hematite

Columnar, colorless apatite

Gold (very rare).
Ritter and Wolff (1958) demonstrated that, in the
Brule sandstones, a suite including abundant magnetite,
lemon-yellow sphene, actinolite, and epidote is charac-
teristic of sediments derived from the northern Black

CLARK AND BEERBOWER: THE CHADRON FORMATION

25

Rate of Deposition of the Chadron

BRULE

This graph is approximate only. Tims intervals of rapid
deposition moy be much shorter, and intervals of slow depo-
sition to non-deposition longer, than shown here.

Fig. 9. Rate of Deposition of the Chadron Formation.

Hills. The staurolite and black tourmaline in the Weta
sample may indicate an admixture of materials by con-
fluent streams from two sources, or may indicate re-
working of southern-Black Hills-derived sediments of
Slim Buttes or Ahearn member age by these later
Chadronian streams. It is not possible to determine
whether the rutile and zircon of the Weta sample are
primary or are inherited from the Slim Buttes forma-
tion in the immediate area, which is known to have been
eroded by Chadronian streams (see Fig. 9).

The light minerals consist primarily of quartz and
feldspars. The quartz comprises angular grains, both
milky and clear, with minor quantities of citrine, but
also many clear grains bearing yellow surface stains. A
few of the larger grains contain biotite inclusions.

The feldspars are divisible, like those of the south-
ern-derived sediments, into creamy to greenish plutonic
feldspars, mostly deeply weathered, and fresh, sharp-
edged oligoclase-andesine grains, presumably volcanic.

Abundant chert grains of several colors characterize
the Scenic specimen, but the Weta sand includes only
occasional grains of gray chert.

General Conclusions

The lithology of the Chadron Formation suggests
that deposition occurred in two major episodes, one
represented by the Ahearn Member, the other by the
Crazy Johnson and Peanut Peak Members (see Fig. 9).
In each case, deposition started with generally coarse
sediments which became progressively finer, and ap-
parently were deposited more slowly. During the periods
of slow deposition in late Ahearn and Peanut Peak
times, the interfluve sediments underwent mild, in-
cipient Iateritization, just sufficient to develop pale

orange colors. In the later period, Iateritization was
confined to the immediate neighborhood of the stream
channels, and formation of caliche was the more wide-
spread surface process. We conclude therefore that
Peanut Peak time was probably drier than late Ahearn
time, and was characterized by seasonal groundwater
fluctuations accordant with periodic changes in volume
of stream flow. During late Ahearn time streams were
much larger, so that no part of the valley plain lay far
away from them, and Iateritization may have occurred
as a result either of seasonal, local rainfall or of seasonal
increases in stream flow from the Black Hills.

The heavy mineral suites give clear indication of a
shift in location of major drainages during Crazy John-
son time. Specimens G 3942 from the base of the Crazy
Johnson northeast of Scenic, and G 3733, from the same
horizon east of Imlay, are composed of elastics derived
entirely from the southern Black Hills. The southward-
trending channel fill west of Scenic, from the top of the
Crazy Johnson, (specimen G 4019) is composed of sedi-
ments from the northern Black Hills. This represents a
westward shift of at least 12 miles, of the southwestern-
most stream from the northern Hills, at the expense of
the southern Hills drainage (see map, Fig. 3).

SYSTEMATIC PALEONTOLOGY

Introduction

The known Chadron fauna of the Big Badlands com-
prises fish, amphibians, reptiles, birds and a variety of
mammals. Although fossils other than titanotheres are
not common in the Ahearn and Crazy Johnson members,
a sufficient number have been found to provide a reason-
able sample of the successive vertebrate faunas. Because
the Peanut Peakian fauna was treated in an earlier
paper (Clark, 1937) only a summary of those forms will
be given here except where additional materials demand
a reinterpretation, or where a restudy has shown that a
modification of earlier opinions is proper. The "micro-
fauna locality", in Peanut Peak sediments, lies in the
north part of the SE M of Sec. 3, T. 42N., R. 45W.,
Shannon Co., South Dakota. In general, we have as-
signed the various genera in accordance with the clas-
sification of mammals proposed by Simpson (1946).

Class Osteichthyes
Subclass Actinopterygi
The only fish materials known from the Chadron are
a collection of small bones, probably in a non-evident
coprolite, from the lower part of the Ahearn Member
(PM 13601) and several ganoid scales from the Ahearn
and Crazy Johnson Members.

Class Amphibia

Order Anura

Family Pelobatidae

Genus Eopelobates

26

FIELDIANA: GEOLOGY MEMOIRS, VOLUME 5

Eopelobates grandis

Specimen. — PM 16441; partial skeleton; Ahearn
member.

Discussion. — Close relatives of this frog now inhabit
the typical lowlands in Indonesia and southern China,
and the upland forests of the southeastern Tibetan rim.

Class Reptilia

Order Chelonia

Although turtle material from the Chadron is not of

great systematic importance, it is of significance in

interpretation of Chadron environments. Most of the

following list is based on field identification of scrap.

Family Emydidae
Genus Graptemys
Graptemys cordifera

Specimen. — PM 13838; carapace and plastron;
Crazy Johnson member.

Graptemys sp.
Specimens.

Scrap from Ahearn Member.

Genus Trachemys
Trachemys antiqua

Specimen. — PM 13839; plastron, right bridge, right
posterior border of carapace; Crazy Johnson Member.

Trachemys sp.

Specimens. — Scrap from Ahearn Member.

Family Trionychidae
Genus Amyda

-Scrap from Ahearn Member.

Amyda sp.
Specimens.

Family Dermatemydidae
Genus Pseudanosteira or Anosteira
Sp. indet.

Specimens. — Scrap from Ahearn Member.

Anosteirine, g. indet.

Specimens. — PM 16302; marginal plate; Crazy John-
son Member. PM 16303; two marginal plates; Crazy
Johnson Member.

Discussion. — The discovery of anosteirine material
in the Chadron represents an upward extension of the
time range of a typically Eocene group.

Genus Xenochelys
Xenochelys formosa

Specimens.- PM 13686; skull; Crazy Johnson Mem-
ber.

Discussion.- This specimen was described and fig-
ured by Williams (1952).

Order Squamata
Suborder Lacertilia

Family Anguidae
Genus Peltosaurus
Peltosaurus sp.

Specimen. — PM 16304; fragment of skull; Ahearn
Member.

Peltosaurus

Specimen. — Maxillary; microfauna locality; Peanut
Peak Member.

Discussion. — This specimen might represent Glypto-
saurus rather than Peltosaurus.

Order Crocodilia
Family Alligatoridae
Genus Alligator
Alligator cf. prenasalis

Specimen. — PM 16273; skull and jaws; Ahearn Mem-
ber. PM 13799; skeleton; Crazy Johnson Member.
Fragmentary specimens from Ahearn and Crazy John-
son Members. Not collected.

Class Aves

Order Passeriformes

Genus indet.

Specimens. — Two scapulae, three ulnae, distal end

of tarsometa tarsus; microfauna locality: Peanut Peak

Member.

Class Mammalia

Order Marsupialia

Family Didelphidae

Genus Peratherium

Peralherium sp.

Specimen. — One jaw with fragments of teeth; micro-
fauna locality; Peanut Peak Member.

Order Insectivora
Family Apternodontidae
Genus Apternodus
Apternodus altitalonidus

Specimen. — PM 13774; fragment of left mandible
with P4-M3; Peanut Peak member; microfauna locality.

Apternodus mediaevus

Specimen. — CM 8669; skull and jaws; Peanut Peak
Member.

Genus Clinopternodus

Clinopternodus gracilis

Specimen. — PM 13835; mandibular ramus with C,
P3_4, Mi, Peanut Peak Member; microfauna locality.

Family Leptictidae
Genus Ictops
Ictops dakotensis

Specimens. — PM 13605; mandibular ramus; Peanut
Peak Member. PM 13773 ; six partial maxillae and four
mandibular rami; Peanut Peak Member; all from micro-
fauna locality.

CLARK AND BEERBOWER: THE CHADRON FORMATION

27

Family Metacodontidae
Genus Metacodon
Metacodon magnus

Specimen. — PM 13835A; partial lower jaw with
P4-M3; Peanut Peak Member; microfauna locality.

Order Primates

Family Apatemyidae

Genus Sinclairella

Sinclairella dakotensis

Specimen. — PM 13585; crushed skull and lower
jaws; Peanut Peak Member; microfauna locality.

Order Rodentia
Family Ischyromyidae
Genus Ischyromys
Ischyromys sp.

Specimen. — CM 9493; mandible; Ahearn Member.

Family Eomyidae
Genus Adjidaumo
Adjidaumo minutus

Specimen. — PM 13832; two lower jaws; Peanut Peak
Member; microfauna locality.

Adjidaumo sp.

Specimen. — CM 9400; partial mandible with two
molars; Crazy Johnson Member.

Genus Paradjidaumo

Par adjidaumo minor

Specimen. — PM 13831; eight lower jaws; Peanut
Peak Member; microfauna locality.

Genus indet.
Specimen. — PM 16306; incisor; Ahearn Member.

Family Eutypomyidae
Genus Eutypomys
Eutypomys cf. thomsoni

Specimen. — Two molar teeth; (specimen not located
1956; see Clark, 1937); microfauna locality; Peanut
Peak Member.

Order Lagomorpha

Family Leporidae

Genus Megalagus
Megalagus turgidus

Specimens. — Two sets of lower molars; no numbers;
microfauna locality; Peanut Peak Member.

Order Carnivora

Suborder Creodonta

Family Hyaenodontidae

Genus Hyaenodon

Hyaenodon cf . cruentus

Specimens. — CM 9090; two milk molars, possibly
DP4-M1; Crazy Johnson Member. PM 12745; mandibu-
lar ramus; Peanut Peak Member.

Hyaenodon cf. montanus

Specimens. — CM 9098; two upper molars, possibly
P4-M>; Ahearn Member. PM 12970; skull and part of
lower jaw; Peanut Peak Member.

Hyaenodon cf . horridus

Specimen. — PM 16288; M3; Ahearn Member.

Suborder Fissipedia

Family Canidae
Subfamily Caninae
Genus Hesperocyon
Hesperocyon gregarius

Specimen. — PM 13630; mandibular ramus with
P4-M3; Peanut Peak Member; microfauna locality. No
number, Princeton; mandibular ramus with P4-M1;
Peanut Peak Member.

Genus Daphoenus
Daphoenus sp.

Specimen. — CM 8799; fragment of right lower jaw
with alveoli of P4-M1; Crazy Johnson Member.

Discussion: This fragment, about the same size as
D. hartshornianus or Daphoenocyon minor, has a narrow,
shallow jaw, and, as indicated by the alveoli, well-spaced
narrow teeth. The specimen is a typical Daphoenus and
cannot be referred to Daphoenocyon. This is the only
specimen of Daphoenus from the Chadron, and we be-
lieve that Daphoenus was probably rare at this time.
Certainly it was uncommon even during Brule time. Its
place in the Chadron ecosystem was probably pre-
empted by Daphoenocyon which resembles it closely in
post-cranial osteology.

Subfamily Amphicynodontinae
Genus Parictis

Discussion. — The discovery of a new species of
Parictis in the Ahearn Member poses a difficult taxono-
mic problem. The species (P. parvus) strongly resembles
P. dakotensis and an undescribed species from Pipe-
stone Springs but also shows very close affinities to
Campylocynodon personi and indeed may be synonymous
(see Discussion under P. parvus below). We believe that
Campylocynodon is best relegated to the status of a
subgenus of Parictis with C. personi as type species
because recognition as a separate genus of slender-jawed
parictines would require a taxonomic grouping of the
two below the subfamily rank. Parictis then comprises
two subgenera, Campylocynodon and Parictis, with a
trend from early forms with trenchant sectorial but
rounded tubercular dentition to later species with the
entire dentition low, broad, and rounded.

Comparison with Daphoenus, Hesperocyon, Dapho-
enocyon dodgei, and the European genera Cynodictis,
Cynodon, and Cephalogale demonstrates that the de-

28

FIELDIANA: GEOLOGY MEMOIRS, VOLUME 5

6

v«

876f - r

)mmifiii|iii|iii|iii|iii|iiyiiiliii|iff[[iiiiii|[|[|ii[[iiTiii

Blllllll!lllllllllllfllllllllll1]|||||ll|i

Fig. 10. The genus Parictis: lateral view of specimens. 1. Un-
described species from Montana. 2. Type specimen of Parictis
(Campylocynodon) personi. 3. Type specimen of Parictis (Campy-
locynodon) personi. 4. Type specimen of Parictis (Campylocyno-
don) parvus. 5 & 6. Undescribed species from Montana. 7. Geno-
type specimen of Parictis (Parictis) primaevus. 8. Cast of type
specimen of Parictis (Parictis) dakotensis.

velopment of an accessory cusp on the protoconulid of
Mx and the crowding of the trigonid of Mi are valid
characteristics distinguishing the genus Parictis. Fig-
ures 10 and 11 illustrate the known specimens of the
genus.

Subgenus Campylocynodon
Parictis (Campylocynodon) personi Chaffee, type species.

Diagnosis. — Characteristics as for the genus, but the
jaw is relatively slender and the sectorial dentition
somewhat trenchant.
Parictis (Campylocynodon) parvus, n. sp. (Figs. 10-12).

Type. — PM 16265; right lower jaw with Ps-M2 and
alveoli of C, P1-2, M3; red layer near base of Ahearn
member; west flank of Quinn Draw, SW 1/4, sec. 25,
T 43N, R. 46W, Shannon Co., South Dakota.

Description. — Ramus of jaw slender, mental fora-
mina below anterior root of P2 and posterior root of P3;

P3 short, outline shaped like a parallelogram, principal
cusp high with a ridge extending down to the postero-
external angle of the tooth but with no accessory cusp;
anterior and posterior cingula heavy, but not connected
across labial and lingual faces of the principal cusp.
Principal cusp set anteriorly, but directed vertically or
slightly posteriorly. P4 ovoid in plan, slightly larger than
P3. Principal cusp vertical, with a blade-shaped, well-
developed accessory cusp half-way up the postero-ex-
ternal ridge. Cingula heavier than on P3 but not meeting
laterally. Mx relatively low-crowned, shorter and broader
than Mi of P. (Campylocynodon) personi; trigonid with
paraconid more lingually placed than in P. personi, but
otherwise similar; pronounced cingulum across anterio-
external face of trigonid. Talonid much shorter than in
P. personi; hypoconid a cutting ridge; entoconid-hypo-
conulid forming a lingual ridge, separated from the
hypoconid by a trench which notches the posterior rim
of the tooth; small but marked cingulum on exterior
face of hypoconid. M2 two-rooted, broad, short, very
low-crowned. Cusps of trigonid well-developed, stout,
triangular, very closely crowded at the antero-internal
corner of the tooth. A low, heavy accessory cusp on the
postero-external corner of the protoconid lies directly
posterior to the antero-external cingular shelf. Hypo-
conid slightly lower than in Mx; talonid broadly basined,
with entoconid-hypoconulid very weakly developed.
M3 single-rooted.

Discussion. — This species may be synonymous with
P. (Campylocynodon) personi. The dental characters of
the type of P. personi are, unfortunately, almost inde-
terminable. Major differences between the two type
specimens are: (1) P. parvus is smaller overall; (2) P.
parvus has teeth that are shorter antero-posteriorly but
equal in width and in height of crown; (3) talonid of
Mi is proportionally much smaller in P. parvus. Since
only one specimen of each species is known, the range of
variability of these characters cannot be determined at
present. The tendency toward low-crowned teeth with
heavy cingula is evident in P. parvus, but the propor-
tionally heavy jaw characteristic of the subgenus Paric-
tis is not apparent.

Subgenus Parictis
Parictis (Parictis) dakotensis

Diagnosis. — Characteristics as for the genus but with
proportionally heavy jaw and low crowned, blunt
dentition.

Parictis (Parictis) dakotensis

Specimen. — So. D. School of Mines; right mandible
with P2^ Mi_2; probably Peanut Peak Member.

Genus Daphoenocyon Hough

Daphoenocyon dodgei (Scott), type species.

Diagnosis. — Strongly brachycephalic; flaring zygo-
mata; deep thick jaw. Molar series relatively long, 85-
97% of premolar length as compared with 64-71% in
Daphoenus and 78-89% in Hesperocyon. Canines verti-

CLARK AND BEERBOWER: THE CHADRON FORMATION

29

INCHFS

ULLTL

ill

[TTTynTJTTTpr[TTT|lirill

METRIC 1

nn ; mi , , h 1 1 1 f , 1 1 ;■ .'. i ,-i : , i.7f i : . 1 1 , i Tl , i ; 1 1 Tl : ; : 1 1 1 1 i?l ■ i ' 1 1 1 1 nl 1 1 i l 1 1 m°1 i > 1 1 ! i m?I 1 1 1 1 h i m I7i 1 1 1 1 1 n Tn in

Fig. 11. The genus Parictus: crown view of mandibles. 1. Type specimen of Parictis (Campylocynodori) personi. 2. Type specimen of
Parictis (Campylocynodori) parvus. 3 & 4. Undescribed species from Montana. 5. Genotype specimen of Parictis (Parictis) primaevus.
6. Cast of type specimen of Parictis (Parictis) dakotensis.

cal, trenchant. Pre-molars short and heavy. General

facial character simulates that of felids (Figs. 14-16).

Discussion. — Hough, (1948) described the genus

Daphoenoeyon with Daphoenus dodgei as the type spe-

Meosurements of Parictis (Campylocynodori) parvus

Length

Width

Height

p3

4 mm.

2.1

3.4

P4

4.8

2.6

3.6

M,

6.4

3.6

4.6

M2

3.3

2.8

1.4

Fig. 12. Measurements of Parictis (Campylocynodori) parvus.

cies. She based her description on the type material
(PM 11422), which comprises three partial jaws from
the Chadron formation, White Horse Creek, Nebraska,
and on two skulls with lower jaws (USNM 17847 and
Walker Museum 1456) from the Orellan. Restudy of the

type material and of the specimens referred by Hough,
and study of new material from the Chadron, demon-
strates that this reference of USNM 17847 and WM
1456 is incorrect. Hough states (p. 594) :

"A brachycephalic, short-faced carnivore about the size of a
large coyote but very differently proportioned. Frontal sinuses
prominent, giving the upper profile of the skull something the
appearance of that of a domestic dog. Basicranial region very short
in comparison with other daphoenid genera. Auditory region short
and narrow. Promontorium large. Mastoid produced into a rugose
knob projecting downward. Facial nerve apparently leaving the
skull through a groove posterior to the mastoid process.

"Type species: D. dodgei Scott 1898

"Specific characters: Lower jaw exceptionally short, maximum
length 51 mm. Premolars crowded. P2 and P3 as well as Pt with
accessory cusps. M, much more primitive with anterior cusp
(protoconid) more rounded than in other genera; metaconid also
higher and more rounded, talonid with high median cusp."

Also, p. 578:
"Genus Daphoenoeyon

"This genus, described below, is closely related to Brachyrhyn-
chocyon but is larger and has a very short cranium as well as face.

30

FIELDIANA: GEOLOGY MEMOIRS, VOLUME 5

The basicranial region is also very short. The lower jaw is short and
thick. The premolars are crowded but not reduced. The frontal
sinuses are prominent and this together with the rounded cranium
goes the skull some resemblance to that of a domestic dog. The
promontorium is large and rounded. The mastoid produced into a
rugose knob. The facial nerve apparently left the skull posteriorly
of the mastoid process. Accessory cusps are present on P, and P,
as well as P,. The molars are triangular with little development of
the internal cingulum. M, is a small rounded tooth situated in the
angle of the jaw."

USNM 17847 has no lower teeth except the canine,
Pi, and anterior half of P2. The canine is long and slender
as in Daphoenus vetus and utterly unlike the short, mas-
sive canine in the type of D. dodgei. Since Pi is missing
from the type, no comparisons of this tooth are possible.
The anterior portion of P2 is slender with the principal
cusp over the middle of the tooth and the anterior ridge
sloping. This is the typical structure of P2 in Daphoenus
vetus and Daphoenus harlshornianus, but in Daphoeno-
cyon dodgei the tooth is broad with the principal cusp
set forward and the anterior ridge almost vertical. The
jaw in Daphoenocyon dodgei is deep and heavy; in
USNM 17847, it is long and slender as in D. vetus. It
must be concluded that there is no basis referring this
specimen to D. dodgei.

Posterior Accessory Cuspule on Premolars
of Daphoenus

Key: O absent

® present but very small

X present

— broken or warn: indeterminable

Species

Specimen Number

Ps

Pz

Protemnoeyon in flatus <A»
* Daphoenus vetus v

CM

552

(S)

Daphoenus hortshornianue

CM

3697

—

O

PM

12650

—

®

vetus (S3J list as
hortshornianus )

PM

13560

L® RO

L® RO

PM

13600

X

O

PM

13792

®

O

(type felmus)

PM

11425

X

®

vetus (Saj list os
hortshornianus)

PM

12635

—

®

PM

12651

X

®

Fig. 13. Posterior accessory cuspule on premolars of Daphoenus.

The skull associated with the lower jaws in USNM
17847 is that of a small Daphoenus vetus with facial
proportions, length of the basicranium, and details of
basicranium and auditory area like those of Daphoenus
vetus. Hough, in her description, mistook features pro-
duced by distortion and breakage for generic characters.

The lower dentition of WM 1456 resembles the type
of D. dodgei only in the development of posterior acces-
sory cuspules on P2 and P3. Accessory cuspules are,

however, present in some typical Daphoenus vetus and
Daphoenus harlshornianus as indicated in Figure 13, and
so are not diagnostic of D. dodgei. The lower dentition
otherwise is that of a typical Daphoenus vetus, with
long slender premolars having a central principal cusp.
The skull of WM 1456 is badly crushed, but the pro-
portions are those of a normal, dolichocephalic Dapho-
enus vetus. The only basicranial character in which this
specimen differs from typical Daphoenus vetus is en-
largement of promontorium.

Since neither WM 1456 nor USNM 17847 can be as-
signed to D. dodgei and since they do not differ from
Daphoenus vetus in any significant way, a complete re-
description of the genus Daphoenocyon based on the
Chadron specimens is necessary.

Description. — The lower jaw is thick with a deep,
sharply-rimmed temporal fossa, a sub-triangular coro-
noid process shorter than in Daphoenus, a heavy, in-
flected angular process with a deep dorsal furrow, and a
definite chin (Fig. 14). The anterior mental foramen is
below P2 and is large; the posterior below P3 and small.
The canine is vertical and heavy. Pi has a single root, is
vertical, and ovoid-cone-shaped. P2 is sub-quadrangu-
lar and is inflated posteriorly, with a posterior accessory
cusp. P3 and P4 are like P2 but are successively larger
with larger accessory cusps; P4 is less inflated than P3.
The premolars are all vertical to slightly raked poster-
iorly and Pi_3 are echeloned or are out of line. Mi is like
that of Daphoenus but heavier and with lower crown
and a quite shallow, basined talonid. M3 is missing from
all specimens but was large and had a single to double
root. The molar series is relatively long and the entire
lower dentition is broad and massive.

The skull is strongly brachycephalic — the muzzle
includes only P1. The narial opening is large, with its
anterior rim vertical. The nasals slope upward posterior-
ly at an angle of about 25° to the tooth row. The in-
cisor row is moderately arched, the teeth vertical with
posteriorly raked points, and not crowded. I1-2 are
small and transversely spatulate with I2 larger than I1.
I3 is much larger and caniniform with short anterome-
dian and long postero-external ridge.

The canine is straight, heavy, and vertically directed.
The surface between the antero-internal and postero-
external ridges is nearly flat and blade-shaped, but the
external face is cone-shaped with a slight flattening
posteriorly. There is no diastema between the canine
and P1. The latter tooth is small, single rooted, and high,
and was probably functional. P2 is simple, quadrangular
and massive. The principal cusp is high and directed
posteriorly. The cingulum is complete. A faint ridge
runs from the cusp to the postero-external angle of the
tooth and another to the antero-internal angle. P* is
like P2 but is larger and more sharply quadrangular,
has a tiny accessory cusp on the postero-external ridge,
and a heavy postero-external cingulum. P2 and P3 are
not in line with each other. P4 is low and simple with
the cingulum complete but weak antero-internally. The
protocone is low and heavy; no cusp is present on the

CLARK AND BEERBOWER: THE CHADRON FORMATION

31

«*4flg£

Fig. 14. The genus Daphoenocyon: mandibles.

antero-external cingulum ; and the deuterocone is prom-
inent with a low lateral crest extending to meet a
similar crest on the protocone. M1 is large and like that
of Daphoenus. M3 is small, narrow, and double-rooted.
Occlusal surfaces show wear, indicating that it was
functional. ""

Daphoenocyon dodgei (Scott)

Daphoenus dodgei Scott, 1898, Notes on the Canidae of the
White River Oligocene, Trans. Am. Philos. Soc, 19, (n.s.), p. 362.

Fig. 15. The genus Daphoenocyon CM 9256, paratype specimen.

Daphoenocyon dodgei Hough, 1948, A Systematic Revision of
Daphoenus and some allied genera. J. Paleo., 22, pp. 578, 594.

Type— PM 11422; Lower Jaw; Whitehead Creek,
Nebraska; "Titanotherium Beds."

Paratypes. — CM 9256; anterior part of skull with
P-M1; Chadronian; Pipestone Springs, Montana. CM
9287; lower jaw; Chadronian; Pipestone Springs, Mon-
tana. CM 9508; pair of lower jaws; Chadronian; Spring
Gulch, West of Sage Creek, Beaverhead Co., Montana.

Referred specimens: PM 13601; lower jaw; Peanut
Peak Member. PM 16280; P3-Mx; upper part of Crazy
Johnson Member. CM 9825; lower jaw; Chadronian;
Pipestone Springs, Montana. CM 9057; lower jaw
broken and rehealed; Chadronian; Pipestone Springs,
Montana. CM 9573; lower jaw with periostitis, Chad-
ronian, Pipestone Springs, Montana. CM 9358B; lower
jaw fragment, P3_4; Chadronian; Hadcock Ranch, W.
side of Missouri River, S. end of Sewell Lake, Broad-
water Co., Montana. CM 673; lower jaw; Chadron
Formation; Near Sugar Loaf, Sioux Co., Nebraska.

Diagnosis. — Large size, Pi-M3 58-63 mm. Jaw very
deep and heavy. Other characters as in description of
genus.

Discussion. — Daphoenocyon dodgei apparently repre-
sented a highly variable population — specimens from
Pipestone Springs show a variability of about 10% with
little or no association between the varying characters.
The D. dodgei specimens from South Dakota may repre-

32

FIELDIANA: GEOLOGY MEMOIRS, VOLUME 5

Comparotive Measurements of Dophoanoc/on, Dophoenus,
and Parictis

length

P3 length
P3 width

"3-f

P4 length

P4 width

P4

w
C

M,

length

M

width

u

JL

Dophoenocyon
D. {lodge! type, PM 11422
B C

8.2

4.9

.597

II. I
5.2

.468

14.3

7.3

.510

6.5
4.3
.661

8.2

5.3

.646

10.7
5.5
.514

14.5
7.2
.496

A

8.1
4.9
.605

II. I
5.2
.468

14.4
6.9
.479

D. vetus
WM 1456

8.5
3.8

.447

9.9

4.4

.444

12.4

4.6

.371

16.0

7.5

.469

D. *etus
PM 13560

8.4

3.6

.429

9.9

4.2

.424

12.0

4.7

.392

16.2

7.7

.475

sent a different species than the Montana material, with
a more shallow jaw, but the samples are too small to
justify such a division.

Daphoenocyon minor n. sp.

Type. — CM 9506; partial left mandibular ramus
with P4-M1 and alveoli of Cx Pi_3 and M2_3; two miles
above main forks of Indian Creek, Pennington Co.,
S.D.; upper third of Ahearn Member (Fig. 19).

Referred Specimens. — PM 16284; partial right man-
dibular ramus with alveoli M2_3; associated; base of
Crazy Johnson Member. PM 16279; partial ramus with
alveoli of P2-Mx; upper part of Ahearn Member.

Diagnosis. — Smaller than D. dodgei, length P1-M3
51 mm. Jaw thick, heavy, but not deep. Tooth row
almost straight.

Discussion. — D. minor appears to be near the
ancestral population of both D. dodgei and a small spe-
cies of Daphoenocyon from Pipestone Springs. D. dodgei
then would represent a trend toward increased size and
specialization and the small Montana Daphoenocyon a
population diverging toward the carnivore size niche
occupied by the present day Fisher.

Family Mustelidae
Genus Mustelavus

Mustelavus priscus

Specimens. — PM 13775; skull and lower jaws; Pea-
nut Peak Member. PM 13776; partial lower jaw; Peanut
Peak Member. PM 13777; maxillary fragment; Pea-
nut Peak Member.

The genus Mustelavus is certainly very close to and
may be synonymous with the European genus Plesictis.

Daphoenocyon dodgei

Parictis

PIPE. PIPE. PIPE COOKR NEBR. S.O.
9267 9573 9825 9508 373 9506

PIPE.
8851

S.O.

PIPE. PIPE.Ofeflon
9068 9571

MANDIBLE, DEPTH BELOW
M, PR0T0C0NI0

26.0

22.0

22.7

2 3.5

22.3

19.3

14.7

11.0

8.8

6.9

DEPTH BELOW Pg

23.0

22.3

21.4

24.0

22.0

—

—

9.9

7.1

T.S

THICKNESS BELOW P3

II. 0

10.7

II. 0

L-14.0
R-10.5

9.2

10.4

7.6

4.9

39

4.5

10.0

P.-M3

60.5

60.0

58.0

61.9

623

31.3

®44i

37.1

29.0

30.7

32.0

P,-P4

33.0

31.2

31.0

331

324

262

®243

219

165

17.7

SECTORIAL DENTITION

42.0

—

38.5

41.0

41.0

35.6

»I.T

260

20.5

21.7

TUBERCULAR OENTITION

19.0

—

19.0

203

20.3

16.8

13.3

10.5

8.4

6.6

1 W

33

3.0

2.4

1.6

i. w

6.7
3.T

—

7.3
4.0

7.5
34

6.5
43

—

5.5

3

5i

4.1
2.4

45
2.4

5
3

p3-!=-

w

9.5
5.3

84
48

8.5
4.8

8 8

4.8

7.2
3.8

6

4.0
2.6

4.5

2.4

P4-L

w

12.4

"ST

11.6
6

11.6
5.7

1 . 4
6

11.3
S3

10.9
48

83
4.4

7.5

5 9

3.0

6.1

2.7

M. t

w

15.0
7.6

15.0
7.9

14.9
7.4

14.7
7.4

13.9
6.5

11.6
6

9

7.6
4.0

7.9
38

M, _L_

* W

8.5
8T

—

9.5
7.3

88

6.3

9.5

60

—

68
5.2

£

32.

27

40.
3.2

M,-«3

29.0

29.3

29.3

296

29.8

25.6

—

17.0

13.0

13.8

ratios:

TUBERCULAR
SECTORIAL

.452

—

.493

495

.495

.469

426

.404

.410

.397

M,-M3
P,-P4

.979

.939

.943

.894

920

.977

856

.791

.791

.780

THICKNESS
P, -M3

.182

.178

.189

L-226
R-.I70

.148

.207

.170

.132

.134

.150

DEPTH AT M|
pl -«3

.430

.367

.391

.380

.361

.376

326

.297

.303

.290

THICKNESS
OEPTH AT M,

.423

.466

.484

Rt447

.409

.539

.517

.445

.443

.506

B

Fig. 16A, B. C. Comparative measurements of Daphoenocyon,
Daphoenus, and Parictis.

Periostitis

Daphoenocyon dodgei

PIPE.

PIPE.

PIPE.

PIPE.

9373

9287

9506

9825

9057

DEPTH OF JAW BELOW
PR0T0C0NID, M|

22

26

2 3.5

22.7

broken

DEPTH OF JAW BELOW P2

223

23

24.0

21.4

MAXIMUM THICKNESS AT P?

10.7

II

(|4.0)
10.3

II. 0

i°«

P| - «3

60.0

60.5

61.9

58.0

P,- P«

31.2

33

33.1

31.0

31.0

M,-M3

29.3

29

296

29.3

*k

—

3.3

3.0

-

—

—

P j=_

2 W

—

6.7
3.7

7.5
3.4

7.3

4

7
4

'•*

6.4
4.6

9.3
5.3

—

8.3
4.8

9
5.5

*fr

11.6
6

12.4
5.9

11.4
6

11.6
5.7

110

5.2

".*

IS

n

14 9

ISO
7.9

".*

65
6.3

8 8
6.3

95
7.3

3 W

—

—

—

SECTORIAL
TUBERCULAR

42 .436
183

41 .495
20.3

36.5 .493

19.0

P
M

31.2 939
29.3

33 .879
29

33 1 694
29.6

31 943
293

P, - "3

60 .178
10.7

605 .182
II

619 2"

-?L?-.I70
14-10.3

36 0 .189
11.0

THICKNESS

Pl-Mj

DEPTH

60 .367
22

605 .430
26

61.9 .380
235

58.0 .391
22 7

OEPTH
THICKNESS

22 .486
10.7

26 423

1 1

23.5 .447
10.5

22.7 .484
1 1.0

CLARK AND BEERBOWER: THE CHADRON FORMATION

33

However, the type species of Plesictis is late Miocene
in age, and probably is generically distinct from the
early Oligocene P. pygmaeus, which M. priscus closely
resembles. Unravelling the tangled synonymy of Plesic-
tis would require both access to European material and
time, which are at present not available. We therefore
retain Mustelavus, in the expectation that the species
M . priscus will be declared a member of whatever genus
the species P. pygmaeus is eventually assigned to.

Family Felidae
Subfamily Machairodontinae
Genus Eusmilus
Eusmilus sp.

Specimen. — PM 16271; fragment of left mandibular
ramus with alveoli of P4-M!, Crazy Johnson Member
(Fig. 17).

Fig. 17. Measurements of Eusmilus sp.

PU 16271 mm.

Depth of jaw below M, 22.3

Maximum thickness of jaw below M, 12.4

Length of P4 alveolus 12.9

Length of Mi alveolus 17.4

Discussion. — This specimen probably represents an
undescribed species of Eusmilus, but it seems best to
await the discovery of more complete material before
erecting a new species. The size of the alveoli indicates
that the teeth were as large as those of a small E.
sicarius with the backward rake characteristic of the
genus, but the jaw is little more than half as deep and,
apparently, about two-thirds as long. Thus the animal
had large teeth, fully specialized in the direction of
Eusmilus, set in a small, relatively light jaw.

As this specimen and PM 16272 are the oldest known
machairodontines, the high specialization of the denti-
tion indicates that the major morphological evolution
of the subfamily occurred much earlier.

Eusmilus?

Specimen. — PM 16272; Fragment of mandibular
ramus with root of Mx; Ahearn Member.

Discussion. — Since the premolar alveoli are missing,
assignment to the genus Eusmilus is questionable. The
fragment is very slightly smaller than PM 16271.

Genus Hoplophoneus

Hoplophoneus sp.

Specimen. — PM 13596; right mandible with P8;
Crazy Johnson Member.

Holophoneus oharrai

Specimens. — S.D. School of Mines Museum No. 2417;
partial skull; Crazy Johnson Member. PM 13593; par-
tial skeleton; Peanut Peak Member.

Genus Dinictis
Dinictis fortis

Specimen. — PM 13638; assorted teeth; Peanut Peak
Member.

Order Perissodactyla

Family Equidae
Genus Mesohippus

Nine species of Mesohippus have been described from
the lower Oligocene:

M . celer Marsh, 1874, Nebraska

M. westoni Cope, 1889, Cypress Hills, Sask.

M . latidens Douglas, 1903, Thompson Creek, Mont.

M. montanensis Osborn, 1904, Pipestone Springs,
Mont.

M. portentus Douglass, 1908, Pipestone Springs, Mont.

M. hypostylus Osborn, 1904, Big Badlands, S.D.

M. proteulophus Osborn, 1904, Big Badlands, S.D.

M. precocidens Lambe, 1905, Cypress Hills, Sask.
Sask.

M. viejensis McGrew, 1956, Vieja, Texas.

These species were differentiated in the original descrip-
tions on the following characters:

1) Size

2) Relative length and breadth of upper cheek-teeth

3) Presence and size of hypostyle

4) Degree of reduction of metaconule

5) Hypsodonty

6) Development of internal cingulum

7) Connection of metaloph to ectoloph

8) Connection of protoloph to parastyle

9) Angulation of ectoloph in upper molars
10) Development of metastyle

Preliminary study of a series of over 40 specimens from
the Chadron formation has demonstrated, however,
that these characters are highly variable at any one
time level and that a complete review of Chadronian
horses is necessary. In order to obtain an objective de-
scription of these animals a series of measures was
employed, including the characters mentioned above
and, in addition, anteroposterior length of the ectoloph,
development of the external cingulum, angle of the
external face of the ectoloph to the basal enamel line,
and ratio of the transverse length of the protocone to
transverse length of the paracone (Fig. 18). Those
characters that could not be measured directly were
ranked in a series of classes. Statistical tests of signifi-
cance were used where necessary.

Of the group of 40 specimens, 13 are from the
Ahearn, 23 from the Crazy Johnson, and 4 from the
Peanut Peak Member; these were supplemented by the
collections of M. viejensis (Field Museum of Natural
History), of Montana Mesohippus from the Carnegie
Museum, and of lower Brule (Lower Nodular zone)
Mesohippus from the Princeton Museum. The measure-
ments on this assemblage of specimens are reported in
the accompanying tables (Fig. 18).

Analysis of these data supports these generalizations:
1) The internal cingulum is absent in early Mesohip-
pus and in smaller individuals. Where present, it is best
developed on M3 and the premolars and is weak or
absent on M1-2. Conules are developed irregularly on

Fig. 18. Key

Protoloph

1. Cingulum to parastyle

2. Both to parastyle

3. Protoloph to parastyle

30. Protoloph connected, but interrupted by preceeding
tooth

Crotchet

0. Absent

1. Slight

2. Marked

Height (H on graphs)

Median space present — M
absent — O

Internal cingulum Hypostyl

1. Present 1 cone; or does not close valley 0.

2. Closes valley but not on tooth vails

3. Complete

Metaconule

1. Completely separate

2. Part of metaloph, but large and round 2.

3. Slight swelling on metaloph

4. Absent o

Metastyle

0. Cingulum does not reach tooth crown 4-

1. Cingulum reaches crown as enamel ridge only

2. Dentine included in laterad cingular ridge 5.

3. As above, plus posterior extension

e

Absent

0

11=

Fig. 18. I. — Measurements of Type Specimens of Chadron Mesokippus.

Species

celer
1874

westoni
1889

latidens
1903

montanensis
1904

porentus
1908

hyposlylus
1904

proteulophus
1904

Locality

Nebraska

Cypress
Hills

Thompson
Cr., Mont.

Pipestone
Springs

Pipestone
Springs

South
Dakota

South
Dakota

P'-M3
P«-M3

pi_P4

M'-M3

p*_4

MLJ

P' L

W

P» L

W

H

Hy

Int. cingulum

Metaconule

Metastyle

Hypostyle

Protoloph connected

Crotchet

27

32

75
32

3-
2
0
2

1
0

P*-Ms
37

P3 L

W
H

Hy

Int. cingulum

Metaconule

Metastyle

Hypostyle

Protoloph connected

Crotchet

P« L

W
H

Hy

Int. cingulum

Metaconule

Metastyle

Hypostyle

Protoloph connected

Crotchet

M1 L 10

W 15
H
Hy
Int. cingulum 0

Metaconule
Metastyle 0

Hypostyle
Protoloph connected
Crotchet

10
16

10.5
14

10
13.5

34

Fig. 18. I. — Measurements of Type Specimens of Chadron Mesohippi — Continued.

Species

Celer
1874

Weston
1889

Latidens
1903

Montanensis
1904

Porentus
1908

Hypostylus
1904

Proteulophus
1904

Locality

Nebraska

Cypress
Hills

Thompson
Dr., Mont.

Pipestone
Springs

Pipestone
Springs

South
Dakota

South
Dakota

M» L

W

H

Hy

Int. cingulum

Metaconule

Metastyle

Hypostyle

Protoloph connected

Crotchet

0

2-3
0
0

9.5-10.2
12-13

3
2
0
0-1?

1

2

3 +

0

1

2

0

13.3
18

2

1

4

3

0

0

5

2

3

3

1

0

M' L

W

H

Hy

Int. cingulum

Metaconule

Metastyle

Hypostyle

Protoloph connected

Crotchet

8.5
12

3
4

0-1
1
3
0

3

2

4

3

1

1

1

2

3

3

1

0

Fig. 18. II. — Vieja Specimens

Species

FM-PM 107 PM 151

PM 142

PM 108

PM42

PM35

Horizon

Pi-M3
P.-M,
Ps-P<
Mj-M,

P2-4

M,-,
P.

57.3

27

30.4

.888

L
W

P*

L

W

H

Hy

Cingulum

8.9
5.4

P.

L
W
H

Hy
Cingulum

9.7
7.1

9.3
7.2

P*

L
W
H

Hy
Cingulum

9.6

7.8

9.6

6.4

10.6

7.7

M,

L
W
H

Hy
Cingulum

9.1
7.0

9.7
6.4

11.5

8.8

3

9.9
6.9

M,

L
W
H

H

W

Cingulum

9.4
7.2

1

9.1
7.0

9.6
6.5

9.8
7.0

12.0
8.6
6.6

70.9

3

10.5
7.3
5.6

76.6

M,

L
W
H

H

W

Cingulum

12.3
6.2
5.1

82.3

11.9
6.1

12.3
6.3

35

Fig. 18. III.
M. v. viejensis

-M. viejensis-celer Group

M. v. ah earn

M. celer

Species

FM-PM 36 FM-PM 56 FM-PM 121 PU 16245 CM 9496 PU 13829 CM 9394 PU 16257

Horizon

Peanut Peak

X

Crazy Johnson

Ahearn

X

X

X

X

X

X

X

L
W

35.6

6.3
4.0

37.0

6.5
5.2

31.5

M>

L
W
H

Hy

Int. cingulum

Metaconule

Metastyle

Hypostyle

Protoloph connected

Crotchet

10.0
12.8

L R

9.7-9.0

11.9

M

10.2

14.3

0

2
2

1
3

32.8

P' L

10.3

11.0

10.3

W

9.5

11.8

10.5

H

M

M

Hy

Int. cingulum

0

2

2

Metaconule

3

2

2

Metastyle

0

0 +

0

Hypostyle

1

5

1

Protoloph connected

0

1

1

Crotchet

0

0

0

P' L

9.9

10.7

10.3

10.8

W

11.9

13.0

12.5

13.7

H

0

M

M

M

Hy

Int. cingulum

2

2 +

2

2

Metaconule

2-

2

2

2

Metastyle

1

0

0

0

Hypostyle

1

1

4

Protoloph connected

1

1

1

2

Crotchet

0

0

0

0

P« L

9.7

11.3

10.2

11.3

11.1

W

13.1

12.7

13.4

14.4

14.1

H

M

M

M

M

0

0

0

Hy

Int. cingulum

2

2

2

2

2

2

Metaconule

1

2

2

2

2

2

—

Metastyle

1

0

0

0

0

Hypostyle

0

1

0

5

4

2

Protoloph connected

1

1

1

1

2

30

Crotchet

0

0

0

0

0

0

0

Ml L

9.6

10.7

11.4

W

10.0

11.9

13.9

14.1

H

M

M

M

0

0

Hy

Int. cingulum

2

0

0

2

0

Metaconule

2

2

2

2

3

Metastyle

1

0

0

0

Hypostyle

0

1

0

4

Protoloph connected

1

2

1

2

3

Crotchet

0

0

0

0

0

0

M5 L

10.6

9.3

11.0

11.4

W

12.3

12.6

14.8

14.5

H

M

M

M

0

0

Hy

Int. cingulum

2

2

0

2

0

Metaconule

2

2

2

2

3

Metastyle

1

0

0

Hypostyle

1

1

1

3

2

Protoloph connected

1

3

3

3

2

Crotchet

0

0

0

0

0

11.2

13.5

0

2
4

1
2

36

Fig. 18. Ill (Continued). — Lower Jaws

M. celer

Species

PU 16248

PU 16244

PU 16246

PU 16247

CM 9392

PU 16251

CM 9394

Horizon

Ahearn

X

X

X

X

X

X

X

P,-M,
P,-M,
P.-P*
M,-M,

In

Mr.
P.-P4

66

62.4
33.5
33.4

64.5
60.6
33.4
32.0

65.6
62.8
32.9
33.5

63.8
60.0
32.2
31.9

70.0
38.3

37.4
35.3

67.5
36.2

.895

.862

.854

.893

.843

.918

.878

29.9

28.8

29.8

28.5

32.3

32.4

31.8

10.0
6.9

10.0
6.3

9.5

6.4

10.5

6.8

6.6

P.

L
W
H

H

W

Cingulum

P.

L
W
H

H

W

Cingulum

10.6

7.8

5.4 +

69.2
3

10.0

7.5

10.0
7.8

10.0
7.3

11.7
8.1

11.2

8.2

10.2
7.8

P,

L
W
H

H

W

Cingulum

10.5
8.6
5.4

63.9

3

9.7
8.1

10.4

8.4

9.7

7.8

11.6
10.4

11.0
9.1

10.6
8.6

M,

L
W
H
H
W
Cingulum

10.0
7.6

9.4
7.2

10.5
7.5

9.7
7.1

11.1
9.0

11.3

7.9

10.8
8.2

M,

L
W
H

H

W

Cingulum

10.5
7.2

9.4
6.9

10.4
7.3

9.8
6.9

11.9
9.0

10.9
7.2

10.9
8.0

M,

L
W
H

H

W

Cingulum

12.7
6.5
5.7

87.7

1

13.0
6.6

13.5
6.5
4.8

73.8

1

12.7
6.6
5.0

75.7

2

15.5
7.8

13.0
6.3
5.8

92.0

2

14.6
7.5

37

Fig. 18. IV. — Mesohippus latidens

Species

CM 9078

CM-FY

PU 13830

Horizon

Peanut Peak

X

Crazy Johnson

X

X

Ahearn

P'-M'
P»-M3
Pi-P«
M'-M'

pi_4

M1-3
P'

L
W

67.9
63.2
38.3
32.3

1.01

7.0
5.4

P' L

W

H

Hy

Int. cingulum

Metaconule

Metastyle

Hypostyle

Protoloph connected

Crotchet

10.6

12.7

0

2
1
0

30
0

> L

W

H

Hy

Int. cingulum

Metaconule

Metastyle

Hypostyle

Protoloph connected

Crotchet

11.4

14.4

0

2
1
0

30
0

P4 L 12.2

W 15.0

H 0
Hy

Int. cingulum 0
Metaconule 3

Metastyle 1

Hypostyle 1

Protoloph connected 30

Crotchet 0

12.0

15.9

0

2
2

1

1

30

0

M' L

W

H

Hy

Int. cingulum

Metaconule

Metastyle

Hypostyle

Protoloph connected

Crotchet

11.2
14.7

0
2
0

30

0

M* L 11.0 11.7

W 15.4 15.3
HO 0

Hy
Int. cingulum 2 3

Metaconule 3 3

Metastyle 1 3

Hypostyle 1 2

Protoloph connected 3 3

Crotchet 0 1

11.9

16.4

0

1

1

30

0

M3 L 11.1

W 15.3

H 0
Hy

Int. cingulum 2

Metaconule 2

Metastyle 2

Hypostyle 1

Protoloph connected 3

Crotchet 1

12.9

15.9

M

3

4

3
L R
5 2
30

1

11.4

15.8

0

3
4

1

1

30

0

38

Fig. 18. V. — Ahearn Specimens

M . v. ahearn

M. celer

M. h. hypostylus

Species PU 13829

CM 9496

CM 9394

CM 9392A

CM 9478

CM 9390

Horizon

Peanut Peak

Crazy Johnson

Ahearn x

X

X

X

X

X

P»-M»
P'-M1

pi_p4

M'-M»

It!

P1

L
W

37.0

6.5
5.2

31.5

DP»

7.5
5.4

P» L 10.3 11.0

W 10.5 11.8

H M M
Hy

Int. cingulum 2 2

Metaconule 2 2

Metastyle 0 0 +

Hypostyle 1 5

Protoloph connected 1 1

Crotchet 0 0

15.0

11.8

M

3
2
0
2

1
0

13.8
14.1

P' L

W
H

Hy

Int. cingulum

Metaconule

Metastyle

Hypostyle

Protoloph connected

Crotchet

10.3

12.5

M

2
2
0

1
0

10.7

13.0

M

2 +

2

0

1
1
0

10.8

13.7

M

2
2
0
4
2
0

12.8

15.7

0

3

0

3
0

i L 10.2 11.3 11.3

W 13.4 12.7 14.4

HO M 0

Hy

Int. cingulum 2 2 2

Metaconule 2 2 2

Metastyle 0 0 0

Hypostyle 5 0 4

Protoloph connected 112

Crotchet 0 0 0

13.9
0

3
2
0
2-5

0

12.8
15.8

M1 L

W

H

Hy

Int. cingulum

Metaconule

Metastyle

Hypostyle

Protoloph connected

Crotchet

M

e

10.7

13.9

0

2
2
0
4
2
0

13.4

15.2

0

2
3
0
5
3
0

14.0

14.6

0

M» L

W
H

Hy

Int. cingulum

Metaconule

Metastyle

Hypostyle

Protoloph connected

Crotchet

11.0

14.8

0

2
2
0
3
3
0

M8 L

W

H

Hy

Int. cingulum

Metaconule

Metastyle

Hypostyle

Protoloph connected

Crotchet

10.2

14.3

0

2
2
1
3
3
0

39

Fig. 18. VI. — Base of Crazy Johnson Member
Mesohippus hypostylus and latidens

Species

CM 9395

CM 9078

CM 8775

CM-FY

CM-FA

Horizon

Peanut Peak

Crazy Johnson

X

X

X

X

X

Ahearn

P'-M3
P2-M3
P'-P«
M'-M5

p2_4

M^3
P'

L
W

67.9
63.2
38.3
32.8

1.02

7.0
5.4

' L

W

H

Hy

Int. cingulum

Metaconule

Metastyle

Hypostyle

Protoloph connected

Crotchet

10.6

12.7

0

2
1
0

30
0

PJ L

W

H

Hy

Int. cingulum

Metaconule

Metastyle

Hypostyle

Protoloph connected

Crotchet

11.4

14.4

0

2
1
0

30
0

P« L

W
H

Hy

Int. cingulum

Metaconule

Metastyle

Hypostyle

Protoloph connected

Crotchet

12.2

15.0

0

0
3

1

1

30

0

M> L 13.4

W 16.0

H 0
Hy

Int. cingulum 0

Metaconule 3

Metastyle 0

Hypostyle 5

Protoloph connected 30

Crotchet 0

11.2
14.7

30
0

M5 L 13.3

W 16.3

H 0
Hy

Int. cingulum 0

Metaconule 3

Metastyle 0

Hypostyle 2-5

Protoloph connected 30

Crotchet 0

11.0

15.4

0

2
3
1

1
3
0

12.8

16.1

M

3

3

0
L R
1 2

2

0

11.7

15.3

0

3
3
3

2
3

1

13.5

16.6

0

1
3
0

4
3
1

M' L 11.

W 15.

H 0

Hy

Int. cingulum 2

Metaconule 3

Metastyle 1

Hypostyle 1

Protoloph connected 3

Crotchet 0

11.

15.

0

2

2
2

1
3
1

12.9

15.9

M

3

4

3
L R
5 2
30

1

13.9

16.9

0

3
4
3

4

30

0

40

Fig. 18. VI (Continued). — Lower Jaws

Species

CM 9395

CM 9078

PU 16255

PU 16254

Horizon

Peanut Peak

Crazy Johnson

X

X

X

x . .

Ahearn

P.-M,
P5-M,
P1-P4
M,-M,

Ml_3
P,-P4

65.3

36.2

70.4
38.7

70.4
35.9
38.5

.842

.832

.836

30.5

32.2

32.2

9.8
7.2

10.5
7.0

10.2
6.9

P2 L

W

H

Hy

Cingulum

P3 L 10.0 10.6 11.1

W 8.7 8.7 9.1

H
Hy
Cingulum 112

P4 L 11.9 10.8 11.0 11.9

W 10.0 9.5 9.1 10.1

H
Hy
Cingulum 2 13

M, L 11.8 9.9 11.7 11.9

W 9.0 8.4 8.7 9.3

H
Hy
Cingulum 12 2

M« L 11.2 11.6. 11.4

W 7.6 8.4 8.9

H
Hy
Cingulum 2 1 2

M, L 15.8 16.6 14.9

W 7.4 8.2 8.0

H
Hy
Cingulum 12 3

41

Species

CM 9089

Fig. 18. VII. — Top of Crazy Johnson Member
PU 16256 CM 9034A CM 9034B CM 9034C PU 16258 PU 16258A CM 9036

Horizon

Peanut Peak

X

X

X

X

X

X

X

X

Crazy Johnson

Ahearn

P'-M»

73.2

P»-M»

67.5

pi_P<

41.2

M1 -Ms

34.4

38.7

pi_«

1.02

M1-'

P' L

7.5

W

5.8

P» L

12.2

12.5

12.7

W

12.0

13.2

12.5

H

0

0

0

Hy

Int. cingulum

2

3

3

Metaconule

2

2

2

Metastyle

0

0

0

Hypostyle

2

2

2

Protoloph connected

1

1

1

Crotchet

1-

0

0

P> L

12.6

13.0

W

13.8

15.0

H

0

0

0

Hy

Int. cingulum

2

1

3

Metaconule

2

2

2

Metastyle

0

1

0

Hypostyle

2

2

2

Protoloph connected

2

1

Crotchet

0

0

0

P« L

12.8

12.8

13.0

14.0

W

15.0

15.7

15.8

16.0

H

0

0

M-

0

Hy

Int. cingulum

2

1

3

2

Metaconule

3

3

3

4

Metastyle

0

1

0

1

Hypostyle

2

2

2

2-4

Protoloph connected

20

1

3

Crotchet

0

0

0

0

M' L

12.6

13.5

13.8

13.4

13.0

14.5

W

14.5

14.0

15.6

16.9

14.7

15.0

H

0

0

0

0

0

0

Hy

Int. cingulum

2

0

0

3

1

2

Metaconule

3

3

3

3

3

3

Metastyle

0

1

0

0

1

0

Hypostyle

O

2

2-5

2

2

2

Protoloph connected

30

10

3

3

30

Crotchet

0

0

2

0

1

1

M' L

12.3

13.8

11.9

W

15.0

16.3

15.2

H

0

0

0

Hy

Int. cingulum

1

1

1

0

2

0

Metaconule

3

3

3

4

4

3-4

Metastyle

0

1

1

0

1

1

Hypostyle

2

2

2

4

2

2-5

Protoloph connected

30

30

2

30

Crotchet

0

0

0

0

2

0

M3 L 11.8

W 14.3
H 0

Hy
Int. cingulum 2

Metaconule 4

Metastyle 1

Hypostyle 2

Protoloph connected 30
Crotchet 0

13.2

15.9

0

2

4
1
3

12.9

14.8

0

3
4

1

3

30

1

13.8

15.3

0

3
4

1

2-5
30

0

42

Fig. 18. VII (Continued). — Lower Jaws
Species CM 9087 PU 16256 CM 9034A CM 9034B CM 9034C CM 9036

Horizon

Peanut Peak

Crazy Johnson

Ahearn

P, - M,
Pj-M5
Pi-P,

M,-M3 37.5 38.0

P*-«

M'-1

P2 - P, L

W

P, L 11.2

W 7.7

H
Hy
Cingulum 1

P, L 11.7

W 9.2

H

Hy
Cingulum 2

P« L 14.0

W 11.3

H
Hy
Cingulum 3

M, L 11.5 12.6 12.5 12.3 11.6

W 9.0 9.5 9.6 8.0 8.6

H

Hy

Cingulum

M2 L 11.2 14.0 12.0 12.1 11.6 11.7

W 8.4 9.9 9.0 8.8 7.9 8.7

H

Hy

Cingulum

M, L 15.5 15.5

W 8.0 8.0

H
Hy
Cingulum 2 3

43

Species

Fig. 18. VIII. — Crazy Johnson Member, General
M . hypostylus

PU 13828

PU 13827

CM 8777

CM 8778

Horizon

Peanut Peak

Crazy Johnson

X

X

X

X

Ahearn

P'-M5
P»-Ms
P'-P4

M'-M'

41.5

L
W

9.6
6.5

DP?

P» L 14.5 14.5

W 13.8 15.3

HO M
Hy

Int. cingulum 3 2

Metaconule 3 3

Metastyle 0 0

Hypostyle 4 2

Protoloph connected 1 1

Crotchet 1 0

12.8

13.2

M

2
2
0
2

1
0

P» L 15.0 14.6 14.8

W 15.1 16.1 15.3

HO 0 M
Hy

Int. cingulum 3 3 0

Metaconule 3 3 3

Metastyle 0 0 1

Hypostyle 2 2 1

Protoloph connected 3 3 1

Crotchet 110

13.1

15.3

M

3
3
0

1
1
0

' L

W

H

Hy

Int. cingulum

Metaconule

Metastyle

Hypostyle

Protoloph connected

Crotchet

15.6

17.0

0

2
3
0
2
3
1

13.5

17.0

0

3
2
1
2

1
0

13.7

16.0

0

2
4
0
0
1
0

M1 L 15.0 14.1

W 16.8 18.0

HO 0
Hy

Int. cingulum 0 0

Metaconule 4 4

Metastyle 0 0

Hypostyle 4 2

Protoloph connected 2 2

Crotchet 0 0

M* L 15.0

W 17.4
H 0

Hy
Int. cingulum 1

Metaconule 4

Metastyle 0

Hypostyle 2

Protoloph connected 30

R L

Crotchet 0 1

2-5

12.5

15.8

0

2
4
0
1
3

M3 L

W

H

Hy

Int. cingulum

Metaconule

Metastyle

Hypostyle

Protoloph connected

Crotchet

14.3

16.8

0

2

4

1

2
30

L R
1 0

14.8

18.3

0

2
4
0
2-5
3

0

44

Fig. 18. IX.— Peanut Peak Member

M.latidens
Species PU 13830

PU 16257

Horizon

Peanut Peak x

X

Crazy Johnson

Ahearn

Pl-M3

P»-M3

pi_p4

M'-M3

J>«_4

P1

32.8

L
W

1 L

W

H

Hy

Int. cingulum

Metaconule

Metastyle

Hypostyle

Protoloph connected

Crotchet

» L

W
H

Hy

Int. cingulum

Metaconule

Metastyle

Hypostyle

Protoloph connected

Crotchet

12.4

' L 12.

W 15.

H 0

Hy

Int. cingulum 2

Metaconule 2

Metastyle 1

Hypostyle 1

Protoloph connected 30

Crotchet 1

11.1

14.1

0

2

30

0

Mi L

W

H

Hy

Int. cingulum

Metaconule

Metastyle

Hypostyle

Protoloph connected

Crotchet

11.4

14.1

0

0
3
0

30

M1 L 11.9 11.4

W 16.4 14.5
HO 0

Hy
Int. cingulum 2 0

Metaconule 3 3

Metastyle 1 0

Hypostyle 1 2

Protoloph connected 30 20
Crotchet 0 0

M3 L 11.4 11.2

W 15.8 13.5
HO 0

Hy

Int. cingulum 3 2

Metaconule 4 4

Metastyle 1 1

Hypostyle 1 2

Protoloph connected 30 3

Crotchet 0 0

45

46

FIELDIANA: GEOLOGY MEMOIRS, VOLUME 5

Fig. 18. IX (Continued). — Lower Jaws
M. grandis

Species

CM 9157

CM 9158

CM 8743

CM 8744

Horizon

Peanut Peak

X

X

X

X

Crazy Johnson

Ahearn

Pi-M,
P2-M3
Pi -P.
M,-M,

Pr-<

Mr.

PS-P4

89.5
47.6

43.3

89.0
49.0

42.0

42.6

p*

L

13.8

14.6

13.7

W

8.6

9.1

9.4

H

Hy
Cingulum

2

p*

L

14.0

14.5

14.2

13.8

W

11.2

11.0

10.0

10.8

H

Hy
Cingulum

3

p«

L

14.8

15.0

14.6

14.2

W

12.0

11.5

10.3

11.7

H

Hy
Cingulum

3

Mi

L

13.6

14.0

13.7

W

9.9

10.2

10.2

H

Hy
Cingulum

3

M,

L

14.7

16.8

13.5

W

9.7

10.6

10.3

H

Hy
Cingulum

3

M,

L

W

H

Hy

Cingulum

17.9
9.1

3

the cingulum, varying from tooth to tooth in a single
maxilla.

2) The external cingulum is relatively heavy in the
smaller and earlier forms.

3) The hypostyle, if present at all, is developed in
all the cheek-teeth, but the shape varies irregularly from
tooth to tooth in a single maxilla.

4) The metaloph attaches to the ectoloph in only a
few individuals and may do so in only a single tooth in
a maxilla.

5) The metaconule is relatively least developed on
M3.

6) The metastyle is relatively best developed on M3.

7) Considerable wear is apparent between adjacent
teeth, affecting in particular the metastyle and the
connection between parastyle and protoloph. Since most
of this wear takes place shortly after the tooth is
erupted, teeth with slightly worn crowns differ very
much from unworn teeth in these characters.

8) Apparently hypsodonty did not increase in Chad-
ron Mesohippus. Early Chadron horses show a flat space

between the bases of the three lophs and another be-
tween the metaloph, ectoloph, and posterior cingulum.
The bases of the metaloph and protoloph expand to fill
these spaces; the tips of the protocone and hypocone
rise slightly and thus increase the angle of slope of their
internal faces; and the internal faces of the ectoloph
change from planes set at a slight angle to concave
vertical surfaces. These changes, however, occur inde-
pendently of one another. The apparent hypsodonty in
the molars of later forms results from an increase in
overall size.

9) The following characters vary independently
from specimen to specimen :

a. development of metaconule

b. hypostyle, presence and form

c. attachment of protoloph to parastyle

d. development of internal cingulum

e. development of external cingulum

f. development of metastyle

g. connection of metaloph to ectoloph
h. hypsodonty

CLARK AND BEERBOWER: THE CHADRON FORMATION

47

i. angulation of ectoloph to antero-posterior direct-
ion
j. antero-posterior length of ectoloph
k. angle of external face of ectoloph
1. ratio of transverse length of paracone
transverse length of protocone.

10) The mesohippi from the Lower Nodular Zone,
Orellan age, vary in the same characters and through the
same range of characters.

Since these observations invalidate most of the
criteria used previously in the diagnosis of Chadronian
horses, a revision of these species is necessary. If the
specimens from one member are examined and the
variability of the population or populations from which
they came is estimated, several distinct populations can
be discerned. Comparison among the populations from
all three members then reveals that these populations
fall into a few groups and that the successive popula-
tions within a group intergrade. Since the populations
from the various groups are specifically distinct in the
defined time segments, the population groups are
described as separate species. Those population groups
in which the difference between the end populations
exceeds the differences "normal" between subspecies of
a neontologic species are divided into two or more
species. Distinct populations within any one species are
named informally, employing the specific name and the
name of member from which the sample was collected.

Whenever practicable, boundaries between species
are set at breaks in the stratigraphic record, but single
specimens without stratigraphic assignment may be
indeterminable because the successive populations over-
lap in morphology. Fortunately, the gaps between spe-
cies as here denned are such that the old species types
could be determined and assigned at the species level.
Therefore the species are named on the basis of previous-
ly described types that fell within the limits of vari-
ability of the redefined species.

Mesohippus viejensis McGrew

Mesohippus viejensis McGrew, P., in press.

Type. — Not yet declared.

Referred specimens. — PM 16244; lower jaw; base of
Ahearn. PM 16250, 16248, 16246, 16247; mandibles;
base of Ahearn member. PM 13829, CM 9496; partial
maxillaries; base of Ahearn Member.

Diagnosis. — Primitive mesohippi with large median
space in molars; molars rectangular to ovoid. Size range
indicated by length of M2: 9.1 to 12.0 mm.

Discussion. — Mesohippus viejensis is clearly a dis-
tinct species (except from the succeeding population
of the group, M. celer), but further subdivision is ex-
tremely difficult. The specimens of M. viejensis from
the Vieja Formation of Texas appear to fall into two
morphological groups, and the coefficients of variability
of dental measurements likewise suggest that we have
samples from rather distinct populations. The majority
of the Vieja specimens are relatively small Mesohippi
and somewhat smaller than the M. viejensis specimens

from the Ahearn, but a few large individuals, as large or
larger than any from the Ahearn, are also known. Until
further information can be obtained on the stratigraphic
occurrence of these individuals, it is impossible to de-
termine whether they represent two contemporaneous
subspecies which occupied the area alternately during
Vieja time, or whether they represent successive popula-
tions in an evolving series.

Although the M . viejensis specimens from the Ahearn
overlap those from Texas in size and in details of dental
variation, a separation seems appropriate because the
small differences that do occur foreshadow further
changes in this phyletic line. Therefore we recognize
two populations, M. viejensis-Vieja and M. viejensis-
Ahearn.

The viejensis-Yieja population consists of small,
generally primitive mesohippi. Size is quite variable as
indicated by variation in the length of M2 (range, 9.1-
12.0 mm; mean, 10.07 mm; standard deviation 1.06 mm;
coefficient of variability, 10.5). So far as known, this
population is limited to Vieja time (see below, Strati-
graphic Paleontology).

The viejensis-Aheam population gives an impression
of somewhat greater size than the viejensis- Vieja al-
though this is not shown by the statistics (range in
length of M2, 9.4-10.5 mm; mean, 10.02 mm; standard
deviation, .519 mm; coefficient of variability, 5.16). In
addition, there are very minor differences in the details
of the upper molars. The known stratigraphic range of
this population is early Ahearnian.

Mesohippus westoni lies within the M. viejensis size
range, but the type is indeterminable because of its
fragmentary condition. The topotype (Ottawa Museum
6293) is almost certainly from a M. viejensis population,
but since the topotype is not a name bearer, M. viejensis
is the only available name.

Mesohippus celer (Marsh)

Anchitkerium celer Marsh, O. C, 1874. Notice of new equine
mammals from the tertiary formation. Am. Jour. Sci., (3) 7, no. 39,
Mar. 1874, p. 251.

Type. — Yale Museum 11302, right maxillary with
P*-M3; "Miocene" of Nebraska.

Referred Specimens. — PM 16251; mandible; Ahearn
Member. PM 13829; right maxilla with P3-M'; Ahearn
Member. CM 9394; partial maxillary; Ahearn Member.
PM 16257; partial maxillary; Peanut Peak Member.

Diagnosis. — Molars with median space very much
reduced. Size range, M1-3, 29.5 32.8 mm. Upper molars
approximately quadrangular, slight lateral expansion,
posterior borders straight or slightly convex posteriorly,
metastyle not expanded posteriorly, parastyle very little
anterior to anterior edge of tooth ; protoloph and meta-
loph developed.

Discussion. — The dentition of M. celer is very similar
to that of M. viejensis, and M. celer was probably derived
directly from a M . viejensis population. The M . viejensis-
celer species group thus consists of successive segments
in a continuum of populations, the viejensis-Vieja

48

FIELDIANA: GEOLOGY MEMOIRS, VOLUME 5

population from the late Duchesnean, the viejensis-
Ahearn population from the earliest Chadronian, and
M. celer from the remainder of the Chadron.

Mesohippus latidens

Mesohippus latidens, Douglass, E., 1903. New vertebrates from
the Montana Tertiary. Ann. Carnegie Mus., 2, pp. 161-162.

Type.— CM 751; left PMV13; Thompson's Creek
near Three Forks, Montana; "Lower White River
Beds."

Referred specimens. — CM 9078; complete upper and
lower dentition; base of Crazy Johnson Member. PM
13830; P3-M3; Peanut Peak Member.

Diagnosis. — Upper molars extremely broad laterally,
short antero-posteriorly; anterior and posterior borders
of teeth straight. Teeth not expanded externally.

Discussion. — These specimens are distinctly different
from the other Chadron species lines and show no
particular affinities to any one of them. CM 9087 from
the base of the Crazy Johnson is somewhat smaller and
has the ectoloph less angulated than PM 13830 from the
Peanut Peak or CM 751 from the Thompson Creek,
Montana. The sample is so small, however, that the
difference cannot be considered significant.

Mesohippus hyposlylus Osborn, 1904.

Mesohippus hyposlylus Osborn, H. F., 1904. New Oligocene
horses. Bull. Am. Mus. Nat. Hist., 20, pp. 170-171.

Mesohippus portenlus Douglass, E. 1908. Fossil horses from
North Dakota and Montana. Ann. Carnegie Mus., 4, pp. 268-269.

Mesohippus proteulophus Osborn, H. F. 1904. ibid., pp. 171—
172.

Mesohippus montanensis Osborn, H. F. 1904. ibid., p. 170.

Type. — AMNH 1180; anterior portion of skull with
palate and complete PLM3 on both sides; Cheyenne
River, S.D.; "Upper Titanotherium zone."

Referred specimens. — hypostylus-Ahearn (see discus-
sion below). CM 9390. CM 9392A. CM 9478.

Diagnosis. — Upper molars not expanded laterally or
only slightly expanded; ectoloph longer than internal
antero-posterior dimension of tooth; hypostyle always
present; molars without median space; posterior borders
of molars straight to posteriorly concave.

Discussion. — The M. hypostylus population group
probably branched from M . viejensis during late Duch-
esnean time, divided into two population groups during
the latter Chadronian, and gave rise through one of
these branches to Mesohippus bairdi of early Orellan
time. Within the M. hyposlylus population groups two
populations can be distinguished, hypostylus-Ahearn
and hyposlylus-Crazy Johnson. These populations over-
lap in morphology but show a progressive shift in the
population means.

A.) hypostylus-Ahearn population: The specimens of
M . hyposlylus from the Ahearn member are, as a group,
distinctly smaller than the specimens of this species from
the Crazy Johnson Member. In addition, the average of
ectoloph angle is less in the Ahearn specimens than in
the Crazy Johnson. The populations, however, overlap
in all characteristics and individual specimens cannot be

assigned without knowledge of their stratigraphic oc-
currence.

The species Mesohippus montanensis Osborn is with-
in the limits of variation of the hypostylus-Ahearn popu-
lation.

B.) hypostylus-Crazy Johnson population: Referred
specimens: CM 9035A, 9035B, 9035C, 9093, 9037, PM
13827, 13828. The average size of individuals in this
population is somewhat greater than hypostylus-Ahearn
and the ectoloph of the upper molars is relatively longer
and is set at a larger angle to the anteroposterior direc-
tion. The molar therefore narrows sharply posteriorly
and is much longer at the outside margin than at the
inside margin. The metaconule is more reduced and
metastyle more developed than in hyposiylus- Ahearn.
Most lower teeth have a complete, heavy cingulum. The
range in size is relatively great as exemplified by the
length of M2; range, 11.0-15.0 mm; mean, 12.92 mm;
standard deviation, 1.06 mm; coefficient of variability,
8.21.

The rather high variability of the hypostylus-Crazy
Johnson sample may have resulted from one of several
factors: 1) high variability of the local population, 2)
sampling from two or more subspecific populations
which occupied the area at various times, 3) fluctua-
tions in the local population through time, and 4)
sampling error.

The largest specimens approach the Peanut Peak
M . grandis (see below) in size and are otherwise similar
to that species. These individuals may represent a
separate population evolving in the direction of M.
grandis. The stratigraphic range of this population in
South Dakota is through Crazy Johnson time, but may
extend to early Peanut Peak time in Montana (Pipe-
stone Springs, see below, Stratigraphic Paleontology).

M. proteulophis Osborn falls within the hypostylus-
Crazy Johnson population limits since the type is near
the mean size of the population and the fusion of the
metaloph with the ectoloph is not a significant charac-
ter. Although described as diagnostic by Osborn this
character does not appear to be associated with any
other character difference; it is present on P4 in the
M. proteulophus type but nowhere else in the dental
series and on P3 and M3 in CM 8775 but nowhere else
in that specimen ; it is present on an M . celer specimen
from the Ahearn member on P3 and M3; and it is un-
known on specimens from the lower Brule. Together
these facts indicate that the character was not fixed in
any Mesohippus population and hence is not of taxo-
nomic significance.

Mesohippus grandis, n. sp.

Type.— CM 9157, lower jaw with P2-M3, left (Figs.
19, 20).

Horizon. — Peanut Peak Member, Chadron Forma-
tion.

Locality. — West flank of Quinn Draw, Washington
Co., South Dakota.

CLARK AND BEERBOWER: THE CHADRON FORMATION

49

Fig. 19. The type specimens of Daphoenocyon minor (1) and
Mesohippus grandis (3-4), also an unnamed species of Mery-
coidodon (2).

Referred Specimens.— CM 9158, CM 8743, CM 8744.

Diagnosis. — Animals of large size, length P2-M3
equals 89.5 mm. The external cingulum on the lower
teeth is very heavy. Characteristics otherwise as in the
M . hypostylus-Crazy Johnson population (see above).

Fig. 20. — Measurements of Mesohippus grandis

CM 9157 CM 9158 CM 8743 CM 8744

P,— M,

89.5 mm.

89.0*

P.— 4

43.3

42.0*

42.6

Mr-.

47.6

49.0

DL

13.8

14.6

13.7

p'w

~8~T6

~9~T

~9A

„L

14.0

14.5

14.2

13.8

Flw

lTT2

TTTo

ToTo

T0T8

PL

14.8

15.0

14.6

14.2

F*w

12.0

1TT5

1073

11.7

w

13.6

14.0

13.7

"9T9"

10.2

10.2

mr L

14.7

16.8

13.5

M'w

9.7

10.6

10.3

XM L

17.9

Mjw

9.1

Discussion. — Since these specimens are markedly
larger than M. hypostylus they may properly be ranked
as a separate species. This population however is
closely affiliated with M. hypostylus and apparently
represents a continuation of the trend toward greater
size shown by some members of the Crazy Johnson
population. The species so far as is known is limited to
Peanut Peak time.

Evolution of Mesohippus during chadronian
time: Although most Chadron horses are poorly known,
we feel that sufficient material is available to justify dis-
cussion of their evolution. The early M . viejensis popula-
tions represented by the M. viejensis-Vieja population
appear to be suitable ancestors for most of the Chad-
ronian Mesohippi (Fig. 21). The M. viejensis-Ahearn
sample is distinct from M. viejensis-Vieja only in slightly
larger size (although the difference is not significant
statistically) and minor characters of the upper molars.
Mesohippus celer specimens known from the upper

Ahearn and from the Peanut Peak differ by a very
slight increase in size and by the reduction of the median
space in the upper molars. The M. viejensis-Vieja sam-
ple is quite variable with respect to size (length of M2,
V = 8.9, N = 6) and thus it is difficult to demonstrate
any size changes between M. viejensis-Vieja and M.
celer. The successive populations in this line were quite
conservative in tooth evolution and in size.

Mesohippus hypostylus is distinct from M. viejensis
when the species appears in the Ahearn but might well
have been derived from early M. viejensis populations.
The samples are so small as to preclude a meaningful
statistical test, but M. hypostylus was probably some-
what larger than M. viejensis by Ahearnian time. Fur-
ther, M . hypostylus lacks the median space characteris-
tic of M. viejensis, but this feature cannot be taken as
debarring M. viejensis from an ancestral position since
the viejensis-celer line also reduced this character,
though at a somewhat later time. If M . hypostylus was
not derived from M. viejensis it must have come from
a closely related species, and this would be effectively
the same so far as interpretation of evolutionary pat-
terns is concerned.

The Crazy Johnson specimens of the M. hypostylus
group are on the average slightly larger than those from
the Ahearn, but the overlap is complete. In the M.
hypostylus-Crazy Johnson population variability of den-
tal minutiae was considerable. The shape of the hypo-
style, reduction of the metaconule, development of the
metastyle, attachment of metaloph to ectoloph, at-
tachment of protoloph to parastyle, development of
internal cingulum, and angulation of the ectoloph all
vary considerably in detail and appear in all conceivable
combinations. This suggests a considerable amount of

Phylogeny of

Chadron Mesohippi

z
<
_l

-1

UJ

O

M. BAIRDI

1-

Z<

<UJ
UIO.

1
\

s
\

V\ M GRANDIS

z
<

z
0

X

0

<
X
0

M. CELER

~"---.. /M. latidens

z

<z

rrx
00

->

V i

M. HYPOSTYLUS

z

<
UJ

X

<

M. CELER

M. HYPOS

1

/
/

JTYLUS /

/

s

M VIE

JENSIS

z
<

UJ

z
<n

UJ

X

0

O

<

Ul

>

M. VIE

JENSIS

— -'"

1

Fig. 21. Phylogeny of Chadron Mesohippi.

50

FIELDIANA: GEOLOGY MEMOIRS, VOLUME 5

underlying genetic variability and low selection pres-
sures for these details and for different combinations of
them. The expression of some of these characters, such
as internal cingulum and the metaconule, differs con-
sistently in M1-2 from that in P2-4 and M3. A similar
order is also exhibited in M. bairdi from the lower
Brule. This phenomenon seems to be related to the
different period of development and time of emplace-
ment for the two sets, M1_ 2 and P2-4-M3, but might be
either an "accident" of development or the result of
selection for developmental correlation in those units,
M'~2, which have somewhat similar periods of utiliza-
tion.

Certain other characters such as the attachment of
the metaloph to the ectoloph, the crotchet, and the
shape of the hypostyle vary irregularly from tooth to
tooth within the individual. The hypostyle, however, if
it is present at all, is present in all the teeth of an
individual, but the pattern of the style will differ from
tooth to tooth. This observation suggests that hypostyle
development was controlled by a single gene which af-
fected the developmental field of all the teeth but
whose expression was modified by the particular de-
velopmental pattern of individual teeth.

The known late Chadronian representatives of the
M. hypostylus line are all large mesohippi included
within M. grandis and represent the apparent culmina-
tion of one evolutionary trend in M. hypostylus.

Available material of the Chadron horses is insuffi-
cient to demonstrate with certainty the origins of the
Brule horse M. bairdi. The smaller M. hypostylus speci-
mens cannot be distinguished from M. bairdi on dental
characters, but no small hypostylus are known from the
late Chadron. On the other hand, M. celer is also quite
similar to M. bairdi in both dental characters and size
and is known from the late Chadron. Therefore we can
make alternative hypotheses pending discovery of
skeletal material:

1) The M. bairdi population of early Brule time was
derived from a M. hypostylus population — either from a
population of small M. hypostylus that lived in an area
other than South Dakota during late Chadron time or
by a reversal of size trends represented in the M. grandis
population. The first seems more likely.

2) M. bairdi was derived from M. celer as a con-
tinuation of the evolutionary trend that increased size
and reduced the median molar space in the viejensis-
celer line.

A distinct line is represented by M. latidens, known
from the base of the Crazy Johnson (1 specimen) and
from the Peanut Peak (1 specimen). M. latidens re-
sembles M. hypostylus somewhat more than it does M.
celer, but might have been derived from some earlier
species (possibly M. viejensis) or from an unknown
species.

The probable phylogenetic relationships of Chadron
horses are summarized in Figure 21. Alternate interpre-
tations are possible, but this is offered as the most
probable and conservative.

Family Menodidae (=Brontotheriidae Marsh 1873)
Genus Menodus
Menodus giganteus Pomel, 1846

Synonyms. — Titanotherium Leidy, Megacerops Cope,
Brontotherium Marsh, Symborodon Cope, Brontops
Marsh, Diploclonus Marsh, and Allops Marsh.

Specimens— PM 16266, 16267, 16268, 16269; horn
cores of four individuals; Crazy Johnson Member. PM
16270; horn core, nasals, and frontal sinuses; top of
Crazy Johnson Member. Uncollected partial skulls and
scraps observed in the Ahearn, Crazy Johnson and
Peanut Peak Members.

Description. — The Ahearn Menodus are generally
small- to medium-sized titanotheres. The largest horn
found was 7 in. long, with an ovoid cross-section 3 by
5.5 in. The smallest specimen found was the skull of an
immature individual in which M3 had not erupted.
Horns of this specimen were 2 in. long, and circular in
cross-section with a diameter of 2 in. at the base.

The specimens from the Crazy Johnson include:
PM 16267, a horn core with subtriangular cross-section,
3 in. long and 3.5 in. greatest transverse diameter; PM
16266, a horn core with subtriangular cross-section, 6
in. long, and 4.5 in. greatest transverse diameter; PM
16268, a horn core with ovoid cross-section, 8 in. long
and 5 in. greatest transverse diameter; and PM 16269, a
horn core with flat cross-section over 8 in. long (at least
2 in. missing), cross-sectional dimensions 5 in. by 2 in.

Specimen 16270 consists of the complete right horn
core, the nasals, and both frontal sinuses. The transverse
ridge between the horns is very high and the horn is low
and tapering with a shape in cross-section of a blunt
isosceles triangle with the base posteriad. The horn and
transverse ridge are produced vertically with their
posterior surface sloping up to 85° from the plane of the
roof of the nasal passage, but the horns themselves are
relatively reduced. The frontal sinuses are a pair of
subhemispherical pockets, 2.75 in. in diameter and
flattened posteriorly; apparently they opened ventro-
posteriorly into the nasal cavity.

Discussion. — The four specimens of Menodus from
the Crazy Johnson (PM 16266 through PM 16269) were
collected from a single "graveyard" in a layer 1 ft. thick
and within 30 ft. of each other horizontally. Thus they
demonstrate the marked variation in horn development
in a contemporaneous population of titanotheres. It is
difficult to believe that each individual represents a
separate species, and we are inclined to the view that
they are simply variants within a single species popula-
tion, indeed perhaps even in a single local interbreeding
unit. We offer therefore the following hypothesis for
Oligocene titanothere taxonomy and phylogenetics:

1) Development of horns in Menodus was the result
of generalized bone growth in the nasal region.

2) The developmental system that controlled this
growth was not a highly homeostatic system but rela-
tively indeterminate.

CLARK AND BEERBOWER: THE CHADRON FORMATION

51

3) Minor disarrangements in the genetic background
or "accidents of development" would thus profoundly
modify the shape of the horns. Sexual dimorphism
would similarly produce major variations.

4) Differences in species might then show in the
horns with respect to: (a) kind of variability; (b) amount
of variability.

5) Phylogenetic changes might also be of the same
kind. Such differences are impossible to demonstrate
without a large series of specimens with accurate, de-
tailed stratigraphic assignments. Until such a series
becomes available, it seems proper to regard all speci-
mens as cospecific with the first-named species ade-
quately described from the same formation and locality,
which lies within the range of variation of the specimens
at hand.

We recognize but two genera of titanotheres from
the Oligocene of North America: Menodus and Teleodus.
The diagnostic characters of the two genera are:

Menodus

Teleodus

(1)

I

I

0-2

3

(2)

Horns small to large, subconical

Horns small, sub-

to cylindrical or flat in longitudi-

conical in longi-

nal section.

tudinal section.

(3)

Top of skull always concave up-

Boss on top of

ward.

male skulls.

With the exception of Marsh's type, no specimen
referable to Teleodus has ever been collected from the
Big Badlands. All of the abundant Teleodus specimens
collected since 1930 have come from formations known
to be somewhat older than the typical Chadron of South
Dakota. This suggests that possibly either Marsh's type
was collected from the Slim Buttes Formation in the
Badlands, or that the locality data were confused in
some manner and the specimen is not from the Big
Badlands.

Selection of names for the genus and species em-
bracing all other Chadron titanotheres has been most
difficult. Menodus giganteus Pomel, 1846, has obvious
priority. However, the type specimen, which has been
lost, consisted of a fragment of a lower jaw, with RM2_3.
Except for its size, the type did not include characters
which would exclude it from Teleodus or from some of
the larger species of the Eocene Protitanotherium. Fortu-
nately, Osborn (1929) designated a neotype skull,
AMNH 505, which does possess diagnostic characters.
The designation seems to be in accord with all of the
rules of the International Code of Zoological Nomencla-
ture for neotypes. The name is, therefore, available,
and the genus and species are adequately diagnosed.
Our characterization is an emendation of previous
descriptions, in order to include within this genus and
species the various Oligocene forms described as sepa-
rate genera.

PM 16270 lies somewhat outside the accepted limits
of Menodus variants, and the reduced horn cores,
vertically produced frontal area, reduced nasals, and

expanded frontal sinuses suggest specializations like
those of the Asiatic embolotheres. Because of the
fragmentary nature of the specimen, this can, however,
be no more than a suggestion.

Family Helaletidae
Genus Colodon
Colodon occidentalis

Specimen. — CM 9482; lower jaw with P2-M2;
Ahearn member.

Family Hyracodontidae
Genus Hyracodon
Hyracodon cf . priscidens

Referred Specimens. — CM 9129; partial skull, face
crushed laterally and brain-case crushed dorso-ventral-
ly, includes DP1-4, M1-2 with M2 only partly erupted.
Base of Crazy Johnson Member. PM 13395 ; palate with
P-M1, jaw with P2 M3; Peanut Peak Member. CM
8717; two lower molars; Ahearn Member. CM 8787;
associated P2-Mi; Crazy Johnson Member. CM 8716;
partial mandible with DP4-M2; Ahearn Member. CM
8715; maxillary fragment with P2-4; Ahearn Member.
PM 14026; fragments of maxilla, P3-M3; Ahearn Mem-
ber. PM 16275; mandible, DP3-Mi; Ahearn Member.
PM 16305; DP3 4, Ahearn Member. CM 9091; mandib-
ular fragments with symphysis, DP2_4, alveoli of I, C
and DPi, rudimentary P2; Crazy Johnson Member.
CM 9099; jaw fragment with DP2_g and alveolus, DPi;
upper part of Ahearn Member.

Discussion. — All of the Chadron Hyracodon appear
to belong to the small species, H. priscidens. CM 9129
is referred to this species because M1-2 are almost
identical with those of the type (Lambe, 1905). CM 8787
is similarly referred since the size of teeth (P2_4 Mi) is
very small for Hyracodon; PM 16275 has only one
tooth (Mi) that can be compared with the type material
but it is of the same size; CM 9091 and CM 9099 have
only the deciduous premolars but these are sufficiently
similar to the same teeth in PM 16275 that these are
likewise referred to H. priscidens.

CM 9091 and 9099 are also of interest because they
show the alveoli for DPi. This tooth is absent in the
Orellan Hyracodon juveniles (Scott and Jepsen, 1941)
but apparently was still present as a vestigal structure
in some Chadronian individuals.

The deciduous premolars were retained for a rela-
tively long period in some individuals of this genus. In
CM 9129 DP'-4 have long, firm roots with no traces of
P1"2 behind the milk teeth, although M2 has started to
erupt. The full set of deciduous teeth is also present in
CM 9091, in which the mandibular symphysis is
completely fused. A specimen of H. nebraskansis (CM
3523) from the Orellan of Sand Creek, Sioux Co., Ne-
braska, also has DP 1-4 firmly rooted with M2 partly
erupted. In both CM 9129 and 3523 DP2"4 are fully
molariform and DP1 is submolariform. Wear has re-
moved approximately one-half of the crown height in
CM 9129 and about three-fourths in CM 3532. A fully

52

FIELDIANA: GEOLOGY MEMOIRS, VOLUME 5

adult specimen (CM 11723) of H . nebraskansis has P3-4
about as much worn as M3 and much less worn than M2,
but two other members of this species from the Carnegie
Museum collection have P3~4 as worn as M2. This
would indicate considerable variation in time of pre-
molar replacement but more extended studies would
be necessary to determine whether this is of taxonomic
significance. In every adult specimen examined the
entire premolar series was uniformly worn, suggesting
that emplacement of the series was nearly simultaneous.

Hyracodon sp.

Referred Specimen.
Johnson Member.

PM 16287, right P4; Crazy

Family Rhinocerotidae
Genus Caenopus
Caenopus mitis

Referred Specimens. — CM 9153; complete mandible
with P2-M3 of both sides and single small alveolus for
Px; Peanut Peak Member. CM 9396; left mandibular
ramus with P3-M3; Peanut Peak Member. PM 16274;
mandible with P2 M3; Crazy Johnson Member. CM
9100; mandible including Pi-M2; Ahearn Member.

Discussion.— The teeth in CM 9153 and 9396 are
quite similar, differing only in the vestigial Pi in 9153
and a tiny antero-posterior ridge across the posterior
valley of P4 that is present only in 9396. The jaws, how-
ever, are quite different: that of CM 9153 decreases con-
tinuously in depth from the rear of M3 to the chin ; that
of CM 9396 maintains the same depth from the rear of
M3 to below P4 and tapers rapidly forward of P4. The
jaw of PM 16274 resembles that of CM 9396. All three
of these specimens are larger than expected in Caenopus
mitis (Wood, H. E., 1927; Scott and Jepsen, 1941) but
are considerably smaller than Subhyracodon. The type
of C. mitis does not include the lower teeth, so direct
comparison is impossible and none of the characters
used by H. E. Wood (1927) to differentiate Caenopus
from Subhyracodon occur on the lower cheek teeth.
Since C. mitis is a Chadronian species and since these
specimens cannot be excluded from C. mitis on char-
acteristics of the lower dentition, we feel that the most
conservative position is to assign them to C. mitis.
These specimens may not be congeneric, much less co-
specific, but the differences in jaw depth and minor
dental characters do not appear to justify their separa-
tion. CM 9100 is apparently a typical C. mitis, at least
with respect to size.
Caenopus sp.

Indeterminable Caenopus material includes three
single teeth from different localities in the Crazy John-
son, and an M1, CM 9467, from the Ahearn Member.

Genus Trigonias
Trigonias osborni

Rer erred specimens. — CM 9399; jaw fragment with
Mi_3| Crazy Johnson Member. CM 9398; assorted upper
and lower teeth from at least two individuals; Crazy

Johnson Member. CM 9397; Pi~M3; Crazy Johnson
Member.

Discussion. — In view of the known variability in
dental characters of Trigonias, both from specimen to
specimen and from tooth to tooth within one jaw, we
regard T. osborni as the only species satisfactorily
established at present. We therefore refer these frag-
mentary specimens to T. osborni.

Trigonias sp.

In addition to the material referred to T. osborni, an
incisor from the Crazy Johnson and a second upper pre-
molar from the Ahearn are assigned to this genus.

Order Artiodactyla

Family Entelodontidae

Genus Archaeotherium

Archaeotherium cf. scotti

Referred Specimens. — PM 16281; lower canine tip
and root, broken; antero-posterior length at base of
enamel, 19 mm.; upper Ahearn. PM 16282; left M2;
upper Ahearn Member.

Archaeotherium cf. mortoni

Referred Specimens. — PM 16286; anterior part of
right M2; middle Ahearn Member. CM 9412; right
mandibular ramus with Mi^; Ahearn Member. CM
9097; left M1; Ahearn Member.

Archaeotherium cf. coar datum

Referred Specimens.— PM 16285; partial left M3;
middle Ahearn Member. PM 16283; partial left P4;
upper part of Ahearn Member.

Family Tayassuidae
Genus Perchoerus
Perchoerus cf . minor

Referred Specimen. — CM 9492; fragments of man-
dible with heel of right M3; Ahearn Member.

Discussion: The size of this specimen (antero-poster-
ior length of M3, 15.5 mm and depth of jaw below anter-
ior part of M3, 19 mm) would place it within the range
of P. minor and P. nanus. Since the type of P. minor is
from the Chadron of Nebraska, CM 9492 is referred to
this species. The type of P. nanus was collected from
the "White River Miocene" of Nebraska, and Scott as-
signed the specimen to the lower Brule because of the
similarity of the matrix to lower Brule lithology. Parts
of the Nebraska Chadronian, however, resemble the
Brule and the time and specific relationships of P. minor
and P. nanus are problematical.

Family Anthracotheriidae
Genus Bothriodon
Bothriodon cf. americanus

Referred Specimens. — CM 9405; pair of lower jaws
with P2-M3; Crazy Johnson Member. CM 9096; palate
with P2-M3; Crazy Johnson Member.

CLARK AND BEERBOWER: THE CHADRON FORMATION

53

Bothriodon sp.

Referred Specimens. — CM 9498; jaw with P2, Mi_2;
Ahearn Member. CM9500; P3, M'-2 associated; Ahearn
Member. PM 16298; jaw fragment with partial molar;
Ahearn Member. PM 16433; skull and skeleton, juve-
nile; Peanut Peak Member.

Discussion. — These specimens are within the size
range of Bothriodon and are the earliest anthracotheres
recorded from North America.

Genus Heptacodon
Heptacodon

Referred Specimens. — CM 8779; jaw fragment with
P2-3, DP4, Mx; Crazy Johnson Member. PM 16299; Mx;
Crazy Johnson Member. PM 16300; M2; Crazy Johnson
Member.

Family Agriocheridae
Genus Agriochoerus
Agriochoerus cf. antiquus

Referred Specimens. — PM 16278; partial upper den-
tition; middle of Ahearn Member. PM 16277; pair of
maxillae and partial lower jaw; upper part of Ahearn
Member. CM 9128; upper molars and associated lower
jaws; base of Crazy Johnson Member.

Discussion. — The two specimens from the Ahearn
(PM 16278 and 16277) are not distinguishable from a
small A. antiquus (PM 12538), from the lower Brule.
The P3 of 16277 has the deuterocone well-developed and
internal to the protocone, so that the tooth is almost an
isosceles right triangle with the hypotenuse antero-
internal; P3 of 16278 has the inner side of the tooth
compressed with the deuterocone directly posterior to
the internal buttress of the protocone, and the external
face of the tooth indented between the roots. P4 of
16277 has a tiny swelling on the postero-internal cingu-
lum, with a single small ridge projecting from it toward
the center of the tooth, like the leg of a T. P4 of 16278
has the bunodont conule on the cingulum, but a tiny,
trifid crest, unconnected with any of the cones, lies in
the position of the small ridge of 16277. However, the
lower Brule specimens of Agriochoerus in the Princeton
collection show great individual variation in P3 and P4,
so these characters may be of no significance.

Possibly these two specimens are not cospecific with
the Brule forms, but more adequate material would be
necessary to demonstrate any difference.

The Crazy Johnson specimen, CM 9128, is slightly
smaller than the two specimens from the Ahearn Mem-
ber, but otherwise indistinguishable from them.

Agriochoerid, gen. and sp. indet.

Referred Specimen. — CM 9092; maxillary fragments
with DP4-M1; Crazy Johnson Member.

Discussion. — This specimen is within the size range
of Mesagriochoerus, but the protoconule is absent as in
Merycoidodon and the mesostyle is quite low as in
Agriochoerus. The M1 as a whole suggests a very small
Agriochoerus, but the fragmentary condition of the
specimen prevents definite reference to that senus.

Fig. 22. A, Type specimen of Merycoidodon lewisi, CM 9105,
crown views of skull and mandible. B, Type specimen of Mery-
coidodon lewisi, CM 9105, lateral view of skull.

Family Merycoidodontidae
Subfamily Merycoidodontinae
Genus Merycoidodon
Merycoidodon lewisi1 n. sp.

Type. — CM 9105; skull, jaws, and skeleton except
fore limb and part of tail ; Peanut Peak Member 12 ft.
below Chadron-Brule contact; Big Corral-Quinn Draw
divide, NW-1/4 sec. 30, T. 43N, R 45W, Shannon Co.,
S. D.

Referred Specimens. — PM 16289; skull and jaws;
Peanut Peak Member. PM 16434; left lower jaw with
Ii-M2; Peanut Peak. PM 16435; symphysis with I1-P3;
Peanut Peak Member. PM 16436; crushed skull with
P'-M3; Peanut Peak Member. PM 16437; skull with
C-M3; Peanut Peak Member. PM 16438; skull with
P3-M3; Peanut Peak Member.

Diagnosis. — Skull typically Merycoidodon in basi-
cranial structures, but large, robust, with heavy canines,

1 The species is named for Mr. Arnold D. Lewis, who collected
the type specimen.

54

FIELDIANA: GEOLOGY MEMOIRS, VOLUME 5

Fig. 23. — Measurements of Skull of Merycoidodon leurisi,
CM 9105, in mm.

I. SKULL

Length, dorsal, along midline 225 . 5

Width, altered to remove crushing 128.0

Depth, M1 alveolus to top of skull 66.0

Width at infraorbital foramen 45.5

Length, prosthion-basion 206.5

Length, prosthion-choana 120 . 5

Length, bregma-inion 73 . 0

Height, canine alveolus-naso-premaxillary suture 39.0

Palate, width between canines 28 . 5

Palate, width between M R and L 32.0

II. UPPER DENTITION

P'-M» 96.3

P'-P* 46.4

M'-M» 50.5

P' L 11.3

W 6.5

P» L 12.5

W 7.9

P> L 12.4

W 11.9

P« L 9.4

W 14.2

M' L 15.2

W 16.7

M« L 18.5

W 20.1

M' L 20.2

W 20.2

III. MANDIBLE

Greatest length 182.0

Length of symphysis 49 . 5

Depth at P, 33.0

Depth at anterior edge of Mj 39 . 0

IV. LOWER DENTITION
P,-M, 100.0
P,-P« 46.8
P.-P, 35.4

M,-M, 52.3

P, L 11.4

W 5.7

P, L 11.3

W 7.4

P, L 13.1

W 9.3

M, L 13.2

W 10.6

M L 15.5

W 12.7

M, L 23.6

W 14.4

long palate, almost straight transverse row of incisors,
strong postorbital construction, and flaring zygoma.

Description. — The bullae are absent from the type
specimen and if present in life must have been small.
The paroccipital processes are long and slender with
antero-internal ridge; postglenoid processes small; squa-
mosal root of zygomatic arch narrow antero-posteriorly,
and the zygomatic arch is straight. Skull mesocephalic
with greatest width immediately anterior to glenoid;
palate long; and zygomatic arches angulated so that
their anterior extensions would meet about 1 in. in front
of the incisors. Postorbital constriction very narrow,
braincase flaring posteriorly to a triangular shape;
foramen rotundum reduced to one-half the diameter of
foramen ovale; nasal bones posteriorly rounded, not
acute; naso-frontal suture nearly a transverse line.

Upper canines and P1 massive and long; upper in-
cisors ranged in almost straight transverse row; back of

I1 posteriad to front of P. Mandible long with very long
symphysis; massive Pi and relatively small P4.

Discussion. — Of the characters listed above, the
basicranial characters are typical for Merycoidodon, but
the remainder of the skull suggests Eporeodon. The
basicranial characters seem adequate to refer this
species to Merycoidodon — the robust build, long palate,
heavy canines, etc., might be a function of size or very
possibly of sex rather than indicating a relation to
Eporeodon.

This species is probably not ancestral to any Orellan
species of Merycoidodon, but it might be ancestral to
Eporedon. The reduction of the foramen rotundum and
the shape and position of the naso-frontal suture are so
variable in Merycoidodon and Eporeodon that they are
probably not significant in evaluating phyletic relation-
ship.

Merycoidodon sp.

Specimen. — PM 16276; maxillary fragment with P4,
M1-"-, partial P3, M3; Ahearn Member; Big Corral Draw,
main fork about 1 mile north of the Pennington- Wash-
ington County line, Pennington Co., S. D. This speci-
men probably represents an undescribed species, but its
fragmentary nature precludes using it as a type.

Description. — Tooth structure typically Merycoido-
don except that internal cingulum on P4-M3 and prob-
ably on P3, also, is unusually strong and denticulate;
two internal crescents of molars join each other rather
than remaining distinct; anterior end of external cres-
cent of P4 forked, the internal fork supported by a stout
pillar which reaches the anterior face of the tooth but
does not unite with the internal crescent.

The tooth row is moderately arched laterally rather
than straight as in other species. P4 is set one-third of its
width mesially to M1, producing a sharp angulation
along both the internal and external row of tooth faces.
The posterior rim of the zygomatic root lies opposite
the posterior edge of M3 rather than opposite the middle
ofM3.

Discussion. — Although the specimen is incomplete,
it is markedly different in the described characters from
any other species of Merycoidodon, lying well outside
the range of variation of the other Chadron species and
also of the Brule species. More complete specimens may
demonstrate that this species is not referable to Meryco-
idodon, but the fragmentary nature of the specimen
makes this reference advisable at present. None of these
characters resemble those of Protoredon, with the excep-
tion of the posteriorly-placed zygoma. This species,
whatever its affinities, is as highly evolved in the parts
preserved as is M. culbertsoni.

Its apparent rarity is probably real rather than an
accident of collection. The relative abundance of other
small artiodactyls in the Ahearn collections, plus the
fact that for several seasons the senior author has col-
lected every generically identifiable scrap discovered,
makes it seem probable that Merycoidodon was rare in
South Dakota during Ahearnian time.

CLARK AND BEERBOWER: THE CHADRON FORMATION

55

Merycoidodon sp.

Referred Specimen. — CM 9391; fragments of lower
jaw with Mi_3, poorly preserved; Ahearn Member.

Discussion. — The specimen is of the size and general
character of M. lewisi. Since it is indeterminable, it
serves only to suggest the presence of M. lewisi or a
related species in the Ahearn Member.

Family Camelidae
Subfamily Poebrotheriinae
Genus Poebrotherium
Poebrotherium cf . andersoni

Referred Specimens: PM 16260, two tibiae, one com-
plete and one partial pes, one fore limb, partial scapula,
and pelvis; Ahearn Member. PM 16261; upper right
P4-M3, Ahearn Member. PM 16262; jaw with P^,
DP4, Mi_2) M2_3 and fragments of other teeth ; Ahearn
Member.

Discussion. — These specimens may all be parts of
the same individual inasmuch as they were collected
from a 2-in. lamina within 6 ft. of each other. They were,
however, part of a mechanical association including four
Mesohippus, a hypertragulid, Parictis (Campylocynodon)
parvus, and another camelid, and so could be from dif-
ferent individuals.

The limb bones are somewhat larger than those of
the described species of Poebrotherium but are otherwise
typical. The metatarsus and metacarpus are about the
same length. The humerus measures 134 mm, but a
portion of the proximal end is missing. Allowing 10 mm
loss, which is very generous, the length of the humerus
becomes 145 mm; the metacarpus is 140 mm and the
metatarsus 145 mm. This is about the limb proportion
of the modern camel and llama. Since these specimens
are from the base of the Chadron formation, they are
the oldest known Poebrotherium. That the oldest known
specimen of the genus is the largest and is also the "most
advanced" in limb proportions is a sure indication that
the actual evolution of the genetic line is still unknown.
Many of the really significant steps in camelid evolution
had already been accomplished before earliest Chad-
ronian time, when these specimens lived. Only fusion of
the already-appressed metapodials, development of
hypsodonty, and increase in size remained to be ac-
complished.

Camelid, n. g. and sp.

Specimens.— PM 16313; right maxilla with DP3"4,
M1-2; PM 16263; left forelimb with humerus, partial
radius and ulna, left hind limb with femur, tibia, pes;
Ahearn Member.

Discussion. — A more complete specimen from Mc-
Carty's Mountain, Montana, congeneric with these, is
being studied at the Frick Laboratory and so this ma-
terial will be left unnamed pending publication on the
McCarty's Mountain specimen. These two Dakota
specimens, which are probably from the same indi-
vidual, are of an animal about the size of Eotylopus but
with four complete metatarsals and four toes on the

hind foot. The tibia is slightly longer in proportion to the
humerus than in Poebrotherium and the metatarsus is
much shorter in proportion to both the humerus and
tibia. This camel is very apparently adapted to a moist,
forested or brushy habitat in contrast to the dry plains
adaptations of Poebrotherium.

Camelid, genus indet.

Specimen. — CM 9023; M^; Crazy Johnson Mem-
ber.

Discussion. — This specimen although generically in-
determinable is in the Eotylopus size range.

Superfamily Hypertraguloidea

Several species of small selenodont artiodactyls are
common in the Chadron of South Dakota. The tax-
onomy of the hypertraguloids is, however, so badly
confused that we are reluctant to assign these speci-
mens to the recognized species and genera of hyper-
traguloids. Until an extensive review of these Chad-
ronian artiodactyls is made, based on modern taxonomic
principles, it seems best simply to refer them to the
superfamily. For the present study, the most significant
points are: 1) the abundance of medium to large hyper-
traguloids in the Ahearn and Crazy Johnson Members;
2) absence of these types from the Peanut Peak Mem-
ber; 3) presence of a few small hypertraguloids in the
Peanut Peak and Crazy Johnson Members.

The following specimens have been collected from
the South Dakota Chadronian: Ahearn Member: PM
16290; PM 16291; CM 9490; PM 16292; PM 16293
PM 16295; PM 16294; CM 9468; CM 9491; PM 16296
CM 8707; CM 8708; CM 8709; CM 8710; CM 8711
CM 8712; CM 8713; CM 8714.

Crazy Johnson Member: CM 9393; CM 9033; CM
9022; CM 9032; CM 9088.

Peanut Peak Member: PM 13834; CM 8703; CM
8704.

STRATIGRAPHIC PALEONTOLOGY

INTRODUCTION

The faunules of the three members of the South
Dakota Chadron Formation are now sufficiently known
to allow detailed correlation of other Chadronian de-
posits with this standard section. Unfortunately, one
large element in these faunules, the small artiodactyls,
needs taxonomic revision before it can be used in cor-
relation, and so is not now available as evidence. In ad-
dition, the rodents and some genera and species in other
orders are so extremely rare in the Chadron that their
occurrence in one member does not give satisfactory
evidence of their stratigraphic range.

In general, correlations are based on the relative
numbers of identical species. In making correlations of
deposits that accumulated within these relatively brief
intervals of time, however, we recognize that local
ecological differences are likely to modify significantly
the local stratigraphic ranges. This effect is most ap-

56

FIELDIANA: GEOLOGY MEMOIRS, VOLUME 5

parent among the reptiles, whose geographic distribu-
tion is sharply limited by temperature, but it must also
exist in most mammalian species.

On the basis of general environmental requirements
and associations, two groups of mammals can be dis-
tinguished within the Chadronian faunas. One of these
consists of elements derived from a pre-Chadronian wet-
forest chronofauna (see p. 68). The other group con-
sists of invading elements from a savannah-savannah
forest1 chronofauna. It seems probable that the wet-
forest chronofauna or a large number of its elements
might survive longer in some areas than in others and
that the invading species would appear at different
times in different areas responding to changes in the
local environment. Therefore, major faunas as well as
individual species would overlap in time.

If the interdependence of species in a chronofauna
were complete, then this chronofauna would behave as a
unit and the species that compose it would always be
found in association. If, as seems more likely, the species
are only partially dependent on one another they would
be associated in most occurrences but not invariably.
Further, in this more flexible chronofauna, some species
might be replaced either through evolution into new
species or by immigration of a different species. In high-
ly integrated (interdependent) chronofaunas, this re-
placement would probably be almost simultaneous
through the area occupied by the chronofauna.

Therefore, although neither the appearance or dis-
appearance of chronofaunas nor of species independent
of the local chronofaunas forms a critical time-marker,
the replacement of species within the local chronofaunas
does constitute such a marker. In addition, we accept
the first appearance of a species as a critical time-
marker if the chronofauna of which it is part has
previously occupied the local area, and the disappear-
ance of a species if the chronofauna of which it is a part
persists locally. The latter two types of occurrence, how-
ever, are considered less significant than the replace-
ment of a species within a chronofauna.

Relation to Pre-Chadronian Faunas

The known fauna of the Ahearn Member differs
strikingly from that of the Duchesne River and Sespe
Formations, and there can be little doubt that the
previous concensus (Simpson, 1946; Scott, 1945; etc.)
was correct in considering the oldest Chadronian ap-
preciably younger than the Duchesnean. (Fig. 24).

The Vieja fauna, however, is much closer to that of
the Ahearn. Although much of the Vieja fauna has not
been studied critically (as of 1957), Agriochoerus and the
horse, Mesohippus viejensis, are common to both the
Vieja and the Ahearn. On the other hand, the mesohippi
differ at the intra-specific level; the agriochoerids may
be of different species; and Teleodus, though present in
the Vieja, is unknown from the Ahearn. Therefore, the
Ahearn is probably a little younger than the Vieja.

1 Henceforth in this paper, the term "savannah" will be used to
include true savannah and the associated savannah-forest environ-
ment.

Comparison of Yoder and Ahearn faunas supports
the belief, expressed previously by the senior author,
that the Yoder is the time equivalent of the Ahearn.
Those species which have been described as distinct are
based upon fragmentary specimens, at best doubtfully
determinable, and in our opinion within the range of
variation of known Chadronian species. Preliminary
studies of the small artiodactyls of the two formations
supports the idea of contemporaneity, and Wood (1955)
concluded from a study of the rodents that the Yoder
was slightly younger than Vieja — an assignment that
might correlate the Yoder with the Ahearn. Studies on
the Yoder collections of the South Dakota School of
Mines should test this correlation and establish the age
of the Yoder more definitely.

Relation of the Pipestone Springs Fauna

Figure 25 makes apparent the similarity of the Pipe-
stone Springs fauna from Montana to the Peanut Peak
fauna. Five species, based on good material, are limited
to these faunas and are unknown from pre-Peanut Peak
members or from the Brule Formation. These are
Apternodus mediaevus; A. altitalonidus; Metacodon mag-
nus; Daphoenocyon dodgei; and Merycoidodon lewisi. A.
altitalonidus and M . magnus are known in South Dakota
from the microfauna locality in the Peanut Peak mem-
ber; Apternodus mediaevus from a single specimen in the
Peanut Peak ; and, therefore, these limited stratigraphic
ranges may be accidents of sampling. They, however,
along with the more common D. dodgei and M. lewisi do
indicate that the Pipestone Springs formation and the
Peanut Peak member are of approximately the same
age.

Four other species, Hyaeonodon horridus, Mesohip-
pus latidens, Hyracodon priscidens, and Caenopus mitis,
are known from good specimens in both Peanut Peak
and Pipestone Springs deposits but are also known from
either middle and lower Chadron or from the Brule.
Hence they indicate a general age similarity but are not
as definitive as the foregoing species.

The Peanut Peakian Mesohippus grandis is un-
known from Pipestone Springs, though large individuals
of the closely related species M . hypostylus are found
there. In itself this might suggest a younger age for the
Peanut Peak Member, but because the Brule horse,
M. bairdi, apparently represents a continuation of the
M. hypostylus line, we conclude that M. hypostylus per-
sisted through Peanut Peakian time though unknown
from South Dakota during this time. Therefore, Meso-
hippus hypostylus cannot be used for precise correlation
within the Chadron ; the less common M. grandis, how-
ever, indicates Peanut Peakian age.

On the other hand, the dog, Daphoenocyon dodgei,
succeeds the ancestral and more primitive Daphoeno-
cyon minor in the savanna chronofauna from the Peanut
Peakian of South Dakota. Thus its appearance in the
Pipestone Springs fauna and the absence of D. minor
suggests very strongly that the Pipestone Springs fauna
is very little, if any, older than Peanut Peakian.

CLARK AND BEERBOWER: THE CHADRON FORMATION

SUGGESTED CORRELATIONS OF CERTAIN OLIGOCENE

FORMATIONS

57

sw

SAS K AT-

CENTRAL

S E

STAGE

INDEX FAUNA

UTAH

MONTANA

CHEWAN

TEXAS

S.DAKOTA

WYOMING

WYOMING

OR E L L AN

o

SCENIC

CANYON

f

MEMBER,

UPPER

z

HI

w
0

o

FERRY

Z

BRULE

WHITE RIVER

BRULE

CHAD-

PEANUT
PEAKIAN

T

NORWOOD

"C
Pll

OOK
RANCH"

'ESTONE
SPRINGS

PEANUT
PEAK

BEAVER
DIVIDE
CGL.

Z
O

ee
a

<

CRAZY

tfi

C RAZ Y

BIG SAND

RON -

JOHN-

Z io

DRAW

z

-1

1 AN

SONIAN

- w

TUFF

1

z

m

JOHNSON

LENTIL

u

o

a

to

AHEA
IAN

•

u.

HI

oc
a.
>-
U

AHEARN

Y ODE R

Teleodus , With

s

Mesohippus v. vie

Ik

X

1

V 1 E JA

DUCHESN E AN

jensiSjAgriochoerus

■

<

SLIM BUTTES

Teleodus , With

>

HI

•

Epihippus, Diplo-

lapointIb

■

lus

HALFWAY. _

R ANDLETT 2
1 in

HI

X

WAGONBED

HI

F

M.

U

o

1

<

Ul

c

u.

z

<

Ul

Ul N T AN

B

t-

«r

u

A

z

o

3

111

Fig. 24. Suggested correlations of certain Oligocene formations.

The oreodont, Merycoidocon lewisi, is common in
both Peanut Peak and Pipestone Springs deposits, and
the only earlier oreodont is known from a single speci-
men from the Crazy Johnson.

The replacement of M. lewisi by M. culbertsoni (re-
placement rather than succession as M . lewisi probably
did not give rise to M. culbertsoni) in this chronofauna at
the beginning of Orellan time suggests that the replace-
ment is a critical time-marker and thus that the Pipe-
stone Springs fauna is no later than Peanut Peakian.

Parictis parvus from the Ahearn is probably ances-
tral to P. personi, from Beaver Divide and Pipestone
Springs, and to P. dakotensis, from the Peanut Peak.
P. personi in turn is more primitive than P. dakotensis
and is close, structurally at least, to the ancestry of P.
dakotensis. This relationship suggests that the Pipe-
stone Springs is older than the Peanut Peakian but
since Parictis is rare as a fossil and since nothing is
known of the group during the middle Chadron, the two
species might well have had a common middle Chadron-

ian ancestor from which P. dakotensis diverged more
rapidly.

Peratherium, Ictops, Menodus, Hopophoneus, Dinic-
tis, and Paleolagus occur both in the Pipestone Springs
Formation and in the Peanut Peak Member. The ranges
of these genera, however, are too long for precise cor-
relation, and the species taxonomy of Ictops and Pera-
therium is too confused or the specimens too inadequate
to allow specific determinations. The few rodents known
from the Chadron Formation appear to be conservative
groups with high dental variability within the species.
Therefore, they are of little value as guide fossils for
restricted time zones. The taxonomy of the small artio-
dactyls is also too confused at present to allow their use
in correlation.

The sum total of the faunal evidence favors a cor-
relation of the Pipestone Springs Formation with the
Peanut Peak member. Certainly the Pipestone Springs
is older than Orellan and almost as certainly much
younger than Ahearnian (see also Wood, A. E., 1955.)

ECOLOGY OF THE KNOWN CHADRON FAUNA

common; rar«: — — — not known, but probably existed in area at thia tima

ZONE

NICHE

GENUS

A H E A R N

CRAZY
JOHNSON

PEANUT
PEAK

U

<

0
<

?

OSTEICHTHYES

CARNIVORES

GR APTEM YS

TR ACHEMYS

I

TR ION YX

ANOSTEIRA

t

CARNIVORE

ALLIGATOR

?

EOPELOBATES

;

<

3

0

4

i

HI
IA

LARGE HERBIVORE

MENODUS

LARGE HERBIVORE

TRIGONIAS

LARGE HERBIVORE

BOTHRIODON

SMALL HERBIVORE

HEPTACODON

f

M

MJ

at
0

■

HI

0
at
O
■

at
>

at

SMALL INSECTI-
VORES, ARBOREAL
TO GROUND

PER ATHERIUM

7

APTERNODUS

CLINOPTERNODUS

METACODON

SMALL HERBIVORES

SOME OF SMALL
ARTIOOACTYLS

SMALL TO
MEDIUM - SIZED
HERBIVORES

MESOHIPPUS

COLODON

AGRIOCHOERUS

4-TOED CAMEL

EOTYLOPUS

MEDIUM -SIZED
CARNIVORE

DAPHOENUS

SMALL AND MEDIUM-SIZED CARNIVORE NICHES PROBABLY FILLED BY
SAVANNA FOREST CARNIVORES

M

HI

U

0

u.

<
z
z
<

>

at

M
<

z
<
>
<

M

MEDIUM, GRAZING
HERBIVORE

MERYCOIDODON

SMALL HERBIVORES

SOME OF SMALL
ART IO DACTYLS

SMALL, BROWSING
HERBIVORES

STIBARUS

LEPTOCHOERUS

MEDIUM, BROWSING
HERBIVORE

PERCHOERUS

LARGE. BROWSING
HERBIVORE

ARCHAEOTHERIUM

MEDIUM. GRAZING
HERBIVORE

POEBROTHERIUM

LARGE. GRAZING
HERBIVORES

HYRACODON

C AENOPUS

SMALL
CARNIVORES

HESPEROCYON

PARICTIS

DAPHOENOCYON M

c

MUSTEL AVUS

MEDIUM SIZED
CARNIVORES

DAPHOENOCYON D.

EUSMILUS

HOPLOPHONEUS

DINICTIS

LARGE CARNIVORE

HYAENODON

Fig. 25. Ecology of the known Chadron fauna.

58

CLARK AND BEERBOWER: THE CHADRON FORMATION

59

The persistence of several primitive species in the Pipe-
stone Springs fauna might suggest that the fauna is
slightly older than Peanut Peakian but might equally
well be ascribed to local survival of these species in a
more favorable environment.

Correlations with Other Chadronian Formations

The scattered and localized Chadronian deposits of
Wyoming, Montana, and Saskatchewan present diffi-
cult problems in correlation because of the rarity of
specimens, the local occurrence of many otherwise un-
known species, or the difficulty in establishing the local
stratigraphic sequence.

The relative ages of the various Chadronian forma-
tions and "local faunas" from southwestern Montana
will be studied by the senior author in the future. Since
these faunas resemble the Pipestone Springs fauna
ecologically, they are best compared with that fauna
rather than with the South Dakota faunules.

The Beaver Divide Conglomerate of central Wyom-
ing may be middle or late Chadronian on the basis of
occurrence there of Parictis personi; probably it is no
older. The Big Sand Draw Lentil may be earlier
Chadronian, but the known fauna does not definitely
suggest such a dating.

The Cypress Hills fauna from Saskatchewan in-
cludes Teleodus, (a characteristic Duchesnean genus),
Mesohippus celer (Chadronian), Hesperocyon lambe (late
Chadronian), and several Orellan species. We suspect
that this is a mixed fauna resulting either from re-
working of the fossils or from continuous deposition
with cross-bedding and channeling which would pre-
clude zoning.

The complex stratigraphy at Sage Creek, Montana,
will be discussed at length when studies on Montana
Chadronian deposits are completed, but the bulk of this
section is Uintan and Chadronian with a thin stratum
which has yielded Teleodus and is probably Duchesnean.

Conclusions

The Chadron includes three distinct rock-time units
with characteristic faunal associations (see Fig. 25) . The
oldest, the Ahearn, is apparently younger than the
Vieja which may be regarded as late Duchesnean. The
youngest, the Peanut Peak, is distinct faunally from
the typical Orellan and underlies the Orellan Brule
formation in the Big Badlands.

The late Eocene-early to middle Oligocene time
sequence in western North America thus comprises the
Uintan, Duchesnean, Chadronian, and Orellan (see
Fig. 24.) The Uinta formation includes three distinct
members, A, B, and C,1 which represent the bulk of
Uintan time. The lower two members of the Duchesne
River Formation, however, are probably Eocene and

1 Wood's (1934) division of the Uinta into Myton member (C)
and Wagonhound member (A and B) is not followed here because
we feel that the A and B members are much more distinct, both
lithologically and faunally, than the B and C.

include the typical Eocene species of Protoreodon and
Diplobunops but are somewhat younger than Uinta C.
The upper Duchesne River, the La Point, includes the
titanothere, Teleodus, which is associated with Mesohip-
pus viejensis and Agricochoerus in the Vieja and with
Mesohippus in the Cypress Hills. The rest of the known
La Point fauna is either transitional between Uintan
and Chadron forms or is represented by specimens so
inadequate that their relationships cannot be precisely
determined.

The La Point Member of the Duchesne River For-
mation may then be properly considered typical of
Duchesnean time (type locality, 12 miles west of Vernal,
Utah — Kay, 1934) with the understanding that no rock
section described up to the present represents all of
Duchesnean time, that the La Point fauna is early
Duchesnean, and that the Vieja Formation of Texas
(Stoval, 1948) is late Duchesnean and should be con-
sidered the type for this part of Duchesnean time.
Teleodus then becomes the index fossil for the Duches-
nean. Teleodus, associated with Poabromylus, Epihippus
uintensis, and Diplobunops, indicates La Pointian or
early Duchesnean; and the Teleodus- Agriochoerus-
Mesohippus viejensis association indicates Viejan or
late Duchesnean age.

PALEOGEOGRAPHY

Topography

The character and distribution of Chadron deposits
in South Dakota demonstrate the paleogeographic and
and topographic relationships of that time. The Black
Hills were the dominant element in the Chadron land-
scape and were the source of all the Chadron clastic de-
posits in the area. Their topography apparently was
very similar to that of the present time, since the pre-
Chadron valleys are aligned with notches in the hog-
backs, and later Oligocene valley-fills extend up these
valleys through the hogbacks and into the Triassic
"Racetrack." The petrologic character of the Chadron
sediments also indicates sources in the present water-
shed areas.

The Chadron formation was deposited on a surface
of gentle to moderate relief cut into deeply-weathered
Pierre shale. This pre-Chadron surface consisted of
three major elements:

1. The valley of the "Red River", which was about 5
miles wide and some 70 ft. deep and trended ap-
proximately ESE.

2. The level upland surface into which the Red River
Valley was cut, which ranged in width from 6-15
miles.

3. Low ridges that formed the divides to the north
and south of the "Red River." These ridges were
about three-quarters of a mile wide; the northern
ridge or Sage ridge rose about 70 ft. above the
upland surface; the southern about 40 ft.

60

FIELDIANA: GEOLOGY MEMOIRS, VOLUME 5

During Ahearnian time deposition was limited to the
"Red River" valley but thereafter spread onto the up-
land, and in Peanut Peakian time extended over the
whole upland area. Figure 4 summarizes the topography
of the pre-Chadron surface and the general drainage
pattern.

Climate

A fairly precise estimate of average temperature
immediately preceeding Chadron time can be made.
Since lateritic weathering of rocks other than limestone
requires an average annual temperate of 60° F or
higher (Krynine, 1949), the slight to moderate lateritiza-
tion of the underlying Pierre shale and the slight lateriti-
zation of some Chadron sediments suggests a tempera-
ture of 60° F or very little higher. A mean annual tem-
perature of 60-65° corresponds with the present situa-
tion in southeastern United States.

Comparison of the pre-Chadron weathering in South
Dakota with that in Montana throws light on the
vertical temperature gradient. Near Whitehall, Mon-
tana, at present elevations of 4000-4500 ft., the Paleo-
zoic limestones show pre-Chadronian lateritization, but
all other rock types were weathered to limonitic or
kaolinitic debris. This indicates (fide Krynine) a mean
annual temperature of about 55-60°, like that of West
Virginia or of Cairo, Illinois. Assuming the average an-
nual temperature in Montana to have been 57-58°, and
the difference in elevation between the two stations to
have been the same as at present, 1300 ft., the vertical
temperature gradient in pre-Chadronian time would be
about 3° per 1000 ft., or the same as the present gradi-
ent. Since the adiabatic gradient depends upon un-
changing laws of physics, the accordance of the esti-
mated temperatures to the gradient supports the ac-
curacy of the estimates.

The presence of small alligators in the Ahearn and
Crazy Johnson Members can be taken to mean that
winter minimum temperatures did not long remain be-
low freezing during much of Chadron time. It does not,
however, indicate that the average annual temperature
was as high during Chadron time as it had been before.
The very incomplete lateritization of Peanut Peak
sediments in the neighborhood of the major Chadron
stream courses, and the lack of oxidation in Chadron
sediments away from those stream courses, indicates
that temperatures were not as high during Chadronian
time as they had been before. Furthermore, the decrease
in amount of weathered upland debris deposited in
Peanut Peak sediments as contrasted with that in the
upper Ahearn suggests that weathering processes in the
uplands progressively declined through Chadronian
time. This would suggest a drop in temperature, or
precipitation, or both. Dorf (1959, pp. 185-189) cites
paleobotanic evidence indicating that the later Oligo-
cene was cooler than late Eocene.

An estimate of annual precipitation is much more
difficult. Lateritization is accomplished under condi-
tions of abundant but highly seasonal precipitation — 40
in. or more, with an alternation of rainy and dry sea-

sons. It may be safe to assume a pre-Chadronian annual
rainfall of over 40 in.

The Chadron sediments themselves include algal
limestones, casts of Unio and pond snail shells, bentoni-
tized ash, and predominately reduced disseminated iron
(greenish to bluish color), all of which indicate abundant
water. They also contain scattered zones of gypsum
crystals, manganese dioxide, and, especially in the
Peanut Peak member, light tan sediments and cal-
cereous nodules or zones, which suggest aridity. The
fauna is a mixture of wet-forest forms such as alligators,
small artiodactyls, insectivores, and Mesohippus with
such dry-plains animals as the camel, Poebrotherium.

This apparently conflicting evidence resolves itself
into a co-ordinated picture when it is noted that the
indications of moisture occur generally near the bottom
of the section or concentrated in the vicinity of the chan-
nel fills. The evidences of aridity, on the other hand, are
to be found away from the channel fills and near the top
of the section. Abundant run-off from the Black Hills,
with the streams passing through a relatively dry plains
area, could produce this pattern of evidence, especially
if desiccation became progressively more severe.

INTERPRETATION OF CHADRONIAN
SEDIMENTATION

General Review of Tertiary Sedimentation

A. Description.

The Oligocene epoch was a time of transition — this
concept has been well documented by faunal studies,
but the evidence to be derived from study of major
lithologic changes has not been systematically pre-
sented. Figures 26 and 27 review the senior author's
observations on Tertiary continental deposits and in-
clude distribution, thickness, lithology, and evidence of
depositional environments of most of these sediments
from the High Plains west to the Nevada-Utah bound-
ary and from the San Juan basin north to the Canadian
boundary. Much of the data on which this chart is based
can be verified in the literature. However, we have
omitted data which could be gleaned from the literature
but which have not been personally observed because it
seems unfair to quote in this context descriptions made
without anticipation of this interpretation.

The chart shows that Upper Paleocene and Wasa-
tchian sediments are primarily fluvial, of wide distribu-
tion and moderate thickness. Sediment colors are gen-
erally dark reds, greens, and purples, and the fossil
bones are heavily impregnated with iron and manganese
oxides. Middle and upper Eocene deposits on the other
hand occur in only a few of the intermontane basins, are
usually thicker than the underlying Tertiaries, and
contain a large proportion of lacustrine sediments.
Middle Eocene sediments are generally pale greenish to
grey or tan, with bones colored pale tan by limonitic im-
pregnation or black with manganese dioxide. The color
of upper Eocene rocks and impregnation of the fossils is
generally similar to that of the Wasatchian. Deep pre-

CLARK AND BEERBOWER: THE CHADRON FORMATION

61

Chadronian weathering in South Dakota, Nebraska,
eastern Wyoming, and Montana suggests that these
areas were exposed surfaces during late Eocene time and
that there were no late Eocene sediments deposited in
this region.

In contrast, Chadronian sediments occur in north-
eastern Colorado, western Nebraska and the Dakotas,
many of the basins of Wyoming, and most of the inter-
montane valleys of Montana. They rest discomformably
on eroded, deeply weathered older rocks — only at two
places, Beaver Divide and Sage Creek, were they de-
posited conformably on late Eocene beds. Bentonitized
ash is always present in the Chadronian sediments and
increases in amount and freshness toward the top of the
section. The Chadronian sediments are thin, commonly
less than 200 ft. thick, and occur only on the eastern side
of the Continental Divide. The sediments are primarily
fluvial and have relatively pale colors. Fossil bone is
slightly impregnated, but sometimes heavily coated,
with limonite.

Post-Chadron deposition in general parallels that of
the Chadron in distribution and characteristics. The
color of Orellan rocks is somewhat paler and more tan
than Chadronian; the ash content is relatively higher;
and fossil bone occasionally has a coating of hematite
but otherwise shows little more iron-manganese im-
pregnation than modern bone found on the surface of
the High Plains. In Figure 26 these changes from the
Paleocene through the later Tertiary are summarized
and the transitional nature of Chadronian deposits is
demonstrated. The development of reddish tan sedi-
ments, immediately followed by non-deposition, at the
close of Chadron time, should be noted.

B. Interpretation

The thickness and distribution of the deposits must
reflect both tectonic and climatic controls. The propor-
tion of lacustrine to fluvial beds must be the result
primarily of tectonic activities by which the basins were
blockaded to form lakes. On the other hand, sediment
color if syngenetic, and bone impregnation, should be
functions of climate rather than of tectonics.

Sediments may, of course, inherit their color from
the parent rock or weathering of that rock, or may de-
velop it epigenetically. The senior author has de-
termined certain field criteria for the recognition of
derived, epigenetic and syngenetic colors (Clark, 1962),
and we limit our discussion to the latter, since they
alone are significant in climatologic interpretations.
Highly colored fluvial sediments are generally deposited
under conditions of high temperature and humidity
with abundant vegetation; pale fluvial sediments indi-
cate aridity but not necessarily cool temperatures.

The nature of impregnation of fossil bone is also
partially controlled by local climate. The senior author
has found that in most semi-tropical to tropical forest
environments fossil bone is heavily impregnated with
hematite, limonite, (in general, hydrous iron oxides) and
manganese oxides. In moist and somewhat cooler en-
vironments there is an impregnation of brown limonite.

Several horse skeletons buried for 35 years in forest
mould near Princeton, New Jersey, showed such im-
pregnation to depths of one-eighth inch. Burial under
somewhat drier conditions seems to produce incrusta-
tion without impregnation; for example, a woodchuck
skull recovered from swamp mould in northeastern Illi-
nois was heavily encrusted. In contrast, burial on the
semi-arid high plains in Dakota and Colorado produces
no iron impregnation. The precise factors that control
impregnation are not known but are probably related to
the acidity of the local ground water, and will vary with
the nature of entombing sediments, porosity of the bone,
and speed of burial. However, the generalization as to
climatic conditions fits all the available evidence.

On the basis of the evidence cited above, a general
interpretation of early and middle Tertiary sedimentary
environments can be made. The thickness and wide
distribution (on the plains and in all of the basins on
either side of the divide) of late Paleocene and early
Eocene rocks suggests widespread uplift of the positive
units with respect to the intermontane basins and the
high plains. The restricted distribution of middle and
late Eocene beds and the development of deep weather-
ing profiles elsewhere indicates cessation of general
tectonic activity and a long period during which the
streams were at grade or cutting shallow valleys in the
basins and on the High Plains. The high proportion of
lacustrine sediments presumably results from local
tectonic adjustments which blocked egress from the
intermontane basins. The geographic association of
most of these basins with the Uinta Mountains suggests
that these local movements were related to that tec-
tonically active unit. Color of sediments and mode of
bone impregnation suggests a middle Eocene cool epi-
sode followed by a late Eocene warm, moist climate.
Paleontological evidence supports but does not prove
this conclusion.

Resumption of widespread deposition in Chadron
and post-Chadron time indicates the operation of a
factor or factors of regional extent which resulted in
overloading of streams east of the continental divide
and in intermittent deposition over this area throughout
the later Tertiary. The paler sediments, the change in
type of fossil impregnation, and faunal evidence demon-
strate rather conclusively a change toward a drier and
somewhat cooler environment. It may well be that the
reddish color of latest Chadron sediments, followed by
non-deposition with abundant swamps, represents a
temporary warming; this is possible but no satisfactory
evidence is known.

Factors Controlling Chadronian Sedimentation

As indicated above, local or regional tectonic factors
apparently controlled pre-Chadronian sedimentation.
On the other hand, Oligocene and Miocene beds are not
so clearly related to tectonic movements and further
discussion of their origin is necessary.

Any one of five possible hypotheses or some com-
bination of these hypotheses may be the true explana-

TERTIARY

ABSOLUTE TIME. IN M/YR
(HOLMES SYMPOSIUM 1964)

65

6.0

GEOLOGIC AGE K

5,5

PALEOCENE

53-54
58.5

LlU PALEOCENE

5.0

49

EOCENE LlM

4,5

4.0

3

4(5

M|U EOCEM

l ACUST • IN
D i POS ITS

FIISH AND ALTEIfD
VOLCANIC ASH

'• IMPREGNATION
OF BONE

ROCK
DATA

INTENSITY OF COLOI

THICKNESS OF
DEPOSITS

AREA OF DEPOSITION

HIGH

TEMPERATURES BASED UPON
PALEOBOTANY (AFTER DO*F,
1»S5)

IN EACH CASE, THE TOP OF

THE GRAPH REPRESENTS MANY,
THE BOTTOM, FEW.

THESE GRAPHS SHOW QUAL-
ITATIVE TRENDS ONLY. THE
DATA CANNOT AT PRESENT
• E QUANTIFIED.

LOW

Fig. 26. Tertiary Paleogeographic Data-Sedimentary.

62

EOGEOGRAPHIC DATA - SEDIMENTARY

63

64

FIELDIANA: GEOLOGY MEMOIRS, VOLUME 5

tion of the thin, widespread deposits across the central
High Plains. These are:

1. eustatic rise of sea level;

2. uplift of a structural barrier across the lower
reaches of the depositing streams;

3. down-warp of the entire region;

4. overloading of streams with volcanic ash; and

5. climatic change sufficient to cause a shift in
stream regimen.

Eustatic rise of sea level can be eliminated as a pos-
sible cause for several reasons. First, the coastal Plain
stratigraphy gives no evidence in favor of it, and sug-
gests rather that Oligocene sea level was lower relative
to the land than was Eocene sea level. Second, the
Atlantic drainages south of central Colorado did not
begin deposition simultaneously with those to the north,
as they should have done had there been a change in sea
level. Third, the Chadron of South Dakota thins both
mountainward and seaward from its maximum thick-
ness about 20 miles east of the rim of the Black Hills. A
rise in sea level might be expected to result in a deposit
thinning progressively upstream from the river mouth.

Uplifting of a structural barrier across the lower
reaches of the depositing streams would also produce a
lens of sediment thickest near the downstream shore of
the resulting lake and thinning upstream. Further, the
Oligocene streams of southwestern Montana probably
did not enter the same master stream as did those of
central Colorado. Therefore, a tremendously long struc-
ture, no trace of which is known, must be called into
being. This second possibility can also be discarded.

A gradual downwarp of the entire area from Idaho to
the Break of the Plains and from the Canadian border
to central Colorado, has been postulated in oral discus-
sions as the cause of mid-Tertiary (Oligocene-Miocene)
deposition. Three main points of evidence have been
regarded as supporting this hypothesis:

1. Presence of Oligocene and Miocene outliers at
elevations up to 7500 (Montana) and 9000 ft. (Big Horn
Mts.). If these levels are projected out into the inter-
montane basins, the late Miocene bottoms of the basins
as reconstructed would lie 3500-4000 ft. higher than
the present ones. Such great masses of fill, in every basin
and trough, could only be explained by filling of a down-
warped area.

2. The small amount of erosion of mountain slopes
since early Oligocene time. Basal Chadronian sediments
everywhere contain pebbles derived from rocks exposed
in the present cores of the adjacent ranges. Chadronian
sediments choke the mouths of present canyons de-
bouching from the various mountain ranges, and thus
indicate that early Chadronian mountain topography
must have been like the present topography. This pres-
ervation of an ancient surface can be explained by as-
suming burial under thick mid-Tertiary sediments,
which were then removed during Quaternary time.

3. Widespread superposition of streams on low
ranges or near the ends of long ranges can be explained

by assuming a cover of Tertiary sediments on which the
streams meandered before their recent incision.

These are strong evidence, but other reasoning, equally
cogent, opposes them:

1. There is no direct structural evidence of any such
downwarping.

2. There are no known angular unconformities
within the mid-Tertiary section (except in the Slim
Buttes, where the movement is either non-tectonic or
extremely local). If such extensive downwarping oc-
curred, differential movements causing angular uncon-
formities would be expected.

3. Basal Chadron sediments are everywhere (except
at Beaver Divide) composed of reworked, deeply weath-
ered material, but later elastics usually consist of fairly
fresh rock, showing that once the pre-Chadron soil had
been stripped, weathering did not keep pace with ero-
sion. An area undergoing downwarp would certainly not
be expected to show accelerated erosion.

4. Depositional dips of 3-5° are usual in the montane
mid-Tertiary sediments, and demonstrably initial dips
of up to 25° have been noted. Deposition of uniform,
thin strata on a porous surface at even greater dips can
be demonstrated experimentally.

If average depositional dips of 2° are presumed,
which is conservative judging from the observed dips,
then the mid-Tertiary fill of the Big Horn, Powder
River, and other major basins need not have been more
than 1500 ft. thick. This thickness accords roughly with
the thickest preserved sections, and, further, the deposi-
tional structure would accord with that of the observ-
able sediments. It is our belief that, with the exception
of a few locally downfaulted areas in Montana, none of
the basins of this area ever contained more than 1500 ft.
of mid-Tertiary sediment.

5. The cases of stream superposition do not demon-
strate a fill deeper than 1500 ft. Several streams are
superposed over very low ranges, with much higher
mountains nearby. Others traverse the lower reaches of
higher ranges. In no case is the top of the ridge at the
watergap more than 1500 ft. higher than the recon-
structed pre-Oligocene surface.

6. Since pre-Chadron topography was probably
very much like that of the Recent, the major stream
systems would have filled rapidly to grade with clastic
materials from the mountains during downwarp. Yet
the post-Chadronian sediments contain smaller clastic
constituents with the major proportion consisting of
ash. Therefore, post-Chadronian deposition may have
been partially controlled by the supply of volcanic ash
and might be expected to show a different distribution
than the Chadronian deposits.

7. Downwarp might be expected to cause Chad-
ronian deposition on both sides of the Continental Di-
vide rather than on one side.

8. Progressive downwarp should cause the youngest
sediments to be thickest near the center of downwarp
but these later Oligocene and Miocene sediments are

CLARK AND BEERBOWER: THE CHADRON FORMATION

65

thickest near the base of the local mountain range that
formed the watershed of the streams depositing each
series. The only regional trend in thickness appears to
be related to an increase in volume of volcanic ash
toward northwestern Wyoming.

These eight lines of evidence weigh against the
probability of a regional downwarp as the basic cause of
deposition.

The fourth hypothesis, the overloading of the
streams by volcanic ash, has two points in its favor.
First, the ash occurs in large volume throughout the
upper part of the section and increases in amount, pro-
portion, and average grain size toward the volcanic
center. Second, Chadron deposits are absent west of the
volcanic center, and south beyond the zone where pre-
vailing westerlies might be expected to carry ash.

Several strong lines of evidence militate against ash
as a primary cause of mid-Tertiary deposition. First,
the basal 20 ft. of Chadron sediment nowhere include
volcanics. In South Dakota, ash first becomes an im-
portant part of the sedimentary mass in the Crazy
Johnson member, and does not become dominant below
the Peanut Peak Member. Overloading with ash cannot
reasonably be regarded as the primary cause of an epi-
sode of sedimentation which began without ash and did
not receive significant quantities of ash until the old
topography had been buried. Therefore, increase in
volume of ash in the streams could not have been the
initial or the primary cause of deposition, but it prob-
ably influenced the rate of deposition.

The remaining hypothesis, that of climatic control
of stream regimen, is thus supported by the default of
the other four suggested; furthermore, the positive evi-
dence for this interpretation is substantial.

The occurrence of a major climatic change at the
beginning of Oligocene time has been demonstrated
adequately in this and other papers. The change from a
moist, warm, and possibly monsoonal climate to a semi-
arid, cool climate would profoundly alter vegetational
cover, weathering, and stream discharge and conse-
quently would modify stream regimen.

In late Eocene time weathering must have been
primarily chemical and resulted in a predominance of
clays and solutes. The heavy vegetational cover would
restrict surface run-off and thus reduce the amount of
sediment relative to stream discharge. The master
streams — adjusted to this comparatively small sediment
load — would have low gradients over all outcrops sus-
ceptible to chemical weathering.

On the other hand, the rocks resistant to chemical
weathering in the mountain cores would rise abruptly as
sharp ridges, and consequently stream profiles would
change rather abruptly near the divides. (Cotton, 1941
p. 155-156; Davis, 1923 p. 21; Lawson, A. C, 1932 p.
706). The landscape would then comprise two sets of
features; broad river valleys with gentle gradients de-
veloped on the shales and weakly cemented sandstones,
and bold mountain ranges on the more resistant rocks.

In Chadron and post-Chadron time the cooler, drier
climate must have produced a relative increase in
mechanical weathering with a resultant increase in
supply of coarser elastics. The vegetational cover must
also have been reduced, and surface run-off consequent-
ly increased in relation to total run-off. In turn, in-
creased surface run-off would increase slope wash and
gullying on steeper slopes. The effective load of main
streams would therefore be relatively large at the same
time total water discharge was decreasing considerably.
The net result would be erosion of upland slopes, remov-
ing weathered mantle first and then relatively fresh
rock fragments, and deposition in the major stream val-
leys extending out into the adjacent plains.

The effects of the climatic change would be regional
and thus the pattern of erosion and deposition would
be similar over a wide region. On the other hand, the
boundaries of the area of deposition might be rather
sharp and controlled by the major topographic elements
and by boundaries of wind systems (see p. 66).

The protracted depositional episode (Oligocene and
Miocene) cannot, however, be ascribed simply to a
single brief period of climatic change. If such a change
were the only control, rapid filling following the change
would be succeeded by a long period of equilibrium and
concluded by an even longer period of slow erosion as the
supply of elastics from the uplands decreased. There-
fore, if the climatic hypothesis is generally correct, one
or more modifying factors must also have operated. The
most probable factors are:

1. Small supply of material relative to the area of
deposition. The stream profiles then would be adjusted
very slowly. The evidence of numerous hiatuses and
cut-bank erosion within the Chadron suggests, however,
that the streams were never very greatly out of equi-
librium and that the supply of material for deposition
was an incidental factor.

2. Regional uplifts. Downstream parts of the chan-
nels would be above grade and would actively downcut.
Deposition would then cease or slow as the knick points
shifted upstream. The overall consequences of such
changes are somewhat difficult to visualize but it seems
probable that the streams would come to equilibrium
more rapidly rather than more slowly.

3. Overloading by volcanic ash in post-Chadron
times. The increase in ash-falls in late Chadron and post-
Chadron time was undoubtedly a factor in Oligocene
and Miocene deposition. It could be the controlling
factor only if the amount of ash increased progressively,
since o) the streams were never far from equilibrium,
and b) the streams would tend to come to equilibrium
with the amount of ash in their load and any great
reduction in amount of ash would result in trenching.
The percentage of ash does increase in the later Chadron
and through the Oligocene and Miocene, but it seems
unlikely that this increase is solely responsible for con-
tinued deposition.

4. Progressive climatic change. Further decrease in
rainfall would decrease volume of the streams, both

66

FIELDIANA: GEOLOGY MEMOIRS, VOLUME 5

absolutely and relative to available load. With an
abundant supply of ash largely independent of regional
climate and with progressive decrease in stream dis-
charge, continued deposition would be expected. Such a
progressive overall deterioration of climate is indicated
by: a) paleobotanical evidence; b) increasing dominance
of a savanna fauna during the later Tertiary; and c)
general characteristics of the sediments.

We conclude, therefore, that climatic change ini-
tiated deposition in the Oligocene, that it was the pri-
mary factor in continuance of this deposition, and that
the supply of volcanic ash was a major but accessory
factor in later Oligocene and Miocene deposition. The
fluctuations in deposition then may have been the re-
sult of either climatic fluctuations imposed on the gener-
al trend or of fluctuations in supply of ash.

Climatic Patterns of the Eocene and Oligocene

The distribution of fossil plants, (Dorf, 1955), of
invertebrates, and of vertebrates, as well as petrologic
evidence, indicates that late Eocene climates were warm
and equable. The temperature differential between the
Equator and the North Pole must have been low. Under
such conditions hemispheric wind current systems
would be weak, of small horizontal extent, and much
influenced by local or regional temperature differentials.1

Assuming the present prevailing westerlies, the
Eocene jungle-forest environments in the intermontane
basins are extremely difficult to account for as the
basins at present are arid or semiarid. All available
evidence indicates that during the Eocene the moun-
tains maintained, in general, approximately their pres-
ent elevations above these basins (although the entire
area was nearer sea level) so that an explanation based
on changes in local or regional topography is inadequate.
If, however, the westerlies were much weaker, local tem-
perature differences between the Western Interior and
the two major adjacent seaways, the Gulf of Mexico and
Hudson's Bay, might be expected to produce a mon-
soonal climate.

The area of the southwestern and central United
States must have been quite warm during the summer
and thus would function as a thermal low. (Trewartha
[1954, p. 99] describes the occurrence of such a summer
thermal low at present). The Gulf of Mexico would be
somewhat cooler, and Hudson's Bay or the Arctic would
be considerably cooler. They would become thermal

1 Communication from Dr. David Fultz, Department of
Meteorology, University of Chicago: "A number of qualitative
considerations, both theoretical and empirical, such as the observed
seasonal differences in circulation between summer and winter,
suggest that the smaller the general horizontal temperature differ-
ential in a rotating convective fluid system like the atmosphere, the
weaker will be the currents and the smaller the horizontal dimen-
sions of the predominant current systems. The more this is the
case, the more such systems will be influenced by purely local
temperature gradients. Laboratory experiments over a wide range
indicate by comparison with the present 20-30° C between tropical
and polar regions that if the differential were, say 2°C, the current
systems would be of the order of size of a few degrees of latitude."

monsoonal highs. A summer monsoon, with prevailing
northerlies, would bring moist air in from Hudson's Bay
to cause heavy monsoon rains throughout the Northern
and Middle Rockies. (There was probably also a south-
east monsoon across Texas, but that does not enter into
the present problem.) Heavy rains probably fell on all
places above 1000 ft. elevation, as they do in the Punjab
today.

The winters were probably dry and cool, with almost
no wind. It is possible that the slightly stronger winter
westerlies may have modified the winter monsoon, but
no evidence of this is known at present.

Bradley (1948) noted the evidence of alternately wet
and dry seasons in the middle Eocene Green River
Shales, and suggested that the climate was monsoonal.
He lacked, however, the supporting evidence given by
modern knowledge of conventional dynamics, and sup-
posed the summers to be dry and warm, the winters cool
and wet. The present hypothesis makes necessary warm,
wet summers and cool, dry winters.

Additional evidence of prevailing northerlies in
Wyoming is offered by the distribution of volcanic ash.
Ash constitutes a high proportion of the mass of middle
Eocene sediment in the Green River and Washakie
basins, and a much smaller proportion of middle and
late Eocene sediment in the Uinta Basin to the south.
Houston (1964, p. 18) considers that the volcanics of the
Green River and Washakie Basins "may have come
from the Absoroka source but the petrography is not
definitive. . . . Petrographically these units equate to
the acid breccia of the Yellowstone-Absaroka source,
but in fine-grained rocks this far from source one might
expect some loss of heavier more mafic minerals espe-
cially if the major transport was aerial." We consider
that the Yellowstone-Absaroka district is the most
probable source, because: (1) it is less than half as
distant as the next nearest possible sources in Oregon
and Nevada; (2) the proportion of ash decreases notably
from the Green River Basin southward to the Uinta
Basin, as it should if the source lay to the north, and
should not if the source lay to the west; and (3) as
Houston has pointed out, the petrography of the vol-
canic sediments in the Green River and Washakie
Basins is compatible with Yellowstone-Absaroka vol-
canics. If this is the case, the tremendous volume of ash
in the two basins, 300 miles from the source, indicates
that the prevailing winds during late Eocene time were
northerly.

The authors feel that the evidence justifies the hy-
pothesis that during late Eocene time the climate was
warm and equable, with a low temperature differential
between the Equator and the North Pole. Consequent
weakening of the prevailing westerlies permitted a mon-
soonal circulation to develop, with prevailing northerlies
bringing moist air from Hudson's Bay and the Arctic
toward the thermal low in southwestern United States.

The rainfall pattern produced by a monsoonal cir-
culation with prevailing northerlies would be strikingly
different from the present rainfall map. By analogy with

CLARK AND BEERBOWER: THE CHADRON FORMATION

67

the present situation in the Punjab, one may presume
that little precipitation would occur below elevations of
1000 ft. This means that most of the Central Lowlands
would be dry plains or even desert, analogous to Delhi
or the Sind, while the northern High Plains and the
Wyoming basins enjoyed heavy summer rainfall. There
is, of course, no evidence bearing on Eocene climates in
the Central Lowlands, but the deeply-weathered, lat-
eritic pre-Chadron surface in South Dakota certainly
suggests warm, highly seasonal rainfall.

The general lowering of temperatures in the Oligo-
cene and early Miocene (Dorf, 1959, and preceding sec-
tions of this paper) would represent an increase in the
Equatorial-Polar temperature differentials and thus
would increase the strength and extent of the hemi-
spheric wind system. The monsoonal system would be
greatly modified or destroyed, and replaced by pre-
vailing westerlies and a cyclonic storm system similar to
the recent pattern. Rainfall distribution would then
come to approximate the Recent with a marked decrease
in total rainfall on the eastern slopes of the central and
northern Rockies. This area is, of course, precisely that
in which stream regimen was profoundly changed at the
beginning of Chadron time.

The relationship of this early Oligocene climatic
change to the general pattern of Cenozoic climates has
significant bearing on the causes of that change. A brief
review of Cenozoic climatic fluctuations is needed to
understand this.

Dorf (1955, p. 587) has published a series of geologic
thermometers based upon studies of paleobotany. Fig-
ure 26 shows his temperature data graphed on the most
recent absolute time scale (Phanerozoic Time Scale:
Symposium; 1964, pp. 179-191). Parallels between tem-
perature changes and the lithologic and paleontologic
changes shown in the graph suggest a causal relation-
ship. Three major warm (and presumably equable) epi-
sodes have occurred: 1. late Paleocene-early Eocene;
2. late Eocene; and 3. middle Miocene. Four interven-
ing cool periods are represented by: 1. middle Paleo-
cene; 2. middle Eocene; 3. Oligocene-early Miocene;
and 4. late-Miocene-Recent.

A periodicity of about 10 million years of warmth,
with shorter, intervening cool episodes, seems to have
obtained until the beginning of Oligocene time. A gener-
al cooling trend began at that time and continued until
its nadir in the Pleistocene glaciations, interrupted only
by the partial warming of mid-Miocene time. This sug-
gests a rough periodicity of about 20 million years,
during the cooling trend. It is possible either that the
evidence for warm episodes in about mid-Oligocene and
late Pliocene times has been obscured by the general
cooling, or alternatively that the 10-million-year perio-
dicity was suppressed by the longer trend cooling. The
senior author intends to investigate this problem fur-
ther.

The Pleistocene glacial and interglacial episodes
represent temperature fluctuations on a lesser order of
magnitude. Estimates of the absolute length of the

Pleistocene vary from 1,000,000 to 2, 700,000 years and
estimates of the absolute length of the various stages
depend upon the basis used. Whatever absolute time
scale is used, these stages are fairly rhythmic fluctua-
tions with a periodicity of a few thousands (cold) to
many tens of thousands (warm) of years.

Brooks (1948) and Ahlman (1953) have summarized
the evidence for a series of climatic fluctuations of the
order of magnitude of 150-500 years. These plainly
represent a third mode of temperature fluctuation not
related to known Chadronian phenomena.

Finally, the 13 to 22-year cycle presumably related
to sunspots is apparent in most weather records but is
not evidenced in Chadron sediments.

It appears, therefore, that temperatures have fluc-
tuated on four modes :

1. 10,000,000 year, and possibly a 20,000,000 year
during long-term cooling;

2. 10,000-100,000 year;

3. 150-500 year;

4. 22 year, approximately.

The Chadronian cooling plainly represents the ini-
tiation of the cool phase of a 10,000,000 year fluctua-
tion. It is not the equivalent of a Pleistocene glacial
epoch, because the time involved is too long. The Oligo-
cene is now generally regarded as having a duration
of about 7-10 million years. Chadronian time was a
major portion of this, and potassium-argon dating indi-
cates that the Chadron probably represents 3-5 million
years. That is, Chadronian time was longer than the
entire Pleistocene.

Dorf's graph (Fig. 26 of this paper) makes apparent
an even greater significance of the Chadronian cooling.
At this time, the entire series of long-term fluctuations
began to grow cooler, as mentioned above. Whether or
not this very long-duration cooling is part of a fifth
series of temperature fluctuations on a mode of more
than 100 million years is not known. It is, however, of
importance that coincidences of minima in the 10 mil-
lion year mode and the 10,000-100,000 year mode, plus
this very long-duration temperature depression, were
required to produce glaciation in middle latitudes.

The Chadronian cooling should, therefore, be studied
carefully, with the understanding that it may represent
a simultaneous, coincidental cooling on two different
systems. Once the modes of fluctuation are recognized,
geologic and geophysical research can be devoted to
determining their history and causes.

PALEOECOLOGY

Introduction

The rarity of fossils, the absence of paleobotanical
materials, and the usual difficulties of determining
habits and habitat from occurrence and morphology

68

FIELDIANA: GEOLOGY MEMOIRS, VOLUME 5

restrict interpretation of Chadron paleoecology. Inter-
pretation is further complicated by the apparent geo-
graphic proximity of quite different major communities
among which many of the mammalian species probably
ranged at will. Figure 25 summarizes our knowledge of
Chadron paleoecology and includes our interpretation
of habits and environment preference of some forms. As
is obvious from the table, only a few species can be as-
signed with certainty in this fashion. Other interpreta-
tions are less certain but probable, and many are ex-
tremely dubious. In constructing this table and the lists
that follow, we have been guided by the following
priorities in evidence:

1. Adaptations for particular modes of life evident
in skeleton and teeth;

2. Lithologic associations and our interpretation of
the environment of sedimentation :

3. Association with other species whose ecology is
relatively clear from 1 above.

4. Taxonomic affinities to contemporaneous popula-
tions whose ecology is more certain.

5. Ecology of ancestral and/or descendant popula-
tions.

Points 2 and 3 involve major risks because of trans-
portation of the animals after death and because of the
likelihood that they have died in a place other than
their normal range, e.g., plains animals along a stream or
near a waterhole. In general, if a species is numerous in
channel fills and/or channel margin deposits and is
absent elsewhere, we interpret it as living in an aquatic
or riparian environment; if a species is found commonly
in the flood plain clays, we interpret it as living in a
flood plain-upland environment. Since the temporary
movement of animals from one local habitat into anoth-
er is an important ecological factor, it has been con-
sidered separately in making our evaluations.

Points 4 and 5 possess an even greater uncertainty,
particularly when the affiliated populations are relative-
ly distinct species of different genera. We have attri-
buted little weight to these taxonomic or phyletic affini-
ties except where they tend to substantiate interpreta-
tions based on other evidence.

Environments of Deposition

Four distinct environments of deposition and of
fossil occurrence can be distinguished in the Chadron
Formation :

1. Streams, represented by channel fills of the "Red
River" and its tributaries;

2. Ponds, usually represented by limestones near
channels and channel margin facies;

3. River banks, composed of channel-margin sands
and silts, including a chemically reduced zone of sedi-
ments from 100 yards to one-half mile in width, margin-
al to the channel fills;

4. Flood plains, represented by massive to bedded
clays and silts.

Flood plain deposition did not begin until late
Ahearnian time, but the remaining facies are recognized
in all three members. The width of the channel fills and
channel-margin zone is reduced in Peanut Peak sedi-
ments. Fossils are rare in the flood plain deposits below
the Peanut Peak Member, where they are relatively
common.

General Biotic Structure

Four more or less distinct ecological habitats, the
aquatic, the semi-aquatic, the river-border forest, and
the savanna forest-savanna, can be distinguished with-
in the Chadron vertebrate fauna. Some species were
probably restricted to one habitat; others, such as some
of the carnivores, ranged through several; and still
others, although largely limited to one habitat, must
have spent some time in other habitats.

The aquatic habitat is defined as comprising those
vertebrates limited to streams, ponds, and their banks.
The following genera are characteristic of this habitat:
Indet. osteichthyes, Graptemys, Trachemys, Amyda,
Anosteirids, and Alligator. These genera are found al-
most exclusively in the channel fills, show aquatic
adaptations, and their closest living relatives are
aquatic.

The fauna of the semi-aquatic habitat includes those
species which probably spent a large portion of their
time in the water, but which also foraged in the swamp
and river-border forest areas adjacent to the streams
and ponds. Four genera are assigned to this habitat,
Menodus, Trigonias, Heptacodon, and Bothriodon, and
the assignments are based on absence from flood-plain
deposits, on abundance in channel fills, pond deposits,
and channel-margin beds, and on their short, heavy
limbs which suggest semi-aquatic or swamp habitats.

The river-border forest habitat is more difficult to
define sharply. Stratigraphically it consists of the re-
duced zones adjacent to the channels, and probably
represents a wet-forest habitat. The occurrence of some
forest-adapted animals, Agriochoerus, Peratherium, Colo-
don, Eotylopus, and the four-toed camel, suggest this
conclusion, as do the nodular limy algal deposits, the
chemical reduction of the sediments, and the frequent
occurrence of the semiaquatic genera in these channel-
margin beds. In addition to the genera named above,
Mesohippus, Hesperocyon, and possibly Daphoenocyon
are characteristic of this habitat. The association of
these genera with the river-border forest is based largely
on occurrence within the channel margin deposits in
association with forest animals, and on their rarity or
absence from the flood-plain deposits. The general
adaptations of these later genera are also such as would
fit them to a forest life (though not exclusively). Very
possibly they also ranged in some numbers into the
savanna, and almost certainly the larger carnivores of
the savanna hunted also in the river-border forest.

The fauna of the savanna cannot be sharply de-
limited from that of the river-border, although some
elements can be clearly assigned to one or the other. In

CLARK AND BEERBOWER: THE CHADRON FORMATION

69

general, the habitat is defined by occurrence in flood-
plain clays, by subcursorial (Hyracodon, Caenopus,
Perchoerus) or cursorial (Poebroiherium, Archaeotherium)
habitus, by tooth adaptations to mixed browsing and
grazing (Hyracodon, Caenopus, Merycoidodon, Poebro-
iherium) and by the abundance of many of these forms
in the fauna of the Brule. The following genera are
considered to belong to the savanna habitat: Hyaeno-
don, Parictis, Daphoenocyon, Mustelavus, Eusmilus,
Hoplophoneus, Dinictis, Hyracodon, Caenopus, Archaeo-
therium, Leptochoerus, Stibarus, Merycoidodon, and Poe-
broiherium. The influence of the carnivores from this
habitat on the river-border forest habitat must have
been very great.

Figure 25 summarizes the various faunal habitats
and indicates the probable overlap of genera into differ-
ent habitats. Because of this overlap nearly all larger
mammals were parts of a single natural community at
any given time and may therefore be considered as a
unit which we shall term "forest-savanna." The micro-
faunal elements and some of the small selenodont artio-
dactyls might be more definitive of the habitats, but the
former (insectivores and rodents) are so rare as to be
useless and the taxonomy of the latter is so confused as
to obscure their ecologic relationships.

Evolution of the Biotic Structure

It is probable that, in common with modern biotas,
the successive forest-savanna faunas had a degree of
internal integration, that the elements of a fauna inter-
acted, and that as a consequence of integration and
interaction the fauna had a structure of occupied niches.
Figure 25 summarizes our concept of that structure and
of the position of the various genera within it for these
three successive times — Ahearnian, Crazy Johnsonian,
and Peanut Peakian. The figure also includes an esti-
mate of the relative abundance of these genera during
these periods.

Figure 29 shows a marked change in the abundance
and variety of animals from the various ecologic habi-
tats during Chadron time. The differences are shown in
the number of genera present in each habitat, and in an
"index" based on generic occurrence and abundance. In
calculating the index, each genus known from large
numbers of specimens was weighted by a factor of three;
and those genera which are unknown from the member
but which have known or probable stratigraphic ranges
extending through this member are weighted by a factor
of one. In general, the first or last occurrence of genus is
not taken as its maximum stratigraphic range but in
calculating the index that genus is given a weight of one
in the preceding or succeeding member. The index thus
tends to minimize large differences in fossil abundance,
part or most of which may be due to differential preser-
vation, and also to minimize the effects of non-occur-
rence which might well be due to sampling error.

The Ahearn is marked by a relative variety and
abundance of aquatic (6 genera comprising 27% of the
fauna and an index of 13), semi-aquatic (3 genera, 21%,
index of 11), and river-border forest forms (5 genera,

23%, index of 12). The savanna fauna is varied but in-
cludes a smaller percentage of the total fauna (9 genera,
40%, index of 26). The Crazy Johnson member includes
four aquatic genera (21%, index of 11), four semi-aquatic
genera (21%, index of 10), three river-border forest
genera (16%, index of 10), and nine savanna genera
(47%, index of 28). The fauna of the Peanut Peak
member consists of one aquatic (7%, index of 5), one
semi-aquatic (7%, index of 5), two river-border forest
(13%, index of 6), and twelve savanna genera (73%,
index of 31).

Inasmuch as the fauna of the Ahearn member, which
consists primarily of channel-fill and river-border facies,
does not differ greatly from the fauna of the Crazy
Johnson member, which includes flood-plain deposits,
and because the Peanut Peak member includes all three
sedimentary facies, the differences in fauna between the
Peanut Peak and the earlier members cannot be at-
tributed to general sedimentary facies differences.
Therefore, they must result either from changes in the
physical environment, evolution within the Chadron
fauna, immigration, or most probably, a combination of
these three processes.

Even making a conservative measure of faunal
change with the index described above, there is a great
reduction in the aquatic and wet-forest elements in
Peanut Peak time — from a total index of 31 in Crazy
Johnson time to a total index of 16. In contrast, the
index for the grasslands and dry forest elements in-
creases only from 28 to 31. The change, therefore, lies
primarily in the three habitats related to the streams
and the wet forest.

The principal known changes in the physical en-
vironment from the Ahearn to Peanut Peak are (1) a
reduction in mean temperature and (2) a reduction in
the size of the streams and in the width of the irrigated
areas adjacent to the streams. Temperature changes
would affect critically the turtles and the alligator, and
the reduction of streams would modify the aquatic en-
vironment considerably. In turn, the reduction of the
irrigated areas would reduce the extent of the wet forest
and thus affect the semi-aquatic and river-border forest
animals, both in total habitat area and possibly in the
number and kind of available niches.

Further, the reduction of the areal extent of the
river-border forest would probably affect the exploita-
tion of the herbivores by the savanna carnivores.
Finally, the increase in the savanna fauna, although
relatively small, might have altered to some extent the
predation pressure and amount of competition by the
savanna animals on the semi-aquatic and river-border
forest animals. Inasmuch as the simple presence of the
savanna fauna during Crazy Johnson time did not great-
ly affect the aquatic, semi-aquatic, and river-border
forest faunas, the climatic changes appear to be the
critical factors, and these probably called into play the
biotic factors.

Olson (1952) defined a chronofauna as: "... a geo-
graphically restricted, natural assemblage of interacting

TERTI ARY

65

GEOLOGIC AGE K

53-54

58.5
t-iU PALEOCENE

P A L EOGEOG R APHIC DAT

37-38

31-32 ? oli

OLIGOCENE L|M M-U Ct

ABSOLUTE TIME SCALE AS IN
FIGURE 26.

IN EACH CASE, THE top OF
THE GRAPH REPRESENTS MANY,
AND THE IOTTOM, FEW.

THESE GRAPHS SHOW QUAL-
ITATIVE TRENDS ONLY.
THE OAT CANNOT AT PRE-
SENT BE QUANTIFIED.

ER T E B RATES.

Fig. 27. Tertiary Paleogeographic Data- Vertebrate Fossils.

SAVANNA. SAVANNA
AQUATIC GENERA
SEMtAQUATtC GEN
RIVER BORDER FQ
TOTAL NUMBER OF
THAT CAN BE ASS

ERA

REST

—

HONED

^^^

~"~— --^

^~~^~^

to

i n

____- —

_ — — "

.

-^_~^

___

-=fc-*-'iZ!

~55ii^__

""**■*»<*!,

.^

AH E ARN

CRAZY

J OMNSON peanut peak"*

Fig. 28. Chadronian paleoccology by genera.

animal populations that has maintained its basic struc-
ture over a geologically significant period of time." The
Chadronian fauna of South Dakota fails to meet the
requirements of this definition because the disappear-
ance of many genera without replacement and the ap-
pearance of some new genera in new niches demonstrate
an instability of structure. If, however, the Chadron
fauna is subdivided into an aquatic-wet forest com-
ponent and a savanna component, these components
appear to bear a clear relationship to at least two
chronofaunas.

If we examine the ancestral affinities of the aquatic-
wet forest genera we find three rather indistinctly sepa-
rated groups. One of these groups has its closest relatives
in the North American Eocene and includes Mesohip-
pus, Menodus, Agriochoerus, the four-toed camel, and
Eotylopus. The second group has close relatives in the
Eocene of both Eurasia and North America. This hol-
arctic group includes Anosteira, Trionyx, Alligator,
Trigonias, Colodon, and Daphoenus. The third group
comprises two genera, Heptacodon and Bothriodon,
known from the Eurasian Eocene only. The ancestry of
these groups then lies in the well known Eocene faunas
which are almost certainly jungle chronofaunas. It
seems probable, therefore, that these genera were sur-
vivors of Eocene jungle chronofaunas. As the rather
limited evidence of pre-Chadron environments in South
Dakota suggests a widespread jungle or wet forest,
these may have been survivors in place and in particu-
larly favorable local habitats.

The savanna displays quite a different pattern. Two
origins can be distinguished, a North American and a
Holoarctic or Eurasian (see Fig. 25). Some of these
genera have affinities with the Eocene jungle forms but,
in general, represent different families. In turn, many of
these families appeared rather suddenly during the late
Eocene or early Oligocene in one or the other of conti-
nental areas. Further, in the Chadron, the biotic struc-
ture of the savanna fauna was relatively stable with no
generic extinctions and a continuation of nearly all of

72

FIELDIANA: GEOLOGY MEMOIRS, VOLUME 5

90

80

SAVANNA SAVANNA

FOREST

AQUATIC GENERA

SEMIAQUATIC GENERA

RIVER BORDER FOREST

TOTAL NUMBER OF

GENERA THAT CAN
BE ASSIGNED WITH
CERTAINTY

Chadronian Paleoecology Index

AHEARN CRAZY JOHNSON PEANUT PEAK

Fig. 29. Chadronian paleoecology index.

these genera into the Orellan. From this we conclude
that the savanna fauna is part of an invading chrono-
fauna derived from at least two savanna chronofaunas,
one North American, the other Eurasian, which evolved
during the Eocene in areas from which we have no fossil
record.

The faunal changes in South Dakota may then be
considered the Armageddon of these two different fau-
nal systems. From this battle, decided by a climatic
shift, developed the Orellan savanna chronofauna. This
included most of the Chadronian savanna group plus a
few survivors of the wet-forest chronofauna (Mesohip-
pus) and some new genera which filled the newly avail-
able niches on the savanna and in the river-border
forest.

Obscured in the Chadron but of great evolutionary
importance were the interactions between the Eurasian
savanna genera and the North American. This resulted
in an Orellan fauna in which the majority of the her-
bivores are distinctly North American and the car-
nivores distinctly Eurasian, and thus suggests the great-
er adaptability of the carnivores.

INTERPRETATIVE SUMMARY

Review of Physical History

During late Eocene time, the streams of the Great
Plains were flowing about at grade in broad, flat-bot-
tomed valleys. The Black Hills had been eroded nearly
to their present topography, and the granite core had
been extensively exposed. The average annual tempera-
ture in western South Dakota was slightly above 60° F.,
and the average annual rainfall over 40 in. The weather
system was monsoonal, with prevailing northerlies and
northeasterlies during the summer bringing in abundant
moisture from Hudson's Bay and the Arctic, and still
air or gentle prevailing southwesterlies during the dry,
cool winters. The coldest winter temperatures probably
did not drop much below freezing, and summer tem-
peratures probably did not rise much above 100° F.

Western South Dakota stood high enough to inter-
cept the moist northerlies and received abundant rain-
fall. Eastward in the Interior Lowlands, however,
probably only a few areas (e.g., the Ozarks) stood high
enough to cause heavy precipitation, and the climatic
regimen very possibly was arid or subarid.

By the beginning of Chadron time, polar cooling had
established a hemispheric circulation strong enough to
break up the monsoonal circulation. Prevailing wester-
lies and northwesterlies, with cyclonic storms and fronts
gradually replaced the monsoonal weather. This shift
resulted in a progressive drying of western South Da-
kota with a concomitant increase in rainfall over the
Interior Lowlands. South Dakota changed from a place
of extensive semi-tropical forests to a savanna with
forest belts following the larger streams.

The increasing aridity changed the regimen of all
the streams in the Northern Plains. From a typical wet-

CLARK AND BEERBOWER: THE CHADRON FORMATION

73

tropic slope system like that of southeastern China,
with steep mountains of resistant rock rising above
almost flat lowlands of weathered strata, the streams
shifted to assume the more uniform profile normal under
semi-arid conditions. This forced the deposition of
lenses of alluvium from the edge of each mountain range
outward some tens of miles into the Plains. The lenses
were necessarily thickest very close to the foot of the
ranges, feathering rapidly upstream into the mountain
valleys, and thinning more gradually downstream.

The sediments included in the Ahearn member of
the Chadron formation represent the first depositional
episode in the development of the lens east of the Black
Hills. (The relationship of deposition of the Slim Buttes
Formation to this general history is not known.) The
materials of the succeeding Crazy Johnson and Peanut
Peak members represent the second episode.

Rainfall in western South Dakota decreased during
Chadronian time, and the mean temperature probably
dropped a little. By the end of Chadron time, the sur-
face was a broad depositional plain, with the old topog-
raphy buried mountainward to the base of the lower
Cretaceous hogback that rimmed the Black Hills.

The confluent waters of Battle Creek and Spring
Creek received French Creek as a southern tributary
and flowed eastward as the "Red River." The northern
Black Hills were drained by an east-flowing stream
which lay north of Rapid City and north of the present
Wall of the Badlands, but commingled its sediments
with those of Red River once the old divide between
them was buried.

Origin and Evolution of the Chadron Fauna

The fauna of the Chadron has some genera in com-
mon with the earlier Oligocene Duchesnean faunas but,
in general, is much more similar to the succeeding Orel-
Ian fauna. Analysis of the phyletic relationships and of
the evolving biotic structure suggests a multiple origin
for this fauna. Most of the genera associated with the
aquatic, semi-aquatic, and wet forest habitats are related
rather closely to genera from the jungle or wet forest
faunas of the late Eocene from both North America and
Eurasia. The majority of the Chadronian genera, those
associated with the savanna environment, are less close
to well-known Eocene genera and appear to be derived
from some unknown late Eocene savanna faunas. The
savanna probably represents an intermixing of North
American and Eurasian (or at least holarctic) elements
by intermigration during the earliest Oligocene.

The oldest Chadron fauna apparently is slightly
younger than the Vieja but the river-border forest por-
tion of the Chadron fauna shows very close affiliations
with the Vieja fauna. The Yoder fauna, on the other
hand, would seem to be the same age as the Ahearn but
its precise stratigraphy and paleoecologic relation to the
Ahearn awaits additional study. The Pipestone Springs
fauna is correlative with that of the Peanut Peak and is
very probably Peanut Peakian. The Pipestone Springs
fauna, however, has a larger proportion of forest animals

and demonstrates the persistence of several species that
disappeared earlier in South Dakota.

Of the Chadronian genera only a few show de-
terminable evolutionary series; the remainder are either
too poorly known or fail to show any distinct change
during this time. In general, the rate of evolution ap-
pears to be relatively slow — for example, the early
Orellan horse, Mesohippus bairdi, is barely distinguish-
able from the early Chadron horse, M. hypostylus. The
late Chadronian Parictis dakotensis and Daphaeonocyon
dodgei are, however, distinctly different from what ap-
pear to be their early Chadronian ancestors, P. parvus
and D. minor.

Aside from the trends in these latter genera, no clear
evolutionary patterns appear in the Chadron. This is
very probably an artifact of our data inasmuch as we
are unable to determine the adaptive significance of
minor structural changes.

The fauna as a whole, however, shows rather marked
development. During Ahearnian time, over half of the
fauna was of aquatic, semi-aquatic, or river-border forest
types. This part of the fauna was slightly reduced in
variety and abundance in Crazy Johnson time, but the
difference may not be significant. In Peanut Peak time,
however, many of these genera disappeared and the
remainder were less abundant. On the other hand, the
savanna genera increased slightly in numbers and in
relative abundance from Ahearnian to Peanut Peakian
time.

The relation of these changes and the faunal origins
discussed above suggest interaction of a surviving
Eocene wet-forest chronofauna with immigrants from a
savanna chronofauna. The increased aridity of Peanut
Peak time gave the final decision to the savanna genera
and only a few wet-forest forms survived to form part of
the Orellan savanna chronofauna of the South Dakota
area. We conclude that the critical factor was the cli-
matic shift but that once initiated, competition and
predation became increasingly important in the final
destruction of the wet-forest chronofauna. In this re-
spect, some forest genera apparently survived longer in
the forested valleys of Montana and were represented in
the late Chadronian Pipestone Springs fauna.

Comparison with chronofaunal development in the
Texas Permian (Olson, 1952; Olson and Beerbower,
1953) discloses a number of significant differences and
similarities. The equivalent of the Eocene wet forest
chronofauna appears to be the early Permian (Wichita
and Clear Fork ages) delta chronofauna. The disappear-
ance of the Chadron wet forest chronofauna and its
replacement by the savanna savanna-forest chrono-
fauna is analogous to the disappearance of the delta
chronofauna in late Clear Fork time and the appearance
of the uplands chronofauna in San Angelo and Flower
Pot time (early middle Permian). The extinction of the
Permian delta chronofauna, however, must be attri-
buted solely to deterioration of the physical environ-
ment to an arid, salt-pan floodplain characteristic of late
Clear Fork time, rather than to biotic pressure as the

74

FIELDIANA: GEOLOGY MEMOIRS, VOLUME 5

upland fauna is unknown prior to late San Angelo time.
The Chadronian thus is equivalent to both late Clear
Fork and San Angelo times and must, therefore, re-
flect a somewhat more complex faunal development
than occurred in the Texas Permian.

Even with these differences, the general pattern of
extinction and development are much the same. In
neither case is there any evidence of rapid evolution of
any of the component genera. The new genera, on the
contrary, appear by immigration and thus may repre-
sent slow evolution in some other area. These new gen-
era probably evolved somewhat more rapidly than the
genera in the delta or wet forest chronofaunas as they
diverge rather more from their common ancestors, but
many differences would appear to be in the direction
rather than in the amount of evolution.

The extinction of a chronofauna appears a gradual
process though groups of closely integrated genera
within the chronofauna may become extinct almost
simultaneously and thus produce discontinuities in the
downward curve. During this period some immigrant
genera appear briefly to fill new niches or to crowd indi-
geneous genera from changing niches. These forms
however disappear also as the wet-forest or delta en-
vironment continues to change. The development of
the new chronofauna likewise appears to be gradual,
although documentation of this stage in the Texas
Permian is poor. Again, groups of genera may appear
suddenly and produce discontinuities in the upward
curve. The entrance of genera into the new niches can
proceed no faster however than the appearance of the
niches with the changing environment and the occu-
pancy of potential niches is restricted by the need for
the genus to adapt to the peculiarities of the new en-
vironment.

CONCLUSIONS

The following conclusions seem justified by the
characteristics of the Chadron and of its fauna:

1. During late Eocene time monsoonal air circula-
tion prevailed over the Great Plains and Rockies. The
mountains and higher basins received heavy rainfall,
but the Interior Lowlands were probably arid or sub-
arid. Temperatures were warm-temperate or sub-tropi-
cal, and the low temperature differential between Pole
and Equator subordinated the hemispheric circulation
system to local systems, in this case a monsoon.

2. The relief of the Black Hills and other ranges
under this climatic regimen was very like that of the
present.

3. A major decrease in temperature starting at the
close of Eocene time established a stronger differential
between the North Pole and the Equator and [conse-
quently a stronger hemispheric wind system with pre-
vailing westerlies. This period of cooling appears to be a
minimum in an older 10-million-year climatic cycle,
coincident with the beginning of a broader downward
temperature trend that culminated in the Pleistocene
minimum. Periodicity of 20 million years seems to
characterize the downtrend period.

4. As a result of this global climatic change, the local
climate became drier and somewhat cooler. The stream
regimen was altered and deposition initiated adjacent
to the mountain ranges.

5. Deposition continued into late Chadron time be-
cause of continued climatic deterioration. Fluctuations
in deposition during this period are probably related to
minor climatic fluctuations which may be the result of
the same rhythms shown in the Pleistocene by glacial
advances and retreats.

6. The earliest Chadron is somewhat younger than
the Vieja. The Yoder fauna is Ahearnian and the Pipe-
stone Springs is Peanut Peakian.

7. During middle and late Eocene time, a semi-trop-
ical rainforest chronofauna developed in the swampy
woodlands of Utah, Wyoming, Colorado, and South
Dakota. This is recorded in Bridger and Uinta fossils.

8. Concurrently, a savanna to arid chronofauna
developed elsewhere, possibly in the Interior Lowlands.
The history of this chronofauna, however, is unre-
corded.

9. During Chadronian time, the forest chronofauna
lingered along the stream-margins in Dakota. It under-
went gradual, partial replacement by the immigrant
savanna chronofauna.

10. A few of the genera of the forest chronofauna,
among them Trigonias, Mesohippus, and Pseudopro-
toceras, managed ultimately to evolve into savanna and
plains form.

11. Study of the Chadron chronofauna tends to
substantiate conclusions drawn earlier from the Texas
Permian by Olson.

Chapter VI

PALEOGEOGRAPHY OF THE SCENIC MEMBER

OF THE BRULE FORMATION

by

John Clark

INTRODUCTION

Widespread, excellent badlands exposures combine
with sharply denned sedimentary facies to make the
Big Badlands an almost ideal area for the study of
fluvial sedimentation through time. Brule channel de-
posits, unlike those in the Chadron Formation, are
restricted to definite courses and separated from each
other by considerable distances. This makes possible
the mapping and study of individual Brule streams
in detail which cannot be approached in studies of their
Chadron predecessors. Recognition of differences be-
tween heavy-mineral suites from the northern and the
southern Black Hills has further enhanced the precision
of paleogeographic studies in the area.

This chapter in part summarizes a series of prelimi-
nary studies of Oligocene sedimentation which the
author and his students conducted at the South Dakota
School of Mines and Technology from 1958 through
1961, financed by grants from the National Park Serv-
ice, through the Badlands National Monument. The
final studies reported here were completed under the
auspices of Field Museum of Natural History during
1963 and 1964. The author is indebted to his colleagues
at the Museum for numerous ideas and conferences.
Those students who made helpful contributions (Ritter
and Wolff, 1958; Seefeldt and Glerup, 1958) have
fortunately published their work, and hence have al-
ready received proper credit for it.

This research has proven useful, not only in revealing
details of Oligocene paleogeography, but also in eluci-
dating certain aspects of fluvial sedimentation which
have not, to the best of my knowledge, been previously
recognized.

TOPOGRAPHY AND GENERAL
STRATIGRAPHY
The Brule Formation characteristically erodes to
badlands slopes of 30-75°, varying to barren flats and
occasional vertical cliffs where the local situation pro-
duces special conditions. The Brule can be differen-

tiated at once from the underlying Chadron by the
change from convex-outward Chadron "haystack"
slopes to steep, concave-outward slopes, and usually by
a change from predominantly pale tan or greenish-gray
color to yellowish tans. Gully texture is also very much
finer on Brule slopes than on the Chadron.

The actual contact usually consists of a silicified
pond limestone a few inches thick, containing Chara-
gonia, snails, and ostracods. The limestone is nowhere
continuous for more than a mile, and usually not for
over one-half mile. Between these definite pond lime-
stones and replacing them in the area north and east
from Scenic, the contact consists of a few feet of mud-
stones with intercalated limy laminae which often
crosscut individual mudstone strata.

The Scenic Member of the Brule everywhere directly
and conformably overlies this remarkably flat limy zone.
Only at Chamberlain Pass (Sec. 25, T. 3S., R. 13E.,
Pennington Co.) is there an appreciable relief, and
there it does not exceed 10 ft. (see Fig. 30). The Scenic
Member consists of alternate layers a few feet thick of
buff, tan, red, or gray mudstone with laminated gray
to green siltstone and occasional channel fills of greenish
sandstone.

Bump (1956) described as the top of the Scenic
Member a prominent dark band which separates banded
Scenic sediments from the much more massive, buff-
colored, Poleslide tuffaceous clays which conformably
overly them. The change to more massive bedding in the
Poleslide is everywhere apparent, but this change
does not everywhere occur at the same horizon, and the
dark marker bed extends only from Bump's standard
section south westward. The apparent dividing horizon
between Scenic and Poleslide is not, therefore, exactly
coeval throughout the Big Badlands. This situation
resembles that of the Orella and Whitney members of
the Brule in Nebraska (Schultz and Stout, 1938, p.
1921; Schultz et al., 1955, p. 4; Schultz and Stout, 1955,
p. 44). The Orella-Whitney contact at their standard
section, Toadstool Park, can be observed to change

75

a

at

a.
os
—
(J

T3

a
o

C

o

c

76

CLARK: PALEOGEOGRAPHY OF THE SCENIC MEMBER

77

position stratigraphically by 5 ft. in 200 yards. Obvious-
ly the lithologic characteristics of these members reflect
changes in the nature of deposition, rather than cessa-
tion and resumption.

LITHOLOGY

Detailed field observations, plus laboratory study
of 111 specimens in the FMNH collection, have sup-
plied the data for this section.

A. Type of Sediments
The Scenic Member comprises five types of sedi-
ment: 1) limestone, 2) heterogeneous mudstone, 3) lami-
nated clay, 4) laminated siltstone and sandstone, and
5) cross-bedded sandstone.

1. Limestones. These are of two types. The first are
local, inconspicuous limestones never more than 2 in.
thick and 100 yards in diameter, which consist almost
entirely of algal crusts and threads. The thicker ones
are laminated. No fossils other than algae have been
observed in them. These limestones occur interbedded
with the laminated siltstones and sandstones. They are
uncommon, and apparently represent ephemeral stages
of rapid algal growth in small, temporary swamps.

The other type of limestone is massive to flaggy, and
frequently shows some silicification. Lenses up to al-
most 3 ft. thick occupy areas of 200 yards to more than
a mile in extent. These limestones contain Chara, ostra-
cods, gastropods, and occasional fish bones. They occur
rarely, and only at the base and at the top of the Lower
Nodular zone, in the areas southwest of Scenic and
southeast of Wall. The fauna and flora leave no doubt
that these were shallow ponds of some permanence.
Certain of these ponds were significant elements of
Oligocene flood plain morphology, as will be discussed
later.

2. Heterogeneous mudstones. Complete hetero-
geneity, embracing particles of all sizes from 5^ up to
clay pellets 2 cm. in diameter, characterizes these rocks.
They occur as layers a few inches to 40 ft. thick; the
usual thickness is 3-20 ft., composed of poorly-separated
increments 3 to at least 18 in. thick. Contacts with the
subjacent and superjacent sediments are always sharp.
A matrix of tan to gray mudstone invariably includes
numerous sharp-edged chips and pebbles of darker and
lighter mudstone which generally differs from the ma-
trix in being less cemented, (see Fig. 31)

Microscopically, the mudstones consist of tiny quartz
fragments mixed with devitrified glass and a few fresh
shards of glass, interspersed with clay which is largely
montmorillonite or mixed-layer. Calcite cement pene-
trates the mass, which varies from completely farctate
to moderately porous, depending upon the extent of
cementation. Neither thin-sections nor skiagraphs re-
veal any trace of arrangement of even the finest par-
ticles. Grains of very fine sand are occasionally included,
especially in the vicinity of channel-fill sandstones.
These consist of quartz, fresh microcline and sanidine,
and biotite. The biotite flakes are usually fresh, but

sometimes show a tiny cloud of hematite surrounding
the frayed edges or rising from a single spot on a cleav-
age face, indicating incipient intrastral alteration.

The chips and pellets of mudstone differ very little
from the matrix. Rarely, they show a higher percentage
of calcite, indicating a higher initial porosity. Usually
they yield to weathering more readily than does the
groundmass, producing the rough, pitted surface char-
acteristic of "Lower Nodular Zone concretions." Treat-
ment of fresh-cut surfaces with water causes the chips to
swell, while the groundmass remains unaffected, sug-
gesting that the chips are more clayey than the ground-
mass.

Although the contact surfaces of these mudstones
with other sediments are always sharp, they usually
show ragged irregularities. Chips and particles of under-
lying laminated clays occur in profusion within the
lower few inches of any mudstone stratum (Specimens
G 3743, G 3745, G 3999, G 4058, G 4077).

3. Laminated clays. Layers of laminated clay, from
less than an inch thick to an extreme of 2 ft. thick, form
a notable proportion of the total mass in the Sage Creek-
Dillon Pass area, decreasing southwestward both in
thickness and in total bulk. South of Scenic they occur
as laminae less than 3 in. thick, interbedded with silt-
stone and fine sandstone laminae; west of Sheep Moun-
tain they have not been observed, although a few
laminae probably occur there. They are present and
conspicuous to the southeast of Cuny Table (see Figs.
30, 32.)

The laminae in these clays are not apparent on field
inspection. A variety of laboratory staining techniques
have been tried upon them, none of which has proven
completely satisfactory. The best results to date have
been obtained by cutting a smooth surface, soaking the
specimen overnight in kerosene, and placing the
smoothed surface face down on a hot metal plate. The
laminations usually show up beautifully after this treat-
ment, but the resulting stain fades almost completely
within a few weeks.

Laminations vary from about 2 mm. to less than 0.5
mm. in thickness. They are exceedingly regular, showing
no depositional structures. Post-depositional deforma-
tion comprises: (1) small, vertical faults with vertical
displacement of 1 cm. or less; (2) small, curved, slicken-
sided surfaces developed during compaction by over-
lying sediments; (3) intricate systems of more or less
vertical cracks, with intrusion of heterogeneous mud-
stone from above or below, depending upon the local
situation. (See specimen G 3743, G 3745, G 4046, and
others) . The crack-fills form a network with a spacing of
1 to 4 in. They vary from a fraction of an inch to more
than an inch wide. The wider ones show considerable
foundering of small blocks of laminated clay into the
surrounding mudstone, also small hollowed-out, eroded
excavations along the walls, and all stages of assimila-
tion of clay particles into the mudstones. The structures
are exact homologues of igneous roof-structures periph-
eral to a batholith.

78

FIELDIANA: GEOLOGY MEMOIRS, VOLUME 5

^^^^^T*

1^^ :

• In

Wr .'" *

'■?':

■

z^

1

B^ -Jr .

' - '

*\

7v^^^

K

i

! 4;

i

■■: .'-

y 1

1

* */ ' 1

^B * CT^V

1

.* • i

• H

' 4 ■

**.*.'" '^1

h

-* '^1

■ ' * •

* ^ '*'

^^^^^^^^^^^^ 1

Fig. 31. Photograph of G 4077, a sample of heterogeneous
mudstone, showing heterogeneity in both horizontal and vertical
sections.

The laminated clay must have accumulated in shal-
low, temporary ponds, and cracked due to desiccation
shrinkage. Absence of any curved laminae indicates that
intrusion of fluid mud from below, or in some cases
burial by fluid mud, occurred before dessiccation was
far advanced. All of the cracks observed completely
penetrate their respective laminated sequence; none
have been found which affect only the lower moiety of a
laminated stratum. Any one laminated sequence, there-
fore, accumulated during a time of continuous existence
of a pond.

The laminae need not represent annual increments,
nor even those of individual storms or floods. Experi-
mentally, a lump of this laminated clay was thoroughly
mixed with water, permitted to settle, then disturbed by
one brisk shake: a sequence of four distinct laminae
developed on the bottom. Presumably, therefore, a
strong wind, or one of the larger Oligocene mammals
wading, could have disturbed a shallow pond enough
to have produced multiple laminae after a single influx
of sediment.

The clays vary from somewhat calcareous to non-
calcareous, and generally are much lower in montmoril-
lonite than are the clay fractions of the heterogeneous
mudstones. Whether the laminations reflect differences
in clay-mineral composition, or texture, or cementation,
is not known.

4. Laminated siltstones and sandstones. Laminae 1
to 20 cm. thick characterize these sediments. Except in
the proximity of channel-fill sandstones, they show no
cross-bedding. Individual layers thicken and thin,
change texture, or lens out, so gradually as to be per-

ceptible with difficulty. The layers are gray to greenish
siltstones and fine-grained sandstones, characteristically
well-sorted but thoroughly mixed with abundant mont-
morillonite. Any one layer shows no sorting or stratifica-
tion whatever, but each layer has sharp contacts with
the adjacent underlying and overlying ones. The clastic
minerals of these sediments always resemble those of the
nearest channel-fill sandstone.

Thin, algal limestones occur as lenses of restricted
areal extent interbedded with these siltstones and sand-
stones, as mentioned above.

5. Cross-bedded sandstones. Cross-bedded sand-
stones occur as linear channel fills (see Figs. 33-36),
which occupy areally restricted zones. Westward from
Sheep Mountain, the channel fills generally have defi-
nite top and bottom contacts with the surrounding finer
sediments; to the east and northeast, neither the lateral
nor the top and bottom contacts are sharp. The sand-
stone zones show distinct lateral restriction, but extend
vertically through a considerable thickness of beds with
which they intergrade and interfinger. These sandstone
masses, therefore, represent deposits in the beds of
individual aggrading streams through a considerable
period of time, rather than the fill of channels which
once existed as trenches reaching from top to bottom of
the present sandstone mass.

Cut-and-fill structures are absent, and foreset cross-
beds nearly so. Cross-beds with dips up to 20°, usually
not over 10°, strike characteristically parallel to the
direction of flow (see Figs. 34, 35). In parts of every
channel-fill outcrop, the sands are disposed in a series

Fig. 32A and B. Photograph of G 3743, showing laminated
clay with intrusive heterogeneous mudstone. Stained specimen.

111

m

UJ

1

z

UJ

o

•SI

>

I

0.
<
a:
0

o

UJ

ID
O

UJ

<

0.

. >

C£ UJ
UJ D

I
I-

O

I >

I E

- uj

UJ Q

I
(-
D

o

fl I IT -

n ■ t

J2

J3

a

C4

a!

Pn

as

— ^

79

80

FIELDIANA: GEOLOGY MEMOIRS, VOLUME 5

Fig. 34. Riosome No. 3, near Cottonwood Pass. Looking downstream. Shows cross-bedding normal to direction of flow.

of low, rounded crests and troughs, with distances of 2-5
ft. between crests, whose axes parallel the direction of
flow. The height is usually about 4 in. Any one crest can
generally be traced for a few tens of feet. The sands in
these sediments are usually coarse to fine grained, and
somewhat more micaceous and better-sorted than the
material in the completely unbedded portions of the
sandstone mass. Obscure ripple marks of low height and
wave lengths of 3-6 in. occasionally V downstream in
the troughs, but linguoid and well-developed oscillation
ripples have not been observed.

The total width of a "channel-fill" at any one hori-
zon is 100-300 ft., and the apparent depth at any one
time, estimated from the greatest vertical height of any
one set of crossbeds, was not over 3-6 ft. Therefore,
these longitudinal troughs and crests probably do not
represent islands in a typically braided stream, (Orr,
1964, p. 1) in that they probably did not protrude above
the surface. This is further borne out by the fact that
individual sedimentary increments can be traced across
several troughs and crests.

The troughs and crests, plus the margins of the
channel-way, produce cross-bedding of which over 90%
dips normal to the direction of flow. The remaining
cross-beds dip predominantly downstream, with a small
fraction dipping upstream. Were it not for the linear
shape of the deposits and the proximity of the Black
Hills as the known source of elastics, the direction of
flow would be very difficult to ascertain. Shingling, or
any definite orientation of pebbles, has not been ob-
served.

Interbedded with these sands are others showing
little or no sorting and no visible bedding. Cementation

is usually notably less in the massive layers, and the
percentage of montmorillonitic clay is notably higher.
The coarsest gravels of any one outcrop usually, but not
always, occur in the massive layers.

Seefeldt and Glerup (1958) studied the clastic min-
erals of the channel-fill which lies (Fig. 33) in Sec. 25
and 36, T. 43N., R. 45W., and Sec. 30, T. 43N., R. 44W.
They found that the average mineral composition is:

Barite 2%

Limonite 1 %

Hematite 1%

Garnet 63%

Black Tourmaline 26%

Orange Sphene 3%

Zircon 2%

Occasional actinolite, chlorite, glauconite, red-brown
hornblende, biotite, and muscovite grains were also
noted. Individual samples vary widely from these
averages, but all of the channel-fills mapped as "South-
ern-derived" (Fig. 33) show heavy percentages of garnet
and black tourmaline, with small amounts of orange
sphene and brown hornblende; none include magnetite,
lemon-yellow sphene, or greenish hornblende. Samples
of recent sand from French Creek near the town of
Custer, and from Battle Creek near Hermosa, resemble
the Oligocene sediments in their heavy-mineral as-
semblages.

Since these channel-fills trend generally east-south-
eastward from the southern Black Hills, possess granite
pebbles and mineral assemblages known to occur in the
southern Black Hills, and resemble recent stream sands
from the same area, they are believed to represent
Oligocene streams whose head-waters lay in the south-
ern Black Hills.

The channel-fills east and northeastward of Sheep
Mountain differ notably from those just described.
They have much less definite boundaries and are much

CLARK: PALEOGEOGRAPHY OF THE SCENIC MEMBER

81

less thoroughly cemented than the Southern-derived
sandstones. Any one outcrop includes individual lami-
nae or cross-beds which are thoroughly indurated, inter-
bedded with others soft enough to crumble between the
fingers. The channel-fills weather into greenish-gray
columnar zones several feet thick, which look at a
distance like over-accentuated areas of laminated silt-
stone. They grade laterally into laminated siltstones,
differing from them in possessing cross-bedding and
much coarser sediment. Those which occur within mud-
stone zones are very much smaller and finer-grained
than those within the siltstone zones.

Ritter and Wolff (1958, p. 189) found the following
heavy-mineral assemblage in the large, northeastern-
most channel-fill which they traced for 12 miles (#11,
Fig. 33):

Hornblende 22%

Barite 19%*

Magnetite 23%

Sphene (lemon-yellow) . 8%

Biotite 6%

Limonite 6%

Epidote 5%

Garnet 4%

Hematite 2%

Tourmaline 2%

Minor, variable amounts of chlorite, glauconite, gold, pyrite*,
and zircon were also noted.

* Grains of these two minerals are subhedral to euhedral and
neither would withstand long transport. Hence they were probably
locally derived and have no bearing on ultimate headwaters of
the Oligocene streams.

The lemon-yellow sphene is exactly like that known
to occur in certain of the Tertiary intrusives of the
northern Black Hills. This sphene is not known to occur
elsewhere. The magnetite is believed to have come from
these intrusives also, but definite evidence is lacking.

Abundant magnetite and greenish hornblende, low
but consistent percentages of lemon-yellow sphene and
greenish epidote, low percentages of tourmaline, and
generally low percentages of garnet characterize north-
ern Black Hills-derived sands. The three channel-fills
nearest Sheep Mountain contain much higher percent-
ages of garnet than do those farther northeast, but
otherwise they are typically northern-derived. Since
garnet is locally abundant at many places in the north-
ern Black Hills Precambrian, this variation is not re-
garded as significant.

The largest pebbles observed in northern-derived
channel-fills are less than 1 in. in diameter; the more
easterly contain nothing larger than a coarse sand. This
suggests that the velocity of the transporting streams
decreased rapidly as they travelled farther from the
Black Hills and is in agreement with observations on
individual channel-fills (Seefeldt and Glerup, 1958).

The northern-derived sediments contain minerals
known to occur abundantly in the northern Hills, and

Fig. 35. Riosome No. 3, near Cottonwood Pass. Looking south, normal to direction of flow which is from west to east, or right to left
in the picture. Note lack of cross-bedding parallel to flow.

82

FIELDIANA: GEOLOGY MEMOIRS, VOLUME 5

not known to occur in the southern Hills. Geographical-
ly, the northern-derived channel-fills lie to the north of
the southern-derived ones, and are roughly aligned to-
ward the northern Black Hills.

i

Fig. 36. Drenajesome No. 4, looking north, or upstream. Pick
marks the site of a sandstone dike of Chadron sand intrusive into
Brule. The drenajesome lies between mudstone I and III; III is
overlain by well-developed laminated siltstones.

B. Sedimentary Lithotypes.

The individual rock types described above occur
associated into four sedimentary lithotopes: (1) silty
mudstones; (2) laminated siltstones; (3) laminated mud-
stones; (4) channel -fill zones. These lithotopes grade
into one another laterally, but not vertically.

1. Silty mudstones. This lithotope consists almost
entirely of the heterogeneous mudstones described.
Rarely, it includes discontinuous limestones. Commonly
it grades laterally into sandstone along the borders of
the northern-derived channel zones.

The mudstone lithotype makes up the famous Lower
Nodular Zone (Wanless, 1923, p. 213, et al.) and several
zones higher in the Scenic Member. It consists of mas-
sive buff, gray, bright yellow, or reddish mudstone,
usually with greenish mottling. Weathered surfaces de-
velop a crust with an intimate system of cracks, which
outline polygons about an inch in average diameter.
Less-indurated strata weather to "pop-corn" surfaces
quite like those of the underlying Chadron.

Within \l/2 miles of channel-fills occurring at the
same horizon, these mudstones are irregularly indurated
into nodular concretions by a calcite cement. The con-
cretions occur usually in the buff or gray phases, rarely
in the bright yellow, and never in the red. They are
generally 4-12 in. thick vertically, and of slightly great-
er horizontal than vertical diameters. At a distance from
the channel-fills, individual concretions occur thinly
scattered at separate, discontinous levels within the
mudstone zones, but they increase to quite regular,
massive beds nearer to the sandstones. None at all occur
within 100 yards of the sandstone channel-fills proper.

Individual concretions contain numerous clay-peb-
bles which weather out producing an irregular pitting.
A heavy incrustation of limonite stains the weathered
surfaces umber brown, but freshly broken faces are pale

gray or tan, with either greenish or tan clay-pebbles.
Concretions frequently enclose fossil bones or entire
skeletons, but coprolites occur rather in the uncemented
clays. Nodular concretions are widely distributed
through the mudstone layers from the Sage Creek basin
southwestward to Cedar Creek (west-southwest of the
area covered by this report) ; they lie in one thin, dis-
continous band within the Lower Nodular Zone as far
east as Dillon Past. They never occur in the red zones,
and die out quickly in the pale-yellow zones 3 miles
northeast of Scenic, and also southwest of the area of
channel fills (Fig. 33). They occur, but are less well-
developed, in the mudstone zones above the Lower
Nodular.

Sinclair (1921, p. 463) proposed that the nodular
calcareous concretions in the Lower Nodular zone repre-
sent a limy caliche, produced by carbonate-laden water
soaking out of streams into the surrounding flood-plain
deposits and evaporating at or near the surface. Wan-
less (1923, p. 216) concurred. The distribution of the
concretions near channel-fills, and the fact that they
occur in layers which are horizontal but occur in other-
wise massive, unbedded sediment, could support this
hypothesis, or could equally well result from the action
of groundwater soaking out from the permeable chan-
nel-fills at any time after deposition.

However, the fact that they are absent within 100
yards of channel fills at the same horizon strongly sup-
ports Sinclair's hypothesis. A stream-border zone, even
in a dry climate, might well be the site of fairly rapid
groundwater movement and little precipitation, while
the areas farther out would be evaporating water as
rapidly as it infiltrated, producing a cement. Still
farther from the nearest streams, influent groundwater
would not penetrate in sufficient quantity to develop a
continuous cement. On the other hand, groundwater
soaking from sandstones into neighboring sediments at
some time after deposition should have deposited its
dissolved load either as a concentrated "front" zone at
the outer boundary of its area of penetration, or de-
creasingly outward from the sandstones, depending
upon the cause of the precipitation.

Distribution of the concretions, therefore, supports
Sinclair's hypothesis that they are a penecontempora-
neous caliche produced by influent groundwater. Since
the development of caliches is normally a phenomenon
of dry climates, the nodular concretions suggest deposi-
tion under conditions of aridity more severe than during
the preceding Peanut Peakian (in which caliches were
not developed). It is my belief that the calcite formed
below rather than at the surface, but I have found no
definitive evidence of this.

The general yellow-tan to gray color of the mudstone
indicates a higher state of oxidation than the more
greenish clays of the Chadron, which also suggests that
these mudstones were deposited on essentially dry sur-
faces with little included organic matter to keep them
reduced.

CLARK: PALEOGEOGRAPHY OF THE SCENIC MEMBER

83

The complete heterogeneity of these sediments does
not readily accord with their apparent fluvial origin and
flood-plain environment of deposition. The presence of
unweathered feldspars and practically unweathered bio-
tite proves that the mudstones are not soils; the hetero-
geneous distribution of large chips gives structural evi-
dence in the same direction. Careful observation reveals
no trace of animal burrows or digging which might have
destroyed all bedding. Even the thickest grass cover on a
depositional surface would have holes and thin spots
where some sorting and lamination would occur. There-
fore, the heterogeneity probably is not related to
phenomena at the depositional interface. Furthermore,
the fairly regular increase in size and angularity of the
larger clay lumps toward the nearest channel fills
suggests that the heterogeneity is a phenomenon related
to conditions of transportation.

Biostratonomic data furnish decisive clues to the
factors necessary to produce heterogeneity.

First, deposition occurred in increments over flood-
plain surfaces which were relatively dry between floods.
The abundant coprolites, hackberry seeds, and par-
tially weathered bones attest to this. Such additional
details as groups of carnivore coprolites around a pre-
depositionally weathered herbivore skull, a rodent
partial skeleton inside a turtle shell, skeletons in tetanic
death poses, and the sheer presence of abundant non-
aquatic mammals over many square miles can only be
explained by assuming inter-depositional sub-aerial
episodes. Presence of twisted, calcareous tubes resem-
bling tree-roots may indicate exposure long enough to
develop forests.

Second, the floods deposited increments 4 to at least
18 in. thick. Thickness of an increment is best revealed
by the vertical thickness of the largest skull (or other
bone) to be buried by that increment. If any bone
should be partially buried by one flood, the temporarily
exposed portion would be much more weathered than
the lower, buried part. This is never the case. Fresh
bone is fresh from top to bottom, and cracked, spalled
bone is equally weathered from top to bottom.

An interesting example of this is FMNH PM 9359,
a skull of Archaeotherium mortoni, which occurred up-
side down, with the lower jaws missing. One upper
canine had fallen from its alveolus before deposition,
and lay enmatriced about 6 in. anterior to the skull,
roughly on the same horizontal plane as that on which
the skull rested. Fragments had broken from the back of
the cranium, and lay a few inches behind and to the left
of it. The skull itself was cracked in many places, but
equally so from top to bottom. There can be no doubt,
o, that the skull was weathered in air for several years
before burial, b, that burial was sudden and complete,
burying an object of more than 6 in. vertical dimension,
and, c, that burial occurred so gently as to engulf both
the skull and its surrounding fragments without moving
them. This specimen is typical of many hundreds from
the mudstones, and none are known which offer con-
trary evidence.

Third, the heterogeneous mudstones have undergone
little compaction during deposition and essentially none
since then. The less-cemented portions show up to 25%
permeable pore space, and the thoroughly-cemented
nodules show, in thin section, that cement occupies
what must once have been 25-30% open pore spaces.
The great majority of fossil skulls show deformation
only by cracking followed by the slight warping of
individual pieces of bone which usually accompanies
post-mortem desiccation. Nowhere do the mudstones
exhibit compaction type slickensides. By contrast, the
Chadron mudstones are extensively slickensided, and
fossil skulls enmatriced within them always show strong
flattening. Finally, the included chips and mudlumps
in the Brule mudstones show no distortion whatever,
as they certainly would if significant compaction or
post-depositional flowage had occurred.

Fourth, the mud which comprises the mudstones
was deposited as a moderately viscous fluid rather than
as material settling from suspension. The skull of
Archaeotherium, FMNH PM 9359, demonstrates the
nature of the entombing fluid very clearly. Apparently
the bone was sufficiently dry at the moment of burial to
abstract water from the enveloping fluid and cause it to
"freeze" or gel instantly. The entire dorsal surface of
the skull (which faced downward as the skull lay upside-
down) is covered with markings of tiny clay ropes 0.7-
3.0 mm. in diameter, twisted into tight, flat spirals
5-17 mm. in breadth, which lie appressed against the
bone (Fig. 37). These can only be interpreted as turbid
flow-currents of influent viscous fluid forcing its way
under the skull. No possible mechanism of settling from
a thin water suspension could produce such structures.
Piece #3 (the matrix from the specimen has been pre-
served and the pieces numbered to show position) of the
matrix, which comes from the right temporal fossa,
shows a series of partial lamellae, shaped like concentric
open cones with the apices downward, pressed into each
other but still distinct enough to cleave apart, although
no lineation or arrangement of grains is apparent. This
structure also can only be explained as engulfment by a
series of pulses of viscous, fluid mud.

In summary, the mudstones consist of completely
heterogeneous material, originally highly porous, which
has not been significantly compacted. They were de-
posited in increments 4 to 18 in. thick, and show struc-
tures indicating that they flowed as thick, viscous fluids
which set instantly when they lost water by contact
with dry surfaces. This is consistent with the high
montmorillonite content, and is the only explanation
which fits the observed evidence. The high initial
porosity favored penecontemporaneous migration of
influent ground-water, which deposited interstitial cal-
cite in the form of nodules and tabular masses.

2. Laminated siltstones. This sedimentary lithotope
consists of alternate laminae of coarse and fine siltstone,
massive, fine-grained sandstone, and laminated mud-
stone (see Fig. 36). Occasionally, a lamina of hetero-
geneous mudstone is also included in the sequence, but

84

FIELDIANA: GEOLOGY MEMOIRS, VOLUME 5

these are always less than 1 ft. thick, occur near chan-
nel-fills, and are quite discontinuous.

Individual laminae vary from less than an inch to
over 10 in. in thickness; they show regular bedding
rather than scour-and-fill structures, and may be traced
for distances of hundreds of yards to over a mile. The

Fig. 37. Photograph of marks of viscous flow on a skull of
Archaeolherium, FMNH-PM 9359.

coarsest siltstones grade into extremely fine-grained,
massive, structureless sandstones; they usually con-
stitute the thicker laminae. These coarser sediments are
greenish in color; the finer siltstones are gray, and the
mudstone laminae buff to grayish. The laminated silt-
stone lithotope occurs in total thicknesses of a few feet
to more than 30 ft. From Scenic eastward, it character-
istically thickens and coarsens to occasional channel-fill
sandstones with gradational borders; southwest of
Sheep Mountain the borders between siltstones and
channel-fills are usually sharp.

The coarser laminae develop flat, greenish concre-
tions which are always as thick as the laminae vertically,
and extend horizontally two to four times their vertical
diameter. At a distance from channel-fills these concre-
tions occur singly and are larger, and of irregular
rounded shapes, but within 1 mile of a definite channel
fill they form almost continuous beds. The cement is
calcite. The concretions weather to a much lighter
brown than do those of the mudstones, and the surfaces
are always smoothly rounded, never pitted. No fossils
have ever been found in these concretions or in the
siltstones proper.

A curious structure of completely unknown signifi-
cance usually characterizes these concretions (Fig. 38).
Series of thin, vertical plates, spaced 1 to 5 mm. apart,
transect them in many areas. The plates are irregularly
bundled into sheaves, but they do not touch each other,
and over any one area of a few tens or hundreds of
yards they have a strong directional trend. The trends
usually vary through the NW-SE quadrants. These
plates weather out to fine, paper-like ridges which evi-
dently represent structures penetrating the concretion.
However, freshly-broken surfaces show no trace of them.
They do not show up in thin-sections, except in sections
over 50 n. thick, where they appear as light lines.

Concretions formed in coarser, greenish layers generally
have either no plates at all or more widely spaced ones.

Returning to a more general consideration of the
laminated sediments, it is apparent that these are de-
posits from bodies of standing or slowly moving water.
The stratification, uniform for distances of hundreds of
yards, the generally good-to-excellent sorting of elastics
(except for clays; all of these sediments contain moder-
ate to high percentages of uniformly dispersed mont-
morillonite), the relatively thin bedding, and the ab-
sence of channelling or scour-and-fill structures all
indicate subaqueous deposition. The generally coarser
sediment in this lithotope suggests a higher energy
system than that which transported the heterogeneous
mudstones. The larger amount of water needed to
sheet-flood the Badlands area, as opposed to the smaller
amounts needed to produce mudflows from each stream,
is in accordance with this. Generally the channel-fills
interbedded with mudstones are narrower and com-
posed of finer sediment than are the same channel fill
zones higher or lower in section where they interbed
with the laminated siltstones.

However, the siltstone sequences show no varves,
nor any development of graded bedding. Probably they
do not represent lakes that persisted for as long as a
year, but rather flood-lakes which drained off within a
few weeks after each rainy spell. The only fossils ob-
served are algae in a few very thin limestones, and
one trail of a limbed invertebrate in a sandy siltstone.
If these sediments accumulated in semi-permanent
lakes, fossil fish and leaves should have been locally
abundant. Also, there are nowhere any beach or beach-
bar deposits on even a minute scale.

The absence of mudcracks is attributable to the high
percentage of silt and sand in these sediments, which
greatly reduces shrinkage and thereby inhibits cracking.
However, the absence of any root-marks, burrows, or
other disturbance of bedding surfaces indicates plainly
that periods of exposure between floods were not long
enough to allow development of a forest or even of a
well-established plains flora. It may be that the reason
for the absence of fossils is that periods of lake develop-
ment alternated so rapidly with times of emergence
that neither a lacustrine nor a terrestrial biota had time
to establish itself.

3. Laminated mudstones. The lithotope consists
solely of the rock type described above (p. 77) under
the name "laminated clay." Occurrence is limited to the
area bordering the old Sage Ridge (see Figs. 32 and 33)
and to the district southeast of Cuny Table, which has
not yet been thoroughly studied.

In the Sage Ridge area, laminated mudstones occur
at the top of the Lower Nodular Zone, resting directly
upon heterogeneous mudstones, and also intercalated
with laminated siltstones at various horizons through
the Scenic Member. They range in color from bright red
to flesh colors and pale buff. The color distribution is
geographic rather than stratigraphic: if at a particular
place one laminated mudstone is bright red, then all

CLARK: PALEOGEOGRAPHY OF THE SCENIC MEMBER

85

Fig. 38. Photograph of layered concretions in a silty mudstone, showing vertical plates of unknown origin.

laminated mudstones throughout the section are bright
red at that place and even the heterogeneous mudstones
are reddish brown. The paleogeographic significance of
this lithotope is considerable; more complete discussion
will be given in the following two sections, after the
presentation of more evidence.

4. Channel-fill sandstones. This lithotope also in-
cludes only the rock-type of the same name. The first
notable feature of these channel fills is their vertical
continuity, indicating that any one Oligocene stream
remained in one place through the deposition of several
tens of feet of sediment both in its channel and on the
surrounding plains. Meanders, cut-offs, and ox-bow
lakes were nowhere present.

The second feature is the rarity of any structures
which could be interpreted as braiding in the true sense
(Leopold and Wolman, 1957). Actual division of a
channel by a stream-built island, with reunion at the
downstream end, was observed in only one doubtful
case (NE % of Sec. 11, T. 42N., R. 45W., Shannon Co.).

Third is the apparent rapid decrease in stream velo-
city, represented by the decrease in maximum grain
size, with increasing distance from the Black Hills. All
of the major Southern-derived streams, whose courses
were 25-40 miles long from the edge of the Precambrian
to their present area of outcrop, carried pebbles 2 in. or

more in diameter. The shortest Northern-derived stream
(outcropping on Heck Table, SE 34 of Sec. 35, T. 3S.,
R. 13E., Pennington Co.), with a course from its last
Precambrian contact of about 50 miles, carried feldspar
and quartz pebbles not over 1 in. in diameter, with flat
disks and spindles of black quartz-schist up to 1 in.
maximum diameter. All of the channel deposits farther
east, with necessarily longer courses, contain no grains
larger than coarse sand.

Finally, the channel deposits are remarkable for the
lack of evidence of any erosive action. Nowhere are
there signs of ancient cutbanks, or even commonly of
cut-and-fill structure within the channel itself. The
streams must have operated either continuously under
a depositional regimen or, more probably, under a regi-
men of deposition during high water and graded flow
during low. Probably the excessive proportion of mont-
morillonite present on the depositional plains, as well
as in the channel-ways at all times, caused the flood
waters to resemble thick soup; at low-water stages the
water was probably more normal.

STRATIGRAPHIC RELATIONS OF
SEDIMENTARY LITHOTOPES

A. Method of Determination.

86

FIELDIANA: GEOLOGY MEMOIRS, VOLUME 5

The relationships of the sedimentary lithotopes were
determined by detailed field observations, with meas-
ured sections spaced an average of one mile apart.
Individual strata were followed from one butte to the
next by direct visual observation; levelling by instru-
ment was unsatisfactory because the separate beds were
only a few feet thick, varied in thickness locally, and
were everywhere gently warped by low structures.
However, the excellent badlands outcrops plus the large
number of key beds made visual tracing effective. Direc-
tion of flow within channel-fills was determined by
mapping the channel-fills and observing cross-bedding
within them. Source was established by heavy-mineral
analysis.
B. Stratigraphy of Mudstone Zones in Detail.

Over the entire area except in the immediate neigh-
borhood of channel-fills the Scenic member consists of
alternate beds of mudstone and laminated siltstones.
The contacts between these are sharp. Near channel-
fills, the siltstones become sandier and some of the
mudstones grade laterally into siltstones. Channel fills
of Northern Black Hills derivation everywhere grade
laterally into the surrounding sediments without scour-
ing or trenching them. Those of Southern Black Hills
derivation have sharper boundaries and have en-
trenched themselves in a few places to an absolute
maximum of 6 ft.

At the standard section of the Scenic Member (south
edge of Sec. 27, T. 4S., R. 13E.), a total of five mud-
stone strata, each overlain by a siltstone, can be recog-
nized. One-half mile further south and east, the top of
the uppermost siltstone grades into a sixth mudstone,
separated from the overlying Poleslide by a dark band.
Since the standard section is very near a series of chan-
nel-fills, gradation of this sort is to be expected here.

The mudstone zones were numbered I-VI, starting
with the Lower Nodular Zone at the bottom as I, and
effort was made to trace them as far as possible. If they
should be restricted lenses related to particular channel-
fills, they would be of no geologic significance. If, on the
other hand, they should extend across the flood plains
of several Oligocene streams, they would represent
times of flooding by waters less competent than those
which deposited the siltstones from large areas of the
Black Hills and would carry considerable climatological
significance. It might also be possible, if these layers
were continuous, to establish a faunal succession within
the Scenic Member.

Careful tracing revealed (see Figs. 30, 33, and 39,
and Appendix) that Sheep Mountain was during Oligo-
cene time the location of the major divide between
Northern-derived sediments to the northeast and South-
ern-derived sediments to the southwest. The standard
section of the Scenic Member lies near a Northern-
derived channel zone, three miles northeast of the Oligo-
cene divide.

Section #1, well within the Southern-derived deposi-
tional area, revealed mudstones I, III, IV, and com-
bined V and VI clearly evident. All of these were traced

directly from outcrop to outcrop between the two posi-
tions. The siltstone separating V and VI at the standard
section can be observed to grade laterally southwest-
ward into a mudstone, which explains the union of
these two, south westward. Mudstone II is probably
present, but I am unable to distinguish it from other
entirely local mudstone laminae within the siltstone
zone which separates I and III.

It is thus apparent that, with minor changes, these
mudstone zones do extend from one Oligocene drainage
area to another.

Northeastward, the continuity of certain mudstone
zones is even more striking.

Mudstone I, the paleontologically famous Lower
Nodular Zone, has long been recognized as continuous
throughout the Big Badlands (Wanless, 1922, 1923; Sin-
clair, 1921, 1924, et al.). The nodules gradually disap-
pear two miles northeast of Chamberlain Pass, and the
color of the mudstone changes from buff and gray to red
(Sec. 15, 16, 17, 18, T. 3S., R. 14E.) (Fig. 39). A thin,
discontinuous greenish layer, apparently a small swamp
deposit, occurs two-thirds of the way toward the top of
the zone at this place. The red color fades to more nor-
mal buff, yellowish, and gray eastward in Sections 22
and 23 (Figs. 30, 39) . The red reappears in the neighbor-
hood of Sage Creek Pass (Sec. 11, T. 3S., R. 15E.); it
continues thence northward to outcrops near the town
of Wall, and eastward, with few interruptions, through-
out the Monument. The red zone is, therefore, definitely
divided into two areas (Fig. 33, "Igaposomes").

Red color occurs in zones II, III, and in any local
mudstone lenses which may be present, over exactly the
same areas as those described for zone I. Where one is
red, all are red. Zones V and VI, where recognizable,
show little if any red color. Nodular concretions never
occur in the red zones. However, in the neighborhood
of the Pinnacles a thin buff lamina within the red, a few
feet below the top of zone I, is marked by both nodular
concretions and abundant fossils.

Zone II. This mudstone is much less continuous than
is zone I. It can be traced from the northeast flank of
Sheep Mountain east-northeastward for 63^ miles, with-
in the floodplains of two Northern-derived Oligocene
streams, but beyond the second stream it disappears.

As zone II disappears eastward another mudstone
zone (designated IIA in the columnar sections, Fig. 39,
and Appendix) lenses into the section a few feet above it.
This has been traced all the way to the Pinnacles, a
distance of 14 miles, and may extend farther. It is never
more than 12 ft. thick, and usually less than 6. Since
zones II and IIA have been observed in the same hillside
(see columnar section 14 and appendix), it is evident
that they are two strata of slightly different age and of
less areal extent than mudstones I and III.

Zone III is fully as continuous as zone I; it varies
generally from 10 to 20 ft. thick, occasionally thinning
to 7 or 8 ft. This zone shows a strong tendency to de-
velop a middle stratum different from the top and bot-
tom. In the Sheep Mountain area the middle is either

COLUMNAR SECTIONS

M lid Itont ( C| )
Sandstone (St)

S ill > tone ( 5|, )

Fig. 39A.||Columnar sections, Scenic Member.

87

COLUMNAR SECTIONS

13

1 2

POLESLI DE 14

POLESLIDE
VI
VI H20

GC.B VI

16 POLESLIDE
22

1 9
POLESLIDE
VI ■

GC

POLESLIDE

GC

+

vc

GS.

GrS,

GS,

I

8 V

16
POLESLIDE

VI F18

GC,

+

I

POLESLIDE

I

GrS
IV

GC,|

3 GS„

GrS,

18 GS,

GrS,

GrS,
YGC,

YC

I

GS,

1 T
POLESLIDE

S,S,

■

1 8

20

POLESLIDE
25

G-6rS|t

V

GrS,t_

GS„
VC,

110 GrS,t
YGC, I
16 v

GrC,« .

GS,, POLESLIDE V I

s - vc, B V

2S G-YC

8 SS

+

Gr S|,

"L
ve+c"i

YC,

16 Y,-l

GrS,

26 in |i3 ,,, ■-,-. ,„ ■ „, ma in ■:- ,,, |16 „, ■ ,,, I

GrS|,
Ss

I

"oc,M

RBC

RBC,

I

■ 5

V6C,

I

GS

■ 6

RC. ■

GS,

GS,
Mo
RC,

R YC|

TGC,

12

G-GrS,

GS.

I

I

B-RC,

SC,

e-ers.

[

I

y -ec

I "<• I

BGC,

I ■ 18

GS,

YC,

| I

I

YC,

+
GS,

29 ||

R-GC|

GS,,

I

YBC|

G-GrS,t
3
26

I

T

no ms

5
26

-r

G - Gr5|

Gr-GS,

r

e-Grs,

VGC|

SrS,

I ■ 37

I U25

L IGIND

Seal * : 1" : 10

^HbV Mudsfone C ,
' I Sond.t. «.($,)

I I Slltilmt IS|tl

Fig. 39B. Columnar sections, Scenic Member.

88

COLUMNAR SECTIONS

23

POLESLIDE
25

26

POLE SLIDE
20

23

POLESLIDE OSl

|20 2*

POLESLIOE
20

22

21 POLESLIOE

OS!

+

POLESLIOE

OS,,

15

V ■

vc,

3C.

i

v ■

R C,

I

I

v El

vc,

I

■ GS,

I

27
POLESLIDE
20

29

POLE SLIDE
39

2S

POLE SLIDE
35

V ■

oc.

RC

♦

OS

I

I

III YC,| 2 Ml ■'« HI J3 m

20 « vcl

OS,,

II A ■

SC.I

^ 42

OS,,

OrS,

II A H12

QC, ■ OC

= '|a"*'"aC,|

II A ■

VC,

II A ■

v-ec,

I

Rcli2 RC, I

10

HA «2 "H

11 RC' lO

8 OS,

*c,l RC|

10

8,s, :

os,t y

RC, I

■f

MA ■ RC

1a-R-stc

I

OS

1 I

8

25

IA" RC,

OS,,

I

1 1 I

10 RC|B

5
1T

IA ■ 2

RC'^15

O- BC,

I ■

RC

+
G C

RC|

LEGEND

ScaliM": 10

Muditen* ( C , )
I I Sandstone (S»>

I I Slltiton. (Slt)

Fig. 39C. Columnar sections, Scenic Member.

89

COLUMNAR SECTIONS

30
POLESLIDE

POLESLIDE

30 3 2

POLESLIDE

15

ers,

T

15

POLESLIDE

I

i We in |io in Tp i

GcJ SC,I

a-src,|

15

POLESLIDE 36

POL ES LIDE

I'

|8 Ml | III |10 III I' III |10 III | 15 III

T

!C, ■

19

RC, M1

SC,|2 GS„

16

,i:

GS,

1

16

6 RC,

GC

GrSs

14 RC,

16 ,+ _RClfc

SrCl

src.

|8 6S„

II A

es,t

RC,

es(,

RC, I

II A |8

GC,I GC,

ISC,
8 , '

11 scl

'5
'12

2
13

17 RC,
GC,

GS,t

I

BC|

GC|

LIOEND

Seals : 1": 10'

| Mudltone (C|)
i I Sandstone ( St )
I I Slltlton* <S„>

Fig. 39D. Columnar sections, Scenic Member.

90

CLARK: PALEOGEOGRAPHY OF THE SCENIC MEMBER

91

Fig. 40. Photograph of Scenic Member mudstones, looking NW at the SEM of Sec. 28, T 2S, R 16E.
Section 33, Figure 39. Shows numbered mudstones, with red igaposomes.

This is the location of Columnar

white or gray, and silty; farther east it is a grayish or
buff layer between more reddish ones.

Mudstone IV thins and merges with V about two
miles northeast of Chamberlain Pass, a distance of
about 12 miles from Sheep Mountain. Its eastern distri-
bution almost corresponds to that of II; south westward
it extends for about eight miles.

Mudstone V is almost as continuous as I and III. It
merges with VI on the west side of Sheep Mountain
(Section 1, Appendix) and (with VI) merges with the
Poleslide east of Sage Creek Pass, 18 miles to the east-
northeast. This zone, like I and III, continues recog-
nizably across several Oligocene stream courses (Fig. 33)
and therefore cannot be genetically related to any one
of them. It loses identity by the thinning out of the
siltstones which separate it from the overlying VI and
Poleslide, rather than by wedging out itself. East of the
area where V can be recognized, the basal 20 to 30 ft. of
what is apparently Poleslide contains obscure, discon-
tinuous concretionary bands and thin siltstones, which
probably represent the upper boundaries of V and VI.
However, these rocks resemble Poleslide lithology,
without a marker zone to separate the two members.
Away from the main channel zones, apparently, condi-
tions of deposition of the upper Scenic Member and the
Poleslide Member were so similar that the resulting
mudstones are indistinguishable: the apparent bound-
ary between the two members is the top of the laminated
siltstone lithotope overlying Zone III. This means that
the apparent boundary between Scenic and Poleslide is
somewhat older from Sage Creek Pass eastward than
from the Pass southwestward.

North of Chamberlain Pass and southwest of Sage
Creek Pass, V is yellowish and contains fossiliferous
nodules indistinguishable from those of I, the Lower
Nodular Zone. Ischyromys, Eumys, Paleologus, Ictops,
Mesohippus, and Merycoidodon have been collected,
also hackberry seeds, snails, and coprolites. Study re-

veals no differences between these and specimens from
the Lower Nodular Zone.

Mudstone VI merges indistinguishably with V, west
of Sheep Mountain, and with the Poleslide approximate-
ly 14 miles to the northeast. It also merges with the
Poleslide northward, in Sec. 12, T. 3S., R. 13E. The
dark marker band which separates Scenic from Poleslide
at the standard section disappears east of Sec. 18, T.
3S., R. 14E., two miles northeast of Chamberlain Pass.

From the vicinity of Sage Creek Pass eastward, a
number of bright cinnamon-red, laminated mudstone
zones lens in and out of the section. These are never
continuous for distances of more than a mile or two,
and they differ markedly in appearance from the hetero-
geneous mudstones numbered I, Ha, III (IV disap-
peared farther west, and V and VI have here merged
with the Poleslide). The latter are always more brownish
or chocolate-red, and show some variability in color
from bottom to top. These discontinuous lenses are
thinner, bright red, and show no variation from top to
bottom of an individual bed. They always consist of
laminated clays, never of heterogeneous mudstone.
They wedge out from both north and south against the
flanks of the Sage anticline at Dillon Pass (Fig. 40).

The underlying Chadron covers the top of the Sage
Anticline but thins to 9 ft. in doing so (Clark, 1937, p.
264). It is believed that these reds represent local wash
of red Interior zone soil from areas immediately to the
northwest which may still have been exposed during
Brule time. The top of the Interior zone is a very bright
red in this neighborhood. However, it must be under-
stood that this suggestion is hypothetical, because the
Scenic Member cannot anywhere be observed in con-
tact with Interior Zone. This situation could occur north
of Sage Creek basin, where the Oligocene is everywhere
either covered or eroded away, but it is not known to do
so.

92

FIELDIANA: GEOLOGY MEMOIRS, VOLUME 5

In summary, from the Sheep Mountain area north-
eastward the Scenic Member comprises three mudstone
zones, I, III, and V, which apparently were deposited
generally over the whole area, although V merges east-
ward with the overlying Poleslide Member. It also
includes three mudstone members, II, IV, and VI, which
are local and related to not more than three adjacent
channel fill zones.

To the southwest (see Figs. 30, 33, 41) the situation
is somewhat different. As the cross-section diagram
indicates, the siltstone sequence between mudstones I
and III thins south westward. At the Cottonwood Pass
area it is represented only by major channel fills. Con-
comitantly, zone III thins, the nodules within it disap-
pear, and it changes to a brownish color. It appears to
be simply a brownish clay zone a few feet thick, forming
the top of the Lower Nodular Zone (I), separated from
more typical nodular clays of that zone by a few inches
to several feet of greenish clay or by channel-fills.

Meanwhile, the Lower Nodular Zone thickens to
40 ft. at Cottonwood Pass, then gradually thins west-
ward.

Zone IV, which is not persistent eastward, can be
traced westward through most of the area of major
Southern-derived channel-fills (Sec. 16 and 17, T. 42N.,
R. 45W., Shannon Co.). It apparently represents some
situation general over the part of the area of the south-
ern and middle Black Hills, but not the northern Hills.

The characteristic yellow-buff of zone V, with a
thin, relatively dark gray zone at the top, is a good
stratigraphic marker which continues southwestward
somewhat beyond the main area of channel-fills.

Westward along the main flank of Cuny Table, (from
Sec. 32, T. 42N., R. 45W., to Sec. 9, T. 41N., R. 46W.,
Shannon Co.) the entire Scenic Member thins and
changes. Zone I becomes pale tan, and the nodules
grow smaller, fewer, and less well-developed. In many
places it is difficult to distinguish between Chadron and
Brule. A mass of laminated siltstones, some slightly
cross-bedded, replace zones IV and V. The entire section
of the Scenic Member becomes:

Poleslide — yellow-buff

Gray siltstone and gray platy sandstone ... 40'

Zones I J Yellow to brown clay 20'

and III 1 Pale tan clay with small brown concretions . . 30'

Chadron

Still farther southwest, in Sec. 31, T. 41N., R. 46 W.,
the sections thins more, and the sediment becomes even
finer grained:

Poleslide

Pale grayish silty clay 15'

Yellow to brown clay 10'

Chadron

There are no channel-fills, and no laminated silt-
stones in the Scenic Member over the entire area south
to the Slim Butte (T. 36N., R. 48W., Shannon Co.).
Apparently this area received only fine sediment during
times of maximum flooding from the streams to the

north. Much of the fine material is montmorillonite,
which may represent locally weathered ash merely re-
handled by occasional flood waters.

C. Siltstones and Sandstones.

As already indicated, the channel-fill sandstones lie
in definite restricted areas (see Fig. 33) extending ver-
tically almost or entirely through the Scenic Member.
Individual channel-fills usually do not have definite
boundaries, but grade laterally into laminated siltstones.
This makes determination of the size of the depositing
streams at any one time quite difficult. In general, the
sandstones are widest where they lie within siltstone
zones; the sandstone zones are 100 to 500 yards wide at
these horizons, with individual channel-fills a maximum
of 100 yards wide by 6 ft. thick. The sandstones are
notably restricted where they occur within the mud-
stone strata, especially within I, the Lower Nodular
Zone. There they are less than 50 ft. wide, but 4 to 6
ft. thick. The diminution within zone III is almost as
great, but is impossible to measure accurately. Very few
channel -fills continue into V. Mudstones II and IV, on
the other hand, are generally interrupted by the channel
zones so that tracing them across is very difficult.

The sandstone channel-fills likewise show a decrease
in maximum grain size within the mudstone strata. No
sediments coarser than fine sand have been observed
within the Lower Nodular (mudstone I), in the same
channel-fill series which carry pebbles up to 10 mm.
diameter in their siltstone-stratum levels. Briefly, the
mudstone strata are associated with channel-fills of fine
sand, and the siltstone strata with channel-fills of grit.

The siltstone strata, which occur intercalated be-
tween the numbered mudstones, have definite top and
bottom contacts with those mudstones. They grow
noticeably finer, more even-grained, less laminated, and
less concretionary, away from the channel-fills. Near
the channel-fills, they consist of alternate thin laminae
of mudstone with thicker layers of siltstone or fine
sandstone.

Two miles north of Chamberlain Pass, at a few
places in Sage Creek basin, and north of the Pinnacles,
the siltstones grade laterally into gray to buff mud-
stones. At these places, the numbered mudstones are
red, and the contacts between successive layers of
differently colored mudstones remain sharp.

PALEOGEOGRAPHIC INTERPRETATION OF
THE SCENIC MEMBER

A. Areal Distribution of Physiographic Units.
The Oligocene strata of South Dakota and Nebraska
have been recognized as flood-plain deposits for several
decades. Their thickness, their wide areal extent, the
fact that they transgress divides, and their close prox-
imity to their source mountains, indicate conditions
very different from those normal to flood plains. From
the data presented above, interpretation of some of
these special conditions is possible.

NE-SW CROSS-SECTIONAL CORRELATION OF THE MUDSTONES
IN THE SCENIC MEMBER.

NI CUNT TAIL! SW INDIAN C«im

N W CUNT TAIlf

POLE SLI DE

SW CUNT TABU

MU DS TONE

U LAMINATED SEDIMENTS

I _ i VEHTICAL SCALE

HORIZONTAL S C A L E - APPROXI M AT E ONLT - 1" : 3 M I L E S

THE TOP OF ZONE III HAS IEEN CHOSEN AS DATUM

■ ECAUSE IT IS THE ONLY HORIZON EVERYWHERE

RECOCNIZAtlE

DILLON PASS AREA

IS

34 36

POLE S L I D E
L I THOLOGT

Fig. 41. NE - SW cross-section of the Scenic Member, showing relationships of mudstone units.

93

94

FIELDIANA: GEOLOGY MEMOIRS, VOLUME 5

CROSS - S E C T I O N OF A"WHITE WATER" STREAM VALLEY
IN THE AMAZON BASIN.

MARGEM MARSEM

3E

ESTAVEL BARRANCO

TEI
F 1 F

J ME

MATO
CI LI AR

MATO
CI L 1 AR

'

r

LE1TO

\ PO

>

'

DO RIO

' L ASO

C AM PO /f"

14

: _ ^

^vt^SA

i^j-.

^WW'****

TERRA
Fl RME

I G A PO

C A M PO

*S*lt.

L A G O

NIVEL MAXIMO DA ENCHENTE
NIVEL MINIMO DA VASANTE

From Siol i ,1951 (a) , p. 13.

Fig. 42. Cross-sectional diagram of a stream in a depositional regimen, from Sioli.

In an effort to discover a recent situation similar
enough for fruitful comparison, I turned first to the
Punjab of Pakistan. Observation from the air of the
major floods of 1951 revealed several roughly-parallel
rivers in flood simultaneously, producing a sheet of still
to gently-moving, muddy water about 40 miles wide,
with swift, turbulent currents marking the courses of
the main streams. The flood water was generally less
than 3 ft. deep over the entire area. Complete with-
drawal to the original channels occurred within two
weeks; none of the flooding streams changed their
courses and, except where irrigation works created arti-
ficial situations, there was no significant erosion.

Unfortunately, I was unable to make ground obser-
vations immediately after the flood. News reports spoke
of several feet of mud partially burying riverbank vil-
lages, but since the villages were all constructed of
adobe, much of this mud probably derived locally from
crumbling walls. Judging from the appearance of vege-
tation several weeks later, 1 to 12 in. of sedimentation
would be a better estimate.

Garden and cereal crops were widely destroyed, but
trees, shrubs, and taller herbaceous plants suffered only
minor damage. No eye-witness reports of stranded fish
came to my attention. This might have been due to
stranded fish being locally considered un-newsworthy,
but three possible biotic factors seem to offer more
satisfactory explanations. First, the normal fish fauna
of any one stream would, if it were evenly distributed
by the flood, be areally diluted to about .01 to .05% of
its channel concentration. Second, the floodwater lake
was so shallow and so muddy that most fish would
probably have attempted to remain in the turbulent,

better-oxygenated water of the main stream. Third,
any fish trapped and killed by the lake subsidence would
certainly have decayed or been devoured to destruction
within a few days, under the Punjab summer climate,
and carnivore and insect population (Payne, 1965).

The findings having significant reference to Oligo-
cene sedimentation were that parallel streams on a
confluent plain could flood contemporaneously, pro-
ducing a lake tens of miles wide but very shallow, which
would disappear within a matter of weeks, and that the
flooding streams neither meandered nor changed their
courses.

Much more detailed observations on the central
Amazon have been reported by Sioli (1951). His cross-
section diagram of the parts of a stream system in a
depositional regimen (Fig. 42) divides the area into
definite physiographic provinces; as amplified in his
text, several of these units seem to be represented by
Scenic Member sediments. Figure 43 shows the physio-
graphic provinces as parts of a surface, with the sedi-
ments deposited in each province continuous strati-
graphically below it, as they occur within the Scenic
Member.

There is first of all the through-going stream with its
provenance outside the area, which functions as an
avenue of entry for the bulk of both the water and the
sediments that enter the entire province; Sioli refers to
this as the "Rio." Since another type of stream occurs in
the fluvial province, it seems wise to employ the Portu-
guese term in a technical sense, as defined in the pre-
ceding sentence. The body of sediments deposited by
the rio within its bed through time, structurally con-
trolled by water moving downstream, is here designated

NOUVrthOj

S<JH¥MS

MOUVWHOj
dOOHO

M39W3« OIN33S

NOHtfMbOd NO**Q»MD_

UIH*

a> o

o —

I

■

*

.«

. 5 ,

! '

>

' If)

(J)

o

Hi

► 8

a x

u o

5*

~

Em
CD —

3100IW B3W01

31IH*

NOUVnoOJ N0H0VH3

CM
O)

SC139 ST783D0l0Md -VIN3H3nvid3T

?l! si?.

5*'S

n

>! Hi

¥

mniaSMiONvin

NOUVWHOj 0nG3SOH

to
"o

S038 S5M30010UJ

~ O
01

« S

w Q
* P.

O aj

a o

hOiiv«aOj

5

i z (c

81 g

U 4

5 J

s1

S039 SVb300108d ^ SAV10 N3UUVS ^

saja Nooosao

g £

o„

S038

3l'MW

mniM 3MioN»iii

CO
"O If)

§5

jc C
« V

5 o

I

? in
<u CO

o in

3iVM«a3l3GNON

C OJ

o in
> CO

;i!

31VNIMH3130NON

COMPONENTS OF AN ENVIRONMENT OF
FLUVIAL SEDIMENTATION

D R E N A J E

1GAPO WITH LAGO

OROERLANO

DREN AJE SOME

RIOSOME VARZEASOME I6APOSOME TERRA FIRME

Fig. 43. Components of an environment of fluvial sedimentation.

95

96

FIELDIANA: GEOLOGY MEMOIRS, VOLUME 5

"riosome." A riosome differs from a channel-fill physi-
cally in that it does not represent the fill of a single chan-
nel or depression which ever existed as one entity. It is
rather the body of sediment deposited within one stream
throughout a period during which the stream operated
in a dominantly depositional regimen.

Sioli's second fluvial sub-province is named "varzea."
This, Sioli makes plain, includes the relatively high
ground (natural levee) along the rio banks, and the
entire back slope away from the banks. This area,
which is exposed except during floods, receives its sedi-
ment primarily by outward flow of flood water from the
rio. The mass of sediments accumulated through time in
a varzea should be referred to as a "varzeasome." They
are composed primarily of allochthonous material, but
may be deposited either as mudflows or by settling from
suspension in slow-moving to temporarily still water.
The heterogeneous mudstone and laminated siltstone
lithotopes of the Scenic Member constitute a varzea-
some.

Finally, Sioli recognizes valley walls, normally com-
posed of rock different from that of the stream's pro-
venance, which he calls "terra firme." Where the varzea
back-slope meets the foot of the terra firme, a linear
hollow is produced, which he terms "campo" when dry,
"lago" if occupied by a lake, and "igapo" if occupied by
a slow-moving stream composed primarily of run-off
from the bordering terra firme. For geologic purposes,
the term "campo" may be omitted, and "igapo" used
to mean the shallow trough, intermittently occupied by
still or moving water, which lies at the foot of the terra
firme. The sediments accumulated in an igapo environ-
ment come primarily by wash from the terra firme; they
should be called "igaposome." In the Scenic Member,
the laminated mudstone lithotope, derived from neigh-
boring hillocks of Cretaceous shale rather than from
the Black Hills, is an igaposome.

Sioli was dealing with one very large river in a wide
valley. In the Scenic Member we have at least seven
rios (#2, 3, 5, 7, 8, 11, Fig. 33) on a confluent flood plain,
not separated by terra firme ridges. The varzea back-
slopes of any two adjacent rios must have met in a
winding, relatively linear depression. Floodwaters mov-
ing down the varzea slopes would deposit much of
their load on those slopes, and meet at the intervening
depressions (see Fig. 43) to form first a series of ponds,
then a slow-moving, temporary stream with no head-
waters. These streams should be called "drenaje"
( = "drain"; I have used a Portugese word in order to
remain consistent with Sioli's terminology), and the
body of sediments deposited by such a stream through
time a "drenajesome."

Fortunately, a recent, small-scale development of a
drenaje system can be observed in the Badlands. In
Sections 4, 9, and 16, T. 42N., R. 45W., Shannon Co.,
a series of gullies debouch from the north flank of Cuny
Table onto flats which have not yet been incised by the
present erosional activity of Big Corral Draw. During
major rainy spells these gullies, heavily charged with

Oligocene mud and concretions, cross the flats as a
series of overloaded rio streams at relatively low gradi-
ents. Their courses are almost straight, with flat-bot-
tomed channels 2-6 ft. wide and up to 6 in. deep,
flanked by low natural levees and long varzea back-
slopes. The streams overflow as sheet floods, breaching
their banks first along the straightest stretches. Maxi-
mum deposition takes place at the curves. Break-
through channels, heading from the main rio course to
the drenaje depressions, never develop. Such break-
throughs would inevitably produce a sharp change in
direction; the resulting energy loss in a stream already
heavily overloaded makes impossible the erosion neces-
sary to cut a breakthrough channel. Sheet-overflow
remains the only way in which excess water can escape
from the rio channel.

The overflow water, still charged with sediment, first
forms ponds and then true drenaje streams, as already
described. The drenaje comprise the apparatus neces-
sary for sedimentation in the inter-varzea hollows. The
recent small-scale rios and drenaje very quickly dis-
charge into eroding gullies of Corral Draw. Without
such a downstream drainage, they would presumably
continue as separate entities until they reached some
change of gradient.

The channel deposits numbered 1, 4, 6, and possibly
9 and 10 are drenajesomes. They are notably finer-
grained and smaller than the adjacent riosomes. Drena-
jesome #4a rests directly upon a pond limestone with
which it is almost coextensive, for a distance of over a
mile (Sec. 35 and 36, T. 43N., R. 45W., and north edge
Sec. 2, T. 42N., R. 45W., Shannon Co.). Apparently it
started as a "lago" (pond formed by sheet-flood waters
in either a drenaje hollow or an igapo hollow), and with
increased water supply became a drenaje. Drenajesome
#4 illustrates the rambling, relatively unpredictable
course resulting from the unrelated depositional slopes
of neighboring varzeas.

The nature of channel deposits #9 and #10 is doubt-
ful. They are composed of notably finer sediment than
that in the neighboring riosomes, #7, 8, and #11, and
are much less cross-bedded, which suggests that they
are drenajesomes. However, they are fully as wide and
large as riosomes #8 and #11, which is improbable for
drenajesomes.

In summary, the Scenic Member comprises deposits
laid down on a confluent flood plain, in an orderly sys-
tem of physiographic subprovinces designated as rio,
varzea, igapo, and drenaje. The sediments were derived
from three sources: (1) The Black Hills; (2) neighboring
outcrops of weathered Cretaceous marine shale; (3) air-
borne volcanic ash and its weathering products.

Control of the regimen of these streams was not
structural, since overlap of both Chadron and Brule
sediments into the Black Hills shows that no downwarp
of the Badlands relative to the Black Hills occurred at
this time. Depositional activity by all Brulean streams
from Canada to Colorado further precludes local struc-
tural control.

CLARK: PALEOGEOGRAPHY OF THE SCENIC MEMBER

97

Excessive deposition of volcanic ash could not have
functioned as a control of the entire regimen. Streams
temporarily overloaded by an ash-fall would immediate-
ly have entrenched themselves during times between
falls, and no such entrenching occurred. The existence of
a depositional regimen must have been determined by
climatic factors; the nature of the deposition and the
sedimentary structures produced were strongly influ-
enced by the volcanic materials.

If the temperature control of climatic patterns sug-
gested on page 72 is correct, the dry episodes of het-
erogeneous mudstone deposition and restricted stream
flow represent cooler periods, and the times of increased
precipitation indicated by the laminated siltstones
represent warmer periods.

Surprising evidence in support of this hypothesis was
discovered by a South Dakota School of Mines collector
in 1965. He found an alligator skeleton within the lami-
nated siltstones between the Lower Nodular (Zone I)
and Zone III, in the area near riosome #3, Fig. 33.
(NWM of SEJ4, Sec. 13, T. 42N., R. 45W., Shannon
Co.). This is the first alligator reported from the Brule
Formation, anywhere in its area of outcrop. The com-
pleteness of the specimen rules out any possibility of its
being re-worked from the Chadron. Furthermore, this
is the first fossil vertebrate reported from the laminated
siltstones.

The enormous amount of collecting from the under-
lying Lower Nodular Zone leaves little doubt that, had
alligators been present during Lower Nodular time,
they would have been found. It is therefore a reasonable
presumption that: (1) alligators were present in South
Dakota during Chadronian time but uncommon during
latest Chadronian (known); (2) alligators were absent
during Lower Nodular time; (3) at least one individual
returned during the next succeeding warmer and more
humid episode.

Since streams continued to exist throughout the
time, the temporary withdrawal of alligators and their
even more episodic return must be ascribed to some
cause other than lack of waterways. The most logical
assumption is that: (1) late Chadronian time was cool
enough that only a few small alligators survived ; (2) fur-
ther cooling which initiated the Lower Nodular deposi-
tion caused a southward retreat of the north border of
their range; (3) temporary warming, with associated in-
crease in precipitation, permitted at least one venture-
some individual to repenetrate northward as far as the
Badlands.

This bit of paleontologic evidence gains credibility
as it tends to reinforce conclusions already drawn from
the interpretation of the sedimentary history. We have
two quite independent lines of evidence (albeit one rests
upon a single fossil) pointing in the same direction.

B. Developmental History of Scenic Member.

At the beginning of Scenic Member time the entire
area was a depositional plain composed of Chadron
fluvial sediments which were thick over the old Red

River Valley (see p. 22), but very thin over the Sage
Ridge and the Pine Hills. It is probable that deeply
weathered Pierre shale was exposed in places along these
bordering uplands, although no such places have been
found.

At least six major streams (#2, 3, 5, 7, 8, 11; Fig. 33)
flowed in a general southeasterly direction across this
plain; two others (#9, 10; Fig. 33) may have done so,
although these were possibly merely drenajes. The
southwesternmost three streams had their headwaters
in the Harney granite; the others arose in the meta-
morphics and Laramide intrusives of the northern
Black Hills.

It is possible that streams #2, 3, and 5 represent,
respectively, the courses of ancestral French Creek,
Battle Creek, and Spring Creek, while #7, 8, and 11
were Rapid Creek, Boxelder Creek, and Elk Creek re-
spectively. This suggestion stems from the fact that
all of the major creeks flowing from the Black Hills
today pass through large gaps in the Cretaceous hog-
backs, most or all of which are structurally controlled.
Oligocene sediments lap up to and, in some cases in the
southern Hills, into these gaps, indicating plainly that
the present topography along the front is essentially
an exhumed Oligocene landscape. No evidence of ma-
jor post-Oligocene changes in the drainage pattern of
the Black Hills has been reported. Association of
Oligocene streams with their possible Recent descend-
ants is justifiable in the case of the Southern-derived
streams, where riosomes can be distinguished from
drenajesomes. However, due to question about the na-
ture of channel deposits #9 and 10, the relationship of
Northern-derived streams to Recent ones is not at all
clear. (Pleistocene piracy of all Black Hills drainage by
Cheyenne and Belle Fourche rivers has necessarily
curtailed the courses of all Recent streams, outside the
Black Hills).

The Oligocene streams underwent five periods of
alternate restriction and expansion, followed by a sixth
restriction.

During the first period (that of the Lower Nodular
Zone) the individual streams were less than 100 ft. wide,
generally less than 3 and everywhere less than 6 ft.
deep, and were capable of carrying only fine sand and
silt. In successive floods, they spread a thick layer (15-
40 ft.) of ashy, slightly bentonitic mud over the whole
area.

This mud was added in increments up to at least 18
in. thick, since fossil rhinoceros (Subhyracodon) and
giant pig (Archaeotherium) skulls and one mass of
Hypertragulus skeletons of that vertical depth were
buried quickly enough to preserve them entirely. Mod-
ern bones, if partially buried, show damage or destruc-
tion of the exposed parts within one or two years. Bone
flakes and fractured bones are common in the Lower
Nodular zone, but they are broken from bottom to top,
not better preserved on the bottom as they would be if
partially buried. The vertical thickness of the thickest
well-preserved fossil bone, therefore, is a safe measure

98

FIELDIANA: GEOLOGY MEMOIRS, VOLUME 5

of the thickness of the individual sedimentary increment
which buried it.

Fractured bones, abundant coprolites, hackberry
seeds, and occasional rodent-gnawed bones all demon-
strate that there were many episodes of non-deposition.
During these times an abundant fauna roamed the
plains, which must have been well-vegetated to support
them. Intermittent floods deposited mud in layers up to
18 in. thick, burying both the more recent dead and
bones which had partially disintegrated. The flaking and
splintering type of disintegration which the bones under-
went before burial is typical of bone under temperate
semi-arid to arid climates. So also are the pale tan and
gray colors of the sediments.

When flooding did occur, water from the overloaded
streams either reached the area charged with mont-
morillonite, or rapidly picked up a charge as it swept
out of the channelways, or more probably derived its
load of gelatinous material in both ways. In any case,
the water became a viscous mass, loaded with sand and
silt particles and chips of clay incorporated from the
dry surface across which it was advancing. The mass
advanced by rolling and engulfment, almost in the man-
ner of a lava flow rather than of a more normal flood.
Almost no horizontal movement of engulfed bones was
accomplished : they were simply plastered to the surface
on which they had lain.

In most cases, floods were not attended by much if
any rain in the area of deposition. The angular chips of
surface clay which the advancing flows picked up
would certainly have softened to rounded masses had
they been wetted before incorporation. The igapo area
near Sage Ridge remained free of standing water most
of the time, and only enough "Interior" red soil washed
off the terra firme of Sage Ridge to stain the viscous
sediments of the area slightly.

Following this period came a time of greatly in-
creased stream activity. The rivers tripled in volume
during low water; their channels widened to a low-water
maximum of 150 yards; floods spread frequently over
the entire varzea area from Cottonwood Pass eastward,
producing continuous sheets of shallow, muddy water.
These alternately filled the lower igapo zone near the
Sage Ridge, or dammed back within it the runoff from
torrential rains in the Badlands area to produce lakes
red with fine clay from the "Interior" soils on the ridge.
The igapo lakes thus received alternately varzea-de-
rived muddy silts, and fine-grained, terra firme-derived
red clays. Increased volume of water meant increased
velocity within the channel-ways during floods. Stream
#8, for example, increased its ability to transport clastic
grains from about 1 mm diameter during the dry,
Lower Nodular time to about 10 mm during this time
of moisture. Maximum velocities, based upon this trans-
porting competency, would be 1.0 f/s or 2/3 mph during
the dry time, and 1.3-1.6 fs or 4-5 mph during the wet
periods (Emmons, Thiel, Stauffer and Allison, 1955, p.
171).

The drenaje streams also increased in size and in
velocity, since they functioned as major outlets through
which the flooded areas drained. Apparently the finer
varzea clays and silts packed quickly, permitting con-
siderable coarse material to bypass and reach the
drenaje channels. This is best seen at channel deposit
#4, a drenaje containing x/i inch diameter pebbles,
separated by sandy silts from its neighboring rios, #3
and 5, which were carrying pebbles up to 2 in. in di-
ameter.

From Cottonwood Pass westward, the streams un-
derwent great expansion but apparently slopes were
sufficiently steeper, due to proximity to the Hills, that
there was very little or no flooding, and hence not over
3 ft. of sediment were deposited.

A second period of stream diminution, much shorter
than the first, occurred in the area watered by streams
4, 5, 6, and 7; very shortly thereafter, a period of
diminution occurred also over the eastern area of
streams 8-11. This brief episode is now represented by
mudstone zones II and IIA. Vigorous stream activity,
with frequent episodes of flooding from Cottonwood
Pass eastward, resumed quickly.

The third period of decrease in stream volume was as
widespread as the first, but presumably not as long-
lasting, since only about half as great a thickness of
heterogeneous mudstones accumulated. In all respects
except duration, conditions of deposition must have
resembled those of Lower Nodular (Zone I) time. From
Indian Creek eastward, the succession of:

Zone III mudstones

Laminated sediments and Zone II

Zone I mudstones
is easy to interpret, but westward the time of stream
increase is represented only by extensive channel de-
posits separated areally by a 3-5 ft. layer of greenish
siltstone. Upon cursory examination, zone III appears
to be merely the upper part of zone I, and only by care-
ful tracing can the pinching out of the intervening
laminated zone be recognized. Probably, slopes in the
western area were steep enough to prevent general
flooding during the wet period.

There followed another general episode of stream
revival and flooding. This time, floods spread laminated
sediment over the western part of the Badlands area, as
well as in the east.

The fourth period of relative decrease in stream
activity was general but quite brief: only 1-3 ft. of
heterogeneous mudstones accumulated before the area
as far east as channel-deposit #8 began another sequence
of general floods which laid down more laminated
sediments.

A fifth period of restriction covered the whole area,
continuous with the fourth in the east. The fifth stream
revival, a very shortlived one, seems to have been local
in the neighborhood of channel-deposit #7.

A sixth time of restriction laid down mudstones over
the whole area, distinguishable in the west from the

CLARK: PALEOGEOGRAPHY OF THE SCENIC MEMBER

99

overlying Poleslide mudstones, but not in the east.
Briefly, stream restriction and mudstone deposition
were continuous in the east from Mudstone zone IV
time into Poleslide time, while in the west the streams
underwent one more episode of revival.

A stream can undergo increased activity due either
to increased gradient or to increased volume. Increased
gradient would normally cause a stream to incise, which
did not occur. Furthermore, the normal overlap of
Oligocene sediments on the east flank of the Black Hills
militates against differential uplift of the Hills relative
to the Badlands area during Oligocene time.

A series of uplifts of the Black Hills relative to the
Badlands area would normally produce a series of de-
posits each coarse-grained at the bottom, grading up-
ward to finer. No such gradation exists; all contacts of
mudstone members with siltstones are sharp. Also the
Scenic Member begins with its thickest zone of mud-
stone (the Lower Nodular Zone), rather than with coarse
sediment, as it would if deposition were controlled by
uplift of the source area. All the evidence of the sedi-
ments themselves indicates that the times of greater
stream activity were times of greater stream volume.
Therefore, it seems probable that the Oligocene streams
underwent periods of alternately increased and de-
creased volume.

Since such volume changes could only reflect changes
in amount of precipitation in the source area, Scenic
Member time saw three dry-to-wet climatic changes
over most of the Black Hills, and three more changes
over parts of them.

How long did these periods of alternate aridity and
rainfall last? Admitting that individual laminae several
inches thick were deposited within a few days, how long
did it take to build up 10 ft. of either mudstone or silt-
stone? How long were the lapses between floods? Does
the post-Chadron, pre-Scenic interval of nondeposition
represent a time as long as that represented by the
Scenic Member, or longer?

To the last question we have not yet an answer, but
considerable evidence has accumulated which bears
upon tre others.

Two lines of negative evidence indicate that the total
time period represented by the Scenic Member was very
short. First, none of the mudstone or siltstone strata
show any indication of weathered zones or soils at their
tops. Furthermore, no evidence of interstratal erosion
has been observed. It is unreasonable to assume that
soils could form on surfaces many times, and be com-
pletely eroded down to fresh material without either
leaving relict patches of soil or cutting notable gullies
in the eroded surface. Therefore, it seems reasonable
that no weathered zones have been observed because no
one surface was exposed long enough for significant
weathering to occur.

The second line of evidence consists of several speci-
mens of Ischyromys typus, Mesohippus bairdii, and Mery-
coidodon culbertsoni which were collected from Mud-
stone V, on the south flank of 71 Table, and one speci-

men of Ictops dakotensis collected from Mudstone V at
the standard section, south of Scenic. (These specimens
are at present in the collection of the South Dakota
School of Mines and Technology). In addition, speci-
mens of Ischyromys, Eumys, Paleolagus, Mesohippus,
and Merycoidodon have been collected from Zone V in
Sage Creek, and are now in the FMNH collection.
These specimens, from the top of the Scenic Member,
are specifically identical with specimens from Mudstone
I, at the bottom of the Member. If no observable de-
velopment occurred within six species representing five
different mammalian orders, the time involved must
have been geologically very short. Probably 500,000
years would be maximum, with the possibility that the
entire Scenic Member was deposited within a few years.

Fortunately, biostratonomic data enable us to limit
the maximum and minimum possibilities more closely.

Weigelt (1927) has reported extensive observations
on degenerative processes in animal corpses. The author
has also observed that exposed corpses of medium to
large-sized animals undergo a regular series of degenera-
tive processes which reduce them, through recognizable
stages, to eventual complete dissolution. The processes
vary in nature and in rate of action, depending upon the
climate. It is proposed that the sum total of these
destructive processes be called perthotaxy1, and an
assemblage of bodies in various stages of destruction be
called a perthotaxis.

If an animal community inhabits an area without
interruption or catastrophic death for several years, the
land surface at any one time will exhibit a complete
perthotaxis, that is, bodies in all stages of destruction.
Burial effectively halts perthotaxy, and thereby pre-
serves all stages of the perthotaxis.

If, on the other hand, episodes of sedimentation
occur at intervals shorter than the time necessary to
develop a complete perthotaxis, each sedimentary incre-
ment will entomb a partial or incomplete perthotaxis. If
some such catastrophe as an epidemic should cause the
sudden, contemporaneous death of numerous indivi-
duals, the result will be a dilated perthotaxis, with a dis-
proportionate number of individuals at one stage of
destruction. The dilation may be early, middle, or late,
depending upon the time lapse between the catastrophe
and the next depositional episode. Catastrophic death
by burial, as by a volcanic ash fall or a death-dealing
flood, would cause a primary dilated perthotaxis, with
disproportionate numbers of whole skeletons.

Figure 44 illustrates perthotaxy on temperate
steppes, such as the northern Great Plains, at present,
as determined by the author's observations. Naturally,
variations in exposure alter the processes somewhat.
Season of death makes a considerable difference also:
the body of an animal who dies in November undergoes
very little alteration before the following March. How-
ever, the diagram does give an order of magnitude and a
set of recognizable stages.

1 From -KtpBu , to destroy, and to£i{- , an arrangement or order.

100

FIELDIANA: GEOLOGY MEMOIRS, VOLUME 5

PERTHOTAXY ON TEMPERATE STEPPES

o

_!fl_

- o

u
O

o =

o =

o =

+ p-pr e d at ion and vertebrate scavenging
• d-d - distortion by dehydration

Fig. 44. Perthotaxy on temperate steppes.

In applying perthotaxic criteria to any bed or
lamina, one must first determine that the fossils repre-
sent a death assemblage rather than a mechanical ac-
cumulation. Mudstone I, the Lower Nodular zone, has
long been recognized as such an assemblage. The abund-

significant transportation. Skeletons in various tetanic
"death poses" have apparently been buried quickly,
without appreciable disturbance by the entombing
floods. Most significantly, those bones in which pertho-
taxy has progressed to an advanced stage of flaking lie

ant, well-preserved coprolites could not have undergone surrounded by the chips which have spalled from them.

CLARK: PALEOGEOGRAPHY OF THE SCENIC MEMBER

101

Plainly, this fossil assemblage is a true death assem-
blage, suitable for perthotaxic studies.

Fossils occur at several different horizons within the
Lower Nodular zone. Usually any one horizon is several
inches thick, and continuous for distances of a few
hundred feet to over a mile. Any one vertical section
contains from one to three or four of these. In the
richly fossiliferous near-stream area south and south-
west of the town of Scenic, the total number of these
fossiliferous zones, interpolated laterally, is not over 20.
The exact number cannot be determined, because local
variations in thickness make precise correlation of 4-18
in. zones over distances of a half mile impossible.

All of these zones which the author has examined
contain an almost complete to complete perthotaxis.
Flaking is always well developed ; occasionally, one finds
bone which suffered enough solution to become spongy
before burial. This indicates that the surfaces, and the
corpses upon them, were exposed for periods of at least
7-10 years, between floods.

If, as postulated, each layer is the deposit of a single
flood of not more than a few days duration, and if the
period between floods was 7-10 years, then 20 relatively
complete perthotaxies would represent 140-200 years at
a minimum. This sets an absolute minimum but does
not directly help to limit the large maximum of 500,000

years, suggested by the lack of evolutionary develop-
ment during Scenic time.

Still another piece of negative evidence is suggestive
here. Under the warm-temperate steppe climate postu-
lated, recognizable soil zones would almost certainly
develop within 100 years. None of these have been
found. Twenty periods of exposure at 100 years each
would be 2000 years as a maximum period for deposi-
tion of the Lower Nodular Zone; the minimum we have
established is 200.

Allowing roughly the same conditions of deposition
for the other mudstones but taking into account that
they are thinner and consist of fewer beds, we can set up
the following table:

Minimum duration Maximum

Mudstone (years) duration

V 200 2000

IV 30 300

III 100 1000

II 20 200

I .200. 2000

Total 550 5500

The laminated siltstones probably required much
less time to deposit, but we have no direct evidence on
this point. Assuming generously that they represent
about half of the total time of deposition, we have for

-

Fig. 45. Photograph of Archaeotherium skull FMNH no. PM9359, in situ, showing advanced perthotaxy. Note the chip of bone
several inches below the skull, also the matrix-filled crack in the right zygoma. :.

102

FIELDIANA: GEOLOGY MEMOIRS, VOLUME 5

the deposition of the entire Scenic Member a probable
minimum of about 1100 years, and a probable maximum
of about 11,000.

This gives a duration for the climatic alternations
of a few hundreds to a few thousands of years, which
would then be of the order of magnitude of the post-
glacial climatic fluctuations described by Dorf (1959,
pp. 194-205) and others. By circular reasoning, if these
climatic fluctuations are of the order described for post-
glacial climates, the Scenic Member probably represents
more nearly 11,000 years than 1100. However, the
probability is high that at least one among so many
streams would shift temporarily to an erosional regimen
during 11,000 years. The fact that none of them did so
suggests that the time was shorter. It may be a matter
of considerable interest to meteorologists that fluctua-
tions of this order of duration are not, apparently,
related to glaciation. It may also be of assistance in the
interpretation of other pre-Pleistocene sediments to ex-
pect that sedimentation may have been influenced by
such climatic fluctuations, and look for their effects.

The chapter on the Chadron formation demonstrates
(p. 74) that a few mammalian lines show a recognizable
development from bottom to top of that formation in
South Dakota. It is, therefore, probable that the time
required to deposit the Chadron formation was con-
siderably longer than that represented by the Scenic
Member of the Brule.

The much greater total thickness and relatively
much greater proportion of mudstone in the Poleslide
Member suggest that it also represents more time for
deposition than does the Scenic Member. Such evidence
is extremely vague and unsatisfactory; it is no more
than an inconclusive suggestion.

The Poleslide consists mainly of mudstones like
those of the Scenic, with siltstones absent over most of
the area. Channel deposits are finer-grained and more
restricted than those of the Scenic. The proportion of
fresh volcanic ash increases toward the top of the Pole-
slide, and the amount of bentonitic ash is uniformly less
than in the Scenic. Calcite comprises the usual ce-
menting material, but a few of the sandstones have a
mixed calcite-silica cement. The significance of the ce-
ments is not known. However, the Poleslide plainly
represents deposition under a climate generally as arid
as that which occurred only at intervals during Scenic
time. Detailed paleogeographic study is needed before
adequate interpretation of the Poleslide can be made.

RELATIONSHIP TO ORELLAN STRATIG-
RAPHY AND SEDIMENTATION
IN NEBRASKA

Removal by erosion, widespread cover of grass and
Pleistocene sediments, and local changes in Brule lithol-
ogy combine to preclude direct visual tracing of the
South Dakota members of the Brule Formation south-
ward into Nebraska. Faunal correlation confirms the
general impression that the Scenic Member is approxi-

mately correlative with the Orella Member of Nebraska.
However, the very short time represented by the Scenic
Member renders impossible an exact correlation based
upon fossils: an identical fauna might be expected to
have inhabited Nebraska at any time during the 50,000
years immediately preceding or following the particular
1100-11,000 years represented by the Scenic Member.

Examination of the literature describing Oligocene
stratigraphy and paleogeography in Nebraska (Schultz
et al., 1955. and their earlier papers) suggests that
Orellan conditions in Nebraska were strikingly different
from those in South Dakota. In 1951, Falkenbach and
Schultz (1951, pp. 47-50) proposed a division of the
Orella Member into faunal zones A and B. They also
divided the Whitney member, producing the following
correlation:

Member
Whitney

Orella

Falkenback

AND

Schultz

B

A

Wanless, 1923

Leptauchenia
-Upper Oreodon
Middle Oreodon

Lower Oreodon

Since they state clearly that these are faunal zones, one
must presume a faunal succession within the Orella
Member, and also within the Scenic Member in South
Dakota (Wanless' original division was lithologic, not
faunal). My own collections indicate no such faunal
succession within the Scenic Member. Much of what
Wanless included in the "Upper Oreodon Beds" in the
Corral Draw area is actually part of the Poleslide Mem-
ber: due to thinning of the middle part of the Scenic
Member westward, and to local changes within the up-
per Scenic and lower Poleslide, (see cross-section, Fig.
52, and columnar sections, appendix) individual zones
within the Scenic Member have been universally mis-
correlated.

The definitive 1955 paper (Schultz et el., 1955) de-
scribes deep channelling associated with two major and
several minor paleosol complexes. The authors' cross-
sectional diagram is reproduced here for reference (Fig.
46). The authors state (loc. cit., p. 5) that they first
suspected that paleosols might be present due to the
very sudden appearance of oreodont skulls with larger,
inflated bullae only 10 ft. stratigraphically above other
skulls in which the bullae were uninflated.

The following pertinent points were apparently not
considered in this context:

1. The authors had no way of estimating the actual
rate of sedimentation.

2. The sudden change, if actual, might have been due
to immigration.

3. There is obviously no way of estimating the extent
of genetic change required to accomplish the somatic
change in the bulla. Therefore, there can be no way of
estimating the time involved in such a change — it might
have occurred even as a mutation in one generation.

J M inor Pa

leosot s

Major Paleosol Complex

Paleosol Complex
(Associated with Lower Ash)

1 Paleosol Comlex

| (Associ ated with Channel Fills)

Major Paleosol Complex

- Minor Paleosols

Fig. 46. Cross-section of the Toadstool Park area, copied from Schultz, Tanner, and Harvey, 1955.

103

104

FIELDIANA: GEOLOGY MEMOIRS, VOLUME 5

The reasoning which led to the search for paleosols
was based, therefore, upon a discrepancy between the
time necessary to deposit 10 ft. of sediment whose rate
of deposition was unknown, and the time necessary to
develop an anatomical structure whose manner and
rate of evolution was unknown.

The authors' description of the paleosols gives as
distinctive characters the "facts" that (loc. cit., p. 10)
"The Orella deposits of the Brule Formation typically
are massive, with an angular to subconchoidal fracture
when fresh, but these properties are usually lacking in
the paleosol profiles. Many samples do not show the
characteristic subconchoidal fracture at all. Some are
neither massive nor angular. In several places the frac-
ture is distinctly granular, reminding one of the soil
pieces of beds of a modern soil — Certain paleosols and
paleosol complexes tend to weather out as topographic
benches (see figs.). Some of the benches are due to ledges
of secondary lime, and others may be related to concen-
trations of clay representing B horizons of ancient soils.

"Not only is a paleosol complex a break in the other-
wise continuously fossiliferous strata, but careful study
of the faunas has shown that all of the major breaks in
the faunal successions coincide with the occurrence of
major paleosol complexes. In other words, the period
of time represented by such a paleosol complex must be
of considerable length.

"Some paleosols are associated with extensive valley
fills which may be as much as fifty or sixty feet deep."

The criteria offered as diagnostic of paleosol origin
are:

1. Granular fracture.

2. Presence of "ledges of lime" and layers of clay.

3. Absence of fossils.

4. Evidence of breaks in the faunal succession coin-
ciding with the paleosols.

5. Association with deep channel cutting and filling.

The literature on Nebraska presents us with a pic-
ture of Orellan cutting and filling, with development of
thick soil complexes during long periods of nondeposi-
tion, and a total duration sufficient to permit significant
mammalian evolution. This is strikingly different from
the situation, as interpreted in this report, existing at
the same time in the Big Badlands only 120 miles to the
northeast with no major Oligocene topographic barrier
intervening.

I have twice studied the Toadstool Park area of
Nebraska, in 1956 and again in 1964. My observations
of 1956 seemed so at variance with the published inter-
pretation that I have several times discussed them with
the authors quoted, without altering their opinions. It is
therefore necessary to present my own observations and
interpretations at this time.

First, the "paleosols" are actually the sedimentary
lithotope described in this paper as "laminated sedi-
ments." They consist of alternate laminae of fine sand-
stone, sandy siltstone, and laminated mudstone, a few

inches thick, with normal, sharp sedimentary contacts
between the laminae. No graded bedding or gradational
phenomena have been observed. Some of the coarser-
grained strata at Toadstool Park even exhibit some
cross-bedding (see Figs. 47, 48). Naturally, some of the
coarser-grained sediments are better-cemented with

Fig. 47. Photograph of "the Bench," looking south from Toad-
stool Park. The lamination of the sediments, and their similarity
to those in South Dakota (see Fig. 36) is apparent.

Ca C03 than are the finer ones, and weather out differ-
entially as ledges. The clays are normal for the Brule —
chiefly montmorillonite and mixed layer, associated
with abundant quartz and some half-altered shards of
volcanic glass. Microscope study reveals abundant,
unweathered acid feldspar, with some muscovite and
unaltered to little-altered biotite. The rock does fracture
to a granular surface in the coarser grades, due to the
presence of clastic grains of quartz and feldspar set in a
normal calcite cement. These sediments have absolutely
none of the physical, structural, or mineral characteris-
tics of soils, in any of their laminae. Their association
with channel-fills, as succeeding paragraphs will demon-
strate, is not as described or diagrammed by Schultz
et al., but is rather the same as that already described
for the Big Badlands laminated sediments.

Thin bedding and lamination; uniformity within
each lamina from top to bottom; normal sharp, sedimen-
tary contacts between laminae; alternation of finer and
coarser elastics; presence of abundant, unweathered
feldspars; presence of less abundant but frequent un-
altered biotite; presence within the sequence of cross-
bedded members; and presence of crystalline calcite
cement are characteristics any one of which would
militate against a deposit being a soil or paleosol. Cu-
mulatively, they are overwhelming, positive evidence
against it. Furthermore, my rather extensive observa-
tions of Oligocene sediments have failed to reveal any
paleosols of Chadronian or Orellan age anywhere in
South Dakota, western Nebraska, Colorado or eastern
Wyoming. The burden of proof of the existence of any
such paleosols must now rest with those who describe
them, and must necessarily include careful field study
of the sedimentary structures, and laboratory study
with a petrologic microscope.

Far from indicating slow deposition, this lithotope
in Nebraska, as in South Dakota, should be interpreted

CLARK: PALEOGEOGRAPHY OF THE SCENIC MEMBER

105

-

Fig. 48. Closeup view of laminated sediments including one cross-bedded sandstone, }4 mile south of Toadstool Park. The photograph

was taken just out of sight to the east (left) of the view in Figure 47.

as representing exceedingly rapid deposition by suc-
cessive sheet-floods from actively-depositing streams.
Presumably, successive increments were deposited in
such quick succession that no one surface remained
exposed long enough to develop a phanerogamous flora.
For this reason, land mammals temporarily abandoned
the area and hence, naturally, left no fossil record of
their presence.

For evidence explaining the reported faunal differ-
ences in the heterogeneous mudstones above and below
these laminated lithotopes, one must turn to the re-
gional tectonic structure.

The Toadstool Park area is crossed by a major
normal fault trending N 60° E, dipping SE 70°, with a
vertical displacement of 70-75 ft. The fault brings a
lower Orellan channel-fill, north of the fault, into ap-
proximate juxtaposition with a laminated siltstone
lithotope ( = "the Bench") south of the fault. Figures
49, 50, and 51 illustrate the fault. A small block of sand-
stone located in a recent gully along the fault (see Fig.
49) at one place may be either a rotated block within
the fault zone or a recent slump block; in either case, it
is a purely local phenomenon of no significance. The

fault plane is slickensided and is mineralized with chal-
cedony; Figure 50 shows chalcedony casts of slicken-
sides, also laminae of sandstone transected and slightly
dragged by the fault movement. This fault was ap-
parently overlooked by Schultz et al. (loc. cit.), and the
sharp, fault-scarp terminations of the sandstone have
been misinterpreted as edges of deeply-cut channel-fills
(loc. cit., fig. 1.). Faulting was mentioned, but without
acceptance of its stratigraphic effects, by Schultz and
Stout (1961, p. 47). Actually, the base of the large rio-
some in the figure shows a maximum of 3 to 5 ft. of
cutting, which is also maximum for Orellan riosomes of
the entire area. Figure 1 in Schultz et al. (loc. cit.)
represents so many correlations of lithotopes on one
side of the fault with lithotopes of similar nature but
different stratigraphic position on the other, that no
consistent correlation between it and the interpretation
in Figure 51 of this paper is possible.

However, channel cutting such as described in the
paper cited does not exist in the Toadstool Park area.
To the best of my knowledge, it does not exist else-
where in the Crawford-Chadron area, but further recon-
naissance is needed before I can state this positively.

106

FIELDIANA: GEOLOGY MEMOIRS, VOLUME 5

Evidence has now been presented that two of the
major points of difference between the Orellan paleo-
geography of South Dakota and the described paleogeo-
graphy of Nebraska do not exist in fact: there are no
paleosols in Nebraska, and no deep-cut channels, at
least in any area described. Let us now consider the
third: faunal evidence that the Orella Member repre-
sents a length of time sufficient to allow significant
evolution of oreodonts.

The most significant advance in oreodonts reported
during Orellan time (Shultz et al., 1955, pp. 4, 5; Schultz
and Falkenbach, 1954, pp. 153-159, and 1956, pp. 381-
390) seems to be the change from uninflated to inflated
bullae. However, when one remembers that this ad-
vance is based upon stratigraphy which overlooks a 75-
foot fault in a 270-foot section, and uses as marker beds
zones of sheet-flood sediments interpreted as paleosols,
plus zones of volcanic ash (the "purple-white" layers)
which can be demonstrated to be local in South Dakota,
the detailed age reference of the specimens lies open to
question. Also, Schultz and Stout's reference to "the
faulting (and folding) so characteristic of the Chadron
and Orella" (loc. cit., 1961, p. 47) is one in which I
concur — faults of significant magnitude can be found
at many places. (I do not, however, concur at pres-
ent in their dating of the faults.) How many of these
faults were, like the one at Toadstool Park, unrecog-
nized at the time the oreodonts were collected? One
can only hold the age reference of all specimens col-
lected in the Orella Member in question. There can
be little question of these specimens' reference to the
Chadron or to the Orella Member, but any reference to
subzones within the Orella, or to an upper subzone of the
Orella as opposed to a lower subzone of the Whitney
Member, must be questioned.

The use of "purple-white layers", which are actually
bentonitized volcanic ash beds, as markers is subject to
serious question due to the field relations of three similar
beds in South Dakota. The two already described from
the Ahearn Member of the Chadron (see p. 22 of this
paper), show mineral ogic differences sufficient to demon-
strate that they do not represent the same ash fall, al-
though they occur stratigraphically in the upper part of
the Ahearn Member, and geographically only a few miles
apart. A third such "purple-white layer" occurs (NW^
of SE]4, Sec. 13, T. 42N., R. 45W., Shannon Co.) in
the Scenic Member between zones I and III, as a lens
not over 50 ft. wide and about 1 ft. thick in a small de-
pression-fill. Any such discontinuous lenses are obvious-
ly of no value as stratigraphic key beds. It is not clear
whether or not the "purple-white layers" of Nebraska
are discontinuous; in view of the confusion which must
arise from the misinterpretation of faults as channel-
fills and sheet-flood sediments as soil zones, the con-
tinuity of any one horizon in Nebraska must be de-
termined anew by visual tracing.

As a final note on stratigraphic correlation between
Nebraska and South Dakota, the cross-section (Fig. 41)
clearly demonstrates the danger of "marker beds."

Zone III nodules northwest of Sheep Mountain are
usually referred to as Wanless "Upper Nodular Zone."
Eight miles to the southwest, Zone III has come to rest
upon Zone I, and Zone V becomes the "Upper Nodular
Zone." In Sage Creek and upper Cain Creek, both Zones
III and V contain fossiliferous nodules. At Dillon Pass,
Zone V has merged indistinguishably with the Poleslide.
Along the northwest side of Cuny Table (see Fig. 41)
Zone V is the "Upper Nodular Zone," but a new zone,
about 60 ft. up in the Poleslide, has come to resemble
Zone V lithologically. Anyone studying the standard
section of the Scenic Member, then measuring a section
on the north flank of Cuny Table, would almost certain-
ly make the miscorrelation indicated in Figure 52,
rather than the true correlation determined by visual
tracing.

The hypothesis of oreodont evolution, and conse-
quent stratigraphic correlation, offered by Schultz and
Falkenbach, has two alternatives which are equally
probable.

First, the oreodont stratigraphy in Nebraska may be
correct, but the "Orella Zone B" of Nebraska may be
equivalent to the lower part of the Poleslide in South
Dakota. Since the evidence discussed at length above
indicates that the oreodont stratigraphy in Nebraska
cannot be correct in detail, this hypothesis is purely
academic but worthy of future consideration.

Second, in their last publication on the subject,
Schultz and Falkenbach (1956, pp. 381-392) pointed out
that the members of the subfamily Miniochoerinae have
the bulla small and uninflated throughout Chadronian
and Orellan time. Actually, it is difficult in practice to
separate a Scenic Member Miniochoerine from a Mery-
coidodon. Schultz and Falkenbach have, in their mono-
graphic study, attempted a vertical classification of
Oligocene forms supposedly ancestral to definitely sep-

■

"*#

»-7— r

Fig. 49. View westward along the fault at Toadstool Park.
Left (south) side downthrown; the fault plane dips to the left. The
sandstone in the left foreground is part of the same stratum as that
near the top of the Badlands to the north of the fault. Note drag
on the downthrown block. Also note the apparent continuity of
the sandstone north of the fault with the much higher laminated
sandy strata on the downthrown block; apparently this led to the
misinterpretations in figure 46.

CLARK: PALEOGEOGRAPHY OF THE SCENIC MEMBER

107

Fig. 50. Fault face cutting channel-fill sandstone. This is a view of the ledge from which Figure 49 was photographed. Note the cut
edges of sandstone strata. The whitish material is chalcedony, on which casts of slickensides are visible.

arable Miocene groups. It is highly possible that middle
Oligocene oreodonts represent an interbreeding genetic
pool, exhibiting certain genetically controlled, non-
adaptive characters which, during Orellan time, were
interbred freely, but later became separated as diver-
gent groups moved genetically further from each other.
Briefly, an Orellan Merycoidodon with small bullae
might well have mated with a Miniochoerus with large
bullae, although the Whitneyan forms possessing these
characters could not have interbred. This situation,
coupled with misinterpretation of the stratigraphy and
accidents of preservation, could easily produce such an
artifact as an apparent sudden appearance of large
bullae in one "phyletic line."

In summary, careful study of the Chadron-Craw-
ford-Toadstool Park area of Nebraska demonstrates
that:

1. There are no paleosols in the Chadron or in the
Orella Member of the Brule.

2. The lithotopes described in this report also, with
minor variations, make up the mass of Orellan sedi-
ments in northwestern Nebraska.

3. There is no evidence of significant channel-cutting
or erosional episodes during Orellan time: reports of
cutting have arisen through misinterpretation of a
faulted sequence.

4. Due to the misinterpretations listed, the actual
faunal sequence within the Orella Member has not yet
been determined.

5. All available evidence in northwestern Nebraska
is consonant with a paleogeographic interpretation
similar to that given here for the Scenic Member in the
Big Badlands.

6. Due to the extremely short time represented by
the Scenic Member, no exact correlation of the Scenic
Member with the Orella Member is possible at this
time. Lithologic changes within the lower hundred feet
of the Poleslide Member, in the southwestern part of
the Big Badlands, suggest that the lower part of the
Poleslide of South Dakota might be correlative with the
upper part of the Orella Member of Nebraska. This is a
possibility to be explored, not a suggested correlation.

7. The supposed evolution of oreodont bullae within
the Orella Member could be equally well explained as:
(A) an immigration of forms with large bullae; (B) a
mutation, spreading within a few generations over the
relatively small area involved; (c) an artifact due to the
downward expansion of a vertical classification, coupled
with a misinterpretation of the stratigraphic position of
specimens.

CONCLUSIONS

1. The Scenic Member of the Brule Formation com-
prises five types of sediments:

Limestone
Heterogeneous mudstone

108

FIELDIANA: GEOLOGY MEMOIRS, VOLUME 5

I N T C R PRE T A T I ON (

T OA D3 TOOL

F r I OURE
PARK

Fig.' 51. Diagrammatic interpretation of Figure 49, for com-
parison with the interpretation of Schultz el al., 1955, in Figure 46.
1, Basal Orellan channel-fill north of fault; 2, Slice or slump block
along fault; 3, "The Bench," Orellan-Whitneyan contact sand-
stones; 4, Basal Orellan channel-fill, equivalent to no. 1, south of
fault.

Laminated clay

Laminated siltstone and sandstone

Cross-bedded sandstone.

These combine to form four sedimentary lithotopes:

Silty mudstones

Laminated siltstones

Laminated mudstones

Channel-fill zones.

2. These four lithotopes grade into one another
horizontally but not vertically.

3. The silty mudstone zones transgress the deposi-
tional areas of several Oligocene streams, and can be
traced for distances of several miles.

4. Alternating strata of mudstone and siltstone have
sharp contacts, not gradational ones. No instances of
graded bedding have been observed.

5. The channel-fill sandstones are of two types,
Northern Black Hills derived and Southern Black Hills
derived, each with a characteristic suite of heavy
minerals.

6. The silty mudstones represent times of discontin-
uous deposition by mud-flows from flooding streams of
small volume and low energy.

7. The siltstones represent more rapid deposition by
sheet-floods from the same streams at times when their
volume, and therefore their energy, was much increased.

8. The fluctuations in energy of Oligocene streams
in this area were the result of fluctuations in volume
rather than of changes in gradient.

9. The fluctuations in volume of streams resulted
from alternations of wetter and drier climate.

10. The presence of alligators in the underlying
Chadron, and the presence of one alligator in the first
Scenic Member wet-climate deposit, plus the absence
of alligators in the intervening dry-climate deposit, sug-
gest that the times of dry climate were also times of cool
climate.

11. The fact that fossils representing five mammalian
orders show no differences from bottom to top of the
Scenic Member indicates that deposition of the entire
member required not over 500,000 years (see p. 99).

12. Individual increments of 6-18 in. were deposited
within the span of a very few days.

13. The complete absence of weathering or of soil
zones at the top of any one increment shows that never
did a period of more than 100 years elapse between
episodes of sedimentation.

14. The presence within any one fossiliferous incre-
ment of a complete perthotaxis indicates that periods of
10 years or over usually elapsed between episodes of
deposition.

15. Using the data from the last two conclusions,
the total time required for deposition of the mudstones
of the Scenic Member was 550-5500 years. Allowing
equal time for deposition of laminated sediments, al-
though the evidence suggests that they were deposited
more rapidly, the total time represented by the Scenic
Member was 1100-11,000 years.

16. The alternations of dry-cool and warm-wet
climates were, on this basis, of the same general order of
magnitude as Recent post-glacial warm and cool al-
ternations.

17. The geographic distribution of lithotopes within
the Scenic Member at any one time can best be ex-
plained by comparison with the distribution of sedi-
mentary environments within the Central Amazon
Basin, as described by Sioli.

18. Using Sioli's terms with additions where neces-
sary, the following sub-environments of fluvial sedi-
mentation can be recognized:

Rio : the channelway of a throughgoing stream.

Varzea: the area of sedimentation outward from a
rio, including the natural levee and the long backslope
away from the stream.

Drenaje: secondary streams which arise locally as
drainageways in the more or less linear depressions be-
tween adjacent varzeas.

Terra firme: valley walls, composed of material other
than that being handled by the rios.

Igapo: The approximately linear depression lying
between a terra firme and the adjacent varzea, receiv-
ing sediment chiefly by local wash from the terra firme,
but occasionally by sheet-wash from the rio.

The body of sediments formed in these environments
through time are termed, respectively, riosome, varzea-
some, drenajesome, and igaposome.

19. Deposition of the Scenic Member in the Big
Badlands was controlled by three rios with sources in
the Southern Black Hills, and two to four rios with
sources in the Northern Black Hills.

20. Although deposition of the Scenic Member was
episodic, there were no periods of erosion and all streams
were continuously at grade to overloaded.

21. The lower part of the Poleslide Member generally
resembles the mudstones of the underlying Scenic Mem-
ber in lithology and origin. Easterly gradation of the
upper mudstones of the Scenic Member into the basal
Poleslide indicates either that the widespread aridity
of Poleslide time started earlier in the Northern Black
Hills than in the Southern, or that the eastern part of

PROBABLE MISCORRE L ATIONS OF MINOR LITMOLOSIC UNITS

IN THE SOUTHWESTERN PART OF THE BADLANDS

N FLANK OF CUNY TABLE
J_

4

N W "

OF SEC 10, T 42 N,

R 45 w., SHANNON CO

<-

9 MILES

N OF SHEEP MI. ROAD :
NW—OF NE i,SEC10,T4S.,
R13E., PENNINGTON CO.

POLE SLIDE
V

PO L E S I I D E 5^

\

\

\

\

\

\

\

\

\

\

\

\

\

\

\

\

\

\

\

GRAY "MARKER" ZONES

MUDSTONES

I

U LAM I NAT

ED SEDIMENTS

APPARENT BUT ERRONEOUS CORRELATIONS SHOWN BY DASHED LINES.
TRUE CORRELATIONS SHOWN BY NUMBERS BESIDE MUDSTONES.

NOTE THE MISLEADING OCCURRENCE OF TWO GRAY "MARKER" BEDS, EACH OVER-
LYING A BUFF MUDSTONE, AT DIFFERENT ACTUAL HORIZONS IN THE TWO SECTIONS.

Fig. 52. Probable miscorrelations of minor lithologic units in'the southwestern part of the Badlands.

109

110

FIELDIANA: GEOLOGY MEMOIRS, VOLUME 5

the Badlands area was, by late Scenic Member time,
receiving notably less rainfall than the western part,
or both.

22. The flesh-colored to brilliant, discontinuous red
laminated clays of the Dillon Pass and Big Foot Pass
areas represent igaposomes, having as their source
local rain wash from exposed hillocks of weathered Pierre
shale along the Sage Ridge.

23. The dull red-brown colors of the heterogeneous
mudstone zones in the same geographic areas as the
igaposomes represent varzea mudstones mixed with
slight amounts of locally-derived red clays.

24. The sediments in northwestern Nebraska, pre-
viously interpreted as paleosols, are actually laminated
siltstone lithotopes similar to those of the Big Badlands.
They represent rapid rather than slow deposition.

25. The structures in northwestern Nebraska pre-

viously interpreted as deep channel-cutting are actually
due to faulting.

26. Faunal zoning of the Orella Member in north-
western Nebraska is based upon erroneous stratigraphy.

27. The evolution of oreodonts within Orellan time
is not established.

28. No satisfactory correlation of subdivisions of the
Orella Member in Nebraska with subdivisions of the
Scenic Member in South Dakota has been achieved. In
view of the brief time represented by the Scenic Mem-
ber, such detailed correlation is unlikely on paleon-
tologic grounds, although it may be achieved through
paleogeography.

29. Both in South Dakota and in northwestern
Nebraska, Middle Oligocene sediments indicate rapid
deposition of increments several inches thick, with no
periods of erosion and no long periods of non-deposition,
under an alternation of warm-wet and cool-dry climates.

Chapter VII

PALEOECOLOGY OF THE LOWER NODULAR
ZONE, BRULE FORMATION, IN THE
BIG BADLANDS OF SOUTH DAKOTA

by
John Clark and Kenneth K. Kietzke

INTRODUCTION

Determination of Scenic-Member paleogeography
in detail raises the interesting possibility of collecting
vertebrate fossils from different local environments.
Because the Lower Nodular Zone, the most richly fos-
silferous horizon, represents a time interval of not over
1100 years, during which there is no evidence of change
of climate or of significant shifting of stream courses,
the chance of collecting an unmixed fauna representing
one local environment at any one place seems very
good. A series of such collections from different local
environments might be expected to demonstrate faunal
differences controlled by the known environmental
differences. From this information, it might be possible
to determine the environmental preferences of various
Oligocene genera and species.

The only two previous methods of approach avail-
able for the study of Oligocene mammalian ecology
have been interpretation of osteological characters and
analogy with related recent genera. The first of these is
partially vitiated by the enormous adaptability of mam-
mals: for example, a pine squirrel from northern Canada
and a rock squirrel from western Texas give no osteo-
logic evidence of the differences in their environment and
way of life. Tigers from the Amur basin of Siberia are
specifically identical with tigers from Bengal. Analogy
with closest living relatives suffers from the fact that
few if any Oligocene genera have recent relatives close
enough to give any assurance of valid homology. Paleo-
geographic evidence may, therefore, offer a valuable,
independent line of evidence regarding the habits of
individual genera and species.

Collecting was directed toward these purposes in
1956, 1959, and 1962. These collections form the basis
for this chapter. Collections from the same localities,
made in 1964 and 1965, have more than doubled the
number of specimens; however, these have not yet been
fully studied. Preliminary study indicates that they will
considerably modify some of the statistics presented
here, but will not significantly change the basic conclu-
sions.

Figure 54 includes graphs I-XIII, which will be re-
ferred to by graph number rather than by figure number
hereafter.

Acknowledgments

The authors wish first to acknowledge the kindness
of the staffs of the South Dakota School of Mines and
Technology Museum, and of the University of Colorado
Museum, in which these collections are now housed, for
permitting this study. The senior author wishes to
express his appreciation of research grants by the Yel-
lowstone-Bighorn Research Association, in 1956, and
the Badlands National Monument, National Park serv-
ice, in 1959, which financed the field work in those years.
Finally, the senior author wishes to thank the following
for their services as field assistants: 1956 — R. Livingston
and E. Kanesky; 1959— P. Nelson and V. Gunn; 1962—
The authors enjoyed the cooperation of Dr. P. Robinson
and his assistants in making the CU near-stream col-
lection (#IX). The open-plains collection of 1962 (CU
#1) was made by the authors, working with private
funds.

GENERAL PHILOSOPHY OF FIELD WORK

Selection of localities for collecting was based upon
the primary assumption that an area immediately adja-
cent to one riosome and within a short distance of others
might reasonably be considered to have been forested.
Conversely, an area several miles from the nearest
riosome and near an igaposome would represent a least-
watered spot and would have been either prairie or
brush-covered, if any such cover existed anywhere.
Several places answered each of these descriptions;
selection between them depended upon the following
practical considerations:

1. Localities should be abundantly f ossiferous, in
order to yield adequate samples.

2. They should be relatively inaccessible to the
numerous local amateur collectors.

Ill

112

FIELDIANA: GEOLOGY MEMOIRS, VOLUME 5

3. They should have been relatively untouched by
professional collectors for at least ten years.

4. They should be areas in which corresponding por-
tions of the section are well exposed, in order to achieve
the greatest possible degree of contemporaneity. Brule
streams did not shift their geographic positions signifi-
cantly during Lower Nodular time, nor is there any
evidence of climatic fluctuations which might affect
vegetational distribution during that time, so this was
not practically important.

With these qualifications in mind, one near-stream
locality at Cottonwood Pass, one swamp locality in
Sage Creek, and three least-watered localities, one in
Sage Creek and two near Dillon Pass, were selected.
Collections were then made from each. Where possible,
two collections were made from the same locality at
different times by different field assistants to reduce
psychologic and ocular collecting biases to a minimum.
Fortunately, the junior author observes small specimens
more readily than large ones, while the reverse is true of
the senior author. The mutually cancelling effect of
these biases was utilized by prospecting side-by-side
over the same ground, several times.

One major determination had to be established be-
fore the collections could be used : do the fossils at these
places constitute a death assemblage, or are they simply
a mechanical accumulation? For the near-stream lo-
cality and the localities in Sage Creek, a definite answer
can be given, based upon the following evidence:

1. Individuals in every stage of perthotaxy, from
entire skeletons to separate chips, have been found.

2. Coprolites are abundant.

3. None of the bones show water abrasion.

4. Some partially disintegrated specimens have their
broken chips dispersed around them uniformly in all
directions.

5. The varzea sediments in both areas are heteroge-
neous mudstones, which can be demonstrated in both
areas to have engulfed bone without transporting it
appreciable distances.

6. Complete skeletons have been found in death
poses. One herd of Leptometryx found previously in
the Cottonwood Pass area, and a herd of Hypertrag-
ulus in the Sage Creek area, showed the individuals in
both cases obviously lying in death poses.

7. Celtis seeds and bones of very small animals
occur scattered throughout, never accumulated as they
would be along a strand line, or washed against an
obstacle. (Celtis seeds sometimes occur in sausage-
shaped or egg-shaped masses, apparently representing
storage by rodents in burrows or hollows).

8. Bones and skeletons of all sizes occur indiscrimi-
nately in both localities.

These lines of evidence, considered cumulatively,
seem a clear indication that the Cottonwood Pass and
Sage Creek fossils constitute death assemblages, es-
sentially un transported. The evidence for the Dillon
Pass localities consists only of points 2, 3, 5, 7, and 8; no

complete skeletons are known to us from these areas.
The evidence for a death assemblage is not conclusive
here, although there is no evidence against it. The
Dillon Pass localities must, therefore, be considered in
relation to the known assemblage at Sage Creek, before
conclusions are drawn from them.

In order to establish a uniform basis for collecting,
the following principles were followed:

1. All teeth and jaws, including fragments, were
collected.

2. In case two or more anatomically compatible
fragments showing the same maturity and amount of
dental wear were found within a foot of each other, and
no other bones of that species occurred within several
feet, the two were regarded as belonging to the same
individual. Otherwise, each fragment found was counted
as a separate specimen. This may produce a bias in
the direction of increasing the total number of speci-
mens. Probably the commoner small forms would be
favored relative to larger ones.

However, the only other consistent method of
counting, that of computing the least possible number of
individuals, would in this case introduce a larger bias
in the other direction. Where the bones are known to be
essentially untransported, the only scattering of parts of
one individual is accomplished by perthotaxy before
burial and by recent anataxy. Perthotaxic separation
involves distances of a few inches to a few feet, de-
pending upon the size of the corpse. Direct field observa-
tion of all but the commonest small forms can, there-
fore, usually determine whether or not two jaws or teeth
derive from the same individual. Recent transportation
is practically always directly downslope and, except in
tiny rills, destroys the smaller specimens within a few
feet. Field observation, again, can usually determine
whether or not two cospecific fragments are related.
Application of the method of minimum numbers would,
by comparison, ignore these relationships and such
additional clues as amount of wear. It might well require
excluding a senile maxilla found more than 100 yards
away from several juvenile mandibular rami. Therefore,
the method of minimum possible numbers, which may
be properly applied to mechanical assemblages, seems
more biased than the method of studying each situation,
in a case like the present one.

3. Only teeth, jaws, and skulls were included in the
final count.

4. Turtles were omitted from the collection, because
of their bulk. No other group was omitted. Birds and
lower vertebrates were collected, but excluded from the
statistical computations.

PERTINENT FIELD DATA FOR
INDIVIDUAL COLLECTIONS

I. Near Stream Facies.

Locality: N^ of Sec. 10 and 11, and SW34 of Sec. 2,
T. 42N., R. 45W., Shannon Co., South Dakota. Upper-
most part of Corral Draw drainage, Cottonwood Pass
area.

CLARK AND KIETZKE: PALEOECOLOGY OF THE LOWER NODULAR ZONE

113

Stratigraphic data: One riosome passes through the
north Yl of Sec. 11 ; another outcrops in the west half of
Sec. 10, and a drenajesome in the northwest \i of Sec. 2.
The great majority of specimens occur in the upper half
of the stratigraphic section.

Collections: Collection VII, SDSM, 139 specimens,
from the NE^ of Sec. 11; 1959. Graph VII of this
report.

Collection VIII, SDSM, 47 specimens, from the E
border of the SWM of Sec. 2; 1959, Graph VIII.

Collection IX, CU, 115 specimens, from the entire
area described; 1961 and 1962. Graph IX.

II. Open Plains Facies.

A.

Locality: North-central Sage Creek Basin, SEJ4 of
SEM, Sec. 16, and NEM of NEH, Sec. 21, T. 2S.,
R. 15E., Pennington Co., South Dakota.

Stratigraphic data: The nearest riosomes are at Sage
Creek Pass, three miles south, and northeast of the
Pinnacles, five miles to the east. In the field, the matrix
here resembles that at Cottonwood Pass except that
the color is a little brighter yellow and concretions are
very much less developed. Most fossils in Sage Creek
weather free of the matrix rather than remaining in
concretions which weather free, as larger fossils usually
do in the Cottonwood Pass area. In thin-section, the
Sage Creek matrix is less calcareous and has a lower
proportion of coarse silt to finer material ; otherwise no
differences are apparent. A few individual concretions
are fully as calcareous as many at Cottonwood Pass.

The most fossiliferous zones are thin concretionary
layers 3 to 10 ft. below the top of the Lower Nodular
Zone. Most of the specimens come from these zones.

Collections: Collection II, SDSM, 458 specimens,
from the entire area described; 1956. Graph II.

Collection I, CU, 531 specimens, from the northern
two- thirds of the area; 1962. Graph I.

Collection V, CU, 141 specimens, from the southern
one-third of the area. This collection was separated from
I because a slightly greater development of concre-
tionary nodules in this area was thought possibly to
indicate closer proximity to some distributary channel
now eroded away. For this reason it is labelled "Swampy
Plains Collection." Comparison of this collection with
the other Open Plains and Swampy Plains collections
will show the extent to which faunal composition re-
flects this particular petrologic character ; 1962. Graph V.

B.

Locality: Dillon Pass area, SE34 of Sec. 28 and adja-
cent edge of Sec. 27, T. 2S., R. 16E., Pennington Co.,
South Dakota.

Stratigraphic data: The nearest channel-fill is a half
mile to the northwest. However, the Lower Nodular
Zone consists here mainly of pinkish to red-brown
igaposome clays. Lithogically much of it is alternately
finely-laminated reddish clays, presumably derived from

the Sage Upland, and unsorted silty mudstone deposited
as varzea-type sediments by fluvial water. The area,
therefore, represents an outer fringe of varzea flood
plain, more often subject to backwater flooding than to
sheet floods, and usually not flooded at all. The fossils
occur in the varzea-type mudstones, richest of which is
one 3-6 ft. below the top of the Lower Nodular Zone.

Collections: SDSM Collection IV, 114 specimens,
from the entire area described ; 1956. Graph IV.

Locality: Area south of Dillon Pass, between the
centers of Sec. 33 and 34, T. 2S., R. 16E., 1 mile south
of Dillon Pass, Pennington Co., South Dakota.

Stratigraphic data: The same as for the last, except
that the nearest channel-fill is roughly 1 mile to the
northeast, and varzea sediments form a smaller propor-
tion of the total mass.

Collections: SDSM Collection III, 66 specimens,
from the entire area described; 1956. Graph III.

III. Swampy Plains Facies.

Locality: North-central Sage Creek Basin, SW^,
Sec. 15, E^ of NWM, Sec. 22, T. 2S., R. 15E., Penning-
ton Co., South Dakota. This locality lies adjacent to the
Sage Creek Open Plains locality, but due to a combina-
tion of recent topography and relatively unfossiliferous
beds, no fossils were collected in a strip about 200 yards
wide which separates them.

Stratigraphic data: The highly fossiliferous horizon
about 3 ft. thick, near the top of the Lower Nodular
Zone immediately to the west, changes color rapidly
from tan to pale greenish. A prominent pond limestone
4-8 in. thick occupies the middle of this horizon.
Greenish color develops in the sediments beneath, also,
until more than half of the Lower Nodular section is
greenish. The limestone contains at least three genera of
pond snails in great abundance, as well as algal strands,
ostracods, fish bones, and numerous pupa casts up to
1% inches long. The limy mudstones marginal to and
immediately above the limestone contain abundant
fossil mammals. Since the greenish color is due to re-
duced iron, the presence of abundant vegetation plus
enough water to prevent oxidation during deposition of
the muds is indicated. This was an igapo swamp and
pond, which endured long enough for the development
of an aquatic and paludal biota. It is the only such
situation known within the Lower Nodular Zone; the
contrast with the varzea sheet-flood sediments which
make up the laminated siltstone lithotope (p. 83, this
report) is obvious. As the only representative of a true
paludal environment, it might offer a most valuable
clue to the preferences of some Oligocene mammals.
A separate collection was made here in order to deter-
mine whether or not the fauna reflected the situation
suggested by the lithology.

Collections: Collection VI, SDSM Swampy Plains
Collection, 217 specimens; 1956. Graph VI.

114

FIELDIANA: GEOLOGY MEMOIRS, VOLUME 5

CURATING AND IDENTIFICATION

The 1956 collections were made through funds sup-
plied by the Yellowstone-Bighorn Research Foundation.
Originally housed in the Cleveland Museum of Natural
History, they were transferred to South Dakota School
of Mines and Technology Museum in 1957. Here the
senior author unpacked, prepared, identified, and
curated them. The 1959 collections, VII and VIII,
were prepared, identified, and curated at the latter
museum by the senior author. Following the senior
author's departure in 1961, all of these collections were
studied by a graduate student. Unfortunately, he had
no way of knowing the senior author's policy of counting
closely-associated material as one specimen. He there-
fore separated all associated material, giving it separate
numbers. In May, 1963, the senior author and this
student checked the collection against the senior au-
thor's original records, and reassociated scattered speci-
mens wherever possible. However, the present specimen
catalogue of the South Dakota Museum differs in detail
from the author's original specimen counts published
here, due to this confusion.

The University of Colorado Museum collections
were made by the authors jointly, with one day's co-
operation by Doctor Robinson and his party. They were
housed, identified, and curated in the senior author's
home, then sent to the University of Colorado Museum
for final storage. Identification was done without com-
parative material and with only limited access to pub-
lications. The specimen counts are accurate, and the
identifications are generally so, but work with com-
parative material may reveal a few individual mis-
identifications. These are probably the inclusion of new
forms within well-known species rather than assignment
to the wrong genera.

In general, identification to species has not been
attempted, only in part due to the fragmentary nature
of the material. The major reason is that almost every
Brule genus needs careful revision, and these collections
themselves revealed the wisdom of not accepting the
present specific designations. It is to be hoped that the
collections reported here will function as part of the
basis for thorough revisions at the specific level. Prob-
ably, such specific revisions will reveal differences be-
tween the various faunas which the present analysis at
a generic level fails to show. Merycoidodon was separated
into the easily recognizable species, culbertsoni and gra-
cilis, which have been placed in separate subfamilies by
Schultz and Falkenbach (1956).

Early in 1962, one private collector invaded the
near-stream area. He showed the senior author his
collection, which was later sold. It consisted of Meryco-
idodon culbertsoni, 28 specimens; Agriochoerus, 2 speci-
mens; Mesohippus, 19 specimens. Since these were col-
lected from the same horizon and locality as collection
IX, during the same year, they have been included in
this collection although they are not in the University
of Colorado Museum collection. Omitting them would
have produced a serious bias. The authors also found

several broken limb bones of Metamynodon, Hydraco-
don, Caenopus, and possibly Archaeotherium, at this
time, suggesting that the rarity of dental specimens
belonging to large mammals may be due to activities of
amateur collectors than than to any scantiness of large-
animal populations in the Oligocene near-stream faunas.

VARIABLES AND BIASES AFFECTING

INTERPRETATION OF THE

POPULATION STATISTICS

I. Preliminary considerations.

The ideas presented in this section have been ex-
pressed, in relation to studies of various mammalian
assemblages, many times. However, through their un-
usual habitat documentation the present collections
offer an almost unique opportunity for determining the
possible relationship between a collection of fossils and
the mammalian population from which it was derived.
We have evidence, first, that the collections represent,
essentially, death assemblages which have undergone
no transport; second, that these death assemblages ac-
cumulated in different local habitats within a few miles
of each other geographically; third, that they are truly
contemporaneous; and fourth, that no significant cli-
matic changes occurred during the time of accumula-
tion. The only remaining reasons for variation between
collections are the actual differences between faunas of
the different habitats, and the fidelity of the collections
as samples of those faunas. It is, therefore, justifiable
to repeat and organize ideas regarding validity of the
sample which have been expressed in part many times.

We shall consider here those variables and biases
which are known or presumed to affect the accuracy of
our samples. Four preliminary considerations must be
accepted as part of any census analysis of a fossil popu-
lation :

1. No complete census of any living mammalian
fauna has ever been made. We have, therefore, no
norm for direct comparison with a fossil population. In
view of the extreme modification of all modern faunas
by human action, it is improbable that any significant
census could be taken of any area, now or in the future.
Every biologist is aware that such broad general situa-
tions as: a collection consisting entirely of animals
larger than sheep; one containing 80 per cent carnivores;
or one including no rodents, are presumed to be abnor-
mal. Beyond vague generalities like these we simply
have no information.

2. We have no way of estimating the total size of the
population or universe of which the total collection is a
single sample. It is impossible, therefore, to evaluate
the significance of the sample.

3. A fossil collection must represent a death assem-
blage rather than a mechanical assemblage in order to
justify statistical analysis. Any mechanical assemblage
has suffered so much non-random selection as to be
worthless for census purposes. To illustrate this point:

CLARK AND KIETZKE: PALEOECOLOGY OF THE LOWER NODULAR ZONE

115

(o) Among the Glires, a single Ischyromys molar
will, due to its rounded shape, travel very much
farther by rolling than will a Paleolagus molar.
This would produce a sorting by shape rather
than by size.

,b) The enamel thickness and pattern of Merycoi-
dodon teeth make them very much more resistant
to breakage than Hoplophoneus teeth of the same
general size.

(c) One molar of Dapkoenus, if it survives trans-
port, is readily identifiable to genus, but a felid
molar is not, because the molars of the genera
Dinictis and Hoplophoneus are practically in-
distinguishable. Interaction of transport break-
age with a curating bias of this sort can permit
serious distortion of a census.

The predictable biases inherent to transportation
are profound and by no means mutually cancelling, nor
is there any reason to expect that the unforseeable
biases will average out the total. Those studies which
have been performed from time to time upon collections
from obviously-transported quarry assemblages are
interesting exercises in statistics, but they bear no
demonstrable relationship to ancient populations.

4. Finally, a collection of fossils is not, as Olson has
recognized (1957, pp. 312-313), a direct sample of a
mammalian population. Actually, the situation devolves
into three successive sample-universe stages, as shown
in Figure 53. Each circle comprises a sample of the one
above it, and a universe of which the one below is a
sample.

With these four general limitations in mind, let us
consider the variables which determine, either inde-
pendently or through interaction, whether or not a
particular animal will become an identified specimen
in a collection. Olson (1957) has discussed the applica-
tion of several of these factors to size-frequency distri-
butions; their bearing upon a faunal census is sufficient-
ly different to warrant discussion here. The factors fall
naturally into seven groups:

I Biotic

II Thanatic

III Perthotaxic

IV Taphic (related to burial)

V Anataxic (related to erosion and weathering)

VI Sullegic (Collecting)

VII Trephic (Curatorial).

These factors come into play at the stages indicated in
Figure 53. However, they may interact with factors at
any stage in the relationship. It is not possible to con-
sider any one universe-sample relationship in the chain
without reference to factors at other stages.

II. Biotic Factors.

These are features of the life of the individual or the
species, which determine the availability of individuals
for the successive stages leading ultimately to their in-
clusion in a collection.

A. Total range of species. This includes climatic,
topographic, and historic controls. Historic controls are
such features as the physical history which precludes
hippopotami from the Amazon basin, to which they
might otherwise be ecologically suited.

B. Habitat, or ecologic niche. In the present study
habitats are presumed known, and a major purpose is
to determine whether a significant correlation exists
between the collections of fossils and the predictable
faunas of the respective habitats.

C. Population density within the species.

D. Pressure to leave the preferred habitat. During
episodes of drought, animals from dry habitats would
seek swamps or streams for drinking water. They would
necessarily be at a comparative disadvantage while in a
habitat foreign to them, and might therefore be more
subject to predation than when at home. On the other
hand, flooding of drenaje swamp-forests might tem-
porarily drive some species out to the better-drained
slopes.

E. Osteologic construction. The simple equation,
fragile bone — greater destruction, is a serious over-
simplification of a factor which becomes very complex
due to interaction with perthotaxic, anataxic, and cura-
torial factors. For example, the skull of Paleolagus is
probably more fragile than that of any other Oligocene
mammal of equivalent size. Also, Paleolagus mandibles
tended to crack, during perthotaxy, between M2 and M3,
or posterior to M3, releasing those molars; not uncom-
monly, the entire cheek-tooth battery dropped out and
scattered. Moreover, single Paleolagus teeth often yield
to anataxis by splintering. By comparison, Leptomeryx
and Hypertragulus are much more ruggedly constructed.
However, due to unique generic characterization of
Paleolagus, almost every tooth fragment is generically
identifiable, while partial teeth or mandibles of Lepto-
meryx and Hypertragulus may not be. Therefore, al-
though good skulls of Paleolagus are among the rarest of
fossils, identifiable specimens are among the commonest.
To what extent this reflects original abundance of rab-
bits, or alternatively to what extent it represents a
tendency of perthotaxy to enhance the possibility of
finding recognizable specimens, is not known.

F. Body size. Body size, and hence bone size, is
another extremely complex, interacting factor. In cases
where transportation is not involved and where incre-
ments of sediment were thick enough to bury com-
pletely the largest bones, the chief interaction is with
thanatic, perthotaxic, anataxic, and collecting factors.
Some discussion of this will be given later.

III. Thanatic factors.

Thanatic factors are all those circumstances sur-
rounding the death of an animal which influence its
fossilization potential.

A. Cause of death. Predation, disease, physical acci-
dent, poison, starvation, and intra-specific strife com-
prise the overwhelming majority of causes of death.
Authorities differ as to the relative importance of these

116

FIELDIANA: GEOLOGY MEMOIRS, VOLUME 5

causes; certainly animals differ from species of species,
and within one species from time to time and from place
to place, as to the relative proportions dying from these
different causes.

The effects upon a population census of these various
causes of death, as operating upon any one species,
would be quite different. Physical accident would, in
such habitats as those described, be limited to occasional
mass drownings of small or burrowing forms and to falls
by arboreal species. The latter would average out over a
period of time. Mass drownings might produce a pri-
mary dilated perthotaxis of small animals, but no such
situation has been observed. Poisoning by ingestion of a
toxic plant, such as occasionally occurs to cattle who
eat Death Camas, should produce a dilated perthotaxis
of herbivores; it is certainly a rare situation among
wild populations, if indeed it occurs at all. Epidemic
disease might be expected to produce mass deaths with
a high incidence of individuals of whatever age group
was susceptible. This is believed to be the case with
Hypertragulus (see below). Starvation would ordinarily
be a result of protracted drought causing a shortage of
both food and water in the dryer habitats. It would
almost certainly result in an abnormally large number of
dry-plains forms dying in swamp and river-bank forest
habitats. The larger animals would migrate to water
more easily than the smaller ones, so the death as-
semblage in the dryer habitats would be biased toward
smaller animals. Intraspecific strife would be a con-
tinuum which should not affect a fossil census in any
way.

Predation might produce serious biases, which can
be guessed at but cannot be accurately assessed. Owls
usually destroy the calvarium, but regurgitate the
maxillae and mandibles unharmed. Hawks, eagles, and
vultures swallow, digest, and thereby completely de-
stroy the bones of small animals. Heavy predation by
these birds should constitute a definite bias against all
small, diurnal mammals in a census of bones. The rarity
of avian fossils precludes any evaluation of this effect.
Predation by small mustelines (rare in Brule collections)
almost never damages jaws and teeth; presumably this
would introduce no bias. Carnivores the size of Daphoe-
nus and the felids might be expected to crush and swal-
low any mammals up to about 5 per cent of their own
body weight, which would again produce a bias against
everything smaller than Ischyromys. Hesperocyon, the
commonest Brule carnivore, would be expected to
destroy Eumys and everything smaller. Ingestion of
small animals whole does certainly dissociate bones and,
by immersion in gastric juices, remove all organic
matter from them. Such a bone emerges from the
carnivore in a fairly advanced state of perthotaxy within
hours of death, which greatly reduces its chance of
preservation. Predation is, therefore, certainly a factor
operating selectively against small mammals and, in the
case of carnivorous birds, against small diurnal mam-
mals.

In summary, mass drownings and epidemics might
produce dilated perthotaxes, starvation would probably

cause invasion of moist habitats by normally dry-land
forms, and predation would bias a census against the
smaller forms. The other causes of death would prob-
ably not recognizably affect a census.

B. Locus of death. Relationship to the surface of the
ground seems to be a decisive factor. We have found no
evidence of any mammal burrows, or, naturally, of any
mammals who died in them. Supposedly, a burrowing
mammal might be expected to die in its burrow from
any of the causes listed except predation by large
carnivores. Burrows in a flood plain subject to flood
every ten years should, therefore, include large numbers
of skeletons either whole or in early stages of perthotaxy.
The actual situation is exactly opposite: not only have
no burrows been found, but associated skeletal material
of small, probably burrowing, mammals is almost non-
existent. This could mean either that no Oligocene
mammals were burrowers, which seems improbable, or
that individuals living in areas subject to frequent
sedimentation did not dig burrows, which seems much
more probable. The senior author has observed that on
Badlands flats at present, burrows are dug only in those
places not subject to deposition. The effect of this factor
on a population study is not known.

The other peculiar locus of death lies some distance
above the ground, in trees and shrubs. Presumably, an
animal dying of predation in this situation would be
destroyed, and an animal dying of any other cause would
probably do so in a hollow tree, where it could be buried.
However, there is nothing to prevent animals who build
nests among branches, as do many squirrels, from falling
to the ground within a few weeks or months of death.
Whether or not this occurs is debatable: the generally
low percentage of primates and other demonstrably
arboreal forms in most fossil assemblages would suggest
that it does not. One can presume that the present col-
lection has some bias against truly arboreal forms, al-
though a definite assignment to arboreal or terrestrial
habitat is impossible for many of the genera represented.

C. Mortality relative to age. In the Brule collections,
as in practically all other Tertiary collections, infant and
juvenile specimens are exceedingly rare. One may set
up the following arbitrary, but usable, scale:

Stage Terminated by

Infancy Eruption of M1,

Juvenility Eruption of M\

Adolescence Eruption of M3j

On this basis, adolescent individuals are generally un-
common (except in Ictops, Mesohippus, and Hypertra-
gulus, see below) ; juveniles rare, and infants exceedingly
rare. Since most mammalian species experience an
infant-juvenile mortality of 25-75 per cent, the fossil
collections very apparently do not approximate the life
assemblages. Presumably, the fragile bones of young
individuals are more readily destroyed by perthotaxy
and anataxy than are adult bones.

This sets up a strong bias against animals having a
high infant-juvenile mortality, which are usually small
animals with large litters. For instance, if a species has

CLARK AND KIETZKE: PALEOECOLOGY OF THE LOWER NODULAR ZONE

117

within a given area an adult population of 1000 in-
dividuals, which produces 1000 young per year 90 per
cent of whom die in infancy, then due to the differential
preservation of adult bone the population effective in
producing fossils is very little above 1000. A species of
1000 adult individuals in the same area, producing 500
young per year, 20 per cent of whom die in infancy,
would on the other hand have a fossilization potential
of 1400 + individuals.

Thanatic factors combine to produce the death as-
semblage, which is the sum of the fresh corpses that
come to rest upon a surface previous to burial of that
surface by the next episode of sedimentation. As used in
this paper, the death assemblage is not the partial or
complete perthotaxis observable at any one time.

IV. Perthotaxic factors.

Climate and exposure are the chief variables con-
trolling a perthotaxic system. Climate can be presumed
to have been uniform in the case at hand, and exposure
would be not too different within any one habitat.
Physical accident, such as being stepped upon by a
large Oligocene mammal, would have destroyed speci-
mens in a truly random fashion; probably accident
would destroy only a negligible number of specimens.
Body size, and especially tooth volume relative to
enamel thickness, seem to have had much more im-
portant effects upon perthotaxy. Observation upon
recent skeletons under semi-arid conditions demon-
strates that animals of rabbit size and smaller undergo
almost no dehydration-cracking of the teeth, and much
less splintering of limb bones than do the larger forms.
Very large bones — cow and horse sizes — exfoliate but
do not splinter so readily as do the bones of animals of
medium size. Large teeth, on the other hand, splinter
very quickly. These aspects of perthotaxy would, there-
fore, produce a bias favoring preservation of the smaller
forms.

Scavenging-pressure is more difficult to evaluate.
Once the flesh was removed from an Oligocene mam-
mal's bones, apparently, it was of little interest to
scavengers — very few instances are known of a partial
skeleton with limb bones chewed or broken off. Con-
versely, numerous instances have been observed of
undamaged herbivore skulls accompanied by several
droppings of carnivore coprolites. This suggests that the
skulls were of no gustatory interest, and functioned
either as objects of scorn or as markers. Plainly, some
process must have very quickly and effectively removed
the flesh from many corpses, rendering them uninterest-
ing before they suffered damage from scavengers. Either
decay or insects could have accomplished this. The
universality of the removal suggests both, operating
under a climate considerably warmer and somewhat
wetter than the present; Payne (1965) outlines the
process under a summer climate in Virginia. Briefly,
there is no evidence that scavengers destroyed a
significant number of specimens, although they prob-
ably scattered the bones of the larger species.

RELATIONSHIP OF A LIFE ASSEMBLAGE
TO A COLLECTION OF FOSSILS

DETERMINED BY
BIOTIC FACTORS

THANATIC FACTORS

PERTHOTAXIC FACTORS
TAPHIC FACTORS

ANATAXIC FACTORS
SULLEGIC FACTORS
TREPHIC FACTORS

Fig. 53. The relationship of a life population to a collection of
fossils. The interaction of factors has not been shown. Also, since
factors would function differently for each genus in every collection,
the size of the assemblages is not representative.

V. TAPHIC (BURIAL) FACTORS.

Six major factors of burial can, theoretically at least,
result in a difference between the perthotaxic assem-
blage at any one time and the taphic or total fossil
assemblage.

1. Time interval between episodes of sedimenta-
tion.

2. Thickness of sedimentary increments.

3. Velocity of depositional current in contact
with bone or corpse.

4. Nature of sediment: amount of compaction,
grain size, nature of clay minerals.

5. Post-depositional action of roots and burrow-
ing animals.

118

FIELDIANA: GEOLOGY MEMOIRS, VOLUME 5

6. Permeability of compacted sediment and na-
ture of permeating solutions.

In the Lower Nodular Zone, there is no evidence
that any of these functioned to destroy bone selectively,
or in any way to bias the samples. The time interval
was always long enough to permit development of a
complete perthotaxis. Sedimentary increments were
thick enough to bury completely even the largest
bones. The sediments were deposited by fluids that
engulfed without transporting or abrading. The sedi-
ment was a gritty siltstone, with enough montmoril-
lonite to bind and give it adhesive strength but not
enough to cause much damaging compaction. There is
no evidence of bone having been damaged by Oligocene
roots or burrowing animals. Such solutions as filtered
through the highly permeable elastics were thoroughly
charged, at various times, with lime, silica, and iron
salts; they preserved the bones rather than dissolving
them.

Generally speaking, therefore, the total fossil popu-
lation (see Fig. 53) represents a very large proportion of
the death assemblage at the several times of deposition
of the sedimentary increments. It had been altered only
by perthotaxic biases, not by any related to burial.

VI. Anataxic factors.

Anataxy can be defined as the sum of the phenomena
which operate to expose and to destroy a fossil. It oc-
curs in three inter-related states: A). Weathering in situ
of matrix and fossil. B). Exposure by erosion. C).
Weathering and transportation after exposure.

A. Weathering in situ. Under the present climatic
regime, one of the first effects of weathering is the pre-
cipitation of hematite around bone. This occurs in the
near-stream sediments, but not in the Sage Creek or
Dillon Pass areas. Bones of Leptomeryx size and larger
develop a heavy aureole of hematite; smaller one de-
velop less or none. In the Sage Creek area, weathering
tends to break medium-sized bones, within the top
inch of the weathered zone, but does little or no damage
to jaws and teeth of Ischyromys size or less. Since generi-
cally identifiable fragments of the larger bones survive
this weathering, little if any bias results.

B. Exposure by erosion. Erosion is almost always by
rain-wash, rill-wash, or by gravitational sloughing of
weathered material from hillsides during spring thaws.
The latter may seriously damage or destroy specimens
of all sizes, but particularly the larger ones. However,
this type of erosion is of minor importance in the fos-
siliferous areas in question. Rain and rill-wash effective-
ly separate bone particles already broken by weather-
ing, and also separate parts broken by perthotaxy. The
former case can usually be recognized, but the further
separation of a pair of small jaws originally scattered by
perthotaxy may result in one small animal being counted
twice. For instance, a pair olEumys jaws originally 3 in.
apart and exposed at the same time could, if one were
caught by a rill, come to rest 20 or 30 ft. apart during a
single rain. This does produce a bias favoring the smaller

forms, because parts of a larger animal are less easily
moved and are more easily recognized as belonging to
the same individual.

C. Anataxy after exposure. Ultimately, fossils are
destroyed by weathering and transportation. A few
skulls armored by the hardest nodules, in the Cotton-
wood Pass area, have been known to travel up to one-
quarter mile with little damage. However, fossils not so
protected do not usually withstand transport over 100
yards.

Medium-sized fossils not subject to transportation
endure surface weathering for periods of up to tens of
years in the Cottonwood Pass area. Any parts project-
ing from the nodules break off, and the nodules often
break into angular fragments, but the bones develop an
extremely durable patina of iron and manganese oxides.
This does not usually occur to bones of Leptomeryx size
and smaller, which apparently weather to destruction
within years. Due to the recent activities of amateur
collectors, very few identifiable, weathered specimens
were collected, so our sample probably was not biased
toward large forms as it might otherwise have been.

In the Sage Creek area, weathering of exposed bone
is not rapid. Fragments of Archaeotherium bones, up to
2 in. long, exposed in 1956, were apparently destroyed
by 1964; on the other hand, a piece of zygoma 6 in. long
was virtually unweathered. In this area, jaws and teeth
of small animals seem to be at least as resistant as large
ones, and may be more so. Such bias as might occur
would favor the preservation of small specimens.

VII. SULLEGIC OR COLLECTING FACTORS.

At any one time, a relatively small proportion of the
total fossil population of a stratum has been exposed
and is in the process of slow destruction. The collection
of the most complete sample possible of this exposed
group is attended by many difficulties, which may pro-
duce serious biases.

Five factors have proved to be particularly applic-
able to the Scenic Member collecting problem: A).
Method of prospecting. B). Position of collector in
visual prospecting. C). Personal bias. D). Historic re-
sampling. E). Differential cementation and weathering.

A. Method of collecting. The traditional method of
collecting consists of driving, riding, walking, creeping,
or wriggling across fossiliferous terrain, observing fos-
sils visually, and collecting them by whatever means is
thought best. The senior author has personally observed
cases of prospecting from car or horseback; these can
obviously be ignored as serious methods of obtaining a
representation of any but the largest forms. Other
methods of visual collection will be discussed in para-
graph B.

A second, more modern method of collection is by
stealing the grit from ant hills and concentrating the
tooth and bone fragments which the ants have picked
up. Ant hill concentrates are very useful for giving a
representation of the total small-mammal fauna, but
they are subject to three biases, mechanical, psycho-
logic, and ecologic.

CLARK AND KIETZKE: PALEOECOLOGY OF THE LOWER NODULAR ZONE

119

Mechanically, the ant is limited to particles small
enough to carry and large enough to remain on the hill
through rain and wind. Any fossil species with generical-
ly identifiable jaws, teeth, or fragments thereof which
fall within this range is very likely to be represented in
an ant hill. The size bias is obvious and predictable.

A more serious bias, difficult to evaluate, relates to
the psychology of the ants. The senior author has ob-
served two instances of this:

On the south rim of 71 Table, in Sec. 16 and 17,
T. 3S., R. 14E., Pennington Co., the ants select grit
from Pleistocene gravels for their hills. They show a
strong preference for garnets over quartz and feldspar
grains of the same size. Do they prefer garnets because
of the shape, color, texture, or specific gravity?

East of the main road in the SWM, Sec. 29, T. 5S.,
R. 8E., Custer Co., the ants are selecting from materials
available on the Sharon Springs Member of the Pierre
Formation. Two ant hills observed are composed almost
entirely of glistening, blade-shaped selenite crystals 3/4
in. long; the ants have abundant pellets of limonite and
small, gray calcite concretions of suitable size, equally
available. Since the limonite has a higher specific gravity
than the selenite, do the ants prefer the selenite for its
shape, color, luster, or surface texture? One hesitates
to ascribe to them an esthetic appreciation of either
glistening selenite or red garnet.

These two instances clearly document a deliberate
selection upon some basis other than size and, in the
second instance, other than specific gravity. All of the
selenite crystals are at the upper limit of the ants'
carrying capacity, and are much less easily moved than
a round limonite pellet of the same weight, showing
that portability is probably not the deciding factor. If
shape happens to be decisive, might not the ants select
Paleolagus teeth (which are of almost identical size and
shape with the selenite) over round Ischyromys molars?

The psychologic bias of ants is, in the authors'
opinion, plainly demonstrated. Until more is known of
the basis for their preferences, all ant-hill collections
should be regarded as subject to definite but unpre-
dictable bias other than size. Such collections are there-
fore not fit subjects for statistical studies aimed at ap-
proximating a census of fossil communities.

The ecologic bias constitutes a third serious factor
limiting the statistical usefulness of ant-hill collections.
One particularly fossiliferous locality may form the
homesite of 60 or 80 ant colonies, while another of equal
size may support none, or at best two or three. The
distribution of small vertebrate fossils within such a
horizon as the Lower Nodular Zone is non-random but
unpredictable, and the distribution of ant hills is cer-
tainly non-random (since it is controlled by geographi-
cally variable ecologic factors) but is equally unpredict-
able. We therefore have ant-hill collections of widely
variant size from different localities; the collections
share a qualitative psychologic bias, but differ in size-
of-collection bias. We do not know whether or not the
distribution of ant hills might be controlled by some

factors of the local geology which themselves reflect
paleogeography: the distribution of recent ant hills
might actually be influenced by factors interacting as
controls of fossil distribution.

Because of the psychologic and ecologic biases, plus
the interaction of size, perthotaxic, anataxic, and cura-
ting biases, ant hill collections do not constitute samples
suitable for population studies. The difference in method
of sampling from visual prospecting makes it obviously
impossible to merge the two types of collection for any
quantitative analysis. The real usefulness of ant hills
lies in their concentration of very small forms, which
greatly increases qualitative knowledge of the total
fauna.

The third sampling method consists of washing
samples. The matrix itself is washed wherever: (a) fos-
sils are sufficiently abundant; (6) fossils are dissociated
enough that washing will not further damage them;
(c) the matrix dissociates in water. Conditions (b) and
(c) obtain nowhere in the Lower Nodular Zone. Placer
concentrates of bone in small rill-gullies are common,
however, and can be washed out. They frequently in-
clude jaws and teeth smaller than can be reliably re-
covered by visual collecting, and as such are qualitative-
ly useful. The fact of Recent transportation automati-
cally injects mechanical biases. The difference in sam-
pling method precludes using them in combination with
collections made by any other method, to form a "total
population" for quantitative analysis. As previously-
transported fossils, any matrix-washed collections would
be unsuitable for quantitative population analysis.

B. Methods of visual prospecting. During the 1965
field season the authors experimented with different
methods of visual prospecting in the Sage Creek dry
plains locality. We first covered the fossiliferous area
systematically, on hands and knees, with eyes about
2 ft. from the surface, collecting several hundred
identifiable specimens, mostly Leptomeryx and smaller
forms. We next covered the same area systematically by
walking in a stooped position, with eyes 3-4 ft. from
the surface. To our surprise, we found five Merycoidodon
skulls that we had missed on the closer search, plus a
scattering of smaller specimens. We then chose an area
about 20 ft. by 10 ft. which we had prospected repeated-
ly, directly adjacent to our camp, and prospected it by
crawling, with our eyes not over 12 in. from the ground.
We found about 30 specimens, at least 12 of which were
tiny insectivores not previously represented. We then
tried the adjoining space of the same size, and found
only one insectivore plus a few fragments of Paleolagus,
small rodents, and Leptomeryx teeth. This experiment
demonstrated to our satisfaction that an area must be
prospected several times, moving in different directions,
and certainly using all three ocular positions, in order
to achieve an approximately complete collection. It
also demonstrated a high local variability in distribu-
tion of fossils.

C. Personal biases. The training, persistence, psycho-
logic fluctuations, and visual acuity of the collector

120

FIELDIANA: GEOLOGY MEMOIRS, VOLUME 5

constitute obvious and well-known biases. We have
found that physical distractions, such as gnats, or
mosquitoes, or excess sweat dripping into the field of
vision, definitely reduce collecting efficiency, especially
ability to see small fossils. Available time becomes a
factor through reducing the possible number of times
one can prospect over an area. It may also become a
psychologic factor, by causing the collector to hurry and
thereby to miss small fossils; the latter tendency is more
prevalent in the chief of the party than in his assistants.

D. Historic resampling. Collection by most profes-
sional collectors of an earlier age certainly interacts
with size, perthotaxy, and anataxy to produce a bias.
The senior author collected over all of the districts
discussed in this paper, during the decade 1932-1942,
and well remembers several significant points.

Large fossils were at that time almost everywhere in
evidence. During one ten-day stay in Sage Creek in
1933, our party averaged well over one skull per day of
animals Mesohippus size and larger. Such collecting
no longer exists in the Big Badlands. We are forced to
the conclusion that exposure by weathering is not so
fast as has been supposed. Once an area is really
thoroughly collected, several decades must elapse be-
fore it will attain its original concentration of large
bones.

Earlier collecting was strongly biased toward larger
specimens, and especially toward more complete ones.
The senior author recalls that in 1933, a party of four in
Sage Creek collected less than 100 specimens of Lep-
tomeryx size and smaller, in 10 days. By comparison, we
two made collections I and V, totalling 672 specimens of
which over 70 per cent are small, in the same area in
two days in 1962.

The size bias of earlier collectors, plus a slow rate of
exposure by erosion, produces a bias favoring smaller
forms in any recent collections. Since this bias depends
upon interaction of the ratio of small to large forms in
the death assemblage, perthotaxic factors, the actual
rate of erosion, and the extent of previous collection, it
cannot be evaluated.

E. Bias due to differential cementation. Collecting
effectiveness depends in a complex fashion upon ce-
mentation of the matrix. High percentages of montmo-
rillonite with little cementation produce a "pop-corn"
surface. Most small teeth and jaws are destroyed by
expansion and contraction, and the remaining ones
settle between lumps of matrix where they are easily
overlooked.

Tight cementation, as in the near-stream nodules,
tends to interact with collecting biases to favor finding
large fossils. A single nodule containing a single, large
skull is highly visible. Small bones either weather free,
to settle among the nodules, or weather out of a large
nodule a little at a time, suffering destruction in the
process and never becoming easily visible.

VIII. Curatorial factors.

One can assume proper field, laboratory, and cata-
loging procedures in a modern institution. However,

identification introduces a serious bias which the most
conscientious effort cannot remove. A monospecific
genus or monogeneric family can, in general, be recog-
nized from a smaller tooth-fragment than can a species
or genus having close relatives. An outstanding compari-
son can be drawn between Mesohippus, which can often
be identified from less than one-tenth of a molar, and
the rhinoceroses. A small Caenopus and a large Hyra-
codon, on the one hand, or a large Caenopus and a small
Subhyracodon, on the other, can sometimes be distin-
guished only with difficulty on the basis of half of a
molar. A larger fragment is needed to distinguish
Leptomeryx from Hypertragulus, than to determine be-
tween Merycoidodon and Agriochoerus. A few rodent
genera can be determined on the basis of half of a
molar; several cannot be. The curatorial bias is not at
all random nor is it uniformly reinforcing or mutually
self-cancelling to any other bias. In practice, it becomes
a very significant bias.

IX. Summary.

The life assemblage is determined by biotic factors.
Thanatic factors produce a considerable bias in the
death assemblage, which is a sample of the life assem-
blage. Perthotaxic factors considerably, and taphic
factors slightly, bias the total fossil assemblage, which,
in turn, is a sample of the death assemblage. Anataxic,
collecting, and curatorial factors strongly bias the fossil
collection, which is a sample of the total fossil assem-
blage. Many of these biases interact; there is good
reason to suppose that they are not mutually self-
cancelling. A census study of a fossil population must
face, in addition to these groups of biases, the facts that
we do not know the size of the life population, we have
no census of a Recent mammalian population to use as
a norm, a collection must be demonstrably a burial of
bones in situ rather than a mechanical assemblage, and
that the collection is not a direct sample of the life
population.

ANALYSIS OF FAUNAS1

I. Validity of collections as samples.

The bias factors discussed above cast doubt upon
the validity of our collections as representative sam-
ples, ultimately, of the Oligocene faunas of their respec-
tive areas. It is therefore necessary to consider Charts
I-XIII, in order to determine whether sufficient evi-
dence of consistency exists to justify further study.
Foreknowledge of the habitats makes possible a search
for consistent differences between collections from dif-
ferent habitats as well as consistent similarities between
collections from the same habitat. Since only one col-
lection has been made from each of the "swampy plains"
habitats, and since the two have been interpreted from
different geologic bases, attention must focus upon the
near-stream and the open plains collections. Figure 55

1 Throughout this section the following abbreviations will be
used in the text for brevity and clarity: SDSM — South Dakota
School of Mines and Technology Museum: CU — University of
Colorado Museum.

CLARK AND KIETZKE: PALEOECOLOGY OF THE LOWER NODULAR ZONE

121

lists two characteristics shared by all collections, and
six in which the two sets of collections are internally
consistent but different from each other.

Paleolagus, as the ecologic homologue of Sylvilagus,
might be expected to be both ubiquitous and ubiqui-
tously abundant. The abundance of Ischyromys, how-
ever, might be actual or might be due to collecting
biases, since it is one of the largest Oligocene rodents.
This will be discussed later. The low, generally con-
sistent percentage of carnivores throughout suggests
that the collections probably approximate the life as-
semblages in this respect.

The other six characteristics listed in Figure 55,
taken together, suggest that the collections are adequate
to reveal broad generalities about the life assemblages
of the different habitats. Chart XIII offers some sup-
port to this presumption. Of ten proportions computed,
the SDSM swamp collection resembles the near-stream
collections more nearly than the dry plains collections
in the following five:

1. Percentage of Glires.

2. Percentage of perissodactyls.

3. Percentage of small mammals.

4. Percentage of medium-sized mammals.

5. Percentage of perissodactyls to artiodactyls.

Three more show no significant differences between
collections:

1. Percentage of carnivores.

2. Percentage of artiodactyls.

3. Percentage of large mammals.

Of the remaining two, the percentage of Merycoido-
don fluctuates irregularly, and the SDSM swamp is
uniquely high in percentage of Mesohippus. Since the
faunal assemblage of a swamp could confidently be ex-
pected to resemble that of a near-stream fauna more
nearly than the fauna of a grassy plain, this strong re-
semblance is added evidence that our collections are
significant samples.

The Colorado swampy-plains collection more nearly
resembles the open plains collections than the near-
stream collections. The geologic criteria for separating
out the Colorado swampy-plains from the open plains
environment are much less positive than are those for
the SDSM swamp. Again, the faunal resemblance
parallels the paleogeographic situation, reinforcing the
conclusion that the collections constitute significant
samples of the Oligocene life assemblages.

However, the high variance between collections
should also be noted. Variations of up to 10 per cent
between corresponding generic populations in Graph I
(531 specimens) and II (458 specimens) increase to al-
most 20 per cent between collections VII (139 speci-
mens) and IX (115 specimens). Recorded variations of
more than 20 per cent for the smaller collections indicate
that these are of little value except as they may justi-
fiably be combined with others for more general infor-
mation.

Using these observed variations as rough limits of
significance for collections of the approximate sizes
noted, we shall proceed to interpretation of the Oligo-
cene communities of the various habitats.

II. Near stream faunas. Graphs VII, VIII, IX, XII.

A. General. The graphs demonstrate clearly that the
fauna comprises animals of varied sizes, with medium-
sized species in the majority. Mesohippus (24%) and
Merycoidodon (31 %) make up the bulk of the fauna, as
has been known for many decades. Unfortunately, the
turtle Stylemys could not be included in the collections
or the statistics; as a casual observation, there were at
least as many turtles as Mesohippus in the field.

With a total of 296 specimens divided among 26
mammalian genera, the average number of specimens
per genus is 11.4. Mesohippus and Merycoidodon, the
two commonest genera, are represented by 55.4 per cent
of the total number of specimens. The order of frequency
among the commoner rodent-rabbit genera is: Ischyro-
mys, Paleolagus, Eutypomys, Megalagus, Eumys. The
rodent-rabbit population amounts to only 20 per cent of
the total fauna. Small artiodactyls (Hypertragulus and
Leplomeryx) constitute 4 per cent of the total. Carni-
vores comprise 5 per cent, with eight felids, five canids,
and two hyaenodonts. Perrissodactyls account for 28
per cent of the total fauna, and are 64 per cent as
abundant as artiodactyls.

B. Individual collections. The most striking differ-
ences between individual collections are the much
higher percentage of rodent-rabbits in the CU collection
— 42.6 per cent as opposed to 5.7 per cent in the SDSM
collection — and the absence of Hypertragulus and Lep-
tomeryx in the CU collection. Field assistants who helped
make the SDSM collections were relatively inexperi-
enced, which might cause them to miss the smallest
fossils. However, this certainly would not explain why
the well-trained collectors who made the CU collection
found 49 rodents and rabbits but no hypertragulids.
Some extremely local phase of Oligocene geography may
be reflected here, but if so no evidence of it has been
detected in the sediments.

III. Sage Creek open plains faunas. Graphs I, II,

V, X.

A. General. The SDSM Open Plains Collection,
Graph II, was made over the combined areas of CU
Open Plains Collection, Graph I, and CU Swampy
Plains Collection, Graph V. In order to maintain a uni-
form basis for comparison, the two CU collections are
considered here, and both are included in Graph X. The
distinctive characters of Collection V will be noted in
Section B of this discussion.

In general, the Open Plains collections are almost
four times larger than the near-stream ones. This re-
flects an actual greater abundance of specimens. The
senior author collected 60 specimens in 30 minutes on
one memorable occasion, and the total CU collection of
672 specimens represents the authors' combined efforts
for two rather short working days in late September.

122

FIELDIANA: GEOLOGY MEMOIRS, VOLUME 5

Even during the senior author's 1930-1934 experience
in the classic near-stream area, before heavy prospecting
had noticeably modified it, no comparable number of
specimens was found there.

A complete perthotaxis has been observed in Sage
Creek, as at Cottonwood Pass. Proportionally fewer
whole skulls and skeletons, and far more separate jaws,
occur at Sage Creek; the difference becomes less
noticeable among the animals of medium size. Possibly
this phenomenon reflects merely the large number of
rodents and rabbits at Sage Creek rather than a qualita-
tive difference in perthotaxy in two localities. Pertho-
taxy operates disproportionately faster on post-cranial
bones of smaller animals. Conversely, tooth damage by
dehydration in museum osteological collections demon-
strates that teeth of small rodents lie undamaged for
years in the same atmosphere which reduces teeth of
sheep-sized animals to splinters.

A small-animal perthotaxis in a community of
mixed sizes should, therefore, comprise a disproportion-
ately large number of teeth and jaws, with limb bones
and skeletons rare to absent. This is certainly the case
in the Sage Creek Open Plains faunas.

The second notable feature is the great predomi-
nance of small animals over medium-sized and larger
ones — 81 per cent of the total, compared with 27 per
cent in the near-stream fauna. Since both sets of collec-
tions represent complete perthotaxies, this difference is
actual rather than an artifact. Perthotaxy probably did
increase the number of small specimens preserved rela-
tive to large ones, within each area, but it could not
have caused a significantly greater preservation of
small forms in one area relative to small forms in the
other.

With 1127 specimens divided between 31 mam-
malian genera, the average number of specimens per
genus is 33.2. Leptomeryx and Paleolagus, the two com-
monest genera, are represented by 54.8 per cent of the
total number of specimens. This compares with Meso-
hippus and Merycoidodon, which together comprised
55.4 per cent of the near-stream fauna.

The order of frequency among the commoner Glires
genera is: Paleolagus, Ischyromys, Eumys.

The rodent-rabbit group makes up 37 per cent of the
mammalian community, compared with 20 per cent in
the near-stream population. Small artiodactyls have
increased in number of genera from two to five, and in
population from 4 to 40 per cent. Carnivores remain at
5 per cent, two hyaenodonts, with 40 canids, three
mustelids, and eight felids. Artiodactyls dominate the
population, both in number of genera and in number of
individuals. They constitute 50 per cent of the total
population, divided between 12 genera, contrasted with
44 per cent of the population divided between seven
genera in the near-stream fauna. Small artiodactyls ac-
count for most of this; five genera include 40 per cent of
the total population, compared with two genera com-
prising 4 per cent of the near-stream community. Peris-
sodactyls are, conversely, greatly reduced in impor-

tance, but not in number of genera. The same four
genera comprise 7 per cent of the open-plains popula-
tion and 28 per cent of the near-stream population. This
reflects the increase of small animals rather than any
considerable change within the perissodactyls them-
selves; perissodactyls make up 39 per cent of the medi-
um and large-sized animal population in the near-stream
fauna, and 41 per cent of the corresponding community
in the open-plains fauna.

Summarizing, the open-plains population differs
from the near-stream population by the addition of
tremendous numbers of small animals. Leptomeryx
shows the greatest increase, followed by Paleolagus,
Ischyromys, Hesperocyon, Hypertragulus, and Eumys.
Changes of proportion within the medium-to large-
sized elements of the fauna are relatively unimportant,
when they are considered separately from the small
animals.

B. Individual collections. The two open-plains col-
lections show a somewhat higher consistency than do
any two of the near-stream collections, as might be
expected from the greater size of the open-plains col-
lections. The large number of Hypertragulus in collec-
tion I is surprising, and the explanation is not known.

Collection V, the CU swampy-plains collection,
shows several marked differences from the other two.
Although the concentration of Leptomeryx remains as
high as in the open-plains populations, Eumys is rare,
Paleolagus, Ischyromys, and Hypertragulus have de-
creased, and several of the less-common small genera
are absent or rare. Mesohippus + Merycoidodon have
increased, relative to the total fauna. This increase is
not, as the following table shows, a simple gain by
default of the smaller animals:

Ratios to Large and Medium-sized Populations.

Collection

Mesohippus

Merycoidodon

XII Total Near-Stream

33%

43%

X Total Sage-Creek Open Plains

32%

41%

V CU Swampy Plains

41%

36%

Since the population of Mesohippus has risen markedly,
while the population of Merycoidodon has fallen, rela-
tive to their size-group, one may presume that this con-
stituted a most favored niche in Mesohippus' local
range.

These population changes, involving Mesohippus,
Merycoidodon, Eumys, and to a lesser extent Ischyromys
and Paleolagus, support the hypothesis that this locality
differed geographically from the adjacent collecting
area immediately to the east. Collection II, SDSM
open-plains collection, therefore represents a mixed
fauna. Probably the collecting area of Collection V
represents merely a plain with broken forest cover sup-
ported by sub-irrigation, rather than an actual swampy
prairie.

IV. Dillon Pass Area Open-Plains Collections.
Graphs III, IV, XI.
A. General. Varzea and igapo sediments meet in the
locality where these collections were made. Paleogeo-

CLARK AND KIETZKE: PALEOECOLOGY OF THE LOWER NODULAR ZONE

123

graphically, this should mean that the influence of the
nearest through-going streams was minimal; local hill-
side wash brought as much sediment and water into
the area as did the streams. Vegetational cover should,
therefore, also have been minimal, and the population
should be pure open-plains. Since the collections from
this area did not give conclusive evidence that they
represent a death assemblage (although there is no
evidence against it), they have been regarded as doubt-
ful, to be judged in terms of the more satisfactory Sage
Creek collections. Their comparatively small size also
limits their usefulness.

Comparison of Graph XI with Graph X shows im-
mediately that the same population trends are present
but more accentuated at Dillon Pass. Small animals
make up 91 per cent rather than 81 per cent of the fauna.
The commonest forms are Leptomeryx, Paleolagus,
Ischyromys, and Hypertragulus, in that order. Car-
nivores continue between 4 and 5 per cent, with Hes-
perocyon the prevalent genus. Mesohippus and Mery-
coidodon are present, but in greatly reduced numbers.
Between them they make up 3 per cent of the popula-
tion, rather than 11 per cent as in the Sage Creek open-
plains collection. Furthermore, both have been reduced
relative to the total large and medium-sized population,
Mesohippus from 33 per cent to 20 per cent, and Mery-
coidodon from 41 to 13 per cent. The general aspect
of the assemblage is one of trends carried beyond those
at Sage Creek. The accordance of this faunal com-
munity with the facies predictable from the paleogeo-
graphy suggests that these small collections represent
perthotaxic assemblages, just as the larger ones at Sage
Creek do.

B. Individual collections. The individual collections
show no peculiarities worthy of note, due partly to the
high variance induced by their small size.

V. SDSM Swampy Plains Fauna. Graph VI.

A. General. Lithologic evidence that this was once a
swamp is much stronger than the evidence indicating a
swamp environment for Collection V, which actually
shows merely some indication of more sub-irrigation
than the adjacent areas. Collection VI might logically
be expected to resemble the near-stream collections even
more than does Collection V. It should differ appre-
ciably from the open-plains collections, even though the
edges of the two collecting areas are within 200 yards of
each other.

The collection amply bears out this prediction.
Medium-sized animals make up almost half of the
population. The rodent-rabbit group is down to 10 per
cent of the total, with Ischyromys the only genus rela-
tively common. Leptomeryx has dropped to 23 per cent,
while Hypertragulus has risen to 11 per cent. However,
the increase in Hypertragulus may be due to a biologic
accident: this collection includes a herd of at least 22
individuals, all young adults, which died and were
buried within a four-foot radius. The single mass death
has almost certainly unduly weighted the count of

Hypertragulus, and also the percentages of artiodactyls
and small animals.

Mesohippus apparently found its favorite environ-
ment here. It formed 30 per cent of the entire fauna, and
59 per cent of the large-to-medium sized fauna, con-
siderably higher than in the near-stream assemblage.
Conversely, Merycoidodon drops to 12 per cent of the
total and 24 per cent of the large-to-medium-sized
population.

Generally, the swampy plains community was low in
rabbits and rodents (except Ischromys), low in Lepto-
meryx and Merycoidodon, and very high in Mesohippus.
It may also have been high in Hypertragulus, but was
not necessarily so.

B. Individual collections. Collection VI, SDSM, is
quite different from collection V, CU. The latter is
almost invariably intermediate in composition between
the near-stream faunas and the open-plains faunas.
Collection VI always closely resembles the near-stream
fauna in those aspects in which it is intermediate. The
reduction of rodents and rabbits is carried beyond that
of the near-stream assemblage (however, our 1965
collection, as yet unstudied, indicates that the extreme
reduction may be a collecting bias). Merycoidodon is
most abundant in the near-stream community, less so
in the CU swampy plains fauna, and least (of the well-
watered areas) in the SDSM swampy plains fauna. The
percentage of Mesohippus is directly reciprocal of this.
Since the two forms are of roughly the same size, the
figures probably reflect a real preference of Mesohippus
for swamps and Merycoidodon for forested river-banks.

ECOLOGIC RELATIONSHIPS OF
PARTICULAR GENERA

The table, Figure 56, lists the total fauna as known
from our collections. Obviously, this is not the total
known fauna of the Lower Nodular Zone. Our 1964-65
collections have made several additions, which will be
reported as studies progress. The data on organisms
other than mammals are included to assist the reader to
evaluate the total ecologic setting of the mammals. We
have included some, but not all, non-mammal additions
of 1964-65.

I. Plants.

Known fossil plants include Celtis (hackberry) seeds;
algal balls, incrustations, strands, and Charagonia; and
some doubtful root casts. A few thin sections of sedi-
ments have revealed what may be spores; it is hoped
that careful study by a competent palynologist may
reveal much about the flora from study of spores.
Meanwhile, it is apparent that our direct knowledge of
the Brule flora of the area is so meager as to be meaning-
less.

II. Invertebrates.

Molds of Unio shells, frequently encrusted with
algae, occur on the borders of and within the riosomes
and drenajesomes. Pond snails of at least three genera

124

FIELDIANA: GEOLOGY MEMOIRS, VOLUME 5

occur abundantly in the limestone in the SDSM swamp
locality, and less commonly in the varzea sediments
near the riosomes. Ostracods are also abundant in the
swamp limestone. Ovoid, smooth-surfaced masses of
clay reputed to be casts of pupa cases or of larval bur-
rows, varying in size from MX %A in. to giants,
^Xl-H inches, have been collected from all locali-
ties except those of the Dillon Pass. They are largest
and most abundant in the swamp limestone.

The invertebrate fauna is almost as poorly repre-
sented as are the plants.

III. Pisces.

A few scraps of teleost bone, apparently representing
two families, are under study. They have been found in
the riosomes and in the swamp limestone.

IV. Amphibia.

Bufonid humeri occur rarely. They also are under
study.

V. Reptilia.

Rhineura has been found in the near-stream locality,
and Peltosaurus rarely in all localities except Dillon
Pass. Graptemys is the only aquatic turtle included in
the collection; it occurs in riosomes. Stylemys occurs
commonly in all localities except Dillon Pass, where it
is rare. It is most common, however, in the near-stream
localities, and less common in both the dry plains and
the swamp localities. It seems probable that Stylemys
was capable of living in any of the local environments,
but preferred forests to either swamps or open, grassy
plains

VI. Aves.

The birds represented were both, apparently, shore
or wading birds, and both were found along a rio margin.

VII. Mammalia.

A. Marsupialia.

1. Peratherium. This genus is exceedingly rare in
South Dakota, as compared with Chadron and Brule
faunas of Nebraska, Colorado, and Montana. Its
presence at Florissant and relative abundance at Pipe-
stone Springs suggests that it was at least forest-living if
not truly arboreal. Relative rarity in the Big Badlands
might mean either that forests were less dense or less
prevalent there, or that some other element necessary
to the ecology of Peratherium was missing in South
Dakota forests. Consideration of the habits of Recent
opossums leads the authors to favor the first hypothesis.

The relative abundance of Peratherium in Nebraska,
in areas lying presently several hundred feet above the
Big Badlands suggests a possible altitudinal control of
vegetation, with the Big Badlands near the lower limit
of forests. A similar situation is common to most semi-
arid regions today. There is no structural evidence that
the elevation of the Badlands has changed significantly
relative to Nebraska, since Oligocene time.

The alternative explanation, that of a latitudinal
thermal gradient sufficient to discourage oppossums

from the Big Badlands, seems most improbable. Even
the present high thermal gradient is locally obscured by
differences in elevation; the temperatures of north-
western Nebraska, both annual and daily, do not differ
significantly from those of the Big Badlands. Oligocene
world-gradients were certainly less steep than the pres-
ent one. We have no reason to believe, therefore, that
appreciable climatic differences existed between the
two areas, other than the differences due to relative
elevations.

B. Insectivora.

1. Iclops. The numbers collected are so small as to be
insignificant, except as they indicate the probable pres-
ence of Ictops in all local habitats.

C. Lagomorpha.

1. Paleolagus. This genus is ubiquitous, but shows a
strong preference for dry plains. The observed per-
centages in near-stream collections are almost certainly
too low; our 1964-65 collections suggest that Paleolagus
may have constituted between 5 and 10 per cent of the
total assemblage. The contrast between this and 20-30
per cent for the dry plains is certainly significant. The
known preference of Sylvilagus for brush and edge-of-
forest niches may find a homologue here, but we cannot
be sure.

Our collections suggest that two or more species of
Paleolagus are involved. Analysis may reveal a closer
restriction to habitat at the specific level.

2. Megalagus. A few specimens from each major
habitat tell us only that this rabbit was present but rare
in each.

D. Rodentia.

1. Prosciurus.

2. Adjidaumo.

These genera are represented by too few specimens
to be significant.

3. Eutypomys. Although the numbers of specimens
are too small to be significant, there is a suggestion that
Eutypomys may have preferred a near-stream, forest
environment. This is not evidence that they were or
were not swimmers.

4. Ischyromys. Ischyromys was ubiquitous, common,
and amazingly uniform in relation to the total assem-
blage. It varied from 7 to 16 per cent of the entire cen-
sus, generally lower in the near-stream and swamp
faunas, where it was the most numerous of the Glires,
and higher in the plains areas, where it was far out-
stripped by Paleolagus but remained second only to the
latter. Populations are large enough to suggest several
ecologic hypotheses, but not to establish them.

First, since the population of Ischyromys varies ac-
cordantly (although not proportionally) with that of
Paleolagus, the two were not in direct or acute com-
petition for food. Otherwise the overwhelming pre-
ponderance of Paleolagus in their preferred, dry-plains
habitat would have resulted in a decrease of Ischyromys.

CLARK AND KIETZKE: PALEOECOLOGY OF THE LOWER NODULAR ZONE

125

Possibly Paleolagus, like the modern rabbits, fed mostly
upon leaves of grasses and herbaceous plants, while
Ischyromys ate chiefly seeds. The fist-sized aggregations
of Celtis seeds which occur uncommonly in the dry-
plains area might, judging from the size of the seeds and
the aggregations, be hoards collected by Ischyromys.

Second, Ischyromys is commonest in the areas where
trees were least common. It is therefore probable that
these animals were ground-dwellers rather than pre-
dominantly arboreal in habit. Whether they dug bur-
rows is not known; a reported occurrence of an Ischyro-
mys skeleton inside a Stylemys shell (this we have not
seen) would suggest that they found shelter on the
ground, as modern chipmunks do. The cranial anatomy
strongly indicates a terrestrial habitat. We regard it as
probable that they lived in shallow burrows or in grass
nests, but could climb trees.

5. Eumys. Eumys is always less common than Ischy-
romys, varying between 1 and 6 per cent of each collec-
tion, but is quite as ubiquitous and its population size
roughly parallels that of Ischyromys. Unfortunately,
the total populations are too small for statistical signifi-
cance. Its apparent preference for dry plains may be an
artifact and at best is only a suggestion.

E. Carnivora.

Since the total carnivore population is under 10 per
cent of the total faunal assemblage for any habitat, all
evidence of its composition must be regarded as in-
conclusive.

Hesperocyon seems to have been the most numerous,
with Dinictis second. The variety but extreme rarity of
small carnivores other than Hesperocyon is worthy of
note; we do not understand it. The large number of
Daphoenus relative to the cats will be surprising to any-
one acquainted with the rarity of Daphoenus in the
various large previous collections. We believe this de-
monstrates the fallibility of our collections at this nu-
merical level, rather than any characteristic of the life
assemblage.

Anatomical interpretations of the habits of Brule
carnivores have led to a curious anomaly. Dinictis and
Hoplopheneus are obviously cat-like in their adapta-
tions and general way of life. The post-cranial anatomy
of Daphoenus is, except for the phalanges, more cat-like
than dog-like (see Scott and Jepsen, 1937, pp. 55-78).
In its adaptive characters, Hesperocyon strongly re-
sembles the viverrids (ibid., pp. 81-105). Hyaenodon is
generally supposed, due to the massive head, "weak"
feet, and heavy teeth, to have been a scavenger.

The niche of a medium-sized, cursorial, terrestrial
carnivore is apparently open, in a mammalian com-
munity most of whom would be vulnerable to attack by
such a predator. Moreover, this community represents
in large part an invading ecosystem (see p. 69, this
report) which had a long history as a savanna chrono-
fauna in its original home. It is certainly possible that
Brule savannas and prairies may have been well-enough
vegetated with brush and high grass to discourage such

active runners as coyotes — pre-Columbian Illinois and
western Missouri had low coyote populations — but gray
wolves do not restrict themselves to short-grass plains,
and dholes, dingoes, jackals, and Chrysocyon certainly
live in vegetated places.

A brief study of Hyaenodon indicates that the ab-
sence of a cursorial carnivore may be more apparent
than real. All three molars participate in the formation
of a shearing blade which occupies up to 40 per cent of
the total cheek-tooth length, as opposed to under 30
per cent in Hyaena. The musculoskeletal structure of
the calvarium strongly resembles that of Hoplophoneus
and differs so extremely from that of Hyaena as to be
worthy of a separate study. The zygoma is as light
proportionally as that of Martes. The ascending ramus
is exceedingly short, as in Hoplophoneus, and the entire
anatomy of the posterior moiety of the jaw resembles
that of a cat rather than of Hyaena. The feet are, unlike
those of Hoplophoneus, digitigrade. In fact, the name
" Hyaenopus" might have been less a misnomer than
" Hyaenodon."

The only anatomical factors weighing against a
cursorial predator's role for Hyaenodon are the extreme-
ly large skull and relatively small feet.

The balance of the anatomical evidence suggests
that Hyaenodon was an active predator, probably not as
cursorial as the modern wolf or coyote. It might well
have hunted more in the fashion of the lion, while
Hoplophoneus, Dinictis, and Daphoenus hunted more
from ambush, like leopards and jaguars.

Scavenging probably was done by any hungry indi-
vidual of any species who happened upon a carcass, but
no one mammalian group specialized for this activity;
insects probably did the bulk of the scavenging. Abun-
dant stock of available prey might have resulted in a
very low mammalian scavenging pressure, which is in
accord with the perthotaxic evidence previously dis-
cussed.

F. Perissodactyla.

1. General. The four genera listed in the table, plus
Subhyracodon and Metamynodon which are known but
not listed, all represent lines present in known Eocene
wet-forest chronofaunas. The perissodactyl fauna dif-
fers qualitatively from that of Chadron time in the
absence of titanotheres and in the reduction of the re-
maining perissodactyl population relative to the total
fauna. Climatic change resulting in a general shift from
forest to savanna-prairie and prairie is the most evident
cause of the waning of this previously dominant order.
Colodon, Protapirus, Hyracodon, and Metamynodon sur-
vived only barely into Whitneyan time, and became
extinct by the end of it. Mesohippus and the true rhi-
noceroses adapted, albeit with difficulty, to their chang-
ing environment.

2. Colodon. No statistically significant collection of
Colodon is known. Its presence in the open-plains col-
lections of both Sage Creek and Dillon Pass poses a
real enigma. Certainly there is abundant anatomic,
taxonomic, and stratigraphic evidence suggesting that

126

FIELDIANA: GEOLOGY MEMOIRS, VOLUME 5

Colodon was a forest animal. We do not know why the
two individuals represented in the open-plains collec-
tions travelled in such company.

3. Hyracodon. Hyracodon seems to have been an
uncommon but ubiquitous animal. It seems less com-
mon in our collections than in most of the classic col-
lections from the Big Badlands. This may be an er-
roneous impression, or it may represent a bias of pre-
vious collection.

4. Caenojms-Subhyracodon. These two genera have
been lumped and referred to as Caenopus, due to the
difficulty of discriminating between them on the basis
of fragmentary specimens. In any case, two specimens
from the open plains and eight from the near-stream
area can tell us only that large, wide-ranging rhino-
ceroses did occur as rare elements of the fauna. Since
this group shares with Mesohippus the distinction of
adapting to the changing environment, rarity during
Brule time may indicate a period of near-extinction
preceding their successful adaptation.

5. Mesohippus. Mesohippus' predilection for a swamp
and forest habitat receives adequate documentation
from the collections. The genus was, during Orellan
time, meeting the change in climate by restricting its
life to those relicts of the old forests which still remained.
The number of individuals within the swamp and forest
areas suggests that Mesohippi were highly successful
there. The high percentage of late adolescent individuals
in all collections indicates a probable even higher mor-
tality of infants and juveniles, all trace of which is
generally lost. Mesohippus, therefore, was almost cer-
tainly a much more important element of the faunas of
its favored habitats than its numbers in collections
indicate.

In contrast with their abundance in Orellan sedi-
ments, horses are one of the rarer fossils in Whitneyan
rocks of Nebraska and South Dakota. Progressive
cooling and drying during that time almost eliminated
them from the area. Near-extinction may have been
local, with abundant, evolving populations in other
parts of the continent, or may have been general. In
either case, the group recovered within the area during
Miocene time, triumphantly adapted to their new en-
vironment.

G. Artiodactyla.

1. General. Artiodactyls first became dominant ele-
ments of known faunas during late Uintan time. The
Uinta C fauna consists primarily of small and medium-
sized artiodactyls, with lesser numbers of titanotheres,
rhinoceroses, achaenodonts, and carnivores. The Uintan
artiodactyls were, as far as known, all well-adapted
wet-forest forms. Smaller genera filled the niches occ-
pied in modern wet forest by rabbits, large terrestrial
rodents, musk deer, and small jungle antelope. Their
perissodactyl competitors were tapiroids and the exceed-
ingly rare Epihippus. Rodents and rabbits apparently
offered little competition. The number of genera of ar-
tiodactyls was high, reflecting a high adaptability and
diversity of habitat within the wet-forest environment.

Chadron artiodactyls are diverse, and readily divis-
ible into two groups: descendants of the Eocene jungle
chronofauna on the one hand, and immigrants on the
other. Several of the immigrants seem better adapted to
a savanna than to a wet-forest environment. The Cha-
dronian faunas in South Dakota show a dominance of
perissodactyls by number of individuals. The numbers
of genera representing the two orders are nearly equal,
with a preponderance of artiodactyl genera. Exact
counts have little meaning, because of the need of com-
plete taxonomic revision of Chadronian artiodactyls.

Our collections indicate the apparently sudden erup-
tion of the Artiodactyla into a dominance which they
rarely if ever have lost. Thirteen genera compare with
five of perissodactyls. Artiodactyls consistently com-
prise 43-55 per cent (see Table XIII) of the total mam-
malian census including rodents. Even in the near-
stream and swampy plains habitats, favored by the
relict Mesohippus, the total perissodactyl census is not
over 70 per cent of the number of Artiodactyls. Merycoi-
dodon is the most abundant single genus in the near-
stream fauna, and Leptomeryx in the open plains; in the
swampy plains fauna, Leptomeryx is second only to
Mesohippus. In every case, the most abundant artio-
dactyl is more numerous than the most abundant rabbit
or rodent.

The artiodactyl genera can be divided into two
groups, the first consisting of relict genera with known
late Eocene forest ancestors, the second comprising
Chadronian or Orellan immigrants:

Relict genera Oligocene immigrants

Leptockoerus Arehaeotherium

Stibarus Perchoerus

Agriochoems Bothriodon

Hypertragulus Merycoidodon

Bathygenys
Poebrotherium
Leptomeryx
Hypisodus
Leptauchenia

Comparison of this listing with our collections is most
instructive. Leptockoerus and Stibarus occur in the open
plains and swampy plains faunas, but in such small
numbers as to make their presence or absence insignifi-
cant. (The senior author has, in earlier years, collected
both from near-stream sediments). Agriochoerus shows
up in equal numbers, still insignificant, in the open
plains and the near-stream faunas. Hypertragulus makes
up 5-8 per cent of the open plains fauna, 11 per cent of
the swampy plains assemblage, and less than 2 per cent
of the near-stream fauna. The relict group are all un-
common to rare; the distribution of Hypertragulus sug-
gests that it lived in marshes and high grass.

Of the immigrant group, Merycoidodon shows an
obvious preference for a near-stream environment, and
Leptomeryx for open plains. Bathygenys and Leptauchenia
are known from one specimen each, and Perchoerus
from three. Thirteen specimens of Hypisodus, all from
open plains, and four of Bothriodon, all from near-

CLARK AND KIETZKE: PALEOECOLOGY OF THE LOWER NODULAR ZONE

127

stream, are statistically insignificant but interesting as
they may suggest favored habitats in keeping with the
animals' anatomy. Archaeotherium and Poebrotherium,
both forms whose anatomy strongly suggests active
movement over dry ground, occur in insignificant
numbers in both near-stream and open-plains collec-
tions. The immigrant fauna apparently included species
adaptable to all three of the habitats recognized here.

2. Leptochoerus and Stibarus. As already indicated,
our collections can tell us nothing of the habits of these
two forms.

3. Archaeotherium. The geographic distribution of
Archaeotherium in the Lower Nodular Zone is exceeding-
ly sporadic throughout the Big Badlands. Specimens oc-
cur in fairly large numbers at the near-stream locality,
at Princeton "Entelodon Peak" locality (NW 1/4 Sec. 5,
T. 4S., R. 13E., Pennington Co.), just south of the Sage
Creek Dry Plains area of this report (center, Sec. 22,
T. 2S., R. 15E., Pennington Co.), and 1 mile south of
Dillon Pass (center, Sec. 33-34, T. 2S., R. 16E., Pen-
nington Co.). The senior author knows of from 3 to 12
skulls collected from each of these localities, in addition
to the specimens listed in the Sage Creek and Dillon
Pass collections of this report. Conversely, the authors
have seen only one identifiable scrap or specimen of
Archaeotherium in the classic collecting locality of Big
Hollow-Bear Creek, Sec. 3, 4, 9, and 10, T. 4S., R. 13E.,
and Sec. 34, T. 3S., R. 13E., Pennington Co. Leptomeryx
occurs in fair numbers at the classic locality, and
Merycoidodon skulls tend to be smaller on the average
than those at Cottonwood Pass and Entelodon Peak.
It seems probable that Archaeotherium inhabited some
notably restricted band within the spectrum of ecologic
niches available during Lower Nodular time.

4. Perchoerus. Occurrence of Perchoerus in insignifi-
cant numbers both in the near-stream and the open
plains collections indicates merely that we have no
adequate evidence of its habitat. In view of the size
and rarity of Perchoerus in South Dakota, statistically
adequate collections probably cannot be made.

5. Bothriodon. Once more, the number of specimens
is statistically insignificant. However, all of the speci-
mens came from rio borders, and we have collected
others from actual riosome sands. We have never seen
a scrap of any anthracothere bone in the open-plains
environments. This suggests, but does not demonstrate,
that the anthracotheres lived in streams and along their
banks.

6. Merycoidodon. The evidence clearly indicates that
Merycoidodon preferred the river-border forests to
either open plains or swamps. Separation of the smaller
species, M. gracilis, from the more abundant, larger
M. culbertsoni reveals an even sharper preference for
forests on the part of the latter. Although the numbers
lie below the level of significance, there is a strong sug-
gestion that M. gracilis preferred swamps and open
plains.

Neither the limb structure nor the dental equipment
of M. culbertsoni gives any particular support to the

often expressed idea that it frequented plains and grassy
prairies. Its relatively broad, short, tetradactyl feet,
femur and humerus as long as antibrachium and tibia
respectively, complete fibula, and generally heavy build
certainly suggest less adaptation to cursorial plains ex-
istence than do the corresponding structures in Meso-
hippus, which is universally regarded as a "forest
horse." The brachydont dentition, like that of almost
all Orellan mammals, gives no indication of adaptations
for grazing. The distributional evidence therefore rein-
forces the anatomical indications of a forest habitat for
Merycoidodon culbertsoni.

M. gracilis shows no anatomical differences at-
tributable to its presumably different habitus. It might
be either that we are misinterpreting inadequate data,
or that M. gracilis was not yet highly specialized for its
environment.

7. Agriochoerus. Agriochoerus was present in small
numbers both in the gallery forest and on the open
plains. The numbers collected are statistically insignifi-
cant.

8. Bathygenys. One specimen of this rare, small form
was found, surprisingly, in the open plains area. Since
previous specimens all came from Chadronian deposits
in Montana, from presumably forested intermontane
basins, this specimen is anomalous.

9. Leptauchenia. One specimen of Leplauchenia repre-
sents a downward extension of the genus from Whit-
neyan sediments, in which it is abundant, to Orellan, in
which it has not previously been known. Considered
separately, its occurrence in open-plains sediments means
only that one individual died in that area. Taken in con-
junction with its universal presence in varzeasomes and
absence from riosomes of Whitneyan age, the paleogeo-
graphic evidence strongly suggests that Leptauchenia
was not so aquatic as its anatomy supposedly indicates.

10. Poebrotherium. Probrotherium, of all Orellan
genera, might be expected to show a preference for the
open plains. With nine specimens from the open plains,
eight from the near-stream, and one from the swampy
plains, it apparently shows no preference at all. The
samples are not significant, except as they show that
Poebrotherium occurred uncommonly but ubiquitously.

11. Hypertragulus. Our collections reveal that a
strong habitat preference, rather than actual rarity, is
responsible for Hypertragulus' poor representation in
previous collections. The vast majority of the earlier
Orellan collections were made in the near-stream areas
south and southwest of Scenic, where both Leptomeryx
and Hypertragulus are uncommon. Hypertragulus was
proportionally most abundant in the swamp fauna and
apparently was only slightly less important an element
of the open plains community. It varied from half as
abundant as Leptomeryx in the swamp to one-third or
one-sixth as abundant in the open plains. (Total num-
bers of both in the near-stream environment are too
small for statistical significance; our later collections
confirm the relative scarcity of both in the near-stream

128

FIELDIANA: GEOLOGY MEMOIRS, VOLUME 5

area, but suggest that Leptomeryx may have been the
more numerous of the two).

Another interesting sidelight on the environment of
Hypertragulus has been revealed by the distribution of
specimens. In 1937, the senior author collected a group
of over 20 skeletons of Hypertragulus, huddled together
in rather comfortable "sleeping" poses, 18 of them with
their noses east and all beautifully articulated. The
group occurred near Entelodon Peak in the near-stream
area. All were very young adults, with M£ partially
erupted but unworn. The presumption was that this
herd had either died of epidemic or had been over-
whelmed by a windstorm as they lay with noses down-
wind. (The major portions of this find are in the Carne-
gie Museum and the University of Colorado Museum).

In 1956, the senior author discovered another group
of about 20 skeletons, this time in the swamp area. The
bones were partially disarticulated by perthotaxy, but
burial apparently did not move them. Again, they were
all very young adults. This find is in the South Dakota
School of Mines Museum, and the specimens are in-
cluded in our statistics.

In 1965, the junior author found a third group of
Hypertragulus skeletons, this time largely disarticulated
by perthotaxy. The bones covered an area about 2J/£
by 8 ft. in extent, and were disposed through 18 in. of
vertical thickness, in a convex front. One rodent jaw
was found with them, but no other fossils. This find has
not yet been prepared, but we are sure of more than ten
individuals, all very young adults. We interpret this
find as representing a herd assemblage in advanced
perthotaxy, picked up by the rolling front of a mud-
flow and transported not over a very few feet. Longer
transport would have scattered the already separate
bones, and mixed them with those of other forms.

We have represented here three herd assemblages,
all young adults at time of death, in three different
stages of perthotaxy at time of burial. Obviously, the
mechanics of their burial were unrelated to the cause of
their death. Obviously also, they were not killed by
predation or by physical accident, since the individuals
in the first group were in perfect condition. Drought or
famine are improbable, because the animals were of the
age group most resistant to such deprivations. We are
therefore left with the probability that Hypertragulus
was subject to fatal epidemics which struck particularly
at late adolescents. Whatever the disease was, its
victims characteristically died peacefully, without con-
vulsions, and probably quite rapidly. Had the illness
been of long duration, less seriously sick individuals
should have wandered and the tight herd would have
broken up. The animals in groups 1 and 2 (and pre-
sumably also in 3) probably died within hours of the
time when they were still well enough to walk to their
final resting place and lie down there.

The swamp habitat suggested for Hypertragulus
conforms with the anatomical evidence of the relatively
short, tetradactyl feet, and retention of a vestigial tibia.

12. Leptomeryx. Leptomeryx, like Hypertragulus,
shows definite evidence of its preferred habitat. From a
minor element in the near-stream fauna, it increases to
23 per cent of the swamp community, and over 30 per
cent of the open plains assemblage. It is as numerous
as the next two commonest genera, Paleolagus and
Ischyromys, combined. Hypertragulus, which is half as
common as Leptomeryx in the swamp habitat, is less
than one-fifth as common in the open plains. The
generally long limbs and feet, didactyl pes, and loss of
the fibular shaft give anatomical confirmation of the
paleogeographic evidence that Leptomeryx was an active
runner on dry ground.

The abundance of Leptomeryx spotlights the ap-
parent absence of an efficient cursorial carnivore. If, as
seems probable, this genus evolved as part of a de-
veloping savanna chronofauna during middle and late
Eocene time, and invaded the general area as part of
that chronofauna during Chadronian time, one would
expect that an adequate cursorial carnivore would have
evolved and immigrated with it. Such evidently was not
the case. It would have been a difficult prey for Hespero-
cyon, so we must presume that it fell only to the short
rushes of Hyaenodon and the ambushes of Daphoenus
and the feloids.

13. Hypisodus. Here again we have a small artio-
dactyl which, although in this case rare, shows a pre-
dilection for open plains. Again, the didactyly and long,
slender limbs indicate cursorial abilities. In Hypisodus
the plains habitat is indicated even by the teeth which,
as the name indicates, are distinctly long-crowned. To
the best of our knowledge, this is the oldest herbivore
genus with long-crowned molars.

SUMMARY

The paleogeography established in the previous
chapter offers us a record of several adjacent life
habitats, coexistent through a brief time (the 200-2000
years necessary to deposit the Lower Nodular Zone),
without a significant climatic change. Three easily
recognizable habitats, a near-stream zone presumably
occupied by gallery forests; an open-plains area, far
from any stream, which might have borne plains,
prairie, or savanna vegetation; and a swamp area
within the plain, were chosen for ecologic study.

Fossils occurring in all three situations were buried
by engulfment in situ; they were not transported. They
therefore represent buried perthotaxies, modified only
by later bias factors.

Considering the total mammal population of the
three local habitats as the universe to be studied, dif-
ferences between this universe and the collection as an
ultimate sample are produced by seven groups of fac-
tors:

Biotic

Thanatic

Perthotaxic

Taphic

CLARK AND KIETZKE: PALEOECOLOGY OF THE LOWER NODULAR ZONE

129

Anataxic
Sullegic
Trephic
These operate differentially upon individual species or
genera within a fauna. They function both separately
and as inter-related variables. Biotic and thanatic
factors combine to produce the death assembly, the
total number of corpses arriving upon a surface before
the next episode of sedimentation. This is a sample of the
life assemblage. Perthotaxic and taphic factors reduce
the death assemblage to the total fossil assemblage, a
sample which in turn is reduced by the last three groups
of factors to the collection. A collection is thus a sample
of a sample of a sample of the life assemblage or com-
munity.

We have no way of evaluating the various biasing
factors. We also have no complete census of any recent
mammalian community to use as a norm. Therefore, it
is necessary to determine by inspection whether or not
the collections show enough consistency to indicate that
they are adequate samples. Collections of over 300
specimens, from the same habitat, vary from each other
by up to 10 per cent. Between collections of 100 to 300
specimens, the variance is about 20 per cent. Collections
of under 100 specimens are of little value.

The near-stream gallery forests were inhabited by a
community of rich generic variety, primarily of medium-
sized mammals. Merycoidodon was the commonest;
together, Merycoidodon and Mesohippus comprised 50
per cent of the total mammalian population. Small
artiodactyls were present, but relatively unimportant.

Small animals dominated the open-plains habitat.
Leptomeryx and Paleolagus together made up over half
of the community, with Ischyromys third in importance.
Mesohippus dropped to 6 per cent and Merycoidodon to
5 per cent.

The swamp community is not well known, but
Mesohippus (30%), Leptomeryx (23%), and Hyper-
tragulus (11%) were the dominant forms.

The habits of certain individual genera are con-
siderably illuminated by the collections. Paleolagus and
Leptomeryx were ubiquitous but much preferred the
open plains. Hypertragulus was ubiquitous but preferred
swamps; it ran in herds, and young adults were subject
to virulent epidemics. Ischyromys was highly ubiqui-
tous, and probably seeds formed the chief item of its
diet. Hypisodus was a plains dweller. Mesohippus pre-
ferred swamps, although it also freely inhabited the
gallery forests and to a lesser extent the plains. Merycoi-
dodon strongly preferred forests; it constituted a minor
element of the communities in other habitats.

Of the carnivores, Hesperocyon was most numerous.
Hyaenodon was an active predator; it probably filled the
niche of an entirely terrestrial, cursorial predator, al-
though it was not very well adapted for running. The
carnivore population formed 4 8 per cent of the total
community in every habitat.

In general, the invading plains-savanna fauna which
first appeared during Chadronian time had established
dominance by Orellan time. Perissodactyls were de-
creasing in variety and in importance. Mesohippus
maintained itself by clinging to the dwindling relicts of
the old wet-forest environment. The artiodactyl-rodent-
rabbit dominance so evident in mammalian communi-
ties ever since Orellan time was clearly recorded for the
first time in the Big Badlands, although it had probably
developed earlier as a savanna ecosystem in what is now
the North Central States.

Ubiquitous occurrences of anatomically well-adapted
plains forms in the gallery forests, and forest forms in
the plains, may be due to: (1) biases in the samples; (2)
incorrect or overspecific interpretations of habits from
anatomy; (3) the fact that the gallery forests probably
grew on firm ground and may have been quite open, at
the same time that the plains may have been dotted
with small thickets and woods.

COLORADO MUSEUM^ OPEN PLAINS FAUNA- SAGE CREEK

PElTOSAURUS

PERATh£«um

CTOPS
PALEOLAGUS

MEGALAGUS
PROSCIURUS
ISCHYROMYS
EUTYPOMYS
EUMYS

hespcrocyon

daphoenus

mustelavus

OMCTlS

MESOHIPPUS
COlOOON

MYRACOOON

CAENOPUS

ARCHAEOTHCRlUU
AGRtOCMCCRUS

merycgwooon

MERYCOtDOOON m

batmvgenys

POEBROTmERIUM

HYPERTRAGULUS

LEPTOMERYX

HYPISOOUS

53 SPECIMENS

GLIRES
TOTAL

566X

CARNIVORE
TOTAL

3 6%

PERiSSOOACTyl
TOTAL

€6%

ARTlOOACTYL
TOTAL

506%

PERiSSOOACTyl
ARTlOOACTYL

13 OX

PALEOLAGUS - ISCHYROMYS -
LEPTOMERYX - HYPERTRAGULUS

MESOHIPPUS - MERYCOiDODON

47891

5'ZE DISTRIBUTION
<20LB 6<2X

20 -75 LB ,T2X

>75LB 1.5%

SDSM-OPEN PLAINS FAUNA - CENTRAL SAGE CREEK

Hi's o* sic m»s i/h," !i. mmiNiTOK eo.*o«

ICTOPS

PALEOLAGUS a

MEGALAGUS
PROSCIURUS
ISCHYROMYS 6

EUMYS I

HYAENODON
PALE OG ALE

HESPEROCYON I

DAPHOENUS

DINICTIS

HOPLOPHONEUS

MESOHIPPUS 2

HYRACODON

STlBARUS

ARCHAEOTHERIUM

PERCHOERUS

AGRiOCHOERUS

MtKYCOlOOOON CULBCRTSONI

MERYCOIDOOON CraohS I

POEBROTHERIUM

HYPERTRAGULUS I

LEPTOMERYX 17

456

SPECIMENS

SIZE DISTRIBUTION

36.6%

20 -75 LB

TOTAL

TOTAL MAMMALS

CARNIVORE
TOTAL

61%

<20LB

TOTAL MAMMALS

65%

> 75LB

TOTAL

TOTAL MAMMALS

ARTlOOACTYL

TOTAL

48 0%

PERISSOOACTYL

ARTlOOACTYL

136%

PALEOLAGUS - ISCHYROMYS -
LEPTOMERYX -HYPERTRAGULUS

MESOH I PPUS • MERYCOIDODON

Fig. 54. Graphs I-XIII, population statistics.

130

PALEOLAGUS

ISCHYROMYS

HYRACOOON

AGRIOCHOERUS I

MERYCOIDODON CULBERTSONI I

LEPTAUCHENIA

LEPTOMERYX
HYPERTRAGULUS

1.6%

7.6%

I 16%
1.6%

SDSM-OPEN PLAINS FAUNA

25 7%

22.7%

212%

16.7%

I MILE S OF OILLON PASS JCT CCNTCff ,SCC 3S-S4. TIS, RISE

66

SPECIMENS

SIZE DISTRIBUTION

GL1RES
TOTAL

485%

< 20LB
TOTAL

86.6%

CARNIVORE
TOTAL

16%

20-75LB
TOTAL

45%

PERISSODACTYL
TOTAL

7.6%

>75LB
TOTAL

7.6%

ARTIODACTYL
TOTAL

424%

PERISSODACTYL
ARTIODACTYL

17.9%

PALEOLAGUS-ISCHYROMYS

LEPTOMERYX -HYPERTRAGULUS

MESOH 1 PPUS - MERYCOIDODON

5700.0%

SDSM-OPEN PLAINS FAUNA

ICTOPS

PALEOLAGUS

PROSCIURUS

ISCHYROMYS

EUMYS

HESPEROCYON

COLODON

MESOHIPPUS

ARCHAEOTHERIUM

MERYCOIDODON GRACILIS

HYPERTRAGULID

LEPTOMERYX

HYPERTRAGULUS

HYPISOOUS

GRAPH 32

1

34

15
3

■

1

7
1

1

3

1
1

•*■ 26%

I

I 0.9%

1

■

4
1

•H 35%

SEj-SEC28AND ADJACENT E0GESEC27. T2S. RI6E. PENNINGTON CO„S.0AK.

36.0%

114

SPECIMENS

SIZE D

GLIRES
TOTAL

46.5%

•S?0LB

TOTAL

CARNIVORE
TOTAL

6.1%

20-75LB

TOTAL

PERISSODACTYL
ARTIODACTYL

35%

>75LB
TOTAL

ARTIODACTYL
TOTAL

43.0%

PERISSODACTYL
TOTAL

0.8%

PALEOLAGUS- ISCHYROMYS-

LEPTOMERYX - HYPERTRAGULUS

MESOHIPPUS-MERYCOIDODON

4 4%

2350%

Fig. 54 (continued). Graphs I-XIII, population statistics.

131

COLORADO MUSEUM^ SWAMPY PLAINS FAUNA

NE j-SEC 2I.T2S.RI5E PENNINTON CO

PALEOLAGUS
MEGALAGUS

ISCHYROMYS

EUMYS

HESPEROCYON

3

DAPHOENUS

1

BUNAELURUS

1

OINICTIS

1

MESOHIPPUS

17

HYRACODON

3

CAENOPUS

1

LEPTOCHOERUS I

AGRIOCHOERUS 2

MERYCOIOODON CUL6ERTS0NI 13

MERYCOIOODON gracilis 2

POEBROTHERIUH '

HYPERTRAGULUS 5

LEPTOMERYX 51

HYPISODUS I

184%

7.1%

2.1%

3.5%

6
141

21
141

76
141

21
76

92

15

141 SPECIMENS

BUSES

TOTAL

CARNIVORE
TOTAL

PERISSODACTYL

total

ARTIQOACTYL
TOTAL

PERISSODACTYL
ARTIODACTYL

PALEOLAGUS - ISCHROMYS-
LEPTOMERYX - HYPERTRAGULUS
MESOHIPPUS - MERYCOIDOOON

27.0%

4.3%
14.9%
54.7%
276%

6133%

SIZE DISTRIBUTION

100 < 2QLB

141 TOTAL

-5L 20-75LB
141 TOTAL

*Mr

71.9%
26.6%
29%

SDSM SWAMPY PLAINS

SE^SECI3.T2S, RISE. PENNINGTON CO

PALEOLAGUS

MEGALAGUS

PROSCIURUS

ISCHROMYS

EUMYS

HYAENOOON

HESPEROCYON

DAPHOENUS

HOPLOPHONEUS

OINICTIS

HYRACODON

MESOHIPPUS

LEPTOCHOERID I

MERYCOIOOOON GRACILIS 7

MERYCOIOODON CULBERTSONI 20

POEBROTHERIUM I

LEPTOMERIX 51

HYPERTRAGULUS 25

66

GRAPH VI.
■■ 14%

30.4%

217 SPECIMENS

_2i

217

GLIRES
TOTAL

10.1%

IB
217

CARNIVORE
TOTAL

8.3%

217

PERISSODACTYL
TOTAL

332%

105
217

ARTIODACTYL
TOTAL

48.4%

72
105

PERISSODACTYL
ARTIODACTYL

68.6%

92
93

PALEOLAGUS-ISCHYROMYS-

LEPTOMERYX - HYPERTRAGULUS

989%

SIZE DISTRIBUTION

103
217

<20LB
TOTAL

106

20-75 LB

217

TOTAL

6

>75LB

217

TOTAL

484%
488%
2.8%

Fig. 54 (continued). Graphs I-XIII, population statistics.

132

SDSM NEAR- STREAM FAUNA

GRAPTEMYS

BATHORNIS

PALEOLAGUS

EUTYPOMYS

EUMYS

ISCHYROMYS

HOPLOPHONEUS

OINICTIS

COLOOON

HYRACODON

CAENOPUS

MESOH1PPUS

ARCHAEOTHERIUM 7

PERCHOERUS 2

BOTHRIODON 4

AGRIOCHOERUS 3

MERYCOIDODON GRACILIS 5

MERYCOIOODON CULBERTSONI 42

POEBROTHERIUM B

LEPTOMERYX 4

HYPERTRAGULUS 5

NE ± SEC II. T 42N. R45W

29*
I 3.6 %

139

SPECIMENS

SIZE DISTRIBUTION

8

TS9~

GLIRES
TOTAL

5.8%

17
137

<20LB

TOTAL

12.4%

4

T59"

CARNIVORE
TOTAL

2.9%

99

TIT"

20-75 LB
TOTAL

72.3%

45

"rag-

PERISSOOACTYL
TOTAL

32.4%

21

>75LB
TOTAL

15.3%

so

TT5"

ARTIOOACTYL
TOTAL

57.6%

45
80

PERISSOOACTYL

ARTIOOACTYL

563%

13
"80"

PALEOLAGUS

LEPTOMERYX
MESOHIPPUS-

ISCHYROMYS-

HYPERTRAGULUS
MERYCOIDODON

163%

SDSM = NEAR- STREAM FAUNA

E BORDER OF SW^- ,SEC 2. T 42N.R 45W

GRAPH SUL

ISCHYROMYS

CAENOPUS

MESOHIPPUS

MERYCOIDODON GRACILIS 3

MERYCOIDODON CULBERTSONI 15
LEPTOMERYX 2

HYPERTRAGULUS 2

64%

40.4%

319%

43%
43%

4
47

0
47

21
47

-22.
47

_8_

37

47 SPECIMENS

GLIRES

CARNIVORE
TOTAL

PERISSOOACTYL
TOTAL

ARTIOOACTYL
TOTAL

PERISSODACTYL
ARTIOOACTYL

PALEOLAGUS- ISCHYROMYS-
LEPTOMERYX-HYPERTRAGULUS

MESOHIPPUS- MERYCOIOODON

85%

0
447%
46.8%
95.5%
21.6%

SIZE DISTRIBUTION

8

<20LB

47

TOTAL

37

20-75 LB

44

TOTAL

2

47

>75LB
TOTAL

17.0%
78.7%
43%

Fig. 54 (continued). Graphs I-XIII, population statistics.

133

COLORADO MUSEUM NEAR- STREAM FAUNA

AMPHIBIAN

1

RHINEURA

1

AVES

1

PERATHERIUM

2

ICTOPS

1

HYAENOOON

2

HOPLOPHONEUS

1

DINICTIS

3

PARICTIS

1

HESPEROCYON

3

DAPHOENUS

1

PALEOLAGUS

13

MEGALAGUS

4

AOJIDAUMO

1

EUTYPOMYS

10

EUMYS

1

ISCHYROMYS

20

MESOHIPPUS

19

AGRIOCHOERUS

2

MERYCOIDOOON CULBERTSONI 28

GRAPH TX.

26%

I 17.4%
16.5%

243%

49
IIS

II
IIS

19
115

30
IIS

19
30

33

47

115 SPECIMENS
GLIRES

TOTAL
CARNIVORE

PERISSODACTYL
TOTAL

ARTIODACTYL

PERISSODACTYL
ARTIODACTYL

COTTONWOOD PASS AREA-NW j SEC II, T42N.R4SW.

PALEOLAGUS -ISCHYROMYS-
LEPTOMERYX - HYPERTRAGULUS
MESOHIPPUS - MERYCOIDOOON

426%
96%

16.5%
261%
633%

70-2%

SIZE

OISTRIBUTION

56

< 20 LB

112

TOTAL

500%

56
TTF

20-75LB
TOTAL

500%

0
TT2-

>75LB
TOTAL

0

TOTAL FAUNA, OPEN PLAINS, SAGE CREEK

COLLECTIONS 1,11.

236

8

2

120

52
2

33
7
I

■
7

PELTOSAURUS

PERATHERIUM

ICTOPS

PALEOLAGUS

MEGALAGUS

PROSCIURUS

ISCHYROMYS

EUTYPOMYS

EUMYS

HYAENOOON

PALEOGALE

HESPEROCYON

DAPHOENUS

MUSTELAVUS

BUNAELURUS

DINICTIS

HOPLOPHONEUS

MESOHIPPUS

COLODON I

HYRACODON 15

CAENOPUS 2

LEPTOCHOERUS I

ST1BARUS I

ARCHAEOTHERIUM 3

PERCHOERUS I

AGRIOCHOERUS 6

MERYCOIDOOON CULBERTSONI 58

MERYCOIDODON CHACIUS 2 9

8ATHYGENYS I

POEBROTHERIUM 9

HYPERTRAGULUS 61

LEPTOMERYX 383

HYPISODUS 12

68

419
1130

53

1130

86

1130

_565

:30

86

565

800

155

1130 SPECIMENS

GLIRES
TOTAL

TOTAL

PERISSODACTYL
TOTAL

ARTIODACTYL
TOTAL

PERISSODACTYL
ARTIODACTYL

SIZE DISTRIBUTION

371%

916

1127

< 20LB

TOTAL MAMMALS

815%

4.7%

189
1127

20-75 LB
TOTAL MAMMALS

16 8%

20

1127

> 75LB

7 6%

TOTAL MAMMALS

18%

500%

152%

PALEOLAGUS - ISCHYROMYS -
LEPTOMERYX - HYPERTRAGULU S
MESOHIPPUS- MERYCOIDODON 516 1%

Fig. 54 (continued). Graphs I-XIII, population statistics.

134

TOTAL FAUNA - OPEN PLAINS - DILLON PASS AREA

COLLECTIONS m, 12

ICTOPS I

PALEOLAGUS 51

PROSCIURUS I

ISCHYROMVS 30

EUMYS J

HESPEROCYON 7

DINICTIS I

COLODON I

MES0H1PPUS 3

HYRACODON 5

ARCHAEOTHERIUM I

AGRIOCHOERUS I
MERYCOIDODON CULBERTSONI I

MERYCOIDODON GRACILIS I

LEPTAUCHENIA I

HYPERTRAGULIO I

HYPERTRAGULUS 15

LEPTOMERYX 55

HYPISODUS I

1.7%

m 3.9%

17%

0.6 X
061

283%

30.6%

160 SPECIMENS

SIZE DISTRIBUTION

65

180

GLIRES
TOTAL

47.2%

165
ISO

<20LB
TOTAL

91.7%

8

180

CARNIVORE
TOTAL

4.4%

9
ISO

20-75LB
TOTAL

5.0%

9
180

PERISSODACTYL
TOTAL

5.0%

6
180

>75LB
TOTAL

3J%

78
180

ARTIOOACTYL
TOTAL

43.3%

9
78

PERISSODACTYL
ARTIOOACTYL

11.5%

151

S

PALEOLAGUS - ISCHYROMYS
LEPTOMERYX -HYPERTRAGULUS
MESOHIPPUS -MERYCOIDODON

3020.0%

TOTAL FAUNA-NEAR-STREAM FACIES

AMPHIBIAN

GRAPTEMYS

RHINEURA

AVES

BATHORNIS

PERATHERIUM

ICTOPS

PALEOLAGUS

MEGALAGUS

ISCHYROMYS

ADJIDAUMO

EUTYPOMYS

EUMYS

HYAENODON

HESPEROCYON

DAPHOENUS

PARICTIS

DINICTIS

HOPLOPHONEUS

MESOHIPPUS

COLODON

HYRACODON

CAENOPUS

ARCHAEOTHERIUM

PERCHOERUS

60THRI0D0N

AGRIOCHOERUS

MERYCOIDODON GRACILIS

MERYCOIDODON CULBERTSONI

POEBROTHERIUM

HYPERTRAGULUS

LEPTOMERYX

COLLECTIONS VII. VIII, IX.

90%

1.0%
1.0%

23%
20%

301 SPECIMENS

SIZE DISTRIBUTIONS

61
301

GLIRES
TOTAL

20.3%

81
296

<20LB
TOTAL

274%

15

301

CARNIVORE
TOTAL

5.0%

192
296

20-75LB
TOTAL

64.9%

B5
301

PERISSODACTYl
TOTAL

28.2%

23

296

>75LB

TOTAL

7.8%

132

301

ARTIOOACTYL
TOTAL

43.9%

296
26

TOTAL MAMMALS
MAMMALIAN GENERA

114%

85
132

PERISSODACTYL
ARTIOOACTYL

64.4%

54
T56

PALEOLAGUS - ISCHYROMYS -
LEPTOMERYX - HYPERTRAGULUS

346%

Fig. 54 (continued). Graphs I-XIII, population statistics.

135

COMPARATIVE SUMMARY

GRAPH XIII

10

20

50

60

70

80

90

100

TOTAL

TOTAL

PERISSODACTYL-
TOTAL

ARTIODACTYL-

TOTAL

PERISSODACTYL-
ARTIODACTYL

MAMMALS < 20LB-

MAMMALS 20-75LB-
TOT AL

MAMMALS >75LB-

MESOHIPPUS
LARGE-MEOIUM TOTAL

MERYCOIDODON
LARGE-MEDIUM TOTAL

XII 205%

>^ tuiuiimniniiniiiiiiuHiiiiiu m ii»»utuii;ii miMMiiiimnimtteciiiin 1 1 uumi i iiiitiiiiuitiiiiHic 27.0%

VI llllilllllllllllll 10.1% X" NEAR-STREAM, TOTAL.

^mimmmmmmmmmmmmmmmmmm^h 37.1%

xl ^^^^M^^^^^^MM^^^^^^^^^MM^BMI^Ma 47.2% v CU SWAMPY

XII 5%

VMM I 4 3% VI SOSM SWAMPY PLAINS.

VI IllinillllllillllHIII 8 3%

. ' * X OPEN PLAINS. SAGE CREEK, TOTAL.

x Ml 4.4%

Xll • 28% XI OPEN PLAINS, DILLON PASS

V BIIIIBIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII 14.9%

VI IIIIIIII IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIillllllllllllMIIB 33.2%

x MMMMMMMM 7.6%

X MMMMMMi 5.0%

Xll 439%

V MMWaMMMMIIIIHIIMIIIIWHIIIIIIIWIIIIIMHIIMWyiliyWIIIIIMJIIIIMIIIIIIil illMII lllllll III llllllllllllllllllllllllllllllllllllllllll 54.7%
VI IIIIIII1IIIIIIIIIIIIIIIIIIIIPIIIIIIIIIIBIIIIIIIIII1IIIII1IIIIIIIIIIIW 48.4%

!■ M1 liM— M—^MMMMM— 50.0%
XIMMMMMMMMMMMMMMMM 43.0%

Xll 64 4%

V lllllllllll>llllllllllllllllllllllllllllllllllllllllllllillll!UI!ll!lllllllllllllllilllllllllll!!l!IIIIIIIIIIKIIIIillll 27.6%

VI MMMMMMMMMMMMMMMNMRM HiniBiiiliinniiiiiiniiiiliiii BiiiiiiiniiiiiiiiiiiiiiiiiiiiiiiiiiiB inn miiiiiniiiiiii! i niiiim minim 68 6%

X ■ 15.2%

XI 115%

Xll 27.4%

V WiiiiiiiiiHiiiiiiiiiiiiaiinsittiliiui::;^ 71.9%
VI »!|i|lllllllll!!IIIIIIIIIIIIIlll!lilllM Illllllllllllllllllllllllllll 48.4%

IMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMI 81.5%

XI MMMMMMMMMMMMMMMMMMMMMMMMMMMMMMMM1 91.7%

X" 64 9%

v iiiiiiiiiiniiiNiiiiniiiiiiiiiiiiiiigiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiniiiiiiiiiiiiiinttiiuiiiiiiiiiii 26.6%

VI llllllllimimilllllllllllimmillllimSIIB 48.8%

XMMMMMMMMMMM 16.8%

XI MM 5.0%

Xll 7.8%

V Bill 2.9%
VI .•• 2.8%

XM 18%

XIMMI 33%

Xll 330%

V : f>;.i: ;..:i. ■ . ■ : mi;:-, mm;:-. u:!: : M:::. ■:.: ijiiii-- ::.i : ::nili^;rMli:n:.ili -.imi.H-:-: 41.4%

nMMMMMMMMMMMMMMMMMMMMMI 59.0%
XMMMMMMMMMMM 32.0%
'^MMMMMB 20 0%

Xll 43 0%

V uiiltstMiiiillli ■ .Mil : .;iiim -.:::i-,,:iii! :i:u: .11 ■ ■■nii:i!i'=':i! . ;li- illiri-iilinii1:;!!!;:: 36.6%

vi wiiiiiiiiiiiiinigiiiiiiiiiiiiiniiuiiiiuiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiii 24.0%

X MMMMMMMMMMMMMMMMMMMMMMMMMi 41.0%
XI^M^^MIMM 13.3%

Fig. 54 (continued). Graphs I-XIII, population statistics.

136

Fig. 55. Internal consistency of Open Plains and Near Stream collections. Note that the first two items show general consistency;
all others are consistent within each facies, but differ between facies.

a c

£'1

a c

OPEN PLAINS (COLL. I, II, III, IV, V, X, XI)

1. Paleolagus and Ischyromys form the majority of Glires.

2. Carnivores under 10% of total.

NEAR-STREAM (COLL. VII, VIII, IX, XII)

1. Paleolagus and Ischyromys form the majority of Glires.

2. Carnivores under 10% of total.

a
§

a

3
O

&

.9

i

.2

o
O

1 . Mesohippus +Merycoidodon culbertsoni under 20 % of total.

2. Medium-sized animals under 30% of total.

3. Hesperocyon over 50% of total carnivores (except in Coll. I

which has only 1 carnivore specimen).

4. Perissodactyls less than 20% as numerous as artiodactyls.

5. Mesohippus +Merycoidodon=\ess than 50% of Ischyromys

+Paleolagus+Leptomeryx+Hyperlragulus.

6. Leptomeryx approximately = Paleolagus +Ichyromys.

1 . Mesohippus -\-Merycoidodon culbertsoni= 40 %-70 % of total

2. Medium-sized animals 50%-80% of total.

3. Hesperocyon less than 30% of total carnivores.

4. Perissodactyls over 50% as numerous as artiodactyls.

5. Mesohippus +Merycoidodon=more than 100% of Ischy-

romys +Paleolagus +Leptomeryx +Hypertragulus.

6. Leptomeryx much less numerous than Paleolagus + Ischy-

romys.

Fig. 56. Total fauna represented in the collections used in Chapter VII. Some non-mammal additions from 1964-65 collections have
been listed.

I. Plants

Charagonia

Unidentified algal strands and sheets

Celtis

II.* Invertebrates

Pond Snails — 3 genera not yet identified
Unio

III.* Fishes

Amia, not Amia calva
?Ictalurid

IV.* Amphibia

Bufonid humeri

V. Reptiles

A. Testudinata

1. Stylemys

2. Graptemys

B. Lacertilia

1 . Rhineura

2. Peltosaurus

VI. Mammals

A. Marsupialia

1. Peratherium

B. Insectivora

1 . I clops

C. Lagomorpha

1 . Paleolagus

2. Megalagus

D. Rodehtia

1. Ischyromys

2. Adjidaumo

3. Eutypomys

4. Eumys

5. Prosciurus

E. Carnivora

1. Hyaenodon

2. Hesperocyon

3. Daphoenus

4. Parictis

5. Mustelavus

6. Bunaelurus

7. Dinictis

8. Hoplophoenus

F. Perissodactyla

1. Mesohippus

2. Colodon

3. Hyracodon

4. Caenopus

G. Artiodactyla

1. Leptochoerus

2. Stibarus

3. Archaeotherium

4. Perchoerus

5. Bcthriodon

6. Agriochoerus
1. Merycoidodon

8. Bathygenys

9. Leptauchenia

10. Peobrotherium

11. Hypertragulus

12. Leptomeryx

13. Hypisodus

* These groups represented in 1964-65 collections, not in the collections used in this paper.

137

Chapter VIII

INTERPRETATIVE SUMMARY

This chapter presents in as coherent a fashion as
possible the physical and biotic history established in
the detailed discussions which precede it. The intention
is to synthesize, in order to clarify the main threads of a
story which might otherwise be lost in the morass of
detailed evidence. Where alternative interpretations are
possible, the one most favored is used ; the alternatives
have already been indicated in the preceding chapters.

Laramide orogeny elevated the Black Hills, creating
consequent slopes generally eastward in the area of the
Big Badlands. This simple pattern was modified by
movement along a series of parallel basement faults
trending ESE-WNW, which produced an asymmetrical
trough with its deeper portion south. A high ridge of
Cretaceous shale hills formed its southern rim, and a
lower ridge the northern.

The system of basement faults extended northwest-
ward through the Mesozoics and Paleozoics which
formed the east flank of the Black Hills uplift. A series
of streams developed along the fault zones, cutting
watergaps wherever they intersected the encircling
hogbacks. Some of these streams lay to the north of the
trough, but all from Rapid Creek southward through
Fall River entered it. Their coalescent waters first cut a
relatively flat surface on the Pierre Shale of the trough
bottom, then cut a flat-bottomed valley 5 miles wide
and about 70 ft. deep, extending down the middle of it.

The valley deepened slowly westward, and shallow-
ed from the longitude of Scenic eastward.

Meanwhile, Paleocene and Eocene erosion, under
warm-temperate humid climate, carved the Black Hills
almost to their present topography. The Precambrian
intrusives of the southern Hills lay widely exposed, as
did also the Precambrian metamorphics and Laramide
intrusives of the northern Hills. The bounding scarps of
the Badlands trough were concomitantly reduced to
linear zones of low hillocks, which weathered to lateritic
soils tens of feet deep.

This long-continued geomorphic equilibrium came
to a sudden stop with the deposition of mature, quartz-
chert gravels by every stream, accompanying the
development of shallow lakes or swamps along their
courses, at distances of 30-60 miles from the Black Hills.
The streams rapidly shifted from gravel to sand
transportation, and brought in quantities of highly ark-
osic sand. Alteration of the feldspar grains to kaolin,
both before and after deposition, indicates that the
climate was still warm and humid. Deposition occurred

as discontinuous lenses scattered over a very wide area.
The Slim Buttes formation comprises the deposits of
this initial Duchesnean depositional episode.

Non-deposition, during which the sediments cracked
and fissured to depths of several feet, ensued. Some
actual removal of Slim Buttes sediments probably oc-
curred also.

The next, or Ahearnian, depositional episode opened
with the central valley well developed, a deep lateritic
soil blanketing the trough lateral to it, and deep soil
on the boundary hillocks of the trough. Abundant
lenses of Slim Buttes sediments lay scattered within the
trough and for many miles to the north of it.

Ahearnian streams within the valley, coalesced to
form the Red River, brought in a flood of fresh arkosic
grit and sand. Slim Buttes materials became incor-
porated with the new sediments to produce the basal
conglomerate of the Chadron Formation: a coarse,
siliceous gravel with arkosic sands and grit. In some
places, the Slim Buttes material was so little rehandled
as to retain its white color; over most of the valley-
bottom, however, the arkose predominated. Red clays
from older Black Hills sediments alternated with green-
ish, montmorillonitic muds to produce a banded,
mottled sequence above the basal sands. The Red River
apparently underwent frequent changes of regimen,
producing cut-and-fill structures and downstream-ori-
ented crossbedding.

Gradually the streams lost both volume and energy.
Wider areas of the Red River valley became flood plain,
and the streams carried smaller pebbles. Deposition
became progressively slower, and ceased entirely when
the old valley was filled approximately to its brim. The
later flood-plain sediments were more tan than greenish,
indicating more oxidation during deposition. Some areas
near stream channels developed pale orange colors, as
alternate wetting and drying altered a modicum of the
iron content to hematite.

A third depositional pulsation started with coarse,
arkosic elastics and continued to deposit gradually
finer materials at a lessening rate. Greenish to grayish
clays very rich in montmorillonite spread out over the
old valley and across the entire trough. The streams
maintained their approximate geographic positions as
deposition proceeded. Once again, the finer sediments
(which now constitute the Peanut Peak member) near
the top of the sequence and near the stream courses
were oxidized and incipiently lateritized.

138

CLARK: INTERPRETATIVE SUMMARY

139

The Chadron Formation thus represents two epi-
sodes of fluvial sedimentation, each starting with wide-
spread deposition of coarse elastics by streams of high
volume and energy, and progressing generally toward
slower deposition of finer material, with a high per-
centage of altered volcanic ash. The sediments indicate
a lessening of vegetation, and periodic drying, toward
the end of each episode.

Annual temperatures of about 60-63° F in South
Dakota and 55-60° F in southwestern Montana ap-
parently characterized pre-Chadronian climates. Oxida-
tion of sediments, chemistry of bone preservation, di-
rection of ash drift from volcanic centers, and nature
of the fauna all indicate a late Eocene climatic regimen
of warm-humid monsoonal character. Winters were
mild, with clear skies and little or no frost. Summers
were warm, with prevailing northerlies bringing in moist
Arctic air, which precipitated heavy rains as it rose to
higher elevations. The Central States, below 1000 ft.
elevation, were probably dry savannas to semi-deserts.

Studies of Duchesnean climates must await more
evidence. The kaolinization of feldspars as far north as
the Slim Buttes certainly indicates more warmth and
humidity than at present, and probably more than
during the succeeding Chadron time.

Chadronian time was definitely cooler than the
preceding Eocene. Lateritization decreased relative to
its previous level; bone preservation changed; feldspar
grains were carried by streams and deposited in rela-
tively fresh condition. The prevailing westerlies began
to establish themselves as dominant over the old mon-
soonal circulation, bringing showers of ash from the
Yellowstone Park centers. However, the cooling did
not progress very far. Small alligators still lived in the
swampy streams, and the influx of volcanic ash weath-
ered almost completely to bentonite. Two periods of
sedimentation started each with water and presumably
vegetation abundant enough to keep the iron in the
sediments reduced, and ended much drier but probably
not much cooler. Evolution of several phyletic lines
within successive faunas shows that these depositional
episodes lasted at least as long as any one Pleistocene
glacial-interglacial cycle.

The fauna of South Dakota at the beginning of
Chadron time consisted primarily of a relict warm-
temperate forest chronofauna with known Late Eocene
ancestors. To this were added a few immigrant savannah
forms. Throughout Chadron time, the balance shifted
toward the immigrants. Extinction of titanotheres at
the close of Chadron time meant the end of what had
been one of the dominant super-families for several
million years.

Faunal evidence supports the petrologic indications
that Chadronian time witnessed a cooling and drying,
with savanna-savanna forest conditions replacing the
older wet forest.

Since there is no evidence of structural movement of
basins relative to mountains, or of the entire area rela-
tive to sea level, and since Chadronian deposition began

at only slightly different times in every east-flowing
stream from Montana to Colorado, climatic control of
the depositional regimen is probable. The change in
both chemical and physical erosion incident to a general
cooling resulted in a change from warm-humid types of
stream gradients to semiarid types. This necessitated
deposition close to the source mountains, in order to
establish the smooth, fairly steep gradients characteris-
tic of graded stream valleys under semiarid climate. As
the cooling progressed in a series of pulsations, so also
did deposition. Volcanic ash weathered concomitantly
with its fall, and influenced the manner of deposition
but did not control the system.

By the end of Chadronian time, the area outside the
Black Hills had been built to an almost featureless
plain, with a few hillocks of Pierre shale still protruding
along the old Sage Ridge. Streams from the southern
Black Hills were carrying gravel out as far east as the
western part of Indian Creek basin, then losing power
quite rapidly. Marshes and shallow, marshy lakes
spread over much of the interstream plains and east-
ward from Indian Creek. This equilibrium maintained
itself for a considerable time.

During the Chadron-Brule interval, the streams
from the Black Hills established individual courses
within the trough, and at least one from north of the
trough had transgressed the old, partly buried Sage
Ridge divide.

Brule deposition, unlike that of the Chadron, was
initiated during a time of low stream volume and
velocity. (It may have been, and probably was, higher
than during the interval preceding). The change in
regimen from almost perfect equilibrium to active
deposition may have been caused by a climatic shift
toward greater coolness and aridity, or by a change in
rainfall distribution, but most probably both occurred.

Infrequent but widespread floods resulting from
rains in the Black Hills caused sheets of thickly fluid
mud to cover the plains. Laden with gelatinous mont-
morillonite, the material overpassed the banks of the
shallow channelways, thickening as it travelled both by
dehydrating and by picking up chips of dried mud from
the surface. It literally plastered to the surface the
skeletons and bones that lay there, congealing upon
them as the dry bones absorbed water from the stiffen-
ing mass.

Each flood deposited from a few inches to two feet of
sediment. Periods of a few years to less than 100 years
elapsed between major floods, during which the grasses
and smaller herbaceous vegetation re-established them-
selves. Trees and larger shrubs probably were not
killed. The fauna very quickly re-established itself:
larger mammals and climbers were merely inconven-
ienced, and smaller forms moved in from Sage Ridge,
the Pine Hills, the Black Hills, and any unflooded areas
on the plain. By the next time of flooding, the area had
re-established its biota and its perthotaxic assemblage.

The region developed a truly depositional geo-
morphic pattern, with rios, probably forested varzea

140

FIELDIANA: GEOLOGY MEMOIRS, VOLUME 5

slopes, drenaje streams without headwaters, and prob-
ably savanna-prairie igapos between the outermost
streams and the Cretaceous ridges. Even during inter-
flood periods, none of the streams accomplished any
erosion.

The climate, apparently, was considerably drier
than that of late Chadron time. Apparently it was also
somewhat cooler: alligators had disappeared from the
streams, and titanotheres from their borders. The flood-
plain clays were extensively oxidized to limonitic colors,
with only a faint mottling of greenish, reduced zones;
rotting vegetation must have been less abundant and
drying more profound. For the first time, groundwater
became influent from the streams to the plains. It was
apparently very limy, as opposed to the more acid water
which probably flowed down the same streams during
Chadron time. The climate was sufficiently arid to pro-
duce perthotaxy typical of semiarid districts, which
suggests annual precipitation of under 30 in. However,
winter temperatures probably were very much less
severe than present ones, with only occasional frosts.

A sudden change toward warmth and greatly in-
creased rainfall altered the nature of the floods without
causing a shift in fluvial regimen. Individual streams
tripled in size. Their confluent floodwaters spread over
the area as widespread, shallow, temporary lakes which
drained away, probably within a few days or weeks
after each flood. The quantity of water and elastics so
far diluted the available montmorillonite that the floods
acted as normal, muddy water rather than as the
thicker, more viscous fluids of the previous arid times.
At least one alligator worked its way back into the area,
demonstrating that precipitation increased concomi-
tantly with increased warmth. Lakes borne of floods
alternated so rapidly with exposed flats that neither a
lacustrine nor a terrestrial biota had time to establish
itself. Westward from Cottonwood Pass, the streams
increased like those to the east, but slightly steeper
gradients due to proximity of the Black Hills prevented
widespread flooding.

A brief period of diminution of streams in the central
part of the area was followed by a similar decrease in
streams watering the eastern part. Vigorous action with
widespread flooding resumed very quickly.

A third time of aridity with intermittent deposition
by sheet mudflows produced sediments like those of the
first. It was just as intense and widespread, but pre-
sumably not as long-lasting as the first.

Suddenly-renewed, vigorous flooding once more
spread temporary lakes across the entire area. Vigorous
action by the repleted streams brought in abundant
elastics which, mixed with the omnipresent montmoril-

lonite, settled to form wide-spread layers of muddy
sand.

A brief return to aridity at the headwaters of the
streams watering the central part of the area produced a
fourth quite thin, local mudstone layer. Quick resump-
tion of warmer, more humid climate over the whole area
built up more laminated sediments, which covered the
local mudstone and elsewhere formed a sequence with
the sediments of the preceding humid episode.

A fifth cooler, arid period became general over the
entire area. It lasted as long as the first one. Except for
a brief, local return to moist conditions along one
stream, it continued into Poleslide Member time.

These alternate episodes of warmer, more humid and
cooler, drier climate comprised a time span of between
1100 and 11,000 years. Individual warm and cool times
were, therefore, of the same order of magnitude as the
post-Pleistocene warm and cool periods. Chadronian
deposition, by contrast, had been controlled by cli-
matic rhythms as long as Pleistocene interglacial epi-
sodes.

The mammalian fauna which occupied the area
during the more arid times found at least three well-
differentiated habitats. There were relatively well-
drained, open gallery forests on the varzea slopes within
a mile or less of the rios, savanna prairies in the igapo
and near-drenaje areas, and occasional well-vegetated
swamps or ponds in both igapo and drenaje bottoms.

Each of these habitats was occupied by a characteris-
tic mammalian community.

Mesohippus and Merycoidodon together formed 55
per cent of the mammalian population of the forests.
A richly varied assemblage made up the remainder.
Paleolagus and Leptomeryx entered the forests, but
definitely did not prefer them. Relict genera from the
old Eocene warm-forest chronofauna, especially peris-
sodactyls, were significant elements of the community.

The open plains, on the other hand, constituted the
preferred habitat of a dominantly small-animal fauna,
of which Paleolagus and Leptomeryx were far the most
numerous. The great majority of the more successful
genera in this habitat were members of the dry-plains
fauna which had first invaded during Chadronian time.

The swamp fauna somewhat resembled that of the
gallery forests, save that here Mesohippus found its
most successful retreat, while Merycoidodon was less
important. Herds of Hypertragulus also found shelter in
the swamps.

During the times of repeated flooding, lacustrine
conditions alternated with fluvial so rapidly that neither
an aquatic nor a terrestrial biota was able to establish
itself.

Chapter IX

CONCLUSIONS

Most of the conclusions listed here have already been
listed in their respective chapters. They are repeated,
with such additions as the preceding text justifies, for
the sake of continuity and the reader's convenience.

1. Laramide movement produced an asymmetrical
trough extending east-southeastward from the Black
Hills. The deepest portion and highest bounding scarp
lay to the south.

2. Streams departed from the east flank of the Black
Hills through structurally-controlled gaps, flowing
thence east-southeastward in roughly parallel courses.
Several occupied the trough, while others lay to the
north of it.

3. A dominantly erosional regimen with a warm-
humid climate persisted through Paleocene and Eocene
times. The bounding scarps of the trough were worn
down to low ridges; deep, lateritic soil developed on the
ridges, and shallower soils on the floor of the trough.

4. During Duchesnean time, streams of high velocity
brought coarse, maturely weathered gravel out 60-100
miles from the Black Hills.

5. Immediately following channel-bed deposition of
the gravel, large quantities of partially-weathered arko-
sic sand, mixed with clay, filled the channels and numer-
ous temporary lake basins.

6. Kaolinization affected the feldspar grains of the
sands both before and after deposition, indicating a
very warm temperate to subtropical climate in the area
of deposition during Duchesnean time.

7. Sufficient time elapsed after deposition of the
Duchesnean Slim Buttes Formation to permit weath-
ering of feldspar grains, cementation followed by crack-
ing of the rock lenses, and erosional removal at many
places.

8. During late Eocene and Duchesnean time, mon-
soonal air circulation prevailed over the great Plains and
the Rockies. The mountains and higher basins received
heavy summer rainfall, but the Interior Lowlands were
probably arid or subarid. Temperatures were warm-
temperate or sub- tropical ; the low temperature dif-
ferential between the Pole and the Equator subordi-
nated the hemispheric circulation system to local sys-
tems, in this case a monsoon.

9. The Black Hills and other ranges were eroded
under this regimen until their cores were exposed.
Their relief by the end of Eocene time very closely ap-
proximated that of the present.

10. A major decrease in temperature starting at the
close of Eocene time established a stronger differential
between the North Pole and the Equator, and conse-
quently a stronger hemispheric circulation system with
prevailing westerlies. This period of cooling appears to
represent a minimum in a 10-million-year climatic
cycle coincident with the beginning of a broader down-
ward temperature trend that culminated in the Pleisto-
cene minimum.

11. As a result of this global climatic change the
local climate became drier and somewhat cooler. The
stream regimen was altered and deposition initiated
adjacent to the mountain ranges.

12. Deposition continued into late Chadron times
because of continued climatic deterioration. Fluctua-
tions in deposition during this period are probably
related to minor climatic fluctuations which may be
the results of the same cycles shown during Pleistocene
time as glacial and inter-glacial stages.

13. The earliest Chadron is somewhat younger than
the Vieja. The Yoder fauna is Ahearnian, and the Pipe-
stone Springs is Peanut Peakian.

14. During middle and late Eocene time, a semitrop-
ical rainforest chronofauna developed in the swampy
woodlands of Utah, Wyoming, Colorado, and South
Dakota. This is recorded in fossils of the Bridger and
Uinta Formations.

15. Concurrently, a savannah to arid chronofauna
developed elsewhere, possibly in the Interior Lowlands.
The history of this chronofauna is unrecorded.

16. During Chadronian time, the forest chronofauna
lingered along the stream margins in Dakota. It under-
went gradual, partial replacement by the immigrant
savanna chronofauna.

17. A few of the genera of the forest chronofauna,
among them Trigonias, Mesohippus, and Pseudoproto-
ceras, managed ultimately to evolve into savanna and
plains forms.

18. Study of the Chadron chronofauna tends to
substantiate conclusions drawn earlier from the Texas
Permian by Olson.

19. The Scenic Member of the Brule Formation com-
prises five types of sediments:

1. Limestone

2. Heterogeneous mudstone

3. Laminated clay

141

142

FIELDIANA: GEOLOGY MEMOIRS, VOLUME 5

4. Laminated siltstone and sandstone

5. Cross-bedded sandstone.

These combine to form four sedimentary lithotopes:

1. Silty mudstones

2. Laminated siltstones

3. Laminated mudstones

4. Channel-fill zones.

20. These four lithotopes grade into one another
horizontally but not vertically.

21. The silty mudstone zones transgress the deposi-
tional areas of several Oligocene streams, and can be
traced for distances of several miles.

22. Alternating strata of mudstone and siltstone
have sharp contacts, not gradational ones. No instances
of graded bedding have been observed.

23. The channel-fill sandstones are of two types,
Northern Black Hills derived and Southern Black Hills
derived, each with a characteristic suite of heavy
minerals.

24. The mudstones represent times of discontinuous
deposition by mudflows from flooding streams of small
volume and low energy.

25. The siltstones represent more rapid deposition by
sheet-floods from the same streams at times when their
volume, and therefore their energy, was much increased.

26. The fluctuations in energy of Oligocene streams
in this area were the result of fluctuations in volume
rather than of changes in gradient.

27. The fluctuations in volume of streams resulted
from alternations of wetter and drier climate.

28. The presence of alligators in the underlying
Chadron, and the presence of one alligator in the first
Scenic-Member wet-climate deposit, plus the absence of
alligators in the intervening dry-climate deposit, sug-
gest that the times of dry climate were also times of
cool climate.

29. The fact that fossils representing five mammalian
orders show no differences from bottom to top of the
Scenic Member indicates that deposition of the entire
Member required not over 500,000 years.

30. Individual increments of Scenic Member sedi-
ment, amounting to 6-18 in., were deposited within the
span of a very few days.

31. The complete absence of weathering or of soil
zones at the top of any one increment shows that never
did a period of more than 100 years elapse between
episodes of sedimentation.

32. The presence within any one fossiliferous incre-
ment of a complete perthotaxis indicates that periods of
10 years or over usually elapsed between episodes of
deposition.

33. Using the data from the last two conclusions, the
total time required for deposition of the mudstones of
the Scenic Member was 550-5500 years. Allowing equal
time for deposition of laminated sediments, although
the evidence suggests that they were deposited more
rapidly, the total time represented by the Scenic Mem-
ber was 1100-11,000 years.

34. The alternations of dry-cool and warm-wet cli-
mates were, on this basis, of the same general order of
magnitude as post-glacial warm and cool alternations.

35. The geographic distribution of lithotopes within
the Scenic Member at any one time can best be ex-
plained by comparison with the distribution of sedi-
mentary environments within the Central Amazon
Basin, as described by Sioli (1951).

36. Using Sioli's terms with additions where neces-
sary, the following sub-environments of fluvial sedi-
mentation can be recognized:

Rio: the channel way of a throughgoing
stream.
Varzea: the area of sedimentation outward
from a rio, including the natural levee
and the long backslope away from the
stream.
Drenaje: secondary streams which arise locally
as drainageways in the more or less
linear depressions between adjacent
varzeas.
Terra firme: valley walls, composed of material
older than that being deposited by the
rios.
Igapo: The approximately linear depression
lying between a terra firme and the
adjacent varzea, receiving sediment
chiefly by local wash from the terra
firme, but occasionally by sheet-wash
from the rio.

The body of sediments formed in these environments
through time are termed, respectively, riosome, varzea-
some, drenajesome, and igaposome.

37. Deposition of the Scenic Member in the Big
Badlands was controlled by three rios with sources in
the Southern Black Hills, and two to four rios with
sources in the Northern Black Hills.

38. Although deposition of the Scenic Member was
episodic, there were no periods of erosion and all streams
were continuously at grade to overloaded.

39. The lower part of the Poleslide Member generally
resembles the mudstones of the underlying Scenic Mem-
ber in lithology and origin. Easterly gradation of the
upper mudstones of the Scenic Member into the basal
Poleslide indicates either that the widespread aridity of
Poleslide time started earlier in the Northern Black
Hills than in the Southern, or that the eastern part of
the Badlands area was, by late Scenic Member time,
receiving notably less rainfall than the western part, or
both.

40. The flesh-colored to brilliant, discontinuous, red,
laminated clays of the Dillon Pass and Big Foot Pass
areas represent igaposomes, having as their source local
rainwash from exposed hillocks of weathered Pierre
shale along the Sage Ridge.

41. The dull red-brown colors of the heterogeneous
mudstone zones in the same geographic areas as the

CLARK: CONCLUSIONS

143

igaposomes represent varzea mudstones mixed with
slight amounts of locally-derived red clays.

42. The sediments in northwestern Nebraska, pre-
viously interpreted as paleosols, are actually laminated
siltstone lithotopes similar to those of the Big Badlands.
They represent rapid rather than slow deposition.

43. The structures in northwestern Nebraska pre-
viously interpreted as deep channel-cutting are actually
due to faulting.

44. Faunal zoning of the Orella Member in north-
western Nebraska is based upon erroneous stratigraphy.

45. The evolution of oreodonts within Orellan time
is not established.

46. No satisfactory correlation of subdivisions of the
Orella Member in Nebraska with subdivisions of the
Scenic Member in South Dakota has been achieved. In
view of the brief time represented by the Scenic Mem-
ber, such detailed correlation is unlikely on paleon-
tologic grounds, although it may be achieved through
paleogeography.

47. Both in South Dakota and in northwestern
Nebraska, Middle Oligocene sediments indicate rapid
deposition of increments several inches thick, with no
periods of erosion and no long periods of non-deposition,
under an alternation of warm-wet and cool-dry climates.

48. Three ecologic zones can be recognized in Lower
Nodular Zone paleogeography: 1) A near-stream zone,
probably occupied by gallery forests; 2) an open-plains
area, which might have borne plains, prairie, or savanna
vegetation ; 3) a swamp area within the open plains.

49. Fossil assemblages within all three were buried
by engulf ment; they represent buried perthotaxies
rather than transported assemblages.

50. Differences between the life assemblage as a
universe, and our collections as a sample, are caused by
at least 29 factors which can be considered in seven
groups.

51. These factors affected various genera both inde-
pendently and by interaction.

52. Due to the number, complexity, and interaction
of these factors, only a perthotaxic assemblage can ap-
proximate an adequate sample of a life population. The
various mechanical assemblages of fossils bear no more
than an occasional coincidental relationship which does
not justify even the most rudimentary statistical analy-
sis.

53. The collections at hand show a 10 per cent ob-
served variance between collections of 300 or more
specimens; 20 per cent variance for collections of 100-
300 specimens, and much wider variance for smaller
collections.

54. The Orellan near-stream gallery forest fauna was
dominated by Mesohippus and Merycoidodon, with a
large number of other genera represented by relatively
small numbers of individuals.

55. The open-plains fauna comprised mostly small
animals, with Leptomeryz and Paleolagus making up
more than half of the total.

56. The swamp fauna resembled the forest fauna
more nearly than that of the open plains.

57. Mesohippus was much the most numerous medi-
um- to large-sized mammal in the swamp; apparently
the genus was attempting to meet the changing climate
by clinging to the relict patches of swampy forest.

58. In general, the invading plains-savannah fauna
which first appeared during Chadronian time had es-
tablished dominance by Orellan time.

REFERENCES

Ahlmann, H. W., son.

1953. Glacier variations and climatic fluctuations.
Bowman Memorial Lectures, series 3, Ameri-
can Geographical Society.

Bradley, W. H.

1948. Limnology and the Eocene Lakes of the
Rocky Mountain region. Bull. Geol. Soc.
Amer., 59, no. 7, pp. 635-648.

Brooks, C. E. P.

1948. Climate through the ages. Ernest Benn, Ltd.,
London.

Bump, J. D.

1956. Geographic names for members of the Brule
Formation in the Big Badlands of South
Dakota. Amer. Jour. Sci., 254, pp. 429^32.

Clark, John.

1937. The stratigraphy and paleontology of the
Chadron Formation in the Big Badlands of
South Dakota. Ann. Carnegie Mus., 25, art.
XXI, pp. 261-350.

1954. Geographic designation of the members of the
Chadron Formation in South Dakota. Ann.
Carnegie Mus., 33, art. II, pp. 197-198.

1962. Field classification of red beds. Bull. Geol.
Soc. Amer., 73, no. 4, pp. 423-428.

Cotton, C. A.

1941. Landscape as developed by the processes of
normal erosion. Cambridge Univ. Press, pp.
1-301.

Davis, W. M.

1923. The cycle of erosion and the summit level of
the Alps. Jour. Geol., 31, no. 1, pp. 1-41.

Dorf, Erling

1955. Plants and the geologic time scale, in The
crust of the earth. Geol. Soc. Amer., Special
Paper 62, pp. 575-592.

1959. Climatic changes of the past and present.
Contrib. Mus. Paleo. Univ. Mich., 13, no. 8,
pp. 181-210.

Eardley, A. J.

1962. Structural geology of North America, 2nd ed.
Harper and Row, New York.

Emmons, W. H., Thiel, G. A., Stauffer, C. R., and
Allison, I. S.
1955. Geology: Principles and Processes, 4th ed.
McGraw-Hill, New York.

Falkenbach, C. H. and Schultz, C. B.

1951. Suggested correlations of the Oligocene of
Nebraska and South Dakota. Guidebook,
Fifth field conference, Soc. Vert. Paleont., pp.
47-50.

Holmes Symposium

1964. The Phanerozoic time scale, a symposium.
Quart. Jour. Geol. Soc. London, supplement,
120 S, pp. 171-191.

Hough, Jean

1948. A systematic revision of Daphoenus and some
allied genera. Jour. Paleo., 22, no. 5, pp.
573-600.

Houston, R. S.

1964. Non-paleontologic methods of correlation of
rocks of Tertiary age in Wyoming. Pt. III.
The petrographic calendar. Univ. Wyo. Con-
trib. Geol., 3, no. 1.

Kay, J. L.

1934. The Tertiary formations of the Uinta Basin,
Utah. Ann. Carnegie Mus., 23, no. 6, pp.
357-371.

Krynine, P. D.

1949. The origin of red beds. Trans. N. Y. Acad.
Sci., series II, 2, no. 3, pp. 60-68.

Lambe, L. M.

1905. Report of field work in the Cypress Hills,
Saskatchewan, and list of vertebrates col-
lected. Geol. Surv. Canada Summary Rept.
1904 (Ann. Rept. 16) A, pp. 364-371.

1908. The vertebrata of the Oligocene of the Cy-
press Hills, Saskatchewan. Geol. Surv. Can-
ada Contrib. to Canad. Paleont., 3, pt. 4, pp.
1-64.

Lawson, A. C.

1932. Rainwash erosion in humid regions. Bull.
Geol. Soc. Amer., 43, no. 3. pp. 703-724.

144

REFERENCES

145

Leopold, L. B. and Wolman, M. G.

1957. River channel patterns: braided, meandering,
and straight. U.S. Geol. Surv., Prof. Paper
282-B, pp. 39-84.

MacDonald, J. R.

1958. The "Rosebud Formation" of western South
Dakota. Proc. S.D. Acad. Sci., 36, (1957) pp.
113-114.

Malhotra, C. L. and Tegland, E. R.

1959. A new Tertiary Formation in Harding Coun-
ty, South Dakota. Proc. S. Dak. Acad. Sci.,
37, pp. 263-274.

Nicknish, J. M.

1957. Investigation of the basal ash of the Arickaree
Formation in northern Shannon County,
South Dakota. Bull. Geol. Soc. Amer., 68, no.
12, part. 2, p. 1868 [abstr.].

Olson, E. C.

1952. The evolution of a Permian vertebrate chrono-
fauna. Evolution, 6 (2), pp. 181-196.

1957. Size-frequency distributions in samples of
extinct organisms. Jour. Geol., 65, no. 3, pp.
309-333.

Olson, E. C, and Beerbower, J. R.

1953. The San Angelo Formation, Permian of Tex-
as, and its vertebrates. Jour. Geol., 61, no. 5,
pp. 389-423.

Orr, H. T.

1964. Some criteria for the recognition of braided
stream deposits. Univ. Wyo. Contrib. Geol.,
3, no. 1, 1964, pp. 1-14.

Osborn, H. F.

1929. Titanotheres of Ancient Wyoming, Dakota,
and Nebraska. U.S. Geol. Surv., Mono. 55.

Payne, J. E.

1965. A summer carrion study of the baby pig Sus
scrofa Linnaeus. Ecology, 46, no. 5, pp. 592-
602.

Plumley, W. J.

1948. Black Hills terrace gravels: a study in sedi-
ment transport. Jour. Geol., 56, no. 6, pp.
526-577.

Ritter, J. R. and Wolff, R. G.

1958. Channel sandstones of the eastern section of
the Big Badlands of South Dakota. Proc. S.D.
Acad. Sci., 37, pp. 184-191.

Scott, W. B.

1898. Notes on the Canidae of the White River
Oligocene. Trans. Amer. Philos. Soc, 19 (n.s.),
p. 362.

1945. The Mammalia of the Duchesne River Oligo-
cene. Trans. Amer. Philos. Soc, 34 (n.s.), pt.
Ill, pp. 209-253.

Scott, W. B. and Jepsen, G. L.

1937. The mammalian fauna of the White River
Oligocene. Trans. Amer. Philos. Soc, (n.s.),
28, pt. I, pp. 1-153.

1941. The mammalian fauna of the White River
Oligocene. Trans. Amer. Philos. Soc, (n.s.).
28, pt. V, pp. 747-980.

Schultz, C. B. and Falkenbach, C. H.

1954. Desmatochoerinae, a new subfamily of oreo-
donts. Bull. Amer. Mus. Nat. Hist., 105, art.
2, pp. 143-256.

1956. Miniochoerinae and Oreonetinae, two new
subfamilies of oreodonts. Bull. Amer. Mus.
Nat. Hist., 109, art. 4, pp. 373-482.

Schultz, C. B., Tanner, L. G., and Harvey, C.

1955. Paleosols of the Oligocene of Nebraska. Bull.
Neb. State Mus., 4, no. 1, pp. 1-15.

Schultz, C. B. and Stout, T. M.

1938. Preliminary remarks on the Oligocene of
Nebraska. Bull. Geol. Soc. Amer., 49, no. 12,
p. 1921 [abstr.].

1955. Classification of Oligocene sediments in Ne-
braska. Bull. Univ. Neb. State Mus., 4, no. 2.

1961. Guide book for the ninth field conference of
the Society of Vertebrate Paleontology. Spe-
cial publ., Univ. Neb. State Mus., no. 2.

Seefeldt, D. R. and Glerup, M. O.

1958. Stream channels of the Scenic Member of the
Brule Formation, Western Big Badlands,
South Dakota. Proc. S. D. Acad. Sci., 37, pp.
194-202.

Simpson, G. G.

1946. Discussion of the Duchesnean fauna and the
Eocene-Oligocene boundary. Amer. Jour. Sci.,
244, pp. 52-57.

Sinclair, W. J.

1921. The "Turtle-Oreodon Layer" or "Red Layer",

a contribution to the stratigraphy of the

White River Oligocene. Proc. Amer. Philos.

Soc, 60, no. 3, pp. 457-166.
1924. The faunas of the concretionary zones of the

Oreodon Beds, White River Oligocene. Proc.

Amer. Philos. Soc, 63, no. 1, pp. 94-133.

Sioli, Harold

1951. (1) Algunus resultados e problemas da limno-
logia Amazonica ;

(2) Sobre a sedimentacao na varzea do Baixo
Amazonas: Boletin Tecnico Institute Agro
nomico do Norte, no. 24, pp. 3-44, 45-65.

Stovall, J. W.

1948. Chadron vertebrate fossils from below the rim
rock of Presidio County, Texas. Amer. Jour.
Sci., 246, no. 2, pp. 78-95.

146

REFERENCES

Trewartha, G. T.

1954. An introduction to climate. McGraw-Hill
Book Co., pp. 1-402.

Wanless, H. R.

1922. Lithology of the White River sediments. Proc.
Amer. Philos. Soc, 61, no. 3, pp. 184-203.

1923. The stratigraphy of the White River beds of
South Dakota. Proc. Amer. Philos. Soc, 62,
no. 4, pp. 190-269.

Ward, Freeman

1921. Geology of a portion of the Badlands. S.D.
Geol. Nat. Hist. Surv. Bull., no. 11.

Weigelt, Johannes

1927. Rezente Wirbeltierleichen und ihre palaobio-
logische Bedeutung. Max Weg. Leipzig, pp.
1-227.

Williams, Ernest

1952. A staurotypine skull from the Oligocene of
South Dakota. Harvard Mus. Comp. Zool.,
Breviora, no. 2, pp. 1-16.

Wood, A. E.

1937. The mammalian fauna of the White River
Oligocene, Pt. 2, Rodentia. Trans. Amer.
Philos. Soc, 28, pt. II, pp. 155-268.

1955. Rodents from the lower Oligocene Yoder For-
mation of Wyoming. Jour. Paleo., 29, no. 3,
pp. 519-524.

Wood, H. E., II

1927. Some early Tertiary Rhinoceroses and Hyra-

codonts. Bull. Amer. Paleo., 13, no. 50, pp.

1-104.
1934. Revision of the Hyrachyidae. Amer. Mus.

Nat. Hist. Bull., 67, art. 5, pp. 241-242.

Wood, H. E., Chaney, R. W., Clark, J., Colbert,
E. H., Jepson, G. L., Reside, John B., Jr., and
Stock, C.

1941. Nomenclature and correlstion of the North

American Continental Tertiary. Bull. Geol.

Soc. Amer., 52, no. 1, pp. 1^18.

APPENDICES

Appendix I
COLUMNAR SECTIONS

Zone VI

Key:

1. All measurements given in feet, to the nearest foot.

2. Roman numbers refer to mudstones as numbered in text.

3. Abbreviations:
G Gray

Y Yellow
R Red
Gr Green
B Buff

Occasional combined letters indicate either a color intermediate
between two of these, or a color which grades from one to the other
irregularly. Color symbols placed thus q > indicate a vertical
color gradation.

CI Mudstone or laminated clay
Sit Laminated siltstone
Ss Sandstone, either laminated or cross-bedded

IV

III

II

CI, dark gray

2

YC1

10

White concretionary zone

1

Y CI with thin concretionary bands

15

White concretionary zone

1

Grades to dark at top

GCl

5

GSlt

12

YB CI

3

GSlt

2

YB CI (discontinuous)

2

G Sit with many concretions

13

G CI bottom not exposed

Total

67

SECTION 1

SWM of NWM, Sec. 5, T.4S, R.13E.

Poleslide — Massive mudstone

Zone VI
V

IV

III

GCl

22

White concretions

1

i
f

1

GCl

GSlt

GCl ?

24

GSlt
G CI

1

J

GSlt

26

YGCl
GrSlt
YGCl

10
2

4

Banded Gr Sit-

-YC1

25

YGCl

28

Chadron

Total

142

SECTION 2

Center, SEJ^, Sec. 5, T.4S., R.13E.

This section is mostly mudstones, which have replaced the
siltstones by lateral gradation. It lies near the position of the divide
between Northern- and Southern-derived sediments, which may ac-
count for thin section and preponderance of fine-grained sediment.

Poleslide — Massive CI

White zone 1

SECTION 3

SWM of NEM. Sec. 4, T.4S., R.13E.

This section lies just west of a N-S trending channel zone of
Northern-derived sediments; within this channel area, zone II dis-
appears, and zones III and IV merge to one thick mudstone, sepa-
rating again on the other side. The channel zone extends upward
in section from just above mudstone I to the base of mudstone IV.
A few very small, discontinuous channel fills lie higher. A 10-inch
zone of blue-gray siltstone, very prominent and containing numer-
ous biotite flakes, develops near the top of mudstone IV. Mud-
stone II appears on both sides of the channel area, and is at one
place traceable across.

Zone VI
V

IV

III

II

Poleslide

Dark G CI

YG CI, locally white

White zone

3
14

1

GCl

9

G Sit, laminated, some Gr Ss

36

Y CI

10

GSlt

3

YGCl

9

GSlt

19

Y CI

5

Gr Sit & Ss

3

YG CI bottom not exposed

Total

112

147

148

FIELDIANA: GEOLOGY MEMOIRS, VOLUME 5

SECTION 4

SEJi of NWM, Sec. 3, T.4S., R.13E.

A prominent white zone here separates zone VI from the gray
siltstones below. Northward at the standard section of the Scenic
Member (Section 5, this appendix), zone VI grades laterally into
banded siltstones. A dark gray zone immediately overlying sepa-
rates the Scenic from the overlying Poleslide at both localities.

Zone VI

IV

III

II

Poleslide

GC1

YC1

3

15

White concretionary
GSlt

zone

2
10

YGCl

3

GSlt

26

YBC1

4

GSlt

9

YGCl

GSlt

YGCl

3

1

10

14

GSlt

8

YB CI

3

GSlt

12

YB CI bottom not exposed

30

Total

109

IV

III

II

White Ss-Slt

3

YG CI, dark

4

White Ss-Slt

G Sit with some CI laminae

3
30

YC1

10

G to white Sit, concretionary

2

YG CI (some Sit laminae)

12

GSlt

8

GC1

2

G Sit and Gr Ss

11

YB CI, nodular
Base not exposed

12

Total

116

SECTION 5

SEK of SWM, Sec. 27, T.3S., R.13E.

This is the standard section of the Scenic Member (Bump,
1956). Through a printer's error, this section was erroneously
placed in Sec. 23, but Bump personally showed the locality to the
author and mentioned the error. Likewise, the standard section
of the Poleslide is in Sec. 33, not 23, of T.43N., R.44W.

SECTION 7

SE14 of SEM, Sec. 35, T.3S., R.13E.

This section lies on the NE borders of a zone of channel fills,
trending SE. The best-developed sandstones lie on the southwest
flank of Heck Table and on the east side of the butte in the SW^
of the SEJ4, Sec. 35. Maximum diameter of pebbles is 10-12 mm.
Stratigraphically, the greatest development is above mudstone III
in both thickness and areal extent. The sandstones interfinger
with and replace the mudstones in this zone. Zones IV and V
alter completely to siltstones in the SE^ of Sec. 35. They are
recognizable and directly traceable in this columnar section and
in Sections 8 and 9.

Poleslide

Zone VI

Zone V

IV

III

II

Poleslide

G Sit, laminae of CI at top

43

YBC1

9

GSlt

8

YB CI

1

G-B Sit

7

YB CI

6

GSlt

4

YB CI

6

G-Gr Sit & Ss

micaceous

8

YBC1

28

Chadron

Total

120

IV

III

II
I

B & dark CI
YB CI
GC1

3

17

6

GSlt

9

YB CI

2

G Sit & CI

13

BGCl

5

GSlt

10

BGCl

GSlt

BGCl

10

2

13

GSlt

8

BGCl

4

G Sit, some discontinuous
laminae

20

YB CI, nodular
Bottom not exposed

8

Total

130

SECTION 6

SWK of SE}4, Sec. 3, T.4S., R.13E.

In this section, zone IV is indistinguishable from the upper
part of III, but the two are separated by a siltstone stratum to
the north, east, and south.

Poleslide — Laminated siltstone

Zone VI Dark CI 4

Y CI with nodules 15

SECTION 8

NWM of NWJ£, Sec. 1, T.4S., R.13E.

Poleslide

[ Dark CI
Zone VI < YB CI, some concretions

i Light Y CI

GSlt

V YGCl

G Sit, pale

4

15

3

24

APPENDICES

149

IV

III

YBC1 2

GSlt 8

fYB CI 8

I Pale G Sit 3

i GB and YB CI 10

GSlt 6
Bottom not exposed

Total 91

SECTION 11
S edge of Sec. 24, T.3S., R.13W.

Poleslide

SECTION 9
NEM of SWJi, Sec. 36, T.3S., R.13E.

Northeastward, the siltstone between III and IV thins mark-
edly, then thickens again and IV merges upward with V. Mud-
stone III divides into two, with a light sandstone in the middle.

Zone VI

IV

III

II

Zone VI

IV

III

II

Poleslide

( Dark CI
i YB CI
1 Gray CI

3

10

3

GSlt

5

GBC1

3

GSlt

11

GBC1

3

G Sit and CI

15

GB CI, 1 white Sit in
middle

16

GSlt

13

GCl

3

GSlt

22

GCl

Base not exposed

8

Total

115

SECTION 10

NWJi of SEM, Sec. 25, T.3S., R.13E.

Poleslide

Dark CI 1
YCl }
GCl J

22

G Sit with laminae of
YCl

18

YCl

7

Gr Ss-Slt

1

YGCl

5

GSlt

13

fYGCl
< G-white Sit
I YGCl

6
5
5

GSlt

6

G CI, some Gr Ss

4

GSlt

30

YGCl

18

Total

140

Zone VI

gcii

YCl }
GCl J

GSlt

15
11

V

GCl

Gr Ss and Sit

8
3

IV

GCl

10

III

f Gr Ss, coarse
J YGCl
1 GSlt
(YCl

GSlt

3

6

11

6

8

II

GCl

GSlt

2
24

I

BGC1

Base not exposed

1
Total

10

117

SECTION 12

NWM of NWJ^, Sec. 30, T.3S., R.14E.

Poleslide

Zone VI

/GCl
\YC1

5
15

V

/ Gr Sit
\ YCl

GSlt
GrSs

16

7

13
0-5

III

YCl

Gr Sit and fine Ss

YCl

GSlt

:}

13

1
24

II

YGCl

GSlt

5
17

I

YCl 1
GSlt }
GCl J

Chadron

Total

32
152

Note.— The Sit in

Zone I is a channel-fill lens,

quite restricted.

Zone VI

III

SECTION 13
Center, Sec. 19, T.3S., R.14E.

Poleslide

GSlt \
YCl /

GSlt

YCl

GSlt
GrSs

YBC1

16

10

8

24
5

18

150

FIELDIANA: GEOLOGY MEMOIRS, VOLUME 5

II

GSlt

11

RC1

6

GSlt

21

Y CI

18

Chadron

Total 137

SECTION 14

RC1
Chadron

35

Total 148

NEi4, Sec. 19, T.3S., R.14E.

Mudstone I thickens downward from the last section; the base
of the Scenic Member has a noticeable relief from this section
southward to Chamberlain Pass. The nodules of I disappear and
the zone becomes reddish brown, immediately northeast of this
section. Zone V develops a nodular zone here with fossiliferous
concretions like those of Zone I farther south. Zones V and VI
become more massive and buff colored, like the overlying Poleslide,
separated from it only by a thin, discontinuous, greenish zone with
limy laminae.

Poleslide

SECTION 16

SEJi of Sec. 18, T.3S., R.14E.

This section was taken up the SE side of the same ridge of
which section 15 represents the SW side.

Zone VI

Zone VI
V

III

IIA

II

Y CI

22

Y CI and very

fineSs

8

GSlt

30

RB CI

13

GSlt

8

(R CI
GSlt
R CI
GSlt

Ir CI

2
5
3
3
3

GSlt

6

R-YC1

3

G Sit, with some fine
green Ss

29

YBC1

37

Chadron

Total

172

III

IIA

Poleslide

G CI, dark

YCl

GCl

3

10
5

18

GSlt

9

GCl

10

GSlt

16

YGCI

11

G-Gr Sit

12

GCl

4

G Gr Sit

33

YG CI

25

Chadron

Total 138

SECTION 17

NWJi of Sec. 17, T.3S., R.14E.

Between this and Section 16, a greenish zone which apparently
forms the top of the Scenic Member, but actually lies in the lower
third of Zone VI, disappears and VI merges with the Poleslide.
Occasionally, a few concretions several feet up in Poleslide lithol-
ogy may mark the top of VI, but this is not demonstrable.

Poleslide (Y CI, bottom is part of VI)

Zone VI Gr CI 3

SECTION 15

NW'| of SEM. Sec. 18, T.3S., R.14E.

Note. — The base of Mudstone I is here not sharply separated from
the Chadron; there are no limy zones at the contact, and almost no
concretions. The unusual thickness here may indicate that part
of the Chadron has been included through error; the next section
shows only 25 feet of I. Also, Mudstone I here includes a greenish
siltstone channel-fill 3 feet thick, within its upper portion.

Ill

IIA

Zone VI

V

III

IIA

II

Poleslide

YCl

22

Gr Ss, fine "l
GSlt \
YCI J

15

RB CI

8

G Sit and Gr Ss, fine

17

RYC1

7

G Sit and Gr Ss laminae

15

R-G CI

3

GSlt

26

G and Gr Sit

7

YG CI

8

G and Gr Sit

16

YGCI

16

GSlt

10

YBC1

5

GSlt

25

YG CI, one Gr Ss lens

35

Chadron

Total

125

SECTION 18

SE}i of SE^. Sec. 17, R.14E., T.3S.

The buff to yellow mudstones have changed to salmon and
brownish reds between Section 7 and this Section.

Poleslide

Zone V

GSlt
Y>

APPENDICES

151

III

HA

GSlt
j B-R

G-Gr Sit, several discontinuous
laminae of B-Cl

YCl

G to pale Gr Sit

Y-R Cl, greenish zone, not lam-
inated, at 20' above base

14

10

24

6

18

28

Chadron

Total 113

SECTION 19

SEH

of NEJi, Sec. 26, T.3S., R.14E.

Poleslide

Zone VI

fGCl
YCl
[ Gr Ss-SIt

1
12
18

V

G-YC1

Gr Ss and Sit

5
28

III

Y-GCl

GSlt

10
10

HA

G-YC1

3

G-Gr Sit 4

G Cl, discontinuous

Gr Sit 28

<)

36

I

B Sit, pale, very few nodules

25

Chadron

Total 117

Note: (1) This is a near-channel-zone section, with much coarse
sediment and the mudstones notably tan to buff rather than red,
like those both east and west of here.

(2) In this area, channel fills appear intermittently
throughout both Scenic and Poleslide; determining the contact is
difficult.

SECTION 20

Center of Sec. 24, T.3S., R.14E.

Poleslide

(Basal 9 feet, below a gray mudstone, may be Scenic VI)

Zone V

III

G-Gr Sit 25

G-Y Cl 5

G Sit 26

YCl 8

Gr-G Sit and Ss, 3 thin, dis-
continuous Y Cl

laminae 38

Y Cl, with a Gr Ss at

about 20' from base 37

Chadron

Total 139

Zone V

III

II A
IIA

SECTION 21
NE'i of Sec. 19, T.3S., R.15E.
Poleslide

GSlt

YCl

GSlt

Y Cl (thickens
NEto4')

GSlt

G Cl

GSlt

Y-BC1

Chadron

16

5

10

2
20 (estimated)

2
42

Total 130

SECTION 22
SWJi of SE^, Sec. 17, T.3S., R.15E.
Poleslide

ZoneV

III

IIA

GSlt

15

YCl 6
GSlt 3
YCl 2

11

GSlt

8

Y-BC1

14

G Sit, one Cl lamina at
22' from base

36

Y-GCl

3

GSlt

18

Y Cl, a few concretions in
upper part

32

Chadron

Total 137

SECTION 23
SWJ-I of Sec. 16, T.3S., R.15E.
Poleslide

Zone V

III

IIA

Note: One of the Y Cl laminae between I and III may repre-
sent IIA, but this cannot be demonstrated.

GSlt 1
YCl
GSlt
YCl
GSlt ,

20 (estimated)

Y Cl

11

G Sit

22

Y Cl

3

GSlt

11

YCl (d

scontinuous)

1

GSlt

10

YCl

2

GSlt

30

Y-G Cl,

no concretions

24

Chadror

i

Total

134

152

FIELDIANA: GEOLOGY MEMOIRS, VOLUME 5

ZoneJV

III

IIA

IA

SECTION

24

HEH of Sec. 15,

T.3S.,

R.15E.

Poleslide

GSlt

20 (estimated)

YCl

6

GSlt

22

f YCl 5
\ G Sit 5
{ YCl 2

12

GSlt

10

Y CI

10

GSlt

17

III

Y CI, turns red to east,
pinches out N

GSlt

YB CI, no concretions,
lower half highly
bentonitic

36

Chadron

Total 142

Note: IA is a red igapo clay of some areal extent. It wedges
in and out, apparently representing a short time of heavy local
rainfall without extensive flooding from the nearest rios.

SECTION 25

NWK of SWM, Sec. 11, T.3S.,

R.15E.

Poleslide

GSlt

25

Zone V

I}*

8

G Sit, discontinuous
bright R lens at 5
above base

30

III

R CI

Gr Ss and Sit

6
6

IIA

GSlt

12
11

IA

R CI

GSlt

4
10

I

R CI
Chadron

28

Total i

[40

SECTION 26

NEJ*

of Sec. 10, T.3S., R.15E.

Poleslide

GSlt

20

(estimated)

Zone V

GCl

GSlt

R CI, Gat base

GSlt

R CI

GSlt

5
26

4

4
5

4

IIA

IA

GCl

GSlt

RC1

GSlt

R CI

Gr Ss and Sit

RC1

GSlt

GCl

Chadron

10

13
2

10
2
6
2
8

30

Total 121

Zone V

SECTION 27
SWJ4 of NEK. Sec. 34, T.2S., R.15E
Poleslide — Contact irregular
GSlt
GCl
GSlt

III

IIA

IA

R CI, bright, banded,
in discontinuous lenses

GSlt

GCl

GSlt

R CI thins northward to
0.5 foot

GSlt

G-R CI

GSlt

RC1

GSlt

R
G

CI, base not
exposed

20 (estimated)
3
9

4
8
7
8

2
10

8
25

2

15
18

Total 139

SECTION 28
NWM of SE}4, Sec. 35, T.2S., R.15E.
Poleslide — Contact irregular
GSlt
III G-R-Gr CI, mottled

GSlt

R CI, bright
GSlt
IIA G-R-Gr CI, mottled

G Sit
IA R CI, bright

GSlt
I G-B Sit, with base green

35

6

10

1

8

5

10

1

10

39

APPENDICES

153

Chadron

Chadron

Total 125

Notes: (1) Mudstone V merges with the Poleslide from here
northeastward. The numbered zones are bentonitic and dull-
colored, not bright red like the small zones.

(2) A quite persistent zone of siltstone occurs 30-40
feet up in the Poleslide in this area; careful tracing indicates that
the base of this is definitely higher stratigraphically than the top
of the Scenic member at the standard section.

(3) All thicknesses above Zone IA in this section are
slightly inaccurate, due to extreme steepness of the cliff this sec-
tion was measured on.

SECTION 29
NWM of NW^, Sec. 36, T.2S., R.15E.

Poleslide — Contact gradational

Zone V

III

IIA

GSlt

Y CI at base

YCl

GSlt

R-GC1

GSlt

R CI bright

GSlt

R CI bright

GSlt

R CI

GrCl
GSlt
R CI
G
R CI

GC1,

concretions
R-G CI,
bentonitic

Chadron

30 (estimated)

3
15
10
16

5

9

5
17

51

Total 161 (140'

14 mile N
of here, due to pinching out of laminae between I
and IIA)

SECTION 30

NWM of NEM, Sec. 36, T.2S., R.15E.

Note: In this area there are many discontinuous lenses of
bright red mudstone, usually under 2 feet thick, especially near
the top of Mudstone I.

Poleslide

G Sit; upper 20' mostly
G CI, possibly Zone V

45

Zone III

G CI, pink near top

8

GSlt

15

RC1

4

GSlt

26

IIA

G CI, R at top

5

GSlt

23

I

G CI, reddish near top

37

Total 163

SECTION 31

NEM of NEK, Sec. 36, T.2S., T.15E.

Poleslide

Zone V

/YCl
\GC1

30 (estimated)

R CI

2

Gr Ss and G Sit

8

III

G CI, R at top

8

GSlt

22

RC1

2

GSlt

15

R CI

1

GSlt

16

IIA

G Gl, R at top

8

GSlt

20

G-Y CI, R at top

12

Bottom not exposed

Total

144

SECTION 32

SWM of NWK, Sec. 33, T.2S., R.16E.

Poleslide Grading Into

Y CI and Y-G Sit

15

III

IIA

GSlt

G-Gr CI, R at top

18
10

GSlt

15

R CI, banded

3

GSlt

6

R CI, bright

1

G Sit and CI

16

G CI, R at top

7

G CI and Sit

5

G CI, R at top

12

Bottom not exposed

Total 108

SECTION 33

SW^ of SEM, Sec. 28, T.2S., R.16E.

200 yards west of Dillon Pass Tourist Camp

Note. Northeastward, Zone IIA retains its identity but comes
to rest directly upon Zone I. This feature occurs apparently along
an E-W zone only 200-300 yards wide; the two are separated again
to the north of this zone (Section 34).

Poleslide

-Contact gradational

YCl

GSlt

10

R CI

1

154

FIELDIANA: GEOLOGY MEMOIRS, VOLUME 5

GSlt

15

Chadron

Zone III

G CI, R at top

GSlt

RCl

7

20

3

Total 79

SECTION 36

GSlt

4

NE

\\i of NEM, Sec. 20, T.2S., R.16E.

IIA

/R CI
\GrCl

5

This section is immediately east of U. S. 16A, on the north or
Bad River side of the White River-Bad River divide.

I

G-R CI

8

All unnumbered red zones are bright, discontinuous lenses.

Bottom not exposed

Poleslide

Total

73

Y CI, concretionary 8

GSlt 8

SECTION 34

RCl 1

NEM of SEM, Sec. 26, T.2S.,

R.16E.

200 yards

SW of Dillon Pass Road Junction

GSlt 5

Poleslide

Ill

RCI 1

GCI } 8

YC1

5

(estimated)

RCl J

GSlt

18

(estimated)

GSlt 9

III

R CI, white zone
in middle

10

R CI 3
G Sit 10

GSlt

19

RCl 2

R CI

4

G Sit 13

GSlt

14

/R CI

\ GCI 6

IIA

G CI, R at top

8

IIA

GSlt

11

G Sit 18

R CI, banded

5

GSlt 4

Gr Ss and G CI,
banded

12

I

/ R CI 34
\GC1

RCl

Chadron

I

Brown CI
Chadron

25

Total 132

Total

128

SECTION 35

Appendix II

SEJ4 of NEJ4, Sec. 20, T.2S.,

R.11E.

This section is east of U. S. 16A, in White River drainage, just
southeast of the White River-Bad River divide.

Mudstone Zone IIA here rests directly upon Zone I, by wedg-
ing out of the intervening sediments. Mudstone Zone III is a
triple bed throughout this area. At the position of this section, it
consists of a blue-gray concretionary mudstone with red layers
above and below.

Ill

IIA

I

Poleslide

Y CI with concretions

10 (estimated)

GSlt

12 (estimated)

GrSs

2

R CI 1
G CI )
R CI J

15

GrSs

4

G Sit with 3 pale
red laminae

25

G CI, dark

4

GCI

17

ABBREVIATIONS USED IN REFERENCE
TO SPECIMEN NUMBERS

AMNH American Museum of Natural History

CM Carnegie Museum

CU University of Colorado Museum

FMNH Field Museum of Natural History

The following letters designate specimens in the
various Field Museum collections:
G Sedimentary Petrology
PM Fossil Mammals
PE Fossil Invertebrates
PF Fossil Fishes
PP Paleobotany
PM Princeton University Museum.

SDSM South Dakota School of Mines and Technol-
ogy Museum.

APPENDICES

155

WORDS INTRODUCED OR REDEFINED
IN THIS PAPER

1. Anataxic factors. The weathering and erosional

factors which act to destroy a fossil during the
degradation of the rocks in which it occurs.

2. Biotic factors. Those factors which determine

whether or not a population of any particular
species will inhabit an area.

3. Death assemblage. The sum total of corpses

which arrive upon a surface between the deposi-
tion of two successive increments of sediment, or
between the inception of one period of incre-
mentation and its termination.

4. Drenaje. A drainage-way without headwaters,

developed in the trough between two actively
depositing streams; also, the stream which oc-
cupies this drainage-way.

5. Drenajesome. The body of sediments deposited

within a drenaje through an episode of sedimen-
tation unmarked by significant erosion.

6. Igapo. The trough which lies between a valley wall

(terra firme) and the outward-sloping deposi-
tional surface of the stream which occupies the
valley.

7. Igaposome. The body of sediment deposited within

an igapo during one depositional episode.

8. Life assemblage. The total number of individuals

within a specified area during a specified time;
the term may be limited to any group, e.g.,
mammals, or cervids, or plants.

9. Perthotaxis. A death assemblage with the animal

corpses in various stages of destruction by the
set of processes normally operative under the
environment concerned.
10. Perthotaxic factors. The factors which act to
destroy animals' corpses within any particular
environment.

11. Perthotaxy. The more or less orderly destruction

of animals' corpses by the combined action of
natural processes.

12. Rio. A throughgoing stream with its provenance

outside the area concerned.

13. Riosome. The body of sediment deposited within

the channel-way of a rio during one primarily
depositional episode (= period of incrementa-
tion) .

14. SULLEGIC factors. Those factors influencing the

collecting of fossils which determine whether or
not any particular fossil at the surface will find
its way into a collection.

15. Taphic factors. Factors determining whether or

not an animal's bones will be buried.

16. Terra firme. The valley walls bounding a valley-

way in which sedimentation is taking place.
Normally composed of material other than that
carried by the rivers.

17. Thanatic factors. The factors surrounding an

animal's death which determine whether or not
its body will arrive upon the surface as a mem-
ber of the death assemblage.

18. Total fossil assemblage. The sum total of fossil

specimens entombed within any particular sedi-
mentary unit, in an area under question.

19. Trephic factors. The factors incident to curating

and identifying a fossil specimen which deter-
mine whether or not a fossil in a collection be-
comes available for use.

20. VARZEA. The depositional surface, including the

natural levee and the long backslope, developed
upon each side of a depositing stream by out-
ward sheet flow over its banks.

21. Varzeasome. The body of sediment deposited upon

a varzea through one depositional episode.

INDEX

Actinolite, 17, 19, 24, 80

Adjidaumo, 27, 124, 134, 135, 137

Agriochoerus, 53, 56, 58, 59, 68, 72, 114, 120, 126, 127,

130, 131, 132, 133, 134, 135, 137
Ahearn (member of Chadron) - ian, 9, 11, 12, 21, 22 23

(described), 24, 25, 56, 57, 59, 60, 68, 69, 73, 74, 75,

106, 138, 141
Allanite, 17, 19

Alligator, 26, 58, 68, 72, 97, 139, 140, 142
Amazon, 94
Amia, 137
Amyda, 26, 58, 68
Ant (Hills), 118, 119
Anosteira, 26, 58, 68, 72
Apatite, 16, 17, 22, 23, 24
Apternodus, 26, 56, 58
Archaeotherium, 52, 58, 69, 83, 97, 114, 118, 126, 127,

130, 131, 133, 134, 135, 137
Ash, volcanic, 13, 20, 22, 23, 60, 64, 65, 66, 96, 97, 102,

106, 139

Badlands National Monument, 75, 111

Barite, 16, 17, 19, 24, 80, 81

Bathornis, 133, 135

Bathygenys, 126, 127, 130, 134, 137

Beaver Divide, 57, 59, 61

Bentonite, 22, 23, 97, 102, 139

Big Sand Draw Lentil, 57, 59

Biotite, 16, 17, 19, 22, 23, 24, 25, 77, 80, 81, 83, 104

Bone, fossil, 24, 139

Bothriodon, 52, 58, 68, 72, 126, 127, 133, 135, 137

Bunaelurus, 132, 134, 137

Caenopus, 52, 56, 58, 69, 114, 120, 126, 130, 132, 133,

134, 135, 137
Calcite, 22, 77, 82, 83, 84, 102, 104, 119
Caliche, 24, 25, 82
Campylocynodon, 28
Carnegie Museum, 21
Cassiterite, 17, 19

Celtis (Hackberry), 83, 91, 98, 123, 125, 137
Chalcedony, 22, 105
Chara (-gonia), 75, 77, 123, 137
Chert, 17, 18, 22, 24, 25, 138
Cheyenne River, 8, 14, 15
Chicago, University of, 21
Chlorite, 16, 17, 80, 81
Cleveland Museum of Natural History, 114
Clinopternodus, 26, 58
Colodon, 51, 58, 68, 72, 125, 126, 130, 131, 133, 134, 135,

137

Colorado, University of, 111

Cook Ranch (Formation), 57

Coprolite, 22, 25, 82, 83, 91, 98, 100, 112

Crazy Johnson (Member of Chadron), 10, 12, 21, 22,

23 (described), 24, 25, 57, 60, 69, 73, 74
Cypress Hills (Formation), 57, 59

Daphoenocyon, 28-32, 56, 58, 68, 69, 74

Daphoenus, 27, 58, 72, 115, 116, 125, 130, 132, 134, 135,

137
Dinictis, 33, 57, 58, 69, 115, 125, 130, 131, 132, 133, 134,

135, 137
Diplobunops, 59
Drape Structures, 13
Duchesne River (Formation) - Duchesnean Time, 56,

57, 59, 73

Eopelobates, 25-26, 58

Eotrigonias, 16

Eotylopus, 55, 58, 68, 72

Epidote, 24, 81

Epihippus, 16, 59, 126

Eporeodon, 54

Eumys,9\, 99, 116, 118, 121, 122, 125, 130, 131, 132, 133,

134, 135, 137
Eusmilus, 33, 58, 69
Eutypomys, 27, 121, 124, 130, 133, 134, 135, 137

Fairburn, Agate, 17, 18

Feldspar, 16, 17, 18, 19, 20, 22, 23, 24, 25, 83, 85, 104,
119, 138, 139, 141

Ganoid, 25

Garnet, 16, 17, 19, 24, 80, 81, 119

Geological Society of America, 21

Glass, 22, 23, 104

Glauconite, 16, 17, 24, 80, 81

Glyptosaurus, 26

Gold, 24, 81

Granite, 17, 22, 80

Graptemys, 26, 58, 68, 124, 133, 135, 137

Graveyard, 23

Gunn, V., Ill

Gypsum, 9, 24, 60

Hackberry — see Celtis

Harrison, T., 21

Harvard University, 21

Hematite, 16, 24, 61, 77, 80, 81, 118, 138

Heptacodon, 53, 58, 68, 72

156

INDEX

157

Hesperocyon, 27, 58, 59, 68, 116, 122, 123, 125, 128, 129,

130, 131, 132, 134, 135, 137
Hornblende, 16, 17, 24, 80, 81
Hoplophoneus, 33, 57, 58, 69, 115, 125, 130, 132, 133,

134, 135, 137
Humus, 23
Hyaena, 125

Hyaenodon, 27, 56, 58, 69, 125, 128, 129, 130, 132, 134,

135, 137
Hypertraguloidea, 55, 131, 135

Hypertragulus, 97, 112, 115, 116, 120, 121, 122, 123, 126,

127, 128, 129, 130, 131, 132, 133, 134, 135, 137, 140
Hypisodus, 126, 128, 129, 130, 131, 132, 134, 135, 137
Hyracodon, 51, 56, 58, 69, 114, 120, 125, 126, 130, 131,
132, 133, 134, 135, 137

Ictops, 26, 57, 91, 99, 116, 124, 130, 131, 134, 135, 137
Interior Zone (and Formation), 9, 12, 17, 18, 98
Ischyromys, 27, 91, 99, 115, 116, 118, 119, 121, 122, 123,
124, 125, 128, 129, 130, 131, 132, 133, 134, 135, 137

Jasper, 18

Kanesky, E., Ill

Kaolin - ite, 9, 18, 19, 20, 24, 138

Lafayette College, 21

La Point (Member, Duchesne River), 59

Laterite- (ization), 25, 60, 138, 139

Leptauchenia, 126, 127, 131, 135, 137

Leptochoerus, 58, 69, 126, 127, 132, 134, 137

Leptomeryx, 112, 115, 118, 119, 120, 121, 122, 123, 126,

127, 128, 129, 130, 131, 132, 133, 134, 135, 137, 140,

143
Leucoxene, 16, 17
Lewis, A. D., 21, 53

Limonite, 16, 17, 22, 24, 61, 80, 81, 82, 119
Livingston, R., Ill

Magnetite, 16, 17, 19, 24, 80, 81

Martes, 125

Meanders, 24

Megalagus, 27, 121, 124, 130, 132, 134, 135, 137

Megalamynodon, 16

Menodus, 50-51, 57, 58, 68, 72

Merycoidodon, 53, 54, 56, 57, 58, 69, 91, 99, 106, 107,

115, 119, 120, 121, 122, 123, 126, 127, 129, 130, 131,

132, 133, 134, 135, 136, 137, 140, 143
Mesohippus, 33-50, 56, 58, 59, 60, 68, 72, 73, 74, 75, 91,

99, 114, 116, 120, 121, 122, 123, 125, 126, 127, 129,

130, 131, 132, 133, 134, 135, 136, 137, 140, 141, 143
Metacodon, 27, 56, 58
Metamynodon, 114, 125
Mica, 17, 19, 20
Microline, 77

Microfauna Locality, 25, 26, 27
Microperthite, 24
Miniochoerus, 107
Missouri Plateau, 13, 14
Monsoon (-al Climate), 66, 73, 74, 139, 141

Montmorillonite, 18, 19, 20, 78, 83, 84, 85, 92, 98, 104,

118, 120, 138, 139, 140
Muscovite, 22, 24, 80, 104
Mustelavus, 32, 58, 69, 130, 134, 137
Nelson, P., Ill
Niobrara (Formation), 9

Oligoclase (& Andesine), 22, 23, 24, 25
Ostracods, 75, 77, 124
Oversteepened Dips, 24

Paleogale, 130, 134

Paleolagus, 57, 91, 99, 115, 119, 121, 122, 123, 124, 125,

128, 129, 130, 131, 132, 133, 134, 135, 137, 140, 143
Paleosol, 102-107, 110, 143
Paradjidaumo, 27

Parictis, 27, 28, 29, 57, 58, 59, 69, 74, 134, 135, 137
Peanut Peak (Member of Chadron) - ian, 10, 12, 21, 22,

23 (described), 24, 25, 56, 57, 59, 60, 68, 69, 73, 74,

75, 82, 138, 141
Peltosaurus, 26, 124, 130, 134, 137
Peratherium, 26, 57, 58, 68, 124, 130, 134, 135, 137
Perchoerus, 52, 58, 69, 126, 127, 130, 133, 134, 135, 137
Perenyi, Dr. T., 5
Perthotaxy (-is), 99, 100 (illustrated), 101, 112, 118,

122, 142
Pierre Shale, 9, 14, 17, 21, 22, 60, 97, 110, 119, 138, 139,

142
Pine Ridge Structure (Pine Hills), 14, 15, 22, 97, 139
Pinnacles, 8, 14
Pipestone Springs (Formation), 56, 57, 59, 73, 75, 124,

141
Plesictis, 32
Poabromylus, 59
Poebrotherium, 55, 58, 60, 69, 126, 127, 130, 132, 133,

134, 135, 137
Princeton University, 21
Prosciurus, 124, 130, 131, 132, 134, 135, 137
Protapirus, 125
Protoreodon, 59
Pseudanosteira, 26
Pseudoconglomerate, 23
Pseudoprotoceras, 75, 141
Pumice, 19
Punjab, 94
Pyrite, 9, 22, 24, 81

Quartz, 16, 17, 18, 19, 22, 23, 24, 25, 77, 85, 104, 119

(Citrine), 17, 18, 19, 25 (Rose), 17, 18
Quartzite, 17, 18, 22

Red River - Valley, 20, 22, 59, 60, 73, 97
Rhineura, 124, 134, 135, 137
Robinson, P., Ill, 114
Rockyford (Member of Rosebud), 13
Rosebud Formation, 13
Rutile, 16, 17, 19, 24, 25
Sage Creek (Formation, Montana), 57, 59, 61
Sage Ridge (& Fault), 9, 15, 20, 22, 59, 84, 91, 97, 98,
110, 139, 142

158

Sanidine, 77

Scenic Member (of Brule), 12, 13, Chap. VI, VII

Schist, 24

Sedimentary Dikes, 22

Selenite, 119

Sespe, 56

Shale-Pebble Conglomerate, 22

Sharon Springs (Member of Pierre), 119

Sheep Mountain, 8, 14, 80, 84, 90, 91, 92, 106

Silica, 22, 102, 118

Sinclairella, 27

Slim Buttes Formation, 9, 11, 12, 16-20 (described), 24,

25, 57, 73, 138, 141
South Dakota School of Mines and Technology, 111,

114
Sphene, 16, 17, 24, 80, 81
Stagner, L., 21
Staurolite, 16, 19, 24, 25
Stibarus, 58, 69, 126, 127, 130, 134, 137
Stall, H., 21
Stylemys, 124, 125, 137
Subhyracodon, 97, 120, 125, 126
Sylvilagus, 121, 124

FIELDIANA: GEOLOGY MEMOIRS, VOLUME 5

Syngenesis, - etic weathering, 17, 61

Taylor, D. T., 21

Taylor, E. H., 21

Teleodus, 16, 51, 56, 59

Tourmaline, 16, 17, 19, 24, 25, 80, 81 (Indicolite), 17, 19

Trachemys, 26, 58, 68

Trigonias, 19, 52, 58, 68, 72, 75, 141

Unconformity, 9
Unio, 60, 123, 137

Vieja (Formation) - Viejan Time, 56, 57, 59, 73, 75, 141

Woodbury, H. O., 21
Woodland, Dr. B. G., 5

Xenochelys, 26

Yellowstone-Bighorn Research Association, 21, 111, 114
Yoder (Member of Chadron), 56, 73, 75, 141

Zircon, 16, 17, 19, 23, 24, 25, 80, 81

/

4_